Brain Circuits, Modularity, and Childhood Syndromes

Florence Levy

A thesis submitted in fulﬁllment of the requirements for the degree of Doctor of Philosophy.

University of New South Wales

May 2010 COPYRIGHT, AUTHENTICITY, and ORIGINALITY STATEMENT

I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microﬁlms to use the 350 word abstract of my thesis in Dissertation Abstract International. I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation,

and

I certify that the Library deposit digital copy is a direct equivalent of the ﬁnal ofﬁcially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format,

and

I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgment is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project’s design and conception or in style, presentation and linguistic expression is acknowledged.

Signed ...... Date ...... iii

Abstract

The thesis explores the contribution of connectionist, cognitive and biological models to a better understanding of childhood psychopathology at a time when rapid advances in brain mapping techniques are contributing to a ‘brain circuit’ understanding of child development and behaviour. The prototypic structure of all brain circuits is described as originating in the frontal lobes, projecting to striatal structures, thalamic nuclei, and a ﬁnal link back to the frontal lobes. The thesis postulates that early childhood behaviour is characterised by modular stimulus-bound behaviours, which utilize feed-forward responses. Development requires the establishment of feedback-controlled circuits. Failure of integration of higher-order cognition with subcortical gating functions, gives rise to a number of childhood syndromes, discussed in the second part of the thesis. The initial chapters of the thesis examine neural network, and mechanistic cognitive models, evolution of language, and mirror systems, including analogies with birdsong development. In humans, integration of emotion with cognitive circuits is thought to occur at ventral striatal/basal ganglia levels. Higher-order functions are thought to require optimal arousal levels, maintained by complex integration of subcortical responsivity with cortical outputs. Working memory is believed central to control of higher order cognition, but like emotional and motor regulation, subcortical structures including the basal ganglia are crucial for sequencing and ordering behaviour, controlled at prefrontal levels. Here language plays a role in allowing the abstraction of ‘chunks’ held in working memory. Circuits involved in emotional regulation, as well as social cognition in terms of analogue mirror vs executive functions are investigated. Chapters 6 to 12 discuss the development and malfunction of circuits involved in Attention Deﬁcit Hyperactivity Disorder, Conduct Disorder, Autism/Aspergers Disorder, Obsessive Compulsive Disorder, Tourette’s Disorder, and Postraumatic Stress Disorder. The implications of dual cortical-subcortical functions for the frequent comorbidities described in childhood psychpathology are discussed. Cortical/subcortical models may provide a direction for candidate gene investigations where the same transmitter may have differing, but complementary functions. iv Contents

Aims and Hypothesis 1

1 Introduction 5 1.1 Models of consciousness ...... 8 1.1.1 Consciousness theories ...... 10 1.1.2 Comment ...... 12 1.1.3 Global Workspace ...... 13 1.2 Modularity ...... 22 1.2.1 Comment ...... 30 1.2.2 Dual processing circuits ...... 31 1.3 Neural network models ...... 34 1.3.1 Temporal sequencing ...... 37 1.3.2 Discussion ...... 42 1.4 Cortical-basal ganglia loops ...... 43 1.4.1 Robotic models ...... 46 1.5 Discussion ...... 47

2 Evolution, language and social cognition 53 2.0.1 Introduction ...... 53 2.0.2 Language evolution ...... 54 2.1 Speech perception ...... 58 2.1.1 Comment ...... 60 2.1.2 Dual language systems ...... 61 2.1.3 The Language hypothesis ...... 63 2.1.4 Comment ...... 66 2.1.5 Mirror systems ...... 66 2.2 Ontogenesis of social cognition ...... 69

v vi CONTENTS

2.2.1 Language development ...... 71 2.2.2 The RR model of language development ...... 71 2.2.3 Nature nuture ...... 76 2.2.4 Birdsong analogy and neural circuits ...... 79 2.2.5 Comment ...... 83 2.2.6 Genes, circuits and behaviour ...... 83 2.2.7 Cognitive function of language ...... 85 2.2.8 Discussion ...... 90 2.3 Reading, language and neural circuits ...... 92 2.3.1 Neural reading circuits ...... 93 2.3.2 Attention and reading ...... 95 2.3.3 Comment ...... 97

3 Neural circuits and behaviour control 99 3.1 Circuit plans of cortex, cerebellum and basal ganglia . . . 100 3.1.1 Circuit plan of neocortex ...... 100 3.1.2 Circuit plan of the cerebellum ...... 103 3.1.3 Circuit plan of the basal ganglia ...... 103 3.1.4 Basal ganglia and executive functions ...... 104 3.1.5 Comment ...... 106 3.1.6 New Anatomy of the Basal Forebrain ...... 107 3.1.7 Comment ...... 117 3.2 Oscillation networks ...... 118 3.2.1 Comment ...... 122 3.3 Temporally structured replay of behaviour ...... 122 3.4 Prefrontal cortex: cognitive and executive functions . . . . 123 3.4.1 Comment ...... 127

4 Development and working memory 129 4.1 Structural development ...... 130 4.2 Working memory ...... 136 4.2.1 Higher Order Cognition and Working Memory . . . 143 4.2.2 Comment ...... 148

5 The emotional brain 151 5.1 Emotion circuits ...... 155 5.1.1 Comment ...... 161 CONTENTS vii

5.1.2 Basal forebrain organisation ...... 161 5.1.3 Comment ...... 167 5.1.4 Discussion ...... 171 5.2 Social cognition and neural circuits ...... 172 5.2.1 Comment ...... 180 5.2.2 Reappraisal theory ...... 181 5.2.3 Discussion ...... 186

6 Attention Deﬁcit/Hyperactivity Disorder 189 6.1 Theories of ADHD ...... 190 6.1.1 ADHD circuits ...... 192 6.1.2 Comment ...... 202 6.2 Age effects ...... 203 6.2.1 ADHD and early onset bipolar disorder ...... 205 6.2.2 Working memory and ADHD ...... 205 6.2.3 Imaging studies ...... 208 6.2.4 Resting/default state and vigilance variability . . . 210 6.2.5 Continuous Performance Task (CPT) ...... 211 6.2.6 Discussion ...... 213 6.3 Treatment models ...... 215

7 Conduct Disorder and neural circuits 221 7.1 Diagnostic issues ...... 221 7.2 Basic emotion theories ...... 225 7.2.1 Fear circuits ...... 226 7.2.2 Reward circuits ...... 230 7.3 Genetic effects ...... 231 7.3.1 Comment ...... 234 7.3.2 Discussion ...... 236

8 Autistic Disorder/Asperger’s Syndrome 239 8.1 Autism ...... 239 8.1.1 Language in Autism ...... 240 8.1.2 Theory of Mind and Autism ...... 246 8.1.3 Central coherence theory ...... 249 8.1.4 Face processing in autism ...... 254 8.1.5 Neural abnormalities in autism ...... 257 viii CONTENTS

8.2 Asperger’s Disorder ...... 268 8.2.1 Discussion ...... 268 8.2.2 Therapeutic implications ...... 269

9 OCD and Tourette’s Disorder 271 9.1 Neural circuits in OCD ...... 272 9.1.1 The glutamate hypothesis ...... 276 9.2 Obsessive Compulsive Spectrum Disorders ...... 277 9.2.1 Developmental perspective ...... 282 9.2.2 Therapeutic implications ...... 283 9.2.3 Comment ...... 284 9.2.4 Discussion ...... 289

10 Posttraumatic Stress Disorder 291 10.1 Neural circuits in PTSD ...... 292 10.1.1 Neurocircuitry models ...... 303 10.1.2 Nucleus Accumbens/Ventral striatum and Anxiety Syndromes ...... 305 10.1.3 Sensitive period: gene environment interaction . . . 310 10.1.4 Comment ...... 311

11 Comorbidity 313 11.1 Diagnostic issues ...... 313 11.2 Circuit models ...... 317 11.2.1 Dual system model and comorbidity ...... 319 11.2.2 Interaction between cognitive and affective systems 321 11.2.3 Comment ...... 322 11.2.4 Cognitive/motor comorbidity ...... 323 11.2.5 Circuit pharmacology ...... 327 11.2.6 ADHD and Autistic spectrum disorder ...... 331 11.2.7 Comment ...... 331

12 Conclusions 333

Bibliography 343

Author Index 403 CONTENTS ix

List of Figures 413 x CONTENTS CONTENTS 1

The ancient subject matter of psychology - the mind and its various manifestations - is distressingly invisible, and a science with invisible content is likely to become an invisible science. Miller, Galanter and Pribram, 1960.

Aims and Hypothesis

The present thesis aims to explore the contribution of connectionist, cognitive, and biological models to a better understanding of childhood psychopathology, at a time when rapid advances in brain mapping techniques are contributing to a ‘brain circuit’ understanding of development and behaviour. In particular, brain circuit models and cognitive and connectionist approaches are thought to provide insights into the differences between stimulus- bound, rapid modular vs sequential cortico-striatal-cortical control of behaviour. A number of models, including parallel distributed processing (PDP), robotic models, and biological circuits, which integrate information from modular, parallel, subcortical circuits via a prefrontal executive system, are reviewed. The underlying neuronal architecture is investigated in terms of ventral and dorsal cortico-basal ganglia-thalamo-cortical re-entrant circuits. An executive system, with long-range cortical connnections to specialised subcortical modules is proposed, that allows parallel search functions, while co-occurring working memory allows representational higher-order behavioural control. Human development is postulated as achieving a balance between sequential top-down representational control, and fast-acting subcortical modules, that react to specific stimuli. Deficits in this developmental process are believed to underlie a number of childhood behavioural syndromes. An understanding of circuit architecture and functions implies that syndromal behaviour may result from circuit disorder at cortical or subcortical levels in a number of parallel circuits. The thesis postulates that early childhood development is, in general, characterised by modular stimulus-bound behaviours, which utilise feed-forward responses, and that the process of de- 2 CONTENTS velopment requires the establishment of feedback controlled circuits, allowing modulation of earlier prepotent behavioural responses. Failures or delays in the establishment of higher-order control gives rise to a number of childhood syndromes, discussed below. It is suggested that higher-order processing and ‘unconscious’ automatic processing can occur concurrently and/or alter- nately during communication and task performance. This implies a dual process and architecture, where symptoms result from deficits or imbalances in the processes which regulate higher- order modulation of automatic processing. These deficits occur in parallel circuits or systems, giving rise to behavioral syndromes and/or patterns of comorbidity described below. Stimulus-based control is thought to consist of habits, skills and procedures, which allow for high speed of reaction. However, when problem-solving and planning are required, higher-order processing is required. The fronto-striatal system is believed to operate both stimulus-based control when this is advantageous, and higher-order control when automatic processing is not sufficient, and both systems are thought to interact with each other as a dual interactive process. The present thesis proposes that higher functions of language and planning, as distinct from fixed- action patterns, require that the appropriate motor system is linked to appropriate working memories. The thesis links deficits in the development of higher cognitive functions, including language and working memory to deficits in the development of neocortical and subcortical neural circuits. As a result of failures in the development of higher-order control of behaviour, it is hypothesised that the behavioural syndromes discussed below, will manifest modular and stereotypic behaviours, resulting from deficits in cortical/subcortical circuit development. The thesis postulates that the phenomenology of a number of the syndromes described in the DSM-IV (APA, 1994) will reflect the developmental relationship between modular behaviours and the establishment of ‘workspace’ control and integration of modular behavioural responses. The developmental timing of this con- CONTENTS 3 trol will vary according to genetic, environmental and individual factors. For example, the development of language allows conscious labelling of behaviours, while the age-appropriate development of cortical-subcortical circuits allows for increased working memory and attention. Thus a developmental timetable will provide a better understanding of the phenomenology and individual variation in the syndromes examined. Subcortical control of behaviour is more likely to be modular, while cortical control is more likely to be both modular and distributed, according to task requirements. 4 CONTENTS Chapter 1

Introduction

Figure 1.1: Structual anatomy of the brain

While neuroscientiﬁc advances have led to greater appreciation of the importance of neural circuits in the control of behaviour, there have been few attempts to relate these advances to childhood behavioural syndromes, as investigated in the present thesis. Hyman (2007) has raised the question of how neuroscience will contribute to revisions of the Diagnostic and Statistical Man- ual of Mental Disorders (DSM) 1 and the International Classiﬁca- tion of Disease (WHO, 1993). He points out that the substantial

1Diagnostic and Statistical Manual of Mental Disorders 4th edition.

5 6 CHAPTER 1. INTRODUCTION gaps in our knowledge of the neurobiology that underlies mental disorders derive in the large part from the difﬁculty of characterising the circuits and mechanisms that underlie higher brain function, and the complexity of the genetic and developmental underpinnings of normal and abnormal behavioural variation. Nonetheless, Hyman suggests that neurobiological information can, along with family and genetic studies, help to shape a reconsideration of important aspects of the DSM system. He also points out that while current diagnoses can be used to select treatments, the medications used do not necessarily respect the boundaries of disorders deﬁned by the DSM - for example, antidepressants can treat anxiety disorders, as well as OCD and depression. The dimensional nature of many disorders, as well as the application of multiple diagnoses or comorbidity may contribute to this problem. The present thesis will explore the implications of how a circuit-based understanding of symptomatology will contribute to diagnostic phenotypes, their overlap and pharmacological and cognitive approaches to treatment. Brown and Barlow (2005) pointed out that past neurobiological approaches have often been based on adult lesion studies, whereas they suggest that present approaches should, where possible, be developmentally-based. Thus genetic advances, as well as advances in brain mapping techniques combined with cognitive models are postulated as allowing a new ‘brain-based’ model of childhood symptomatology, including poorly understood phenomena, such as comorbidity or the co-occurrence of clinical syndromes. The present thesis postulates that an understanding of brain function derived from both connectionist/cognitive and neural models will provide new insights, and help guide future models and treatments. The work of philosophers such as Dennett (1991) and Churchland (2002a) are considered relevant to questions of brain/behaviour relationships explored in the present analysis, including the nature of consciousness, which may change during the course of development. Pavuluri and Sweeney (2008, p1273-1288) describe several in- 7 terrelated concepts important for research and potential clinical practise in coming years.

The importance of understanding the brain as being composed of multiple functionally organised circuits or systems, rather than isolated regions performing discrete functions.

The potential uses of clinical biomarkers for understanding disorders and monitoring response to pharmacotherapy.

The growing understanding of alterations in the developmental trajectory of functional brain systems in patients with onset of psychiatric illnesses in childhood, adolescence, and young adulthood.

The opportunity provided by understanding at the level of functional brain systems to facilitate a translational integration of animal and clinical research.

The prospect of multimodal approaches for studying brain and behavioural systems in parallel

(Pavuluri and Sweeney, 2008, p1273-1288).

The authors suggest that one of the more exciting directions for future research involves efforts to probe the interface between affective and cognitive processing, allowing an understanding of how “thinking and feeling” affect each other. They suggest that studies in macaque monkeys, humans without disorder, and patients with paediatric bipolar disorder suggest interactions between cognitive and affective brain regions at three hierarchically organised tiers: 1/ Higher cortical interactions between dorsolateral prefrontal cortex and ventrolateral prefrontal cortex, 2/ An intermediate level of interface between dorsal and ventral anterior cingulate cortex, 3/ Interaction at subcortical level betwen dorsal and ventral striatum and amygdala. More recently, Insel (2010) has stated that

Today scientiﬁc approaches based on modern biology, neuroscience and genomics are replacing nearly a century of purely psychological theories, yielding new approaches to the treatment of mental illnesses. [...] An ability to identify the brain circuit malfunctions underlying mental illness could have broad implications for diagnosis and treatment (Insel, 2010, p44-50). 8 CHAPTER 1. INTRODUCTION

1.1 Models of consciousness

While consciousness has traditionally been conceptualised as a cortical process, the present thesis postulates that the cortex and subcortical centers interact continuously during the performance of most activities. The phenomenology of a number of childhood syndromes is understood in terms of deﬁcits, delays, or failures in the establishment of ‘top-down’ modulation of parallel subcortical modules, thus allowing behaviour to be ‘captured’ by random environmental stimuli, or repetitive automatic internal circuits. The principle of random environmental ‘capture’ of behaviour, resulting from failure of representational control, will be applied to the understanding of a number of childhood syndromes, including At- tention Deﬁcit Hyperactivity Disorder (ADHD), Autism/Autistic Spectrum Disorder, Obsessive Compulsive Disorder (OCD), and Post Traumatic Stress Disorder (PTSD). The age-related phenotypes of these conditions are better understood in terms of brain circuit development, and stage of circuit establishment. An in- depth understanding of neural circuit phenomenology and its relation to behaviour has become more accessible, with the development of non-invasive brain imaging techniques and related interventions.

Dennett (1991, p253-254) has outlined a “thumbnail sketch” of his theory of consciousness. He believes that

There is no single deﬁnitive stream of consciousness, because there is no central Headquarters, no Cartesian Theater, where it all comes together for the perusal of a Central Meaner. Instead of such a single stream (however wide), there are multiple channels in which specialist circuits try, in parallel pandemoniums, to do their various things, creating Multiple Drafts as they go. Most of these fragmentary drafts of narrative play short-lived roles in the modulation of current activity, but some get promoted to further functional roles, in swift succession, by the activity of a virtual machine in the brain. The seriality of this machine (its “von Neumanesque” character) is not a hard wired design feature, but rather the upshot of a succession of coalitions of these specialists (Dennett, 1991, p253-254). 1.1. MODELS OF CONSCIOUSNESS 9

While Dennett postulates a series of innate specialist modules, whose “multiple drafts” in consciousness are transiently modulated by a “virtual” serial process, Dennett (1991, p277) grants that there are signiﬁcant problems of self-control created by the proliferation of simultaneously active specialists. He describes one of the fundamental tasks performed by the “Joycean” machine (referring to the parallel literary style of the novelist, James Joyce) is to “adjudicate disputes, smooth out transitions between regimes, and prevent untimely coups d’etat, by marshalling the right forces”. Dennett is unclear about the mechanisms of how this control is accomplished, but refers to “concentration, self- admonition and various mnemonic tricks and rehearsals” (Den- nett, 1991, p277).

The issue of serial vs parallel behaviour control becomes important when considering questions of modularity and the development of cortical control over impulsivity, and capacity for sustained attention and goal-orientation, developed during childhood. Churchland (2002a, p69-71) has taken Dennett to task on his metaphors. He claims that the brains of animals and humans are “most emphatically not vN (von Neuman) machines. Their coding is not digital; their processing is not serial; they do not execute stored programs; and they have no random-access storage registers whatever”. Churchland describes the brain as a massively parallel vector processor, in which its conceptual framework resides in the slowly acquired configuration of its 1014 synaptic connections. He believes its specific understanding of the local world here-and-now (its fleeting thoughts and perceptions) resides in the patterns or vectors of activation levels across its 1011 neurons, with the character of these fleeting patterns dic- tated by the learned matrix of connections that transform peripheral sensory activation vectors into well-informed central vectors and ultimately motor vectors. 10 CHAPTER 1. INTRODUCTION

1.1.1 Consciousness theories

Edelman (2004, p34-39) has described a theory which he calls ‘’Neural Darwinism”, by which processes of re-entry (see Lamme06) and iteration are postulated as fundamental to the control of behaviour, both conscious and unconscious. He points out that sub-regions of the brain, mapping somato-sensation or perception are often functionally segregated. For example, systems for vision such as colour, movement and orientation are functionally segregated giving rise to the “binding problem”, which he postulates can be solved by re-entry processes. Accord- ing to Edelman:

Unlike computer models, the enormous variability of neuronal chemistry, network structure, synaptic strengths, temporal properties and motivational patterns, as well as environmental novelty do not lend themselves to formal rules, governed by explicit and unambiguous instructions or input signals. (1) Developmen- tal selection - epigenetic variations in the patterns of connections among growing neurons create repertoires consisting of millions of variant circuits or neuronal groups. (2) Experiential selection - after the major neuroanatomy is built, large variations in synaptic strength, positive and negative result from variations in environmental input during behaviour. (3) Re-entry during development, when large numbers of reciprocal connections are established locally and over long distances (Edelman, 2004, p34-39).

Thus, Edelman (2004, p39) suggested a theory of neuronal group selection (TNGS), in which re-entry is defined by Edelman (2004, p41) as the ongoing recursive interchange of parallel signals among brain areas, which serves to coordinate the activities of different brain areas in space and time. ‘’Unlike feedback, re- entry is not sequential transmission of an error signal in a simple loop. Instead, it simultaneously involves many parallel reciprocal paths and has no prescribed error function attached to it” (Edel- man, 2004, p39). The concept derives from immunology, where such massive reactive processes occur. Edelman (2004, p49-51) Edelman outlines three main anatomical ’motifs’ in human brains. The first is the thalamo-cortical motif, with tightly connected groups of neurons connected both 1.1. MODELS OF CONSCIOUSNESS 11 locally and across distances by rich reciprocal connections. The second is the polysynaptic loop structure of the inhibitory circuits of the basal ganglia (see (Heimer et al., 2008)). The third consists of diffuse ascending projections of different value systems, which act as constraining elements in a selectional system. These consist of diffuse ascending systems such as the dopaminergic, noradrenergic and cholinergic systems. He describes perceptual categorization as being carried out by interactions between sensory and motor systems linked by reentry. Edelman postulates that while perceptual categorisation is fundamental, it cannot give rise to generalisations required to make maps of perceptual maps, yielding a concept. He suggests that generalization arises by abstracting features of such global mappings by means of ‘higher-order maps’. “Perceptual categorisation and concept formation require a dynamic memory system, operating within the brain’s selectional framework, influenced by changes in neural inputs, that come from that brain’s value systems”. Edelman calls the central memory system a value-category memory system, in which the constraints of value systems can determine the degree and extent of recall and output (Edelman, 2004, p54).

Edelman (2004, p54) postulates that at a point in evolutionary time, corresponding to the transition between reptiles and birds and reptiles and mammals, a new reciprocal connectivity appeared in the thalamo-cortical system.

Massively re-entrant connectivity developed between the cortical areas carrying out perceptual categorisation and the more frontal areas responsible for value-category memory, based on fast changes in synaptic strength [...] The re-entrant connections between thalamus and cortex developed [...] while the reticular nucleus of the thalamus developed enhanced inhibitory circuits [...] This allowed the activity of the reticular nucleus to gate or select various combinations of the activity of these speciﬁc thalamic nuclei corresponding to different sensory modalities [...] This bootstrapping between memory and perception is assumed to be stabilised within time periods ranging from hundreds of milliseconds to seconds (Edelman, 2004, p54-55). 12 CHAPTER 1. INTRODUCTION

1.1.2 Comment

The concept of parallel re-entry processes appears to differ somewhat from Lamme (2006) to Edelman. The former concept emphasises circuits which re-enter the cortex and terminate very near the same region from which they originated, while Edelman has a broader, iterative concept which allows the concept of “degeneracy” or the ability of different structures to carry out the same function. Degeneracy is a central feature of immune responses. Edelman (2004, p44) believes this concept is central in solving the brain’s binding problem.

In correlating different maps for colour orientation and object movement, mutual re-entrant interactions link various neuronal groups, ﬁring in phase with each other, but in the next time phase, different neurons and neuronal groups may form a structurally different circuit, which yields a similar output (Edelman, 2004, p32).

Edelman’s definition of conscious states, as temporally ordered and serial as well as showing intentionality and diverse binding, differs in emphasis from authors such as Damasio (2000), who pays more attention to subcortical ‘body’ processes but both are agreed on the transient and pulse-like nature of consciousness. Damasio (2000) describes “core consciousness” as providing the organism with a sense of self, which is stable across the lifetime of the individual and not dependent on conventional memory, working memory, reasoning or language. He believes core consciousness is not exclusively human. He also describes extended consciousness, which has several levels of organisation, evolves across the lifetime of the organism and depends on conventional memory and working memory as well as language and thus at- tains its maximal development in humans. Extended consciousness is built on a foundation of core consciousness. Thus both Edelman and Damasio ascribe important roles to language and memory in higher-order consciousness. Damasio (2000, p115-117) Damasio believes that the roots of the self are found in the unconscious brain devices which continuously maintain the stability of the body state. These consist of nuclei lo- 1.1. MODELS OF CONSCIOUSNESS 13 cated at the brain stem level, such as the parabrachial nucleus, and peri-aqueductal gray, the hypothalamus, the basal forebrain and somatosensory cortices. Damasio describes three steps behind the assembly of consciousness. The first involves the construction of first-order maps of the object, including form, colour, sound and touch, but also includes the mapping of reactions to the object. Consciousness occurs when a further higher-order map, a “second-order map”, describes the relationship between the object and the organism, and the change induced by the object, which gives rise to a sense of self-knowing. He suggests that the superior colliculus, thalamus, cingulate cortices, and some medial parietal association cortices may all contribute via relays in the thalamus. The distinction between associative, reflexive and representation-guided cognitive processes (Goldman-Rakic, 1987c) is central to the distinction between Primary vs HOC (Edelman and Tononi, 2000). The present thesis postulates that the development of symbolic cortical representations, facilitates working-memory guided integration and control of behaviour. Edelman’s definition of conscious states, as temporally ordered and serial as well as showing intentionality is important for an understanding of behavioral syndromes in which behaviour fails to be temporally ordered or sequential, and often subject to immediate reflexive reactions. Also, Edelman’s concept of HOC requiring language centers, and the development of vocabulary and syntax for planned behaviours underlines the importance of language and/or its deficits in the understanding of childhood behaviour and development.

1.1.3 Global Workspace

Dehaene and Naccache (2001) have discussed the philosophical, empirical and theoretical bases on which a cognitive neuroscience approach to consciousness can be founded. They describe three major empirical observations that any theory of consciousness 14 CHAPTER 1. INTRODUCTION should incorporate:

(1) A considerable amount of processing is possible without consciousness. (2) Attention is a prerequisite of consciousness and (3.) Consciousness is required for some speciﬁc cognitive tasks, including those that require durable information maintenance, novel combinations of operations, or the spontaneous generation of intentional behaviour (Dehaene and Naccache, 2001, p1-37).

The authors propose a theoretical framework that synthesises these postulates, namely a Global Neuronal Workspace. They propose that at any given time, many modular cerebral networks are active in parallel, and process information in an unconscious manner. They also propose that information becomes conscious if the neural population that represents it, is mobilised by top-down attentional ampliﬁcation into a state of coherent activity that involves many neurons distributed throughout the brain.

The long-distance connectivity of ‘workspace neurons’ can, when they are active for a minimal duration, make the information available to a variety of processes, including perceptual categorisation, long-term memorisation, evaluation, and intentional action (De- haene and Naccache, 2001).

The authors postulate that this global availability of information through the workspace is what is subjectively experienced as a conscious state. According to Dehaene and Naccache (2001) the problem of the cognitive neuroscience of consciousness does not pose any greater conceptual difficulty than identifying the cognitive and cerebral architectures for motor action, with the main difference that consciousness is an introspective phenomenon, and not an objectively measurable response. The authors thus consider introspective reports as serious data, but con- cede that reports such as hallucinations may not represent reality. They point out that theories, which posit a residual dualism, are not useful. “Research on the cognitive psychology and neuroscience of consciousness should take into account the many levels of organisation at which the nervous system can be studied, from molecules to synapses, neurons, local circuits, large scale networks, and the hierarchy of mental representations that they 1.1. MODELS OF CONSCIOUSNESS 15 support”. Thus “cognitive neuroscientific approaches to consciousness seem capable of addressing both the cognitive architecture of mental representations and their neural implementation” (De- haene and Naccache, 2001). Dehaene and Naccache (2001) describe three fundamental empirical findings on consciousness:

(1) First, a considerable amount of processing can occur without consciousness, including perceptual, motor, semantic, emotional and context dependent processes. For example, priming studies indicate that a very brief visual stimulus can be perceived consciously when presented in isolation. However the same brief stimulus can fail to reach consciousness, when it is surrounded in time by other stimuli that serve as masks. (2) Attention is believed to be a prerequisite of consciousness. Thus conscious perception is thought to result from an interaction of stimulation factors, including intensity and duration, with the attentional state of the observer. (3) Consciousness is required for speciﬁc mental operations.

The authors identify at least three further classes of computations that appear to require consciousness:

(1) Maintenance of durable and explicit information: Dehaene and Naccache (2001) suggest that the ability to maintain representations in an active state for a durable period of time in the absence of stimulation seems to require consciousness. The authors point out that physiological and behavioural studies in both humans and monkeys suggest that the ability to maintain information on-line independently of the stimulus presence depends on a working memory system associated with dorsolateral prefrontal regions (Fuster, 1989; Goldman-Rakic, 1987b). This suggests that the “working memory system made available by prefrontal circuitry must be tightly related to the durable maintenance of information in consciousness”. (2) Novel combination of operations: The authors suggest that strategic operations which are associated with planning a novel strategy, evaluating it, and controlling its execution, and correcting possible errors, cannot be accomplished unconsciously. (3) Intentional behaviour: The authors describe a third type of mental activity that may be speciﬁcally associated with consciousness, as the spontaneous generation of intentional behaviour. For example, they point out that even blindsight patients who show excellent performance in pointing to objects, but they never spontaneously initiate any visually-guided behaviour in their impaired 16 CHAPTER 1. INTRODUCTION

ﬁeld. The authors also describe ‘reportability’ in which the subject uses language to describe his/her mental life as exclusive to conscious information (Dehaene and Naccache, 2001, p1-37).

According to Dehaene and Naccache (2001) there are both functional and neurobiological definitions of modularity, which postulate that automatic or unconscious cognitive processing rests on multiple dedicated processors or modules (Fodor, 1983; Baars, 1989; Shallice, 1988), characterised by information encapsulation, domain specificity, and automatic processing. “In neuroscience, specialised neural circuits that process only specific types of inputs have been identified at various spatial scales, from orientation-selective cortical columns to face-selective areas. The breakdown of brain circuits into functionally specialised subsystems can be evidenced by various methods including brain imaging, neuropsychological dissociation, and cell recording” (Dehaene and Naccache, 2001, p1-37). Dehaene and Naccache (2001) point out that specialised neural responses such as face-selective cells can be recorded in both awake and anaethsetised animals, thus reflecting that an automatic computation can proceed without attention. They propose that “a given process involving several mental operations, can proceed unconsciously only if a set of adequately interconnected modular systems is available to perform each of the required operations”. The hypothesis implies that multiple unconscious operations can proceed in parallel, as long as they do not simultaneously appeal to the same modular system in contradictory ways. The authors note that unconscious processing may not be limited to low-level or computationally simple operations, and may operate unconsciously as long as it is associated with functional neural pathways, either established by evolution, laid down during development, or automatised by learning. “For instance face processing, word reading, and postural control all require complex computations, yet there is considerable evidence that they can proceed without attention, based on specialised neural subsystems. Conversely, computationally trivial but non-automised 1.1. MODELS OF CONSCIOUSNESS 17 operations such as solving 21 minus 8, require conscious effort”. The second Dehaene and Naccache (2001) hypothesis relates to the postulate that the performance of effortful tasks can temporarily inhibit the automatic activation of some processors, and enter into a strategic or “controlled” mode of processing (Posner, 1994); (Shallice, 1988); (Schneider and Shiffrin, 1977). This mode has been variously called the control executive (Baddeley, 1986); the supervisory attentional system (Shallice, 1988); the anterior attention system (Posner, 1994); the global workspace (Dehaene et al., 1998b); (Baars, 1989); or the dynamic core (Tononi and Edelman, 1998). Thus besides specialised processors, the architecture of the human brain comprises a distributed neural system or ‘workspace’, with long-distance connectivity, that can potentially interconnect multiple specialised brain areas in a coordinated, though variable manner (Dehaene et al., 1998a). These areas do not directly exchange information in an automatic mode, but can nevertheless gain access to each other’s content through the workspace, providing potential for the combination of multiple input, output, and internal systems.

The vast amounts of information that we consciously process suggests that at least ﬁve main categories must participate in the workspace: perceptual circuits that inform about the present state of the environment; motor circuits that allow the preparation and controlled execution of actions; long-term memory circuits that can re-instate past workspace states; evaluation circuits that attribute to them a valence in relation to previous experience; and attentional or top-down circuits that selectively gate the focus of interest. The global interconnection of these ﬁve systems can explain the subjective unitary nature of consciousness, and the feeling that conscious information can be manipulated mentally in a largely unconstrained fashion. In particular, connections to the motor and language systems allow any workspace content to be described verbally (Weiskrantz, 1997) (Dehaene and Naccache, 2001).

. The Dehaene and Naccache (2001) third postulate concerns the role of attention in gating access to consciousness. Dehaene et al. (1998a) postulated that “top-down attentional ampliﬁcation is the mechanism by which modular processes can be temporarily 18 CHAPTER 1. INTRODUCTION mobilised and made available to the global workspace, and therefore to consciousness”. According to this theory, the same cerebral processes may at different times contribute to consciousness. The authors postulate a form of “neural Darwinism” (Edelman and Tononi, 2000), in which transient self-sustained workplace states follow one another in a constant stream.

To enter consciousness, it is not sufficient for a process to have on-going activity; this activity must also be amplified and maintained over a sufficient duration for it to become accessible to multiple other processes. Amplification is thought to occur via modulatory input from the release, in dorsal and ventral circuits, of neurotransmitters such as dopamine, norepinephrine, serotonin and acetyle choline. Without such “dynamic mobilisation” a process may still contribute to cognitive performance, but only unconsciously (Dehaene et al., 1998a, p14529-14534).

Dehaene and Naccache (2001) point out that the conscious availability of information is postulated as determined by two structural criteria, which are ultimately grounded in brain anatomy. The information must be represented in an active manner in the ﬁring of one or several neural assemblies, and second, bi-directional connections must exist between these assemblies and the set of workplace neurons. This “active representation” excludes from consciousness an enormous wealth of latent information.

[...] for example, we can become conscious that a sentence is not grammatical, but have no introspection on the inner workings of the syntactical apparatus that underlies this judgment, and which is presumably encoded in connection weights within temporal and frontal areas [...] an interesting exception to the limit on introspection relates to subjects’ verbal reports on processes that are slow, serial and controlled, and generated dynamically through the serial organisation of active representations of current goals or intentions, where strategic steps may be reported. Language is the most obvious of such slow, serial and controlled processes. [...] Bi-directional connectivity with the workspace, implies that even when representations are encoded by an active neural assembly, consciousness will not be achieved if the connectivity needed to establish a reverberating loop with workspace units is absent or damaged (Dehaene and Naccache, 2001). 1.1. MODELS OF CONSCIOUSNESS 19

Finally Dehaene and Naccache (2001) ask whether the workspace hypothesis is compatible with finer-grained knowledge of brain activity and physiology? They consider the requirement for long-distance connectivity and note that long-range cortico- cortical tangential connections, including interhemispheric connections, mostly originate from the pyramidal cells of layers 2 and 3. These layers, though present throughout the cortex, are particularly thick in dorsolateral prefrontal and inferior parietal cortex (Dehaene and Naccache, 2001, p1-37). Dehaene (2009) more recently points out that prefrontal neurons exhibit a remarkable capacity to remain active, even after the perceived object has vanished. “Their firing can remain intense for several dozens of seconds, and keep the working memory of a past episode for essentially as long as the information will be needed” (Dehaene, 2009, p321). According to Dehaene (2009, p318) human prefrontal neurons show a clear adaptation to this massive increase in connectivity: their dendritic trees, which receive incoming inputs are bushier and synaptic contacts are massively more numerous than those other primates. One category of neuron, the great fusiform cell, seems only to exist in Homo sapiens and the other great apes. Similarly, Goldman-Rakic (1988) has described a dense network of long-distance reciprocal connections linking dorsolateral PFC with premotor, superior temporal inferior parietal, anterior and posterior cingulate cortices, as well as deeper structures, including the neostriatum, para-hippocampal formation, and thalamus. According to the authors, this connectivity pattern, which is probably also present in humans, provides a plausible substrate for fast communication amongst the five categories of processes postulated by the theory for consciousness (Goldman- Rakic, 1988).

Temporal and parietal circuits provide a variety of high-level perceptual categorisations of the outside world. Premotor, supplementary motor and posterior parietal cortices, together with the basal ganglia (caudate nucleus) the cerebellum and the speech production circuits of the left inferior frontal lobe are thought to allow 20 CHAPTER 1. INTRODUCTION

for the intentional guidance of actions, including verbal reports, from workspace contents. Furthermore, the hippocampal region is thought to provide an ability to store and retrieve information over the long term (Goldman-Rakic, 1988, p87-102).

According to Goldman-Rakic (1988), direct or indirect connections with orbitofrontal cortex, anterior cingulate (AC), hypothalamus, amygdala, striatum and mesencephalic neuromodulatory nuclei may be involved in computing the value or relevance of current representations in relation to previous experience. Leiner et al. (1991) suggest that recurrent circuits are able to greatly increase the computing power of modules with limited information processing capability. While Leiner et al. (1991) imply aggrega- tion of information, a connectionist approach would emphasise the power of ‘hidden units’ to recycle inputs received from copies of the output units. Thus while each of the systems described above can, according to Dehaene and Naccache (2001) be activated without consciousness, the authors postulate that their coherent activity, supported by their strong interconnectivity, coincides with the mobilisation of a conscious content of the global workspace. The authors believe that the hypothesis of an attentional control of behaviour by supervisory circuits, including AC and PFC, above and beyond other more automatised sensorimotor pathways, may ultimately provide a neural substrate for concepts of voluntary action and free will. Finally they suggest that at a higher cognitive level, the action perception, verbal reasoning and theory of mind (TOM) modules that are applied to interpret and predict other people’s actions, may combine in one’s conscious workspace to account for the subjective sense of self. On the other hand, several mammals and young human children exhibit greater brain modularity than human adults. Nevertheless, they exhibit intentional behaviour, partially reportable mental states, some working memory, but perhaps no TOM, leaving open the question of the “all-or-none nature of consciousness” in children. This view of consciousness in children is consistent with the present view of a developmental 1.1. MODELS OF CONSCIOUSNESS 21 process of ‘consciousness capacity’, structured during childhood. Dehaene (2007, p28-47) points out that while consciousness is said to operate in an uninterrupted flow, the brain operates in an anticipatory mode, which is ceaselessly active in re-assessing the past, in order to better anticipate the future. This “internal consciousness tends to be neglected by psychologists, who rely on reflex-like stimulus-response paradigms”. He approaches the problem in terms of brain architecture, believing that the laws of psychology will never be deeply understood, until they are related to levels of brain organisation. He describes the brain as behaving like a statistician who collects multiple samples before reaching a firm conclusion. He describes successive steps in processing nu- meric symbols, consisting of visual recognition of the symbol, conversion into an internal quantity that serves to support decision making, and motor programming of a response. Dehaene (2007) suggests that while the coordination of multiple mental operations has been demonstrated to slow down task performance (Pasher, 1994, p220-224), only one stage, the so- called central stage encounters a bottle-neck in which mental operations are done in series and not in parallel. ”Empirically, response time costs related to coordination of multiple tasks have been measured, and systematically associated by brain imaging with prefrontal and parietal regions” (Marios and Ivanoff, 2005, p296-305). Dehaene describes a hierarchical division of cognition, which establishes a direct link with the distinction between conscious and non-conscious operations. While conceptual and motor representations can be activated without consciousness, Dehaene describes operations that rely on cognitive control as impossible to execute without our being conscious of them. “[...] consciousness is associated with a serial cerebral system of limited capacity, responsible for controlling other mental operations” [...] “Thus when a stimulus is presented during cerebral processing of another goal, and followed by a mask, the subject will state that no stimulus has occurred (attentional blink)” (Dehaene, 2007). Ac- cording to Dehaene (2007) neuroanatomy and brain imaging are 22 CHAPTER 1. INTRODUCTION beginning to confirm that the prefrontal areas are implicated in a vast distributed associative network, where sudden and coordinated activation punctuates each access to information.

1.2 Modularity

An important issue in relation to behavioural control, is the question of modularity vs cohesion. A module is defined by Segal (1996, p151) as a component of the mind, or brain, a mechanism, a system or some such that explains a ’competence’. She defines intentional competence, as related to a skill such as vision or language, whereas computational competence relates to representational processing, although both processes may coexist. While philosophers since Descartes, and before, have debated the nature of consciousness, they have generally assumed consciousness to be a unitary phenomenon. On the other hand, modularity has been described in relation to perception and memory (Moscovich, 1995; Fodor, 1983), but the development of consciousness as a modular construction, built by physiological and learning processes remains controversial. Marcus (2006, p443-465) points out that although modules are, by definition computationally distinct (Fodor, 1983; Coltheart, 1999), they need not be genetically unrelated. Although evolutionary pressures can clearly lead distinct physiological structures to diverge and specialize, natural selection tends to be a slow process and many putative modules (e.g., a language facility) are relatively recent, and as such, might be expected to derive from common origins. A corollary of Marcus (2006) concept of ‘descent with modification’ might be that overlapping childhood phenotypes or symptoms such as reading and attentional problems may derive from common genetic influences. Fodor (1983) argued that modules are innate, mandatory, fast and domain specific, with proprietary inputs, and are informationally encapsulated. Thus while peripheral cognitive systems, involving senses and language are informationally encapsulated, the rest of cognition according to Fodor is domain-general or un- 1.2. MODULARITY 23 encapsulated. This distinction is central for the present investigation, which will argue that the process of development involves higher-order integration of encapsulated or relatively encapsulated modules, and that childhood syndromes represent failure or pathology of this integrative process. Simpson et al. (2005, p1-19) point out that cognitive faculties can theoretically exhibit domain-specificity or encapsulation with regard to both the information they draw on when processing, and the computational processes by which such processing is implemented, allowing a distinction between representational and computational modules. They define representational modules as domain-specific, organised and integrated bodies of data, while computational modules are domain-specific processing devices. “A parser might be conceived of as a computational module that displays the contents of a (representational) module devoted to linguistic information in order to generate linguistic information, in order to generate syntactic and semantic representations of physical sentence forms”. Simpson et al. (2005) describe adjustments to the Fodorian notion of modularity. They argue that for the domain-specific facili- ties found in central cognition to be modular, it is clearly the case that input to these faculties must be (at least partly) conceptual and that their output may be much deeper than that of peripheral systems. They also suggest that such faculties may be more open to influence from other faculties (ie less encapsulated), though Fodor’s other criteria of mandatory, fast and domain specificity remain. They raise the question of just how modular is central cognition? While some theorists (Sperber, 1994) defined a ‘massive modularity’ (MM) hypothesis (the claim that the human mind consists entirely of modules), others argued for a less massive system. Thus while ’Theory of Mind’ (TOM) might be modular, the existence of a domain-general central executive or “integrative cognitive mechanism” (Baron-Cohen, 1995; Leslie, 1994) was also postulated. There have been further criticisms of Fodor’s (Fodor, 1983) notion of modularity as “structural and fixed”. For example, Shanon 24 CHAPTER 1. INTRODUCTION

(1988) suggested that modules should be regarded as dynamic and context dependent. Shanon (1988) agreed with Fodor that some central processes must be non-modular, and that modules occurring within the central processes exist only temporarily and are not associated with a fixed neural architecture, but criticises Fodor as operating from a mixture of two perspectives, namely psychological and philosophical. However, Bennet (1990) pointed out that Shanon also resorted to a “totally mixed bag of psychological phenomena”, displaying the same disregard for a distinction between functional and empirical criteria as Fodor. He criticised Shanon’s use of the term modularity as being too broad and that “to the extent that the existence of form/function correspondence is a precondition for successful neuropsychological research, there is not much to be expected in the way of a neuropsychology of thought”. Bennet (1990) pointed out that Shanon agreed with Fodor that some central processes must be non-modular, and that modules, occurring within the central processes, exist only temporarily and are not associated with a fixed neural architecture. Arbib (1989) has also questioned Fodor’s concept of modularity as (i) domain specific (ii) innately specified (iii) associated with distinct neural structures and (iv) computationally autonomous. Arbib asserted that the key concept of computational autonomy means that modules “do not share, and hence do not compete for such horizontal resources as memory, attention, judgment intelligence etc”. Ac- cording to Arbib (1989, p214), Fodor’s concept of informational encapsulation means that while there may be internal feedback between the representations within the module, these internal representations are not involved in paths to or from external modules. “However an interface between perception and utilities must take place somewhere if information from input systems is to be utilised in action” (Arbib, 1989, p214). Carruthers and Chamberlain (2000, p1-12) have pointed out that evidence has gradually been mounting in favour of the modularity of many mental functions and capacities, including central 1.2. MODULARITY 25 systems.

While Fodor’s initial characterization of the nature of modules (as innate, fast, domain speciﬁc, and informationally encapsulated, with proprietary inputs and shallow outputs), was highly restric- tive, other theorists have since liberalised the notion, for example dropping the requirements of shallowness and strict encapsulation in such a way that some central conceptual systems might plausi- bly be thought modular in nature (Carruthers and Chamberlain, 2000).

According to Carruthers and Chamberlain (2000, p1-12), modu- larism has generally been associated with nativism, and its two most famous proponents, Chomsky (1988) and Fodor (1998), have both avoided evolutionary theorizing, seeing the appearance of modules as a “mere by-product of the expansion of the hominid neocortex”. Carruthers and Chamberlain (2000) describe the intellectual movement known as “evolutionary psychology” which postulates that the mind contains a whole suite of modular adaptations (Barkow and Tooby, 1992; Pinker, 1997) (in contrast with the empiricist model of the mind as a large general purpose computer). Carruthers and Chamberlain (2000) describe evolutionary psychology as focusing on features of the mind that are more or less unique to humans (e.g. language, consciousness of self and others, aesthetic preferences, and psychopathology). They describe the “most successful” approach as abandoning direct evolutionary explanations of specific cultural repertoires and behaviors, rather seeking to discover cognitive universals, which under- pin the diversity of human thought and action. Samuels (2005, p107-121) has argued the case for rejection of MM, which maintains that, “in addition to whatever innate representational structure we may possess, central processes also rely on a multitude of innate, special purpose information processing mechanisms or modules”. The ‘tractability’ argument assumes that central cognition must be subserved by modular mechanisms, because the alternatives are computationally intractable. These arguments assume that human cognitive processes are classical computational ones, i.e. alorithmically specifiable pro- 26 CHAPTER 1. INTRODUCTION cesses, defined over syntactically structured mental representations. Fodor (2000, p37-64) has argued that abductive reasoning is rendered tractable by central modularity. However Samuels argues that abduction fails when cognitive demands are too great. He also points out that

Non-modular reasoning mechanisms need to be located within an architecture that contains other mechanisms that are responsible for the production of fine-grained, real-time responses. [...] Since human beings succeed in responding in real-time environmental conditions” [...] “the past few decades of research in robotics makes it plausible to posit an additional kind of mechanism that aids in the production of real-time behaviour: modular reactive behaviours that produce rapid behavioural responses to stereotypic environmental conditions” (Samuels, 2005, p107-121) . Carruthers (2005, p70-73) discusses the issues of “central modularity”, “massive modularity”, and the role of language. Since Fodor (1983), a number of authors have suggested that at least some central processes may be modular in terms of central/conceptual systems, as well as modular input and output systems (Baron-Cohen, 1995; Botterill and Carruthers, 1999). This requires a modification of the notion of a module, since central modules are supposed to be capable of taking conceptual inputs, and generating conceptualised outputs. Carruthers points out that one argument for MM derives from evolutionary biology in general, where new structures are “bolted on” to an existing repertoire. However any general-purpose problem-solver would, according to Carruthers be very slow and unwieldly in relation to domain specific competitors. Thus Carruthers claims that the most important argument in support of MM is that if amodular or holistic processes are computationally intractible, then the mind must consist wholly or largely of modular systems. However the MM hypothesis must explain the distinctively human capacity for linking together in thought, items of information from widely disparate domains. Carruthers (2002a, p70-73) defines thoughts as “discrete, semantically available, causally effective states, possessing component structure, where those structures bear systematic relations to the structure of other related thoughts”. He 1.2. MODULARITY 27 quotes Dennett (1991) as arguing that human cognitive powers were utterly transformed, following the appearance of natural language, as the mind became colonised by memes (ideas or concepts, which are transmitted, retained, and selected in a manner supposedly analogous to genes. For the resolution of the dilemna, Carruthers (2005, p82-87) has proposed a model in which the language faculty serves as the organ of intermodular communication, making it possible to combine contents across modular domains. The language faculty would need to have access to the outputs of any other central/conceptual belief or desire forming modules, in order that those contents may be expressed in speech. He believes that language is ideally placed to combine thoughts, in that natural language syntax allows for multiple embedding of adjectives and phrases. Carruthers points out that modular input and output systems have substantial back-projecting neural pathways that make possible different forms of sensory and motor imagery, which are processed as if they were percepts.

Assuming that the same is true for language, then sentences formulated by the production subsystem could be displayed in auditory or motor imagination, hence become available to the comprehension sub-system that feeds off perceptual inputs, and, via that to all of the various central-process modules (Carruthers, 2005, p87). According to Carruthers, cycles of activity would then become possible, where in response to perceptual or linguistic input, the central modules generate a variety of domain-specific outputs. “These are made available to the language faculty, which com- bines some of them into a sentence that is displayed in imagination, processed by the comprehension subsystem, and made available to the central modules once again”. Thus new non- domain specific ideas and beliefs might be generated. Carruthers describes distinctively human capacities such as creativity, fan- tasy and pretense, as well as insight, which radically extend inputs. Carruthers (2005, p87) points out that the question of how 28 CHAPTER 1. INTRODUCTION some central-modular outputs rather than others are selected for encoding into language would require the existence of some sort of general problem-solving executive system. The selection problem might be solved by a salience selection process, while the executive problem would be addressed in terms of a variety of attentional systems. Thus Carruthers suggests a combination of three ideas (1) cycles of linguistic activity in inner speech (2) the use of mental modules in speech comprehension and (3) access of the practical reasoning facility to perceptual inputs as a basis for action. Karmiloff-Smith (1994) endorses some aspects of Fodor’s thesis, but questions that modules are pre-specified in detail and also Fodor’s dichotomy between modules and central processing. She argues that a crucial aspect of development involves going “beyond modularity” (Karmiloff-Smith, 1992). She argues that contrary to Piaget (1955), development rarely involves stage- like domain-general change, but rather domain-specific predispositions, which give development a small but significant kick- start, by focusing the infant’s attention on proprietary inputs. She implies the possibility of multiple levels at which knowledge is stored and accessible. Thus a fundamental aspect of human development consists of the process by which information in a cognitive system becomes progressively more explicit. Karmiloff- Smith (1994) calls this process “Representational Redescription” (RR model). Karmiloff-Smith (1994) argues for separate domains rather than modules, where information already present in the organism’s independently functioning, special-purpose representations is progressively available via redescriptive processes to other parts of the cognitive system, first within and then at times across domains. Karmiloff-Smith’s emphasis on Representational Redescription and reiteration implies that increasing abstraction is necessary for central coherence (Karmiloff-Smith, 1994, p693- 745). Here, the views of Lamme (2006, p12-18) are useful. An image, according to Lamme, is processed through successive lev- 1.2. MODULARITY 29 els of visual cortex, by means of feed-forward connections. “From the very first action potentials that are fired, neurons exhibit complex tuning properties such as selectivity for motion, depth, colour or shape, and even selectivity to faces”. However despite the rapid extraction of complex and meaningful features from a visual scene, many studies in humans and monkeys indicate that no matter what area of the brain is reached by the feed-forward sweep, this in itself does not produce (reportable) conscious experience.

What seems necessary for conscious experience is that neurons in visual areas engage in so-called recurrent (or re-entrant or reso- nant) processing, where high- and low-level areas interact. This would enable the widespread exchange of information between areas processing different attributes of the visual scene, and thus support perceptual grouping. In addition when recurrent interactions span the entire sensori-motor hierarchy, or involve the fronto- parietal areas, this would form the neural equivalent of task set, attention etc (Lamme, 2006, p12-18).

Thus, Lamme claims that recurrent processing is essential for conscious experience (Pascual-Leone and Walsh, 2001). Lamme’s approach to consciousness is important for the present argument. In particular, his emphasis on the importance of re-entry or recurrent neural systems 2 is central for the investigation of neural circuits involved in childhood symptomatology, as well as possibly providing a mechanism for helping to solve the MM debate, in terms of a central global workspace, where modular information is integrated via an ‘on-line’ recurrent process. Stahl (2008, p223-245) comments that “in order to understand and treat circuit disorders it is necessary to (1) Identify the circuits involved (2) Understand their normal function and (3) Un- derstand how their function is affected in a given disorder”.

2A circuit which re-enters the cortex and terminates very near the same region from which it originated. 30 CHAPTER 1. INTRODUCTION

1.2.1 Comment

The argument about modularity of central systems raises the question of central executive function, and how this relates to modular input systems. In particular the MM hypothesis appears to refer to the capacity of a central executive to integrate diverse unique modular inputs. The question of how this is accomplished remains controversial, but integration is central to present issues. Karmiloff-Smith (1994) pointed out that while Fodor (1985) held that the mind/brain is made up of genetically specified independently functioning, special-purpose modules (or input systems), he considered central processing to be general-purpose and influenced by what the system already knew, and therefore to be relatively encapsulated, slow, controlled, and often conscious and influenced by global representational format (Karmiloff-Smith, 1994, p693-745). This relationship between subcortical modular systems vs central higher-order processing is central to the arguments in the present thesis, in which it is argued that knowledge of the development and functions of frontal-subcortical circuits will provide a structural basis for a biological understanding of childhood neuropsychiatric syndromes. Higher-order cognition is not thought to reside in the cortex, but rather in the interaction between cortex and subcortical centers. The process of reiteration (or re-entry) and sequencing of sub-movements, or sub-elements of thought process via cortical-basal ganglia-cortical circuits is a fundamental argument. Optimal child development is believed dependant on the developing relationship between basal ganglia, thalamus and cortex, including both affective and cognitive circuits. Bechtel (2008, p165-167) points out that Fodor and Pylyshyn (1988) argued that cognitive theories were committed to a “language of thought”, which requires productivity and systematicity. Fodor and Pylyshyn maintained that representations in connectionist networks were insufficient for modelling cognitive activity. “A compositional syntax builds complex structures, while preserv- 1.2. MODULARITY 31 ing the structure of its components”(Fodor and Pylyshyn, 1988). Fodor and Pylyshyn (1988) maintained that representations in connectionist networks failed to exhibit productivity and systematicity. On the other hand, connectionists such as Elman (1993) have developed techniques for analysing structures implicit in the weights of a trained network that enables it to perform a task without involving symbolic representations” (Bechtel, 2008). Bechtel (2008, p192-193) discusses representation as a general feature of control systems. In the present context, it is likely that language greatly extends human representational capacity, and thus control. For present purposes, symptomatology may result from suggested pathology in lower level modular systems such as tic disorders, or may result from deficits in integration of distributed systems, which diminish conscious monitoring of behaviour. Con- scious intentions are made possible by maintenance and updating in working memory via the “global workspace” (Dehaene et al., 1998b), until an outcome is achieved. Goals may be long-term, such as getting through medical school, or a short-term immediate task. A corollary of the present view is that where working memory is impaired or fails to adequately develop or be maintained, the control of behaviour is vulnerable to capture by non- goal-oriented environmental stimuli, and repetition of localised modular behaviours with resultant distractibility and/or rigidity. Thus language and working memory are key early developments.

1.2.2 Dual processing circuits

According to Koziol and Budding (2009, p14-16) most behavioural processing can be categorised as stimulus-based, and higher- order control respectively. Stimulus-based control is thought to consist of habits, skills and procedures, which allows for high speed of reaction. However, when problem-solving and planning are required, higher-order processing is required. Koziol and Budding (2009, p14-16) describe higher-order control as allow- 32 CHAPTER 1. INTRODUCTION ing autonomy by breaking a problem down into stimulus-based characteristics, and thus programming goal-directed activity. The authors believe the fronto-striatal system has evolved to operate both stimulus-based control when this is advantageous, and higher-order control when automatic processing is not sufﬁcient, and that both systems interact with each other. Thus while procedural learning and memory lies at the heart of automatic responding, a major evolutionary trend is the progressive involvement of the cortex in the processing of the thalamic sensory information, projected from cortex to basal ganglia to thalamus and back to cortex, via parallel segregated circuits. “As the neocortex dramatically increased in size and complexity throughout evolution, the basal ganglia are thought to have become more specialised, receiving specialised information from cortex, and sending information back to cortex via thalamic relays” (Koziol and Budding, 2009, p364-365). Koziol and Budding (2009, p364-365) postulate that higher- level functions of the cortex are expressed through lower level systems, and these lower level systems are critical in modulating cortical functions.

Cortical and subcortical functions operate in concert in deciding upon behaviours that are in the best interests of the organism, which ensure survival. This is because the basal ganglia and cerebellum play important roles in deciding what information is or is not used by the cortex. (Koziol and Budding, 2009, p364)

The authors describe a stimulus-based processing system, composed of innate behaviour processes and behaviours that run on the basis of acquired associations, allowing for automatic processing when appropriate. The second system of higher-order control is thought to allow for novel problem solving, thus providing ﬂex- ibility in interacting with changing and unpredictable environ- ments. Lieberman (2008, p527-528) points out that local neuronal populations project to an anatomically distinct neural population in another part of the brain, forming a neural circuit. “ 1.2. MODULARITY 33

[...] within a given neural structure, distinct anatomically segregated neuronal populations may occur, that project to different brain structures, forming multiple circuits that regulate other behaviours”. Lieberman (2008) also describes the process of ‘reiteration’, whereby the basal ganglia

[...] enable humans to generate a potentially unbounded number of motor acts, such as words or dances, by selecting and sequencing a ﬁnite set of pattern generators, that each specify a sub-movement, as for example in walking. [...] The basal ganglia also reiterate cognitive pattern generators, that constitute sub-elements of thought processes. [...] In short,the basal ganglia act in concert with cortical areas of the brain, to reiterate pattern generation in motor, cognitive and linguistic tasks (Lieberman, 2008, p527-528).

While most of the neocortex was known to project to basal ganglia, but not hypothalamus, a distinction between ‘limbic’ (emotional) vs ‘basal ganglia’ (motor/extrapyramidal) systems was common in the second half of the last century (Heimer et al., 2008, p20). However, Heimer and colleagues believe that the description of ventral striatal pallidum, and extended amygdala, and the relationship of these areas to the cortical mantle has provided the basis of a neuropsychiatric systems approach to psychopathology. According to Heimer et al. (2008, p50-51), the discovery that the ventral striatopallidal system projects to the mediodorsal thalamus marked the beginning of the notion of parallel cortical- subcortical re-entrant circuits. Cummings (1993) (based on the work of Alexander et al. (1986); Alexander and Crutcher (1990)), described five basal ganglia circuits, involved in motor control, cognition and attention. “Seemingly unrelated disruptions in behaviour such as obsessive- compulsive disorder, schizophrenia and Parkinson’s disease are thought to derive from impairment of these neural circuits” (Cum- mings, 1993). The prototypic structure of all brain circuits is described by Cummings as originating in the frontal lobes, projecting to striatal structures (caudate, putamen, and ventral striatum, with connections from striatum to globus pallidus and substantia nigra, and projections from these two structures to specific 34 CHAPTER 1. INTRODUCTION thalamic nuclei, and a final link back to frontal lobe.

Within each of these circuits, there are two pathways: (1) a direct pathway linking the striatum and the globus pallidus interna/substantia nigra complex and (2) an indirect pathway projecting from striatum to globus pallidus externa, then to subthalamic nucleus, and back to globus pallidus interna/substantia nigra. Both direct and indirect circuits project to the thalamus (Cum- mings, 1993).

Cummings (1993) describes the brain circuits as progressively focussing into a smaller number of neurons as they pass from cortical to subcortical structures, but maintaining circuit segregation project to striatum, from which output is funnelled to limited frontal areas.

Figure 1.2: Revised Alexander Model, Lichter and Cummings, 2001.Copyright Guilford Press. Reprinted with permission of The Guil- ford Press.

1.3 Neural network models

The Parallel Distributed Processing (PDP) paradigm, which was introduced in the 1980s, mainly by Rumelhart and McClelland (1986) offers an analogy to model the processes in brain architecture. PDP models rely by definition on computational models that are based on artificial neural networks comprising large 1.3. NEURAL NETWORK MODELS 35 numbers of interconnected units, that represent either individual neurons or functional groups of neurons. Harnad (1990) describes connectionism as “dynamic patterns of activity in a multilayered network of nodes of nodes or units with weighted positive and negative interconnections” The strengths of the inter-unit connections are individually modifiable in a way that is analogous to the changes to the synaptic strength in response to activation, following Hebb (1949). Several principal network architectures with distinct properties have been used successfully not only as models of cognitive functions but also as models for brain structures. According to Smolensky (1988), connections carrying positive weights ar called excitatory and those carrying negative weights are inhibitory. “In between the input and output units there may be other units, often called hidden units. [...] The computation performed by the network in transferring the input patterns of activity to the output patterns depends on the set of connection strengths” (Smolensky, 1988, p1). Churchland (2002a, p75-77) discusses the concept of recurrent neural networks. Churchland differentiates “feedforward” networks, which give an invariant “one-shot” response from recurrent networks, which:

[...] Represent the changing perceptual world with a continuous sequence of activation patterns at its second layer. [...] For example, information from the higher levels of any network-information, that is the result of somewhat earlier information processing by the network, can be entered as a supplementary context-ﬁxer at the second layer of the network, 3 serving to prime or prejudice that neuronal population’s collective activity in the direction of one or other of its learned perceptual categories. If recurrent or descending pathways are added to the basic feed-forward architecture, we lift ourselves into a new universe of functional and dynamical possibilities [...] A feedforward network gives an invariant, one-shot response to any frozen ‘snapshot’ pattern entered at its sensory layer [...] The recurrent pathways also bestow on the network a welcome form of short-term memory, one that is both topic-sensitive and has a variable decay time. For the second layer is in continuous receipt of a selectively processed ‘digest’ of its own

3Second layer or hidden unit implies one or more units that intervene between external input and behavioural output, allowing the insertion of weights or rules. 36 CHAPTER 1. INTRODUCTION

activity some t milliseconds ago, where t is the time it takes for an axonal message to travel up to the third layer and then back down again to the middle layer[...] With active descending pathways, input from the sensory layer is no longer necessary for the continued activity of the network [...]. Since the network’s behaviour is now a continuous function of both its current perceptual inputs and its current dynamical i.e. activational state, we are looking at a genuine dynamical system with the capacity to display behaviours that are strictly unpredictable, short of possessing accurate information about all of the interacting variables [...]. Thus emerges the spontaneity we expect of and prize in a normal stream of conscious cognitive activity (Churchland, 2002a, p75-77).

Despite this dynamic view, Churchland (2002a, p71-72) places consciousness within the biological “hardware” of the brain, rather than in the “virtual” software, proposed by Dennett.

[...] Those characteristic recurrent pathways are the very computational resources that allow us to recognise a puppy’s gait, a familiar tune, a complex sentence, and a mathematical proof. [...] Which particular structures come to dominate a network’s cognitive life will be a function of which causal processes are perceptually encountered during its learning phase, but the need for a virtual vN machine, in order to achieve this broader form of cognitive ends, has now been lifted. Its existing hardware is equal to the cognitive tasks, he rightly deems important (Churchland, 2002a, p72).

Elman (1990) also describes an interesting dynamic connectionist model of language, time and memory. “Time underlies many interesting human behaviors. Thus, the question of how to represent time in connectionist models is very important. One approach is to represent time implicitly by its effects on processing rather than explicitly (as in a spatial representation)”. El- man’s approach involves the use of recurrent links in order to provide networks with a dynamic memory. “Hidden unit patterns are fed back to themselves; the internal representations which develop thus reflect task demands in the context of prior internal states” (Elman, 1990). Elman reports a set of simulations which range from relatively simple problems, to discovering syntactic/semantic features for words, in which networks are able to learn internal representations which incorporate task demands with memory demands. While he applied his approach to natural 1.3. NEURAL NETWORK MODELS 37 language, Elman points out that the time variation error signal can be used as a clue to temporal structure, perhaps modeling biological networks with recurrent structures. He suggests that language and cognition in general may be usefully understood as the behaviour of a dynamical system. Networks with recurrent connections are able to process temporal relationships in data, because the previous internal states influence the network’s next state. Recurrent networks contain circuits that allow for part of the information at the output of the hidden layers to be fed back. The feedback mechanism provides a rudimentary short-term memory that allows for the modelling of temporal events (Elman, 1990). Thus, the addition of a relatively simple control circuit for the feedback units will allow for modelling the desired effect. The control input that is used to vary the feedback ratio (the amount of feedback signal in the overall signal to the hidden units) can inhibit the memory decay. Provided a sufficiently high proportion of the output signal from the hidden units is fed back, the network will settle into a stable state that is governed by the feedback loop. However, the signals from the input layer to the hidden layer can effectively be sat- urated or swamped by the feedback signal, if the ratio between input and feedback at the hidden layer is strongly biased toward feedback. This saturation (arousal/re-enforcement) cannot then be overcome by changes to the regular inputs alone because the behaviour is no longer governed by the (regular) inputs, giving rise to a recurrent state (Levy and Krebs, 2006). This model is particularly useful in considering recurrent pathological symp- tomatologies discussed below.

1.3.1 Temporal sequencing

According to Churchland and Sejnowski (1999, p119) a crude ﬁrst step in handling a temporal structure with nets is to map a temporal sequence to a spatial sequence. However for a network to be sensitive to what happened in the immediate past, it must 38 CHAPTER 1. INTRODUCTION

Input nodes

Hidden Layer

Output nodes

Figure 1.3: Recurrent Network have a memory of what happened, which can be spatial (eg left means before) or a dynamical memory, in which for example, recurrent loops keep a signal “alive”, or in biological systems, transient changes in synapses to keep a memory alive. According to Churchland and Sejnowski (1999, p121-124), there are many ways of putting feedback into a network, including lateral interactions between units within a layer, feedback from a higher layer to lower layers, or, in the most general case, any unit may have reciprocal connections. Churchland and Sejnowski (1999) describe the capacities of feedback conceived as internal input to (a) incorporate multiple time scales into the processing units, (b) process temporally extended sequences of inputs, (c) generate oscillations and modiﬁable rhythms of varying durations (d) resolve ambiguities such as ﬁgure-ground ambiguities and segmentation ambiguities. Nets with feedback are also called recurrent nets:

External processes and events are extended in time and for a nervous system to successfully recognise and respond may require temporally extended representation; movements required for behavior involve sets of bodily motions sequenced in time; short- term memory is a trick for allowing present access to the recent past, and longer-term memory to the more remote past; learning is adapting to the present on the evolutionary-tied assumption that the future resembles the past (Churchland and Sejnowski, 1999, 1.3. NEURAL NETWORK MODELS 39

p117).

Churchland and Sejnowski (1999, p121-123) describe a recurrent Jordan Network, which produces a sequence of actions. In addition to feed-forward inputs, an additional set of inputs receives copies of the output units. These input units through self connections, preserve a decaying copy of the last several outputs. The Jordan net copies the values of the output units into the input layer, “a special input layer, which also has positive feedback connections onto itself. Long recurrent connections provide the hidden units with information about the recent history of the output layer via the special units. On a given pass the hidden units will get both brand-new external information for the regular input units together with recent history from the special input units” [...] “At a given moment, therefore, a special input unit has available to it information about the current output, but also has a history, albeit temporally decaying history of previous outputs, (for example, in muscle contractions, muscle C can be influenced by what muscle B was doing 20 msec before, and muscle A 20 msec before that, allowing a smooth sequence of behavior). Thus a Jordan net might theoretically have an infinite temporal sequence unless stopped externally”. According to Churchland and Sejnowski (1999, p123-124), the Jordan net cannot be trained on more difficult tasks, because the temporal memory provided by the recurrent net decays in time.

[...] In a more general recurrent network it may be necessary to train even these feedback, as well as feed-forward weights. This may mean that an older input may not be able to exert a significant effect at a more distant time when it is a needed piece of information. [...] such problems occur routinely in language, both spoken and written, where there may be long-range interactions between words, as when a pronoun refers back to a proper name or a verb must agree with its earlier mentioned subject (Churchland and Sejnowski, 1999, p123-124).

Elman (1995) has pointed out that time is the medium in which all our behaviours unfold. “We recognise causality because causes precede effects; we learn that coherent motion over time, 40 CHAPTER 1. INTRODUCTION of points on the retinal array is a good indicator of objecthood and it is difﬁcult to think about phenomena such as language, or goal-directed behaviour, or planning without some way of representing time”. Elman describes the Simple Recurrent Network (SRN) architecture, where hidden units at time t receive external input, and also collateral input from themselves at time t-1. He describes the net input on any given tick of the clock t, as including not only the weighted sum of inputs and the node’s bias, but the weighted sum of the hidden unit vector at the prior time step, and refers to the state space of the system as k-dimensional space, deﬁned by the k hidden units. He describes the hidden and context layers of the SRN:

The context units are used to save the activations of the hidden unit on any time step. Then on the next time step the hidden units are activated not only by the new input, but by the information in the context units, which are just the hidden units’ own activations on the prior time step. [...] The hidden units do not record the input sequence in any vericidal manner Instead, the task of the network is to learn to encode temporal events in a more abstract manner, which allows the network to perform the task at hand” (Elman, 1995). Thus in the typical feed-forward network, hidden units develop representations, which reﬂect the demands of a task being learned, when recurrence is added, the hidden units provide the network with memory (Elman, 1995, p195-223).

Munakata et al. (2003, p83-111) have utilised neural network models to explore the role of the prefrontal cortex in perseveration. They point out that one of the hallmarks of higher intelligence is the ability to act flexibly and adaptively rather than being governed by simple habit. For example, an infant may continue to search a previous location perseveratively for a hidden toy, even when the object is visibly hidden in a new location (described as the A-not-B task). Interestingly infants may experience temporal decoupling of cognitive success at different ages, depending on degree of task difficulty. According to Munakata et al. (2003) the prefrontal cortex plays a critical role in reducing perseveration and supporting flexible behaviour. The authors distinguish between “active” and “latent” memory traces. 1.3. NEURAL NETWORK MODELS 41

In the neural network framework, active traces take the form of sustained activation of network processing units (roughly corresponding to ﬁring of neurons), while latent traces take the form of changes to connection weights between units (roughly corresponding to the efﬁcacy of synapses). Perseveration is described as being based on a competition between latent memory traces for previously relevant information and active memory traces for current information [...] Latent memory traces occur when organisms change their biases towards a stimulus after processing, so that they may respond differently to the stimulus on subsequent presentations (Munakata and Stedron, 2001).

The authors claim that ﬂexible behaviour can be understood in terms of the relative strengths of latent and active memory traces. They have utilised neural network models to investigate a number of A-not-B tasks, which require recall of a hidden object. The network architecture is comprised of two input layers that encode information about the location and identity of objects, an internal representation layer and two output layers for gaze/expectation and reach. The gaze expectation layer updates the activity of its units to every input during the A-not-B task, while the reaching layer responds only to inputs of a stimulus within reaching distance. Each unit in the hidden output layers had a self- recurrent excitatory connection back to itself. The levels of input activity represented the salience of aspects of the stimulus with more salient aspects producing more activity. The network initially showed strong A-not-B error, but strong recurrent weights allowed an older network to maintain an active memory during delay. Thus an older network was better able to hold information about a recent hiding place in mind, rather than falling back to its biases for a previous location. The authors point out that their active-latent account of perseveration shares much with the approaches of Goldman-Rakic (1987b) to theories of working memory, including the importance of working memory in higher-order executive processes. 42 CHAPTER 1. INTRODUCTION

1.3.2 Discussion

While direct neurological plausibility of artificial neural nets can be questioned (Krebs, 1995), the analogy of feedforward vs recurrent processing with stimulus-based vs higher-order cognition is compelling. According to Houk and Wise (1995, p92-103) a long history of connectionist literature (Hopfield, 1982; Rummelhart et al., 1987) has shown that recurrent networks are an excellent architecture for associative memories, perceptual operations and the solution of optimization problems. The ability to respond to graded inputs allows many individual factors to be appropriately weighted; the presence of many neurons takes advantage of the speed of parallel computation, and the availability of many loops of interconnection, facilitates recursive re-adjustments of multiple factors. Houk and Wise (1995) envision that much of the learning that goes on in a distributed modular network may be module specific, due to the topographic specificity of climbing and dopamine fiber inputs to the cerebellum and basal ganglia respectively. On the other hand, learning in the cortical network appears instead to depend mainly on a local Hebbian learning rule. 4 While Dennett and Churchland may disagree on the computational architecture of consciousness, they do agree that the brain consists of massively parallel specialised systems, which are modulated in a serial manner, but they disagree on the mechanism by which this is achieved. For present purposes, the concept of serial ordering of information derived from parallel modules is thought to be a basic principle of conscious brain organisation. Deficits or delays in the accomplishment of cortical integration of modular parallel processes are postulated as a core feature of a number of childhood syndromes discussed below. The role of language is considered central, but not exclusively involved in the ‘mediation’ process, between rapid stimulus-bound modules and representa-

4Learning by groups of neurons that selectively strengthen their synaptic connections when repeated ﬁring of one cell accompanies ﬁring of other cells, after Donald Hebb, 1949. 1.4. CORTICAL-BASAL GANGLIA LOOPS 43 tional control, implying dual cognitive systems.

1.4 Cortical-basal ganglia loops

Koziol and Budding (2009) have outlined the central importance of cortical-basal ganglia circuits in behavioural control. Accord- ing to Koziol and Budding (2009, p34), the striatum constantly receives a broad sampling of sensory data, providing information about object identiﬁcation, object location, and external environmental states, as well as internal information about goals and purposes.

A high input compression ratio lays the anatomical groundwork for the striatum to be able to generalise re the similarity between one context and another similar context. This pattern recognition is implied by the fact that 10,000 or more cortical neurons can project to a single (striatal) spiny cell neuron. At the level of the cortex, numerous “bits and pieces” of information are represented individually. At the level of the striatum, common features of sensory information are extracted and treated as equivalent (Bar-Gad et al., 2003). The basal ganglia are also sensitive to the reward characteristics of the environment, allowing them to function as instrumental learning mechanisms. The Globus Pallidus Interna (GPi) and Substantia Nigra Pars Compacta (SNPc) receive extremely compressed information, consistent with the basal ganglia, acting as a selection mechanism or “gate”, rather than a sensory or motor processor. Through the re-entrant projection system, the basal ganglia affect initiation, but the details of the activity to be executed remain stored or represented in the cortex (Koziol and Budding, 2009, p35-36).

According to Koziol and Budding (2009, p36), the direct pathway projects from striatum to GPi. The medium spiny cells in the striatum are inhibitory. They use GABA as a neurotransmitter, and as they have a very low spontaneous firing rate, they must be stimulated by the cortex. The GPi projects to the thalamus, and the thalamus sends excitatory projections to the cortex, which completes the loop. Neurons in the GPi have a very high spontaneous firing rate, and are thus considered tonically active. Their neurons are almost always firing. This activity comprises the default condition of the GPi. The activity serves to tonically inhibit 44 CHAPTER 1. INTRODUCTION the thalamus, resulting in the thalamus being unable to excite the cortex. In this way, the basal ganglia can be considered to function as a type of “brake” upon the cortex. By stimulating the striatum, the cortex releases the brake. This occurs because cortical activation of the direct pathway causes medium spiny cells of the striatum to inhibit the firing of cells in the GPi. This releases the thalamus from its tonic inhibition. As a result, the thalamus excites the cortex, so that behaviour is released (Koziol and Bud- ding, 2009, p36). Koziol and Budding (2009, p36) describe the indirect pathway as involving inhibitory connections of the striatum to the globus pallidus externa (GPe). The GPe has inhibitory connections to the subthalamic nucleus (STN). The STN has excitatory connections to the internal segment of the Gpi. Therefore the activity of the indirect pathway causes the STN to actually increase the tonic inhibitory activity of the GPi, which suppresses behaviour. The direct and indirect pathways presumably operate in opposite directions and in balance (Koziol and Budding, 2009, p36). Koziol and Budding (2009, p37) describe certain regions of the motor, premotor and frontal eye fields project directly to the STN and form a “hyperdirect” pathway, which increases activity in the GPi. This increases the tonic inhibition in the thalamus, and rapidly enables the organism not to respond, when problem-solving thought is necessary. Thus stimulating the STN applies a brake which is important for impulse control Koziol and Budding (2009, p38). The thalamus is thought to excite the cortex, so that the cortex can engage in sensory-perceptual information processing, projected to the cortices for the programming of motor activity. How- ever, prioritisation of motor behaviour, requires the capacity for inhibition of unrequired behaviours. While the frontal cortices are “an essential participant in the looped architecture of basal ganglia [...] the first region in which massive inhibitory control can be found is within the basal ganglia.

This elevates the basal ganglia as a major player in cognition and executive control. The inhibitory mechanisms of the basal gan- 1.4. CORTICAL-BASAL GANGLIA LOOPS 45

glia challenge the view of cortical supremacy in cognition. The basal ganglia very likely comprised the brain’s ﬁrst executive system, and they continue to heavily contribute to cognitive and behavioural control. In fact it has been proposed that the functions of the cortex are dependant upon the basal ganglia (Heimer et al., 2008). It is through a body of anatomical evidence that a number of researchers in the ﬁeld have re-asserted the role of subcortical functions in the realm of cognition (Koziol and Budding, 2009, p20).

Graybiel and Rauch (2000) have pointed out that different sets of cortico-basal ganglia loops are thought to have specialised functions, depending on the cortical areas participating in the loops. “Thus in Tourette’s Disorder, the ”motor loop” through the putamen is more affected than it is in OCD” [...] “A large part of the frontal cortex receives inputs from the basal ganglia conveyed via the thalamus. These same cortical regions not only project to the basal ganglia (mainly to the striatum) but also project to other brain regions including the thalamus. Cortico-thalamic loops are thought to be critical for cortical functioning. If the basal ganglia form associations among cortical inputs on the basis of context and evaluative signals, and thereby promote automation of selected behaviours, they could relieve the frontal cortex of a substantial computational load in carrying out executive functions” (Graybiel and Rauch, 2000). Balleine and O’Doherty (2010) have described two different learning processes in humans and rodents, one encoding the relationship between actions and their consequences, and a second involving the formation of stimulus-response associations, which govern goal-directed and habitual actions respectively. The authors describe recent research in both humans and rats implicating homologous regions of cortex (medial prefrontal and orbital cortex in humans and dorsal striatum in humans) in goal- directed actions, and in the control of habitual actions. They suggest that these processes may at times compete and at other times co-operate. Balleine et al. (2007) have pointed out that although the involvement of the striatum in the reﬁnement and control of motor movement has long been recognised, recent description 46 CHAPTER 1. INTRODUCTION of discrete frontal cortico-basal ganglia networks in a range of species, has focussed attention on the role of the dorsal striatum in executive function, through integration of sensorimotor, cognitive, and motivational/emotional information, within speciﬁc corticostriatal circuits.

1.4.1 Robotic models

Bechtel (2008, p149-151) suggests that a central idea of mechanistic explanation is that the operation of the parts enables the mechanism as a whole to behave in a speciﬁc way. “But an understanding of the parts alone is not sufﬁcient to understand why the mechanism behaves as it does” [...] “Having shown that entities at two levels (at least) are involved in mechanistic explanation, and that organization is the bridge that relates levels, the question arises how to describe the relations between levels”.

In feedback control it is information directly about the effects of an activity that is fed back through the system to regulate its behaviour. In order for the mind-brain control system to control behaviour, it requires information about various things in its environment. Moreover, animals often need to coordinate their actions with things that are not immediately present, such as a food source that is out of sight. The challenges are still greater when the control system has to take into account changes occurring in what is represented during periods in which it is out of causal contact with the things represented. This requires the system to represent the dynamic activity of something remote and out of causal contact. The systems themselves (as in the Watt governor) would not work if information about the distal circumstances were not being made available within the control system. (Bechtel, 2008, p166).

Gat (1998) has described three-layer architectures as applied to robots. While an analogy only, it is useful to examine successful robot architectures. These, according to Gat, tend to fall into three distinct categories: 1. reactive control alogorithms, which map directly on to activators, with little or no internal state. 2. alogorithms for governing routine sequences of activity, which rely extensively on internal state, but perform no search, and 3. time consuming (relative to the rate of change of the environment) 1.5. DISCUSSION 47 search based alogorithms or planners. Thus the three-layer architecture consists of: a reactive feedback control mechanism; a reactive plan execution mechanism; and a mechanism for performing time-consuming deliberative computations. Gat (1998) describes these as running as separate computational processes. “The sequencer must be able to remember the past and respond conditionally to the current situation, whatever it might be” (Gat, 1998). Thus, according to Gat, some modules require only sensory input, but no internal state, while some require memory of the past and some require both of these, allowing planning for the future.

1.5 Discussion

The emphasis by Dehaene and Goldman-Rakic on the importance of prefrontal cortical integration of modular inputs contrasts with the equally important contribution of subcortical structures to cognition and control of behaviour. Graybiel (1995) has suggested that the striatum is a structure critical for the representation of serial order of both learned and innate movement sequences. She describes indirect, but consistent evidence from clinical studies of patients with Parkinson’s Disease, which suggests that the striatum and its dopamine-containing inputs from the substantia nigra are important for performance of sequential behaviour and for predictive control. “These clinical observations reinforce the notion that the basal ganglia and associated frontal cortical regions may allow forward planning, on which coordinated action sequences depend” (Graybiel, 1995). The description by Dehaene (2007) of only one stage, the central stage, which encounters a bottle-neck in which mental operations are done in series and not in parallel is important in understanding deficits of behavioural control where integration of cortical sequencing either fails or does not develop. The modularity debate appears to revolve around the issue of computationally specifiable processes versus the tractability or 48 CHAPTER 1. INTRODUCTION feasibility of a generalised central information processor. While Carruthers attempts to solve this problem in terms of cycles of linguistic activity and mental modules, Dehaene et al. (1998a) and colleagues attempt a biological solution. They have outlined an important biologically-based theory, which incorporates concepts of modularity with the concept of a Global Workspace, in which conscious operations utilise diverse modular information for both goal completion and future reward prediction, via associative reverberation of long-range connections. Thus the global workspace and consciousness appear closely related. The theory specifies a top-down energy consuming process consciousness, associated with a serial cerebral system of limited capacity, responsible for controlling other mental operations necessary for consciousness, but allowing for on-line bi-directional connectivity with specialised modular subsystems as well as ongoing reverbatory processing in working memory. The above Dehaene (2007); Dehaene et al. (1998a) Global Workspace hypothesis implies an absence of sharp delineation of the workspace system, with a fixed set of brain areas. However, the authors point out that workspace neurons seem to be particularly dense in prefrontal areas, including prefrontal cortices (PFC’s) and anterior cingulate (AC). Thus, the hypothesis departs radically from the notion of a single ‘Cartesian theatre’ in which conscious information is displayed (Dennett, 1991). “In particular, information that is already available within a modular process does not need to be re-represented for a conscious audience: dynamic mobilisation makes it directly available in its original format to all other workspace processes” (Dehaene et al., 1998a). This differs from the RR theory of Karmiloff-Smith (1992), who posits that the linguistic representations built up during infancy and early childhood serve young children for comprehending their native tongue, but are not available as data for metalinguistic reflection. On the other hand Graybiel and Rauch (2000) attach considerable importance to the role of subcortical cortico-thalamic- 1.5. DISCUSSION 49 basal ganglia loops in the organisation of executive functions. Where this fails as a result of cortico-basal ganglia circuit dysfunction, motor compulsions and/or repetitive thoughts may occur. While all three theories imply that consciousness requires higher-level organisation, Karmilloff-Smith relies on ‘recursive re-representation’ which appears to imply increasingly abstract representations at higher levels of consciousness. Dehaene postulates the development of long range, presumably cortical, integration between automatic modules in a ‘global workspace’. However neither theory excludes the other, the former being primarily cognitive, and the latter primarily biological. Dehaene’s workspace concept implies temporary energetic or arousal processes as necessary for the maintenance of conscious attention. Neither Karmiloff-Smith, nor Dehaene make fully clear the actual mechanism by which different information or representations become accessible to each other, but both imply that a degree of accessibility is necessary for consciousness. Dehaene s suggests a necessity for reverberative association (ongoing parallel re-entry process), which requires long-range connectivity, as well as the availability of “bushy dendritic” fusiform cells in the prefrontal cortex. Graybiel (1998) hypotheses that the basal ganglia help to recode cortical and other inputs into a form that allows actions to be released as “chunks” or sequences of behaviour” (Graybiel, 1998). This has relevance for the present exploration of childhood symptomology, where subcortical/cortical integration may be deficient.

The theories of modularity vs central workspace raise important issues in relation to the biological systems by which modular input is integrated to allow goal-directed behaviour. Here the work of Koziol and Budding (2009) and Balleine and O’Doherty (2010) describes a biological system which integrates automatic stimulus-based control when this is advantageous, and higher- order control when this is needed. Koziol and Budding (2009, p203) discuss the operation of “serial ordering (mediated by striatum/basal ganglia gating), and on-line error correction (mediated 50 CHAPTER 1. INTRODUCTION by cerebro-cerebellar circuitry [...] operate in parallel and are essential for accomplishing coordinated adaptive behaviour”. For the present thesis, the concept of ’on-line’, parallel subcortical gating and error correcting processes, helps to solve the questions raised by consciousness and modularity theorists, in relation to the problem of computing central MM. Both the Koziol and Bud- ding (2009) and Graybiel (1998) models imply that at least some, if not many, sequencing functions are carried out subcortically rather than centrally. This has implications for both phenotypic presentations and pharmacological interventions. The present thesis aims to explore cortical-subcortical (top- down and bottom-up) cognitve relationships. The thesis postulates that optimal functioning of the Global Workspace is ideally accomplished during the course of development by both long-range connectivity and experience, and that childhood behavioural syndromes arise from deficits or difficulties in the establishment of optimal control relationships between the cortical workspace and independant modular subsystems. Specifically, it is postulated that when the development of cortical serial processing is impaired as a result of genetic, developmental, or stress- ful environmental factors, regression to modular, automatic and repetetive behaviour occurs. The architecture and neurobiology of parallel cortical-basal ganglia circuits will be discussed below in relation to a number of childhood syndromes. The thesis will examine symptoms of childhood syndromes as they relate to striato-nigral-striatal and thalamo-cortical-thalamic brain circuits. It is postulated that repetitive symptoms observed in a number of childhood syndromes result from pathological or immature fronto-subcortical connectivity (Levy, 2004; Levy and Krebs, 2006), where affect fails to be integrated with higher cognitive functions, and behaviour remains modular and reflexive, giving rise to the behaviours observed in a number of syndromes discussed below. The maturation of cortical-basal ganglia circuits is believed to be fundamental to normal development. Where this fails to occur, a 1.5. DISCUSSION 51 number of childhood or adolescent phenotypes may become manifest, depending on the stage of development involved. It is important to differentiate ‘early’ vs ‘late’ childhood/adolescent syndromes. The former are likely to manifest multiple modular behaviours as in autism, whereas the latter will show some integration of modular behaviours, but may manifest deficits related to a lack of prefrontal cortical inhibitory (or working memory) processes. 52 CHAPTER 1. INTRODUCTION Chapter 2

Evolution, language and social cognition

2.0.1 Introduction

The question of language evolution has relevance to the present thesis both in terms of ‘dual’ childhood/adult language development, and relevance to behavioural syndromes as well as aspects of reading problems and psychopathology. The ’dual’ function model is reﬂected in Smolensky (1988) discussion of the “proper treatment of connectionism” (in a response to Fodor and Pylyshyn (1988)) Smolensky (1988) has discussed a possible connectionist model of native language, which postulates dual language systems.

The competence to represent and process linguistic structures in a native language is competetence of the human intuitive processor; the subsymbolic paradigm assumes that this competence can be modeled in a subconceptual connectionist dynamical system. By combining such linguistic competence with the memory capabilities of connectionist systems, sequential rule interpretation can be implemented (Smolensky, 1988, p12).

a. Knowledge in subsymbolic systems can take two forms, both res- ident in the connections. b. The knowledge used by the conscious rule interpreter lies in connections that reinstate patterns encoding rules: task constraints are coded in context-independent rules and satisﬁed serially.

53 54CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION

c. The knowledge used in intuiitve processing lies in connections that constitute highly context dependent encodings of task constraints that can be satisﬁed with massive parallelism. d. Learning such encodings requires much experience. (Smolensky, 1988, p14).

According to Koziol and Budding (2009, p167-169) specialised sensory processing became distributed within multiple sensory channels, as the cortex expanded in primates and humans. It thus became necessary to develop a system for higher-level information categorisation, such as the semantic categorisation system of language.

As sensory-perceptual systems developed, during the course of evolution, codes had to develop in order to manage the range and characteristics of input, that were increasing exponentially. In humans, the ability to combine and recombine symbols, the capacity to generalise, the ability to generate symbolic representations, and the capacity to detach representations and symbols from the concrete are features of language and cognition, appear to be unique to us (humans). [...] The process of categorisation evolves through language taking something novel and making it routine, thinking providing a link between higher-order control and automatic processing. That which is familiar or routine no longer requires higher-order processing (Koziol and Budding, 2009, p167-168).

2.0.2 Language evolution

Ardila (2009, p161-162) describes three possible evolutionary stages in the development of language: 1. Use of noises or grunts with emotional communicative intent. 2. The development of a proto-language using universal phonemes such as /p/ and /a/. and 3. The linking of words to the development of grammar or a ‘syn- tagmatic’ relationship between words. This requires the development of a ‘higher’ level unit, where one of the words has to be a noun, and the other is usually a verb. According to Ardila, nouns can be created in Stage 2, but verbs require association with an action, involving frontal areas in the association. Ardila (2009, p164) suggests that mirror neurons (discussed further below) in 55

Broca’s area may contribute to the brain organisation of verbs. 1 Jason Brown (Brown, 1977, p10-11) elaborated a structural cognitive model, through which cognition recapitulates a sequence of evolutionary plateaus. He described four more or less arbitary structural levels, each of which has an anatomic and a psychic aspect, the former involving the distribution and physiology of the system, the latter action perception and inner experience. Brown’s model is based on the idea of encephaliza- tion, which refers to a “process of cephalad expansion and rostral migration of function”. Brown describes Mclean’s (MacLean, 1972) heirarchical levels or stages of brain-behavior development, namely ‘reptilian’, limbic (or ‘paleo-mammalian’) and ‘neo- mammalian’ stages. Koziol and Budding (2009, p169-170) believe that gesturing was likely the ﬁrst form of communication in primates, based on categorisation and automaticity, and likely evolved from motor systems. However, they point out that:

[...] gesturing has three limitations as a primary communication system [...] as sensory information capacities expand, it becomes increasingly difﬁcult to represent complex categories and manipulate them through a gesturing system. Second, in gesturing, the receiver of communication needs to be looking at the communica- tor, in order for gesture to be effective. Third, it is obviously impossible to communicate through gesturing if one’s extremities/limbs are occupied with doing something else. [...] Vocal language was the answer to that adaptive pressure. Vocal speech needed to be organised in proper sequence to convey proper meaning. In this way, oral language was very closely tied to motor sequencing systems. Gestures and vocalisation both rely on serial-order processing or motor sequencing. [...] This fact already provides some hint that subcortical brain regions would likely play a role in language (Koziol and Budding, 2009, p169-170).

While the adult human brain is thought to be three to four times larger than that of a chimpanzee, comparative studies have demonstrated a slower developmental process. Thus increased dendritic and synaptic complexity and increased myelinisation in

1Mirror neurons represent a system that matches observed events to similar internally generated actions. 56CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION the prefrontal and occipito-parieto-temporal junction are thought to permit greater processing ability, functional integration, and coherence between specific functional modules such as memory, language, motor skills and executive function (Changeux, 2005). Evolutionary studies point to the importance of the development of the neo-cortex in modern man, possibly allowing further development of symbolic processes. In ‘A Brief History of the Mind’, Calvin (2004, p34-43) outlines the progression from Homo erectus to modern man. He points out that sustained attention is common in the history of predatory animals. The acquisition of upright posture and the reduction in size of canine teeth freed up the hands for tool use, and also allowed a more variable diet. The use of cooperative behaviour and delayed reciprocity, in which each partner risked short-term costs for long-term mutual advantage slowly developed. While the use of the acheulean handaxe and accomplished hunting and gathering enhanced cognition, there was no evidence of increased creativity or art. Ac- cording to Calvin (2004, p54-48), brain size began to overlap with the modern range of brain size by about 400,000 years ago. About this time there is evidence of staged toolmaking and staged food preparation. Unlike ballistic actions which are automatic because feedback is too slow to modify the movement, accurate throwing and hunting staging required a ‘get ready’ ‘get set, go’ mentality. The advent of Homo-sapiens is indicated by findings of skulls with more vertical foreheads than those of Neanderthals, with the face below the eyes being relatively flat, and a protracting chin representing a reduction in bone needing to support the teeth. According to Calvin (2004, p83-104), brain size was increased but communication was still at the ‘primary process’ stage. That is, it lacked much symbolism and was driven by the immediacy of here-and-now needs. Language at that stage resembled protolanguage. Calvin suggests that it is only in the last 50,000 years that archeologists have seen the type of creativity that we associate with cave paintings, sewing needles, decorative carvings, pendants and beads appearing during the last half of the last ice 57 age. The use of symbolism and syntax mark the emergence of both modern man and “language is all about taking a mental model of relationships in your mind - say who did what to whom”. Calvin likens the structural aspects of thought to multistage planning, games with rules that constrain possible moves, chains of logic and structured music. Minsky (2006, 130-131) described six levels of the mind, including instinctive reactions, learned reactions, deliberative thinking, reflective thinking, self-reflective thinking, and self- conscious reflection. Interestingly, he makes the point that higher semantic levels of theory avoid the problem faced, for example, by ‘real’ objects that cannot move with two speeds at once, whereas the use of abstract high level descriptions, which convert lower- level sensory states to abstractions which suppress details that are not relevant. Here the paradox is that semantic abstractions are simpler, and able to simulate reality (Minsky, 2006, p157). Thus changing a pictorial representation of an arch would require changing a great number of pixels, whereas a verbal description from rectangular to triangular can achieve the same purpose with one word, demonstrating the power of language. During child development, the acquisition of speech is thought by Calvin to be crucial for the development of structured thought. “Syntax is the best-studied case of structured thought, one of the candidates for what “paid for the rest, and certainly the earliest one to appear in modern childhood (kids pick one up between 18 and 36 months). Once you have syntax you convey complicated thoughts. And the acquisition of syntax likely tunes up the brain to do other structured tasks” (Calvin, 2004, p86). The capacity to embed sentences within sentences with logical and creative thought is best served by language. Calvin (2004, p49-51) points out that while stuctured thought and higher intellectual functions have allowed the creativity of modern man, some of the more chastening consequences now possible are preparation for war (not just raids), the finding of hidden patterns where none exist, such as astrology and religious cults, and sometimes blind 58CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION alleys of science and politics. Striedter (2004, p436) has pointed out that although mammals comprise only about one-tenth of all vertebrates, they are a reasonably successful class of animals, distinguished mainly by their neocortex. Striedter states that most data indicate that the six-layered mammalian neocortex evolved from a tri-laminar reptilian precursor by adding several cellular layers (Reiner, 1993), adding an auditory processing region (Puelles, 1993) and mod- ifying the trajectory of incoming sensory projections from tangential to radial. He describes expanded neocortical projections, which allowed modern humans to produce more finely controlled movements of the hands, respiratory muscles, eyes, jaws, lips, tongue and vocal cords, prerequisite for the emergence of human language 50,000 to 100,000 years ago. He suggests the lateral prefrontal cortex became disproportionately large in homo- sapiens, increasing the ability of humans to look or point away from salient stimuli and reach around barriers to obtain food, which improved dexterity, and eventually gave rise to symbolic language. The size-related changes in regional proportions allowed major changes in neural connectivity (Striedter, 2004), allowing more widely connected regions to become disproportionately influential. The consequent folding and expansion of prefrontal cortex gave rise to the ring-like structures (Papez circuit) described by Sanides (1970).

2.1 Speech perception

Liberman et al. (1967) described a model of speech code perception, which suggested that encoding of speech occurred via processes that are also involved in its production. Rather than “analysis by synthesis”, they proposed “that overlapping activity of several neural networks - those that supply control signals to the articulators, and those that process incoming neural patterns from the ear- allow information to be correlated by these networks and passed through them in either direction”. The authors 2.1. SPEECH PERCEPTION 59 proposed a series of stage conversions in encoding and decoding from phoneme to sound and back, including sentences, phonemes or sets of distinctive features, motor commands, characteristic muscular gestures in vocal tract, and sounds of speech. Many phonemes were thought to be encoded, so that a single accoustic “cue” carried information in parallel about successive phonemic segments. “A possible model supposes that the encoding occurs below the level of the (invariant) neuromotor commands to the articulatory muscles. The decoder may then identify phonemes by referring the incoming speech sounds to those commands” (Liber- man et al., 1967). Liberman (1974) showed that there is a developmental ordering of first syllable and then phonemic segmentation abilities in young children. McEvoy et al. (1988) have suggested that in young children, echolalia or repetition of speech serves as a primary strategy for the acquisition of language and social skills, prior to the development of normal spontaneous language. The latter ability requires language analytic and segmentation skills, as distinct from a more gestalt-like learning pattern. Liberman pointed out that writing systems based on meaningless units, syllables and phonemes were late developments in the history of written language, and that the alphabetic system, which requires abstraction of the phonemic units of speech, was the last to appear. Thus phonemic segmentation is more difficult than syllable segmentation, but is a crucial literacy skill for a morpho- phonological language such as English. Liberman and Mattingly (1985) subsequently revised the “motor theory” of speech perception. They attributed speech perception to “a dedicated module, or a piece of neural architecture that performs the special computations required to provide central cognitive processes with representations of objects or events belonging to a natural class that is ecologically significant for the organism”. According to the authors there is little or no possibility of awareness of whatever computations are carried on within the module, as the perceptual process it controls is not cognitive. 60CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION

The module has a “shallow” output, consisting only of rigidly de- ﬁnable, domain relevant representations, which are “informationally encapsulated”. The authors postulated that phonetic information is perceived in a biologically distinct system, specialised to detect the intended gestures of the speaker, that are the basis for phonetic categories. “Built into the structures of the module is the unique but lawful relationship between the gestures and the accoustic patterns in which they are variously overlapped” (Liberman and Mattingly, 1985). The authors consider that language differs from other modules, in that the neural representation of the utterance that determines the speakers’ production, ie. speaking and listening are both regulated by the same constraints and the same grammar. Thus “according to the revised theory, phonetic information is perceived in a biologically distinct system, a module, specialised to detect the intended gestures of the speaker”. Central to the motor theory of speech perception is the biological link between perception and production. The authors suggest that the perception- production link is innately speciﬁed, but “the acquisition of one’s native language is a process of losing sensitivity to gestures it does not use” [...] “Because the perceptual process it controls is not cognitive, there is little or no possibility of awareness of whatever computations are carried on within the module (limited central access)” (Liberman and Mattingly, 1985).

2.1.1 Comment

The concept of levels of linguistic abstraction is important in the investigation of executive and language and reading deﬁcits often described in behaviour-disordered children. The processes involved in acqisition of native phonetic and syllabic perceptions are discussed below in relation to birdsong and neural circuits. 2.1. SPEECH PERCEPTION 61

2.1.2 Dual language systems

Most approaches to higher order cognition have taken a hierarchical corticocentric approach. The influence of MacLean (1972)’s hierarchical theory is also apparent in the work of Panksepp (1998), who studied animal emotions. Panksepp (1998, p316) equates higher cerebral functions with neo-cortical development. He outlines basic emotional “energies” as arising from sub-cortical processes, and evolutionary developments, which are involved in neural representations of time and space at higher cerebral levels. He describes how the frontal lobes are capable of anticipating events and generating expectancies and foresights in the world, while people with frontal lobe damage typically perseverate on old strategies and do not plan ahead effectively. “They are susceptible to living in the present moment in a more animal-like state of existence”. On the other hand affective or emotional dynamics find a stronger focus of control within the cingulate cortex. While Panksepp supports the concept of neural circuits, he tends to dichotomise cognitive and emotional functions (Panksepp, 1998, p319). He asks whether downward cognitive controls or upward emotional controls are stronger. Importantly, he states “Although we can employ our emotions with gradients of subtlety that other creatures simply cannot match, even using them for aesthetic or manipulative purposes, we would probably feel very little without the ancient source processes. And when these ancient sources become truly aroused, our cognitive apparatus shifts into fairly narrow grooves of obsessive ideation”. Thus like Edelman (2004), Panksepp (1998) differentiates primary processes based on subcortical brain centers, from higher cognitive functions related to time and space, and postulates important interactions between these dual neural processes. Penn and Povinelli (2007) and Penn et al. (2008) have argued for a pervasive functional discontinuity between human and non- human minds that pervades nearly every domain of cognition, and runs much deeper than even the spectacular scaffolding pro- 62CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION vided by language or culture can explain. The authors point out that while almost all of the cognitive mechanisms shared with non-human animals are at least moderately modular, comparative evidence suggests that similarities and differences between human and non-human relational reasoning is remarkably consistent across every domain of cognition. They suggest it is highly implausable that disparities in each domain are the result of independant, module-specific adaptations, but surmise that a more general super-module may be responsible for augmenting the relational capabilities of all of the cognitive modules inherited from non-human ancestors. The authors believe that one of the most important challenges confronting cognitive scientists is to explain how the manifest functional discontinuity between human and non-human minds could have evolved in a biologically plausible manner. They describe marked differences between human and non-human animal’s performance on sameness and difference tasks, in terms of distinguishing variability, as well as differences in the capacity to generalise a given relation beyond the feature set on which it was orignally trained. According to Penn and colleagues, the differences between human and non-human relational reasoning is remarkably consistent across every domain of cognition, from same-different reasoning and spatial relations, to tool use and theory of mind (TOM). 2 They suggest that this points to a general super-module as being responsible for augmenting the relational capabilities of all the cognitive modules inherited fron non-human ancestors. Unfortunately they believe the two most popular ‘supermodels’ proposed so far, namely TOM and language, are insufficient to explain the comparative evidence. While Penn and colleagues agree that non-human animals do not appear to possess anything remotely resembling a TOM, they point out that it is hard to see how a discontinuity in social-cognitive abilities alone, could explain the profound differences between human and non-human animal’s abilities to reason

2The attribution of mental states to others. 2.1. SPEECH PERCEPTION 63 about higher-order spatial relations.

2.1.3 The Language hypothesis

Penn and colleagues describe language as the oldest and most popular explanation for the wide-ranging disparity between human and non-human animals. The authors believe that while language plays an enormous and crucial role in subserving the differences between human and non-human cognition, they distinguish three versions of the Language hypothesis. These consist of:

1) The hypothesis that verbalised (or imaged) natural sentences are responsible for differences in cognition. 2) The hypothesis that some aspect of human language faculty, is responsible. 3) The hypothesis that communicative and/or cognitive function of language served as the prime mover in the evolution of the uniquely human features of the mind. (Penn et al., 2008).

1) Penn and colleagues point out that while natural languages play an enormous role in ‘extending’ and even ‘rewiring’ the human mind, they suggest there is compelling evidence that the human mind is distinctively human, even in the absence of normal language sentences. For instance they describe the existance of many “remarkable” cases of congenital deaf children, spontaneously inventing gestural languages, with hierarchical and compositional structure, as providing further conﬁrmation that the human mind is “indominantly human”, even in the absence of normal language enculturation. 2) The proposal by Hauser et al. (2002) is thought by Penn and colleagues to be a more plausible variation in the Language only hypothesis. Hauser and colleagues distinguish between the faculty of language in the broad sense (FLB) and the faculty of language in the narrow sense (FLN). FLB encompasses all aspects of our sensory and cognitive systems that go into the production and comprehension of language, including the sensori-motor system, and the conceptual-intentional system. On the other hand, FLN includes the computational mechanisms, speciﬁc to mapping syntactic representations into the systems of phonology and syntax. 64CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION

According to Hauser et al. (2002), the only aspect of human cognition that is qualitatively unique to our species, specific to FLN, is the computational mechanism responsible for recursion 3. However Penn and colleagues argue that there are many aspects of the human conceptual-intentional system that are unique to humans, but are not specifically linguistic, ranging from our ability to reason about hierarchical social relations, to our ability to theorise about unobservable causal mechanisms and mental states. They also point out that Hauser et al. (2002) suggested that recursion evolved first in a non-communicative domain, so the discontinuity must then have begun before language evolved. Carruthers (2005) has argued for a structured langual form (LF), as the medium in which the distintively human capacity for non-domain-specific cross-modular thinking is carried out. The above authors agree that while there are many human abilities that rely on “langua-form” representations, including the ability to reinterpret our own thoughts in a propositional and domain- general fashion, they dispute the implication that, aside from our language faculty, the underlying cognitive architectures of human and non-human minds are fundamentally the same. 3) Adaptive functions of language: While Penn and Povinelli (2007) agree that the ability to reason about higher-order relations provides enormous adaptive advantage, they believe that the communicative function of language may have been only one of a number of factors that pushed the cognitive architecture of our species in a relational direction, and that regardless of which factors contributed most strongly, language is no longer directly and entirely responsible for the functional discontinuity. Penn and Povinelli (2007) found that non-human animals reason solely in terms of:

ﬁrst-order perceptual relations, rather than in terms of the logical, rule-governed, and/or structural aspects of the relations themselves. [...] The comparative evidence suggests that non-human

3The embedding of clauses. 2.1. SPEECH PERCEPTION 65

mental representations are implicitly structured, but non-human animals are incapable of reasoning about the higher order structural relation between relations in a recursive, systematic or productive fashion (Penn and Povinelli, 2007). Penn and colleagues also reviewed computational models of biological cognition. In particular, they discussed the Physical Symbol System (PSS) (Fodor and Pylyshyn, 1988; Pinker and Prince, 1988; Newell, 1980), which claims to provide a computational account for several of the most spectacular aspects of human thought. The PSS hypothesis proposes that mental representations are composed of discrete, symbolic tokens, which can be combined into complex representations by forming syntactically structured relations of various kinds, accounting for our ability to generalise rule-like relations over abstract categorical variables, to reason in an informative coherent fashion, and to use artificial symbols of a natural language in a systematic, recursive and generative manner. Fodor and Pylyshyn (1988) argued that a major distinction between connectionist and classical cognitive architecture is that only classical architecture is committed to a symbol level of representation. However Penn and colleagues point out that there is little consensus among cognitive researchers on what it means for a representation to be symbolic. They also assert that the available comparative evidence suggests that compositionality is a ubiquitous feature of animal cognition. Penn and colleagues posit a “hybrid alternative” to classical theories of cognition, which they call the Relational Reinterpreta- tion (RR) hypothesis. This suggests that human animals alone possess the capability of re-interpreting perceptually grounded first-order representations, in terms of higher-order, role governed, inferentially systematic, explicitly structured relations. Thus the ability to reason about higher-order structural relations in a systematic and productive fashion is a necessary, but not sufficient, condition for the normal development and full realisation of these capabilities in human subjects. They also suggest that it is likely that higher-order relational reasoning belongs to a single ‘super-module’ which is duplicated, re-used, shared or called 66CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION upon by functional modules, associated with each of the other distinctively human cognitive capabilities. They also suggest that recent symbolic-connectionist models of cognition shed new light on the mechanisms that underlie the gap between human and non-human cognition. For example, LISA (Learning and Infer- ence with Schemas and Analogies) implements higher-order relational capababilities, by grafting an additional representational system on to a simpler system of conjunctive representations used for long-term storage.

2.1.4 Comment

The importance for the present thesis is the distinction between “first order” perceptual relations and rule-governed higher- order linguistic functions. The distinction between representational and associative or reflexive processes is central in providing a framework for investigation of the networks involved in conscious processing, and deficits in behavioural control. From a developmental point of view, the achievement of representational capacity is a gradual process, involving language and cognitive development. Thus deficits in early development will have implications for the full development of higher order cognition and self-consciousness. From the point of view of the present thesis, higher-order representational cognition and reflexive stimulus-based reactivity reflect dual function and connectivity. It is likely that the language deficits described in a number of childhood syndromes below such as ADHD and Conduct Dis- order, reflect deficits in the higher-order language capacity, with undue reliance on more automatic language forms. This deficit is even more evident in autistic syndromes, where language, when present, is stilted and repetitive.

2.1.5 Mirror systems

Arbib (2005) hypothesised that a speciﬁc mirror system - the primate mirror system for grasping, evolved into a key component 2.1. SPEECH PERCEPTION 67 of the mechanisms that render the brain ‘language-ready’. “The monkey mirror system for grasping is presumed to allow other monkeys to understand praxic actions and use this understanding as a basis for cooperation, averting a threat. [...] Similarly the monkey’s orofacial gestures register emotional state, and primate vocalisations can also communicate something of the current pri- orities of the monkey”. Arbib points out that the premotor area F5 in monkeys, and Broca’s area in humans contain a “mirror system” active for both execution and observation of manual actions, and that F5 and Broca’s area are homologous brain regions. He suggests that evolutionary changes within and outside the mirror systems may have occurred to equip Homo sapiens with a language-ready brain. Imitation is seen as evolving via a so- called “simple” system, such as that found in chimpanzies (which allows imitation of complex “object oriented” sequences, but only as the result of extensive practise), to a so-called “complex” system found in humans (Arbib, 2005).

Gallese et al. (2004) describe the human mirror neuron system as formed by a cortical network, composed of the rostral part of the inferior parietal lobule, and by the caudal sector of the inferior frontal gyrus, and adjacent part of the premotor cortex. The authors point out that while the presence of an object (the target of the action), appears necessary to activate the mirror neuron system in the monkey, the observation of intransitive (meaningless), and mimed actions, is able to activate the human system (Iacoboni et al., 1999; Buccino et al., 2001; Grèzes et al., 2003). “Thus motor evoked potentials, recorded from the muscles of the observer, are facilitated when an individual observes intransitive as well as transitive actions. Thus, when humans observe someone performing an action, there is concurrent activation of part of the same motor circuits, that are recruited when we ourselves perform that action, besides the activation of various visual areas. Although we do not overtly reproduce the observed action, part of our motor system becomes active ‘as if’ we were executing that very same action that we are observing” (Gallese et al., 2004). 68CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION

According to Gallese et al. (2004) the human insula has been shown, in brain imaging studies, to be activated by disgust, and by the sight of disgusted facial expressions of others (Phillips et al., 1997; Phillips and Young, 1998; Sprengelmeyer et al., 1998; Schienle et al., 2002; Wicker et al., 2003), which did not occur in a patient with bilateral insula damage (Carr et al., 2003). “The human brain is endowed with structures that are active during both first- and third- person experience of actions and emotions” [...] “Thus the understanding of basic aspects of social cognition depends on activation of neural structures normally involved in our own personally experienced actions or emotions. By means of this activation, a bridge is created between others and ourselves”. According to the authors’ view, both first and third person understanding of social behaviour requires activation of the cortical motor, or viscero-motor centers, the outcome of which activates downstream centers to determine a specific behaviour. “When only the cortical centers, decoupled from their peripheral effects, are active, the observed actions or emotions are ‘simulated’ and thereby understood”. Other neural structures thought to be involved in the experience of disgust include the anterior cortex and the basal ganglia. Gallese et al. (2004) point out that there is evidence that the mirror neuron system, in both monkeys and humans, encompasses both transitive (object directed) and intransitive communicative actions (like human silent speech or monkey lip smacking or dog barking). Buccino et al. (2004) showed that observations of communicative mouth actions led to activation of different cortical foci, according to different observed species. Ac- tions belonging to the motor repertoire of the observer (e.g. biting and speech reading) or closely related to it (e.g. monkey’s lip- smacking) are mapped on to the observer’s motor system. Actions that do not belong to this repertoire (e.g. barking) are mapped and henceforth categorised on the basis of their visual properties. According to Gallese et al. (2004), the dichotomy between a direct motor-mediated action understanding, and a cognitive type, 2.2. ONTOGENESIS OF SOCIAL COGNITION 69 based on visual representations is “most likely true for emotion understanding”. They believe that direct mapping may be more ancient in evolutionary terms, and is experience-based, whereas the second type is a cognitive description of an external state of affairs. They suggest that direct viscero-motor mechanisms ‘scaffold’ cognitive description and experience, and that when this is absent or malfunctioning, there remains “only a pale, detached account of the emotions of others”. The authors describe the mirror mechanism as the first “unifying perspective of the neural basis of social cognition”. Gallese (2006) has discussed the implications of his theory of social cognition for a neurophysiological perspective on “intentional attunement” in autism. He points out that in primates, the capacity to understand conspecifics’ behaviours as goal-related provides considerable benefits to individuals, as they can then predict, influence, and manipulate the behaviour of conspecifics. Similarly the traditional view in the cognitive sciences holds that human beings are able to understand the behaviour of others in terms of their mental states. “The capacity for attributing mental states, beliefs and desires to others has been defined as Theory of Mind (TOM) (Premack and Woodruff, 1978). “According to this perspective, social cognition becomes almost synonymous with mind reading abilities” (Gallese, 2006). Gallese (2006) argues that social cognition is not only “social metacognition” (thinking about the contents of someone else’s mind by means of abstract representations), but there is also an experiential dimension of inter personal relationships, which enables a direct grasping of the sense, emotions and sensations experienced by others. Gallese describes this mechanism as “embodied simulation”.

2.2 Ontogenesis of social cognition

Gallese (2006) describes the mutually co-ordinated activities between mothers and infants, including movements, facial expressions and voice intonations as synchronised in time, and shar- 70CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION ing “proto-dialogic” behaviours from 4 months of age. He suggests that the shared “we-centric” space enables the social bootstrapping of cognitive and affective development. The segregation of information into discrete channels (modules) is thought by Gallese to occur concurrently with shared neural networks such as area F5 mirror neurons, which “observe” and execute transitive actions such as grasping, biting or licking. The mirroring mechanism is thought to facilitate social interactions in monkeys (Ferrari et al., 2005). Gallese (2006) believes that mirror mechanisms form the basis of intention detection and action (”intentional attunement”). “The mechanism of intention understanding appears to be rather simple, and depending on which motor chain is activated, the observer is going to activate what, most likely the agent is going to do”. Gallese (2006) speculates that “the statistical detection of what actions most frequently follow the actions, as they are habitually performed or observed in the social environment, can constrain preferential paths, chaining together different motor schemata”.

Gallese (2006) points out that one important difference between humans and monkeys could be the higher level of recursivity attained by the mirror neuron system for actions in our species. “A quantitative difference in computational power and degree of recursivity could produce a qualitative leap in social cognition” (Hauser and Fitch, 2004; Hauser et al., 2002). Inten- tional attunement is thought by Gallese (2006) to enable a direct grasping of the sense of the actions and emotions experienced by others. Thus empathy “entails the capacity to experience what others do experience, while being able to attribute those shared experiences to others and not the self. This is achieved by “intentional attunement”. Gallese describes “embodied simulation” as a speciﬁc mechanism by means of which our brain/body system models its interactions with the world. He suggests that embodied simulation mechanisms may be crucial in the course of the long learning process required to become fully competent in how to use propositional attitudes, like during the repeated exposure of 2.2. ONTOGENESIS OF SOCIAL COGNITION 71 children to the narration of stories. He suggests “embodied simulation” is also at play during language processing (Gallese, 2005).

2.2.1 Language development

2.2.2 The RR model of language development

Karmiloff-Smith (1992, 33-36) describes a cognitive theory of language development, based on Representational Redescription (RR). She points out that a Piagetian explanation of language acquisition would not grant the neonate any innately specified linguistic structures or mechanisms which are preferentially attentive to linguistic input. ”Piagetians maintain that both syntax and semantics are solely products of the general structural organisation of sensorimotor intelligence”. However, according to Mehler and Dupoux (1994), 4-day-old infants are already sensitive to certain characteristics of their native tongue, including cues that correlate with clause boundaries. Thus, some fairly general features about prosodic (and possibly the syntactic structure of human languages) appear, according to Karmiloff-Smith, to be built into the system, or to be learned exceedingly early on the basis of some linguistic predispositions, which interact with the particular environmental input from the child’s native tongue(s). Particular pathways for representing and processing language are selected, and “by puberty the other pathways are lost, and by then the processing of language in a native-like way has become relatively modularised”. Karmiloff-Smith (1992, 47-54) argues that normal development involves considerably more than reaching behavioural mastery. She believes the linguistic representations built up during infancy and early childhood serve young children for comprehending their native tongue, but are not available as data for metalinguistic reflection. “To become flexible and manipulable as data (level-E1 representations) and thus ultimately accessible to metalinguistic reflection, as well as to cross-domain relationships with other aspects of cognition (level-E2/3 representations), the knowl- 72CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION edge embedded implicitly in linguistic procedures (level-1 representations) has to be re-represented. Karmiloff-Smith argues that whereas 3-year-olds represent and process formal word boundaries as such, they seem to know litle if anything explicit about what counts as a word. The RR model posits that 3-year old’s representations of formal word boundaries are in level-1 format. By age 6, a child’s statable knowledge that ’the’ counts as a word is, according to the model, in the E2/3 format. The intervening level, the E1 representational format involves a redescription of information into a format that is accessible to certain tasks outside normal input/output relations, but not yet to metalinguistic explanation. Karmiloff-Smith (1992) claims that while at age 3, children represent formal word boundaries for both open-class and closed-class words, these representations are inaccessible for purposes outside input/output relations, while in level-1 format. Between ages 3 and 5, something occurs internally, such that by age 41/2 years, children can access the represented knowledge and succeed a partially on-line task. The RR model posits that this is possible because the level-1 representations have been redescribed into an accessible E1 format. “ [...] something must again occur internally beyond age 5 and 6 to explain why, by then, children can engage in more consciously accessible theory construction about what words are and access such knowledge in off-line tasks”. This requires further redescription into the E2/3 format. Thus innate specification does not alone explain language acquisition. Karmiloff-Smiyh believes that to understand how linguistic representations become flexible and manipulable (ie open to metalinguistic reflection), we need to involve several levels of representational redescription beyond the semantic and syntactic bootstrapping that leads to behavioural mastery. “This in my view, differentiates human capacities from those of other species” [...] Children do not simply reach efficient usage, they subsequently develop explicit representations which allow them to reflect on the component parts of words and build linguistic theories.” (Karmiloff-Smith, 1992, p47-54). 2.2. ONTOGENESIS OF SOCIAL COGNITION 73

An evolutionary account of language was outlined by Hauser et al. (2002). Their exploration of the evolution of language deﬁned two senses of the faculty of language. The faculty of language-broad-sense (FLB) was thought to include “sensory-motor” and “conceptual-intentional” faculties. These were thought to be shared with other vertebrates. On the other hand, the faculty of language-narrow sense (FLN) was hypothesised as being an exclusive capacity. The authors proposed that “At a minimum, the FLN includes the capacity of recursion”, which allows the capacity to create open-ended generative systems. According to the authors:

Natural languages go beyond purely local structure by including a capacity for recursive embedding of phrases within phrases, which can lead to statistical realities, that are separated by an arbitary number of words or phrases. Such long-distance, hierarchical relationships are found in all natural languages for which at a minimum, a phrase-structure grammar is necessary (Hauser et al., 2002).

Joseph (2000) has outlined a neuropsychologically based description of language development. He states that “human language and the original impetus to vocalise springs forth from roots buried within the depths of the ancient limbic lobes , eg the hypothalamus, amygdala and cingulate gyrus”. He points out that while infants do not have the capacity to meaningfully communicate in grammatical word sentences, they still vocalise, and that these vocalisations are often limbic and emotional in origin. Initially these sounds consist of grunts and sounds indicative of displeasure, but weeks and then months later include pleasure, fear and the separation cry. Joseph (2000) suggests that the normal pattern of maturational development based on myelination and metabolic activity, is that the brain stem and cerebellum begin to develop in advance of the forebrain, which in turn matures in a rostral and parame- dial to lateral arc. According to Joseph (2000), it is only as the medial and lateral amygdala reach an advanced stage of myelination and development, that the fear response and related vocalisations 74CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION emerge. Thus the infant remains basically fearless in the first eight months, allowing social/emotional bond formations. Thus abnormal activity in the amygdala has been shown to give rise to “nightmarish” fear (Hermann and Chambria, 1980). Joseph (2000) describes the “amygdaloid” phase of early infant development as marked by indiscriminate approach and contact seeking, which if thwarted may lead to contact seeking directed at inanimate objects. By 2-3 months of age he describes the amygdala- brainstem pyramidal fibers as well as corticospinal axons as beginning to myelinate, coinciding with a shift in emotional utterances of the infant, which become progressively more complex and prosodic, and increasingly subject to sequencing and segmentation. “The infant begins to coo, goo and babble”. Joseph (2000) distinguishes between early and late babbling, the latter consisting of sequences of CV syllables in which the same consonent is repeated. Increasingly, phonetically varied multisyllables, which increase in frequency until around the sev- enth to tenth month of postnatal development, contributed to by cingulate development. From 12 months, pyramidal/corticospinal tracts as well as the somatomotor areas continue to mature and myelinate, giving rise to increasing neo-cortical control and a “new unique motor skill.” [...] “Once the neocortical speech areas establish hierarchical control, and begin to program the oral- laryngeal motor areas, a new form of (neural-muscular) vocalisation emerges, which appears distinct from its precursors”. “The infant begins to speak its first words”. Thus, Joseph (2000) describes emotional and babbling precursors to language, which appear to develop and are expressed in a similar manner regardless of culture. “As the neocortex develops and the amygdala and cingulate gyrus establish interconnections with the superior temporal and frontal lobe, emotional speech and limbic language is slowly transformed and segmented units of prosodic speech and grammatical utterances appear”. Acord- ing to Joseph (2000) maturity of the right frontal lobe, via the emotional-melodic speech area comes to express emotional and 2.2. ONTOGENESIS OF SOCIAL COGNITION 75 prosodic nuances. By contrast, the left frontal motor and Broca’s expressive speech areas and the left cerebral inferior parietal lobule increasingly, fractionate, punctate and impose temporal sequences onto the stress, pitch and melodic intonational contours of the infant’s speech outputs, producing consonent elements and then words. “Although language is commonly associated with the left hemisphere and is discussed in terms of grammar and vocabulary, the underlying foundations are and remain emotional and have their source in the limbic system”.

It is these same limbic linguistic nuances which enable speakers to convey and to comprehend the connative and contextual implications of what is being said, even when words have been ﬁltered or eliminated. [...] Language is both emotional and grammatically descriptive. A listener is able to comprehend not only the content and grammar of what is said, but the emotion and melody of how it is said - what a speaker feels, even when that speaker is a booing, gooing, babbling infant (Joseph, 2000, p257).

Conway and Christiansen (2001) have described features of sequential learning, in which they include learning fixed sequences, statistical learning and hierarchical structure. The authors suggest that language and sequential learning overlap, not only in the processing of sequential structure, but also in neural mechanisms. They point out that agrammatic aphasics (typically with damage to Broca’s area) have severe problems with the hierarchical structure of sentences, and also with sequential learning. Con- way and Christiansen (2001) define fixed actions as the learning of arbitary fixed sequences. Statistical learning in language refers to the learning of transitional probabilities between syllables in a continuous sequence of auditory material, as in word segmentation. Finally hierarchical structure refers to the organisation of phrases in a hierarchical manner. According to Conway and Christiansen (2001), primates appear capable of encoding, storing and recalling arbitary fixed sequences, consisting of motor actions as well as visual stimuli. On the other hand, there appear to be limitations in primate sequential learning, with a phylogenetic trend, in which humans per- 76CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION form better than apes, and apes perform better than monkeys (Lock and Peters, 1996).

Despite these potential homologues, it is also clear that humans out-perform non-humans on more complex sequential learning tasks - in particular the learning and processing of hierarchically organised temporal sequences sequences. We speculate that this species-speciﬁc difference is an important piece of the language evolution puzzle. Language fundamentally involves hierarchical structure, as the basis for unbounded productivity, which is one of the hallmarks of human communication. The limitations on primate hierarchical learning might thus be one of the key reasons that other primates have not developed language abilities (Con- way and Christiansen, 2001).

2.2.3 Nature nuture

An empirical approach to language derives from the study of infant speech perception. Werker and Tees (1999) point out that newborn infants begin life with a remarkable sensitivity to the accoustic cues that signify different basic elements of speech. By measuring babies’ sucking response to syllables such as /ba/ versus /da/ or /ba/ versus /da/. Eimas et al. (1971) were able to show that infants discriminated consonents most easily that actually occur in most of the world’s languages. Werker and Lalonde (1988) were able to show that Japanese babies were able to hear the distinction between /r/ and /l/, but Japanese adults were unable to hear this distinction. According to Werker and Tees (1999) infants become relatively more sensitive to the phonetic characteristics of the native language, and also to the syllabic context in which that phonetic variation occurs. Thus “the language-general perceptual sensitivities in newborns undergo a a change and become more language-specific in the first year of life” (Werker and Tees, 1999), thus preparing the infant for the ability to understand and speak his/her native language. Thus during the first 14-15 months, infants learn to extract words from the speech stream, and to recognise word forms they have previously heard, and to associate words with objects. According to Werker and Tees (1999) babbling becomes more 2.2. ONTOGENESIS OF SOCIAL COGNITION 77 language-specific by 10 months of age. By that age, both the timing of syllabic productions and the internal patterns of infant babbling begin to match those of native language. Werker and Tees (1999) discuss the nativist and ecological models of language development. The ‘strong’ nativist model posits a special-purpose speech processing model, “evolved to detect and analyse the essential properties of human language”. On the other hand the ecological model suggests that information is “picked up directly” as an accoustic wave form with direct access to the articulators that produce the sounds. The approach is similar to Liberman and Mattingly (1985)’s motor theory in which specialised computational routines analyse phonetic input in terms of the potential mode of production. Werker and Tees (1999) point out that no attempt is made by these approaches to account for the dropping of phonetic detail in the early stages of word learning. Werker and Tees (1999) propose an epigenetic model in which an early over-production of synapses is followed by selective retention of a subset, allowing re-modeling or sculpting of developing perceptual mechanisms. Werker and Tees (1999) suggest that coincident with the decline in non-native consonant (and vowel) discrimination seen by the end of the first year of life, the ability to co-ordinate two sources of information, such as phonetic detail, and position in a word is developed. Computational modeling studies (MacWhin- ney, 1998) have shown that self-organising systems can “jump” to new levels of analysis by the simple accumulation of probalis- tic and frequency information. Werker and Tees (1999) agree that with the establishment of a new level of representation, a discontinuity is produced. Thus while the knowledge of acceptible sound and grammatical patterning in the native language is present by the first birthday or shortly after, the task for the next year of life is to construct a second-order system to effortlessly and efficiently use the medium of speech to map to meaning. Kuhl et al. (2008) have studied the decline in non-native speech perception at the end of the first year of life, accompanied 78CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION by the improvement in native speech perception, and found this to be predictive of later language development. Previous research by Lalonde and Werker (1995) suggested a link between general cognitive abilities and reductions in non-native perception towards the end of the first year. According to Kuhl et al. (2008), language and a ’critical period’ have long been of interest to language scientists (Werker and Tees, 2005; Diamond et al., 1994; Conboy et al., 2008). Liberman’s motor theory (Liberman and Mattingly, 1985) and Eimas feature detection account (Eimas, 1975) are based on linguistic selection experience, while Conboy et al. (2008) suggested that native language perceptual abilities are associated with cognitive control abilities, which play a specific role in the ability to disregard irrelevant phonetic information, while maintaining attention to relevant information. Using a conditioned head-turn test of native and non-native speech sound discrimination and non-linguistic object retrieval tasks, sequencing attention and inhibitory control, the investigators showed that native speech discrimination was positively linked to receptive vocabulary size, but not to cognitive control tasks, whereas non-native speech discrimination was negatively linked to cognitive control scores, but not to vocabulary size. The results suggested specific relationships between the development of native language, speech perception and vocabulary.

Kuhl et al. (2008) point out that studies of the maturation of the human auditory cortex show that between the middle of the first year of life and 3 years of age, there is a maturation of axons entering the deeper cortical layers from the subcortical white matter; and neurofilament-expressing axons appear for the first time in the temporal lobe, with projections to the deep cortical layers of the brain, providing the first highly processed auditory input from the brain stem (Moore and Guan, 2001). According to Kuhl et al. (2008), the temporal coincidence between this cytoarchitectural change and infants’ phonetic learning provides a possible maturational cause of the opening of a critical period for phonetic learning. 2.2. ONTOGENESIS OF SOCIAL COGNITION 79

For the present thesis, the concept of transition between developing brain and native language phonetic ability, as well as the associated concept of a developmental discontinuity to a second- order representational system, provides a possible basis for the understanding of the role of language in the behavioural syndromes described below. Namely, a basic deﬁcit (possibly genetic) in categorical phoneme perception, or a later failure of transition to the native language second-order representational system may predispose the child to further deﬁcits of monitoring systems, related to language and attention. Here, some of the concepts discusssed above in relation to learning in self-organising neural circuits may be important, requiring both appropriate environmental input and relevant perceptual capacity.

2.2.4 Birdsong analogy and neural circuits

A series of studies of juvenile songbirds by Aronov et al. (2008) provide useful insights into forebrain circuits important for vocal babbling. The authors point out that the young brain learns to control muscle coordination by exploration and feedback, allowing self-organisation of spinal reﬂex circuits, and cortical somatosensory maps. At a higher level, juvenile animals are thought to learn the causal relationship between actions and the effects of these actions, “by producing highly variable behaviours such as infant stepping, grasp-like hand movements, early vocalisations, and play” (Aronov et al., 2008). The birdsong data provide a useful analogy for the developmental processes involved in mature neural circuit control of behaviour. The investigation of birdsong provides a number of analogies with human cognitive development. For example, Adret (2004, p303) describes an auditory template theory, which postulates the construction of a complex sound replica, based on a set of both genetic instructions and environmental imitative processes. Mooney (2004, p476) describes birdsong as a rare non-human instance of a culturally transmitted vocal behavior. He suggests 80CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION that the synapse is the fundamental organizational unit of neural networks that motivate behavior and describes comparator, auditory feedback and cancellation mechanisms. A further birdsong analogy is described by Cheng and Durand (2004, p611), who proposed that neural pathways mediating emotional states remained integrated with the vocal control system as forebrain vocal control pathways evolved to support learned communication patterns, implying a dual system.

Vocalisations are therefore both a motor component of an emotional state and can influence emotional state via sensory feedback during vocal production. By acknowledging the importance of emotion in vocal communication, we are proposing that the song system and limbic brain are functionally linked in the production and reception of song (Cheng and Durand, 2004). Aronov et al. (2008) point out that babbling is an early motor behaviour produced by juveniles of vocal mammals and birds. While much of the brain of birds from spinal cord to midbrain reflects an organisation common to most vertebrates, the “higher” brain regions including the forebrain are different. However, there are large nuclear masses, which resemble the mammalian basal ganglia (striatum), and a laminated isocortex (grey matter) is separated from the underlying basal ganglia by a band of myeli- nated axons (white matter) (Zeigler, 2008). Like humans, songbirds are dependant on hearing early in life for successful vocal learning. Birdsong and language both consist of ordered strings of sounds, separated by brief silent intervals. Song syllables are usually grouped together to form phrases or motifs (Doupe and Kuhl, 1999). In zebra finches, babbling (called subsong) occurs roughly from 30 to 45 days post-hatch (dph). ‘Plastic song’ follows, with the gradual appearance of distinctive identifiable, but variable, vocal elements (syllables). According to the authors, plastic song is by 90 dph gradually transformed into highly complex, stereotyped motifs or sequences of syllables that constitute adult song. The premotor circuit for adult song production is believed to consist of the high vocal center (HVC), robust nucleus of the archopal- lium (RA), and brainstem motor nuclei. This “motor pathway” is 2.2. ONTOGENESIS OF SOCIAL COGNITION 81 crucial for generating stereotyped, learned vocalisations, and exhibits firing that is precisely time-locked to the song output. According to Aronov et al. (2008) another circuit, the anterior forebrain pathway (AFP) is homologous to the basal ganglia thalamo-cortical loops in mammals, and projects to RA through a forebrain nucleus, lateral magnocellar nucleus of the nidopal- lium (LMAN). “Although LMAN is not required for singing in adult birds, it is necessary for normal song learning in juveniles, and plays a role in producing song variability in adult and juvenile birds. Aronov et al. (2008) investigated whether primitive subsong vocalisations result from an immature form of adult motor pathway, or whether they are driven by other premotor circuits. They eliminated the HVC bilaterally (important in adult singing) in nine subsong (33-44 dph), and in three additional birds in which they left the HVC intact, but specifically eliminated its projection to RA. After these manipulations, all young birds continued producing largely unaffected subsong. Also 12 older birds in the plastic-song stage (45-73 dph) and 5 adults also sang after HVC elimination, but lost structure and stereotypy and reverted to subsong-like vocalisations. In addition, when HVC was pharmacologically inactivated this reversion was fast and reversible, suggesting an immediate rather than long- term circuit change. The investigators posited three possibilities in relation to subsong: it is entirely produced by the midbrain or brainstem; it is driven by circuitry intrinsic to RA, even in the absence of HVC and LMAN; and third it is driven by or requires inputs from LMAN or RA. They tested these hypotheses by lesions or inactivations of RA and LMAN. The investigators found that RA lesions entirely blocked singing in juvenile birds (n=5, 35-73 dph), indicating that subsong-like vocalisations required descending inputs from forebrain. Similarly, song production was abolished by lesions of the HVC and subsequent inactivation of LMAN (n=5, 51-75 dph), indicating that RA circuitry without its afferent paths was not sufficient to generate singing. In addition, LMAN inactivation 82CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION entirely abolished subsong production in all birds younger than 45 dph. However, LMAN inactivation did not block singing in most older birds (6 of 7 experiments in 5 birds, 45-67 dph). To- gether, the results were believed to show that RA and its inputs from LMAN were necessary for subsong production. The researchers also tested LMAN activity directly by recording from RA-projecting neurons during subsong production in intact birds, and showed that firing in some LMAN neurons correlated with more rapid changes in song structure, as well as exhibiting significant corrrelates to song structure in HVC-lesioned birds. The authors concluded that LMAN and possibly other components of the AFP constitute an essential premotor circuit for the production of early “babbling”. At the same time, the clasical premotor nucleus HVC was not necessary for the generation of subsong. They proposed two premotor pathways in the songbird function, to produce vocalisations at different stages of development. “In young juveniles, the AFP generates poorly structured subsong, whereas in adult birds, the classical HVC-motor pathway generates highly stereotypic motor sequences. These pathways interact in the intermediate song stage to generate structured but variable vocalisations, upon which vocal learning operates” (Aronov et al., 2008). According to Aronov et al. (2008), the transfer of functional dominance from one pathway to another during vocal learning elegantly parallels their anatomical development. HVC does not reach its adult size until the late plastic-song stage; and establishes synapses in RA later than LMAN does.

Song maturation and the decrease in vocal variability have thus been attributed to the strengthening of inputs from HVC and the concurrent weakening of inputs from LMAN. Curiously, although HVC neurons form synapses in RA around the onset of singing (30 to 35 dph) our results show they do not signiﬁcantly contribute to song production in its earliest stage. It is therefore possible that the HVC-to-RA pathway is active during early subsong, but is not yet functionally strong enough to drive singing by itself, or to in- ﬂuence vocalisation in a detectable way (Aronov et al., 2008).

The authors suggest that rather than a “neuronal group se- 2.2. ONTOGENESIS OF SOCIAL COGNITION 83 lection theory” of development (in which early motor behaviours originate in the same circuits that later produce adult behaviour), their ﬁndings suggest that distinct specialised circuits are dedicated to production of highly variable juvenile behaviour. That is, juvenile singing is driven by a circuit, distinct from that which produces adult behaviour.

2.2.5 Comment

The hypothesis by Aronov and colleagues that distinct subcortical circuits for the production of infant behaviour may be a general feature of developmental learning in the vertebrate brain, is important for the present thesis. The childhood to adolescent development and integration of cortical/subcortical behavioural circuits involved in infant behaviours such as babbling, free play, and activity is fundamental to the present concept of cognitive development. Like song maturation, the mechanisms by which this integration is established may involve pre-linguistic and linguistic stages. Deficits in this process are believed fundamental to a number of childhood behavioural syndromes, which also appear to involve integration between reflexive and sequential systems. The studies which demonstrate “fixed-state” vs “hierarchical” phrase structure grammar, the latter mainly limited to humans, as well as the analogies with children’s language learning, are important in the present context. This duality is is likely to be reflected in the language deficits found in a number of childhood behavioural syndromes.

2.2.6 Genes, circuits and behaviour

Marcus (2004, p115-124) points out that at a molecular genetic level, many of the neural signals that we use are nearly a billion years old and found even in bacteria. The basic pattern of the brain is, according to Marcus shaped in essentially the same way in every vertebrate via regulatory genes, driven in part by possibilities stemming from random gene duplications. Marcus (2004, 84CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION p111) quotes Jared Diamond’s observation that “a zoologist from Outer Space would immediately classify us as just a third species of chimpanzee. “Molecular genetic studies have shown that we share over 98 percent of our genetic program with the other two chimps”. Like many other biologists and linguists he believes that language is a key to what makes our species unique. Marcus (2004, p128) describes the numerous hypotheses of many eminent scientists: the aquatic ape hypothesis; the language gesture theory; the theory that language arose from the neural machinery evolved to control our muscles; the theory that language is an extension of our capacity for representing space; the theory that language evolved as a means of courtship and sexual display. Marcus’s observation (p132) that the ingredients and genes that make up our brains are, like the ingredients that make up the rest of our bodies the product of evolution, places both language and consciousness in an evolutionary perspective. Accord- ing to Marcus, a language module may depend on a few dozen or a few hundred evolutionarily novel genes, but is also likely to depend heavily on genes, or duplicates of pre-existing genes-that are involved in the construction of other cognitive systems, such as the motor control system, which coordinates muscular action or the cognitive systems that plan complex events. “Understand- ing language will be a matter of not just understanding unique bits of neural structure, but also a matter of understanding how those unique structures interact with other structures that are shared across the primate order”. For example, Marcus speculates that the ability to learn the rules of grammar may depend on the circuitry for short-term memory, ie the circuitry for recognising sequences and automatising repeated actions common to all primates, and additional “special circuitry” for constructing hierarchical tree structures unique to humans. The recursive capacity to combine simple elements into more complex ones that can, in turn, serve as elements in further combinations is, according to Marcus (2004), central to the construction of syntactic trees, a unique feature of human language 2.2. ONTOGENESIS OF SOCIAL COGNITION 85

(p138). Marcus argues that unlike the theory (Jackendoff, 2002) of small symbolic and semantic steps leading to the development of modern language, language may have developed relatively quickly by a novel combination of existing elements such as neural structures for memory, and the automisation of repeated actions and social cognition. Marcus (2004) believes that we cannot simply point to a particular spot in the brain and say that it is “the language” or “consciousness” area, rather that modules should be re-thought, in light of evolution. He believes that evolution proceeds from that which is already in place, and often involves novel combinations of pre-existing components. For example, the neural systems supporting fruit ﬂy courtship have properties which might resemble a mental module. Neurons involved in snifﬁng females, and wing movements have many functions, but a relatively small number of neurons may supervise those actions uniquely involved in courtship. Similarly, the evolution of higher cognitive functions is likely to be built on the hierarchical integration of previously evolved functions.

2.2.7 Cognitive function of language

Carruthers (2002a) has explored the thesis that natural language is involved in human thinking. He distinguishes “weak” views that language is necessary for the acquisition of many human concepts and can serve to scaffold human thought processes, vs “strong” forms that language is conceptually necessary for thought, and is the medium of all conceptual thinking. He proposes that language may be the medium of conscious propositional thinking, and the medium for non-domain-speciﬁc thinking. The latter proposition is relevant for the present thesis. Car- ruthers suggests that natural language syntax is crucially necessary for inter-modular integration. That is “non-domain-speciﬁc thoughts” implicate representations in what Chomsky (Chomsky, 1995) calls logical form (LF). Carruthers claims that only natural language syntax-with its associated recursive and hierarchical 86CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION structures, compositionality, and generativity- could play a role in inter-modular cognitive integration. He points out that while there is evidence from cognitive archeology that early humans were “smart”, they appeared incapable of bringing together information across different cognitive domains. Thus although they made stone tools, they did not use such tools for specialised purposes. There are, according to Carruthers (2002a) also no signs of the linking of artefacts as social signals. According to Carruthers (2002a) language, unlike perceptual modules, is both an input and an output module.

Its production sub-system must be capable of receiving outputs from conceptual modules in order to transform their creations into speech. And its comprehension subsystem must be capable of transforming heard speech into a format capable for processing by those same conceptual modules (Carruthers, 2002a).

According to Carruthers, when LF representations built by the production subsystem are used to generate phonological representations in inner speech, those representations will be con- sumed by the comprehension sub-system, and made available to central systems, such as the theory of mind (TOM) module. Carruthers believes that perceptual and imagistic states become phenomenally conscious by virtue of their availability to the higher-order thoughts generated by the TOM system (i.e. thoughts ABOUT these perceptual and imagistic states. “Thus consciousness involves availability to “higher-order thought.” [...] “Specifically, language is the vehicle of non-modular, non-domain- specific conceptual thinking, which integrates the results of modular thinking” (Carruthers, 2002a). Carruthers sees this as the unique distinguishing feature of human cognition. Carruthers (2002a) presents evidence from studies of rats (Cheng, 1986), which demonstrated that rats were incapable of integrating geometric and non-geometric information in locating a food source. Also Hermer and Spelke (1994) showed a similar phenomenon in pre-linguistic human children, who appeared incapable of integrating geometrical with non-geometrical properties, when 2.2. ONTOGENESIS OF SOCIAL COGNITION 87 searching for a previously seen, but now hidden object, whereas older children and adults were able to solve the task without difficulty. While there were a number of criticisms in response to Car- ruthers’ hypothesis, three are most relevant to the current investigation. Wynn and Coolidge (2002), examine the role of working memory in skilled and conceptual thought. The authors suggested that Carruthers’ review of the various hypotheses of the cognitive functions of language would have been aided considerably by reference to Baddeley’s model of working memory (Baddeley, 1986); (Baddeley, 2001). They pointed out that most current models of working memory suggest that some coordination of separate cognitive domains takes place independent of language. Wynn and Coolidge (2002) referred to the phonological store of working memory, which holds words in mind during the construction and comprehension of sentences, and the visuo-spatial sketchpad (VSSP), which holds visual and spatial information within immediate attention. While the latter is important to thinking, it is largely non-verbal, and could be argued by Carruthers as not being conceptual thought. Varley and Siegal (2002) pose a criticsm of Carruthers’ hypothesis in terms of evidence from adult aphasic patients. They point out that their subjects were capable of utilising both land- marks and geometric cues in locating a hidden object, suggesting that spatial cognition is spared. However Varley and Siegal (2002) point out that an impairment of public language does not prove the absence at deep levels of LF representations. However this is not testable via public utterances, and a crucial element of Carruthers’ model “must be whether it generates testable hypotheses” (Varley and Siegal, 2002). On the other hand, Bickerton (2002) criticises Carruthers as barely mentioning the brain, and failing to reference any of the many brain-imaging studies of the last couple of decades. He suggests that the concept of a language module “busily in- putting, outputting, splitting and cohering material from many 88CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION quasi-autonomous cognitive modules” creates a needlessly complex model. Rather, Bickerton proposes that “its semantic component, present in protolanguage, provides a layer of representations, additional to those found in the brains of other species, whereas its syntactic component, unique to true language, builds from those secondary representations, the complex cognitive constructs underlying our equally unique behavioural plasticity”. Carruthers (2002b) response points out that his target article was intended as a start on the “central challenge” for any strongly modularist account of human cognition, to account for the distinctive flexibility and creativity of the human mind. “It is cross- modular contents formulated in language and cycles of processing activity involving language in inner speech that are supposed to account for some of that flexibility”(Carruthers, 2002b). A study by Wallentin et al. (2006) incorporates subcortical parallel memory systems for talking about location and age in precuneus, caudate and Broca’s region. The authors point out that language comprehension relies on context. For example, a question like “The man who was standing in front of you a moment ago, was he older than you?” requires complex perceptual experiences of persons occupying changing positions in the environment relative to each other and relative to the observer, as well as retention of the experience in memory. Also working memory must access stored semantic, episodic and pragmatic knowledge, extract appropriate experience from memory and map sentence elements on to the correct elements of the representation. “The power of language lies in its ability to convey such complex representations by simple means (i.e. strings of sounds). Because general words like he/she/it underspecify meaning, they only become comprehensible within the context of discourse, and thus contextual monitoring relies on memory, over both short and long durations”. (Wallentin et al., 2006). According to Wallentin et al. (2006), studies in humans and animals suggest the presence of multiple complementary memory systems, working in the brain by declarative and procedu- 2.2. ONTOGENESIS OF SOCIAL COGNITION 89 ral memory, or in the spatial domain of “place” systems. Accord- ing to the authors, these systems each have a different style of processing and interact by simultaneous parallel influence on behavioural output and by directly influencing each other.

Such interactions can be cooperative (leading to similar behaviours) or competetive (leading to different behaviours). [...] Parietal regions are thought to be especially involved in short-term storage of spatial information. [...] Human memory as compared to visuospatial information is thought to involve separable subsystems (Baddeley, 1986), with short-term storage occurring in posterior (e.g. parietal) neocortical areas, and manipulation and re- hearsal depending on prefrontal areas (such as Broca’s area for maintaining verbal stimuli and language comprehension) (Wal- lentin et al., 2006). Wallentin et al. (2006) describes a complementary type of memory to the explicit or declarative systems, which exists in the dorsal striatum (e.g. caudate nucleus) and supports the learning of slowly modulated skills or habits in both humans and other animals. They also describe a third memory system, proposed to reside in the amygdala (White and McDonald, 2002), where it modulates direct links between individual stimuli and reinforcers. Thus motivational aspects of complex social behaviour involve the amygdala (Adolphs, 2003). Pulvermuller (2002) describes language as organised in discrete, distributed neuron ensembles, that differ in their cortical topographies. He believes that the massive ﬁber tracts of the cortex connect many of its areas directly. He describes sequence detectors for words as well as cell assemblies representing words, and additional neuronal units mediating between word-related neuron populations.

Every arbitarily selected cortical neuron is likely to be linked through a small number of synaptic steps to any other cortical cell, and it is likely that the cortex allows for merging information from different modalities. [...] Word webs can store the information about a word occurrence for several seconds, so that a sequence detector, fed by word webs, can process this information at the time scale relevant for the processing of serial order of words. [...] A storage or memory device or a so-called push-down store may be helpful in processing center embedded (or recursive) sentences. 90CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION

A readout mechanism fully activating, and then deactivating the most strongly primed unit will lead to the generation of the action words in the reverse order of their corresponding act or expressions (Pulvermuller, 2002).

Pulvermuller points out that many cells experience strong and then exponential loss of activity, and thus assemblies activated ﬁrst, will later be at the lowest activity level, whereas the last unit activated will still maintain the highest level of activity. Thus, according to Pulvermuller (2002) complex syntactic phenomena may depend on the activity dynamics of large neuronal populations, to achieve sequencing and embedding of expressions.

2.2.8 Discussion

The discussions raise the fundamental issue of domain-specific vs domain-general adaptations fundamental to language. All authors appear to suggest that what was initially a modular system becomes more accessible to domain general functions over the course of development. While Hauser, Chomsky and Fitch (HCF) apply a comparative approach to vertebrates, Carruther’s concept of “conscious syntax” may also be applied to children compared with adults. For example, language deficits, together with TOM deficits are shown to occur together in autistic children, as discussed below. The concept of a sequentially organised global workspace, which allows for conscious reasoning or syntax, utilising language, suggests that language disorders may be more fundamental in childhood symptomatology than recognised by current classification systems. The capacity to detach representations and symbols from the concrete, and the process of categorisation, which evolves through language, taking something novel and making it routine, (Koziol and Budding, 2009, p167-168), appears to be the unique higher- order feature of language. This capacity may conceivably occur in relation to spatial and/or temporal features of the environment, and by the “trick” of abstract representation and categorisation, these features enable verbal recreation of the environment, and 2.2. ONTOGENESIS OF SOCIAL COGNITION 91 thus greater control. For the present investigation, it is postulated that parallel modular cortical-subcortical systems, which are integrated in a global workspace, provide the basis of both cognitive and behavioural control. Wallentin et al. (2006) postulate a short-term spatial mapping system, a long-term habitual system of generalised knowledge and procedures, a “verbal” short-term memory system, and a system for modulating incentive. The authors suggest that language in the form of verbally cued recall can modulate these systems in differential ways, because of the ability of language to specify context via the most commonly used words such as personal pronouns. However, while higher order cognition appears able to combine representations, particularly in working memory, in order to achieve a goal, it remains unclear whether language is essential for or expands working memory. It will be argued that the phenomenology of a number of childhood syndromes can be understood in terms of the relationship between specialised parallel modular subsystems and a cortical global workspace, in which serial processing of spatio- temporal and linguistic information is integrated, to allow goal- orientation. Symptomatology results from deficits or developmental failure of these integrative developments. A number of childhood syndromes, including Attention Deficit Hyperactivity Dis- order (ADHD), Conduct Disorder and Autism manifest language disabilities. While language and communication deficits are one of the diagnostic criteria for Autism, language and reading problems are also often reported as comorbid disorders in ADHD and Conduct Disorder (Levy et al., 1987; Levy, 1989). Language appears to be important for the developments involved in sequential processing of the child’s environment. Language deficits appear common to a number of childhood syndromes. As suggested by the work of Eimas et al. (1971) and Werker and Tees (1999) and other colleagues above, the language-general perceptual sensitivities in newborns undergo a change and become more language- specific in the first year of life, thus preparing the infant for the 92CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION developments necessary to begin to understand and speak his/her native language. Genetic and/or environmental deficits in these early stages may set the scene for later pathology. Like song maturation, the mechanisms by which control is established are not entirely understood, but deficits in this process are believed fundamental to a number of childhood behavioural syndromes. For example, reading is believed to be a meta-linguistic skill which depends in part on the capacity of early readers to learn sound-symbol rules. Dysphonetic dyslexia refers to individuals who are unable to break words phonetically into their sound parts (Boder and Jarrico, 1982). Levy et al. (1987); Levy (1989); Levy et al. (2009) has found an association of reading problems and more specifically dysphonetic reading problems with both attention deficit disorder and conduct disorder, (Levy et al., 1987; Levy, 1989), suggesting a possible relationship of these behaviour disorders with a failure of development of the early phonetic and language skills described by Werker and Tees (1999).

2.3 Reading, language and neural circuits

Shaywitz and Shaywitz (2008) have reviewed the implications of a dual language theory of reading acquisition. They suggest that in order to break through what they term the fluency barrier, investigators need to re-examine a more than 20-year dogma in reading research, that the generation of the phonological code from print is modular, that is, automatic and not attention demanding or requiring any other cognitive process. While a number of theories of dyslexia (specific reading disability) have been proposed, the phonological theory (Liberman and Mattingly, 1985) has been most influential, though visual, magnocellular (Livingstone et al., 1991) and rapid auditory processing (Tallal, 1980) have also been suggested. According to Shaywitz and Shay- witz (2008), the language system is conceptualised as a hierarchy of component modules (Fodor, 1983), assembled into words by the speaker, and dis-assembled back into phonemes by the listener. 2.3. READING, LANGUAGE AND NEURAL CIRCUITS 93

Shaywitz and Shaywitz (2008) point out that exposing a baby to a natural language environment, results in the development of spoken language without needing to be speciﬁcally taught. On the other hand, written language (evolutionarily only several thou- sand years old (Lawler, 2001) is acquired and must be taught.

Learning to read requires multiple skills, including an awareness that spoken language can be segmented into smaller elements (i.e. phonemes), identifying letters, learning the rules of how print maps onto sound, recognising whole words accurately and rapidly (automatically), acquiring a vocabulary, and extracting meaning from the printed word(s) (Shaywitz and Shaywitz, 2008).

Shaywitz and Shaywitz (2008) describe two routes elaborated by reading researchers for transforming print into speech: a so- called lexical or more direct route, and a sublexical or rule-based pathway (Coltheart et al., 1985; Coltheart and Rastle, 1994). Ac- cording to Coltheart and Rastle (1994), the lexical route relies on the mental (orthographic) lexicon, where representations of printed words are stored; the reader looks up the word in his mental lexicon, and activates representations in his phonological lexicon. On the other hand, the sublexical route relies on the serial mapping, left to right of each letter in a letter/ string to its corresponding sound (assembled phonology). “Thus with a dual route model, the critical process of phonological recoding (translating letter to sound) can occur via two different mechanisms: one a lexical mechanism that addresses the phonological code directly from the stored lexical representations, and the other a sublexical mechanism that assembles the phonological code serially letter- by-letter (Shaywitz and Shaywitz, 2008). The authors suggest that these mechanisms may be represented in the brain by separate, but interrelated left hemisphere neural systems, addressed phonology by a more ventral system and assembled phonology by a more dorsal system.

2.3.1 Neural reading circuits

Shaywitz and Shaywitz (2008) describe neuroanatomical lesions 94CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION in the parietotemporal system as pivotal in mapping the visual percept of the print on to the phonological structures of the language. A second posterior system, demonstrated as a result of posterior cerebral artery thrombosis to be localised in the occipitotemporal area, termed the visual word-form area (VWFA) (De- haene et al., 2005). According to Shaywitz and Shaywitz (2008) debate continues as to whether visual word recognition takes place serially, in a progressive step-by-step approach (Dehaene et al., 2005), or conversely, if the left anterior lateral occipito- temporal system functions as an interface between bottom-up visual form information and top-down semantic and phonological properties (Price and Devlin, 2004). The authors describe a further anterior reading-related neural circuit in Broca’s area, long associated with articulation and thought, important in silent reading (Fiez and Peterson, 1998).

Shaywitz et al. (2007) have used fMRI to study age-related changes in reading in a cross-sectional study of 232 dyslexic and non-impaired boys and girls as they read pseudowords. Their ﬁndings indicated that the neural systems for reading that develop with age in non-impaired readers differ from those which develop in dyslexic readers. The system developed with age in dyslexics differed from that in non-impaired readers, primarily in being more posterior and medial, rather than a more anterior and lateral system.

Shaywitz (2003) compared three groups of young adults, clas- siﬁed as persistently poor readers (PPR), impaired compensated readers (AIR), and non-impaired readers (NI). They found that NI readers demonstrated connectivity between the left occipito- temporal VWFA and the left inferior frontal gyrus (a traditional language area). In contrast they found that PPR subjects demonstrated functional connectivity between the VWFa and right prefrontal areas associated with working memory and memory retrieval. This appeared to conﬁrm dual, initially orthographic and subsequent language-based systems. 2.3. READING, LANGUAGE AND NEURAL CIRCUITS 95

2.3.2 Attention and reading

According to Shaywitz and Shaywitz (2008) the posterior parietal cortex plays an important role in attention, presumably via connections between the posterior parietal cortex and PFC. The authors describe studies which indicated that children with dyslexia exhibited reduced activation in a left hemisphere network, involving the inferior parietal lobule (Cao et al., 2006; Booth et al., 2003; Xu et al., 2001). Nakamura et al. (2006) used visual masking and the disruption of specific components of the reading system by transcranial magnetic stimulation to demonstrate a dorsally placed neural system, involving the left hemisphere inferior parietal lobule and premotor cortex, in print to sound conversion by linking to the word recognition systems in the occipito- temporal cortex. A more ventral system, involving the occipito- temporal, middle temporal and premotor cortex are involved in lexical decision. Nakamura postulated that attentional systems in the PFC activate the more dorsal system, because brain activation triggered by conscious perception of word-like stimuli should be subject to top-down attentional amplification by the prefrontal cortex, providing a distributed activation of the fronto-temporal- parietal network. Further insights into the development of reading skills has been described by Dehaene (2009). Dehaene (2009, 277-293) describes the neural basis of “symmetry perception” as it relates to reading. He points out that one-to-one projections link symmetrical projections of visual areas directly via the corpus callosum and/or anterior commissure, giving rise to perceptual symmetry. However, “ [...] if our visual system made everything it learned symmetrical, we would constantly make mirror errors - we would find it impossible to distinguish “p” from “q” or our left shoe from our right” (Dehaene, 2009, 277). Dehaene (2009, 285) points out that the visual cortex is divided into two functional streams - a ventral pathway which focusses on invariant object recognition (identity, shape and colour), 96CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION and a second visual pathway, the dorsal route through the parietal cortex, which is primarily concerned with space and action (Goodale and Milner, 1992). According to Dehaene, patients with ’mirror blindness’ do not see any difference between mirror shapes. Dehaene suggests that in the early stages of reading, we initially need our dorsal system to distinguish letters such as “b” and “d”.

Progressively, however, our ventral system learns to break with symmetry. It stops considering “b” and “d” as two views of the same object. Ultimately, it assigns them distinct neural populations that cease to generalise across mirror reversals [...] In this way, the occipito-temporal cortex acquires an assymetrical neuronal hierarchy for visual word recognition. Unlike its neighbouring cortical regions for object or face recognition, which continue to generalise over left-right changes, our reading architecture ceases to confuse mirror images (Dehaene, 2009, 288).

Thus according to Dehaene, good decoding skills do not arise from associations between letters and speech sounds alone - letters must also be perceived in their proper orientation, at the appropriate spatial location, and in their correct left-right order. According to Dehaene, integration is limited prior to the full development of the prefrontal cortex.

In the young reader’s brain, collaboration must take place between the ventral visual pathway, which recognises the identity of letters and words, and the dorsal pathway, which codes for their location in space and progresses eye movements and attention. [...] The invention of reading, in particular, did not merely consist of the creation of a set of signs that efﬁciently stimulated the visual cortex. It relied, above all, on association of these signs with auditory, phonological and lexical representations of spoken language. [...] Only human beings invent radically new ways to use their ancient brain processors and string them together to come up with innova- tive rules. Our prefrontal cortex functions like a primitive Turing machine. 4 It operates slowly, makes frequent mistakes, but the novel syntheses it generates can be genuinely creative (Dehaene, 2009; Turing, 1938, 322).

4Turing, 1938,described a theoretical device that manipulates symbols contained on a strip of tape. 2.3. READING, LANGUAGE AND NEURAL CIRCUITS 97

2.3.3 Comment

While Dehaene’s account of ’Reading and the Brain’ is complex it provides some insight into a number of anecdotal clinical observations. For example, the present writer has sometimes been puzzled by the observation of clearly talented 2-year olds, with advanced language skills, who are unable to inhibit prepotent hitting and kicking responses despite being able to clearly ar- ticulate “Lily kicked Tom” or with more complexity, “You cant come to my place”. A language-based inhibition theory might ﬁnd the absence of inhibition despite well-developed language skills somewhat puzzling. However Dehaene’s account of the association of the need for developmental “unlearning” of symmetry, via inhibitory capacity of the PFC, in order to obtain the letter invariance required for reading, (in association with phonological and lexical representations of spoken language), provides an important link between the clinically observed associations of reading problems and a number of childhood behavioural syndromes (Levy et al., 1987; Levy, 1989). It is possible that an incapacity to distinguish “b” from “d” beyond the preschool age group could be an early marker of continuing prefrontal immaturity and syn- chronicity, with consequent reading and behavioral difﬁculties related to lack of inhibitory capacity. 98CHAPTER 2. EVOLUTION, LANGUAGE AND SOCIAL COGNITION Chapter 3

Neural circuits and behaviour control

Crick and Koch (2005) have pointed out that “In biology, if seeking to understand function, it is usually a good idea to study structure”. According to Tau and Peterson (2010), the construct of self- regulation refers to “a set of mental processes responsible for the execution, guiding and monitoring of desired behaviours, while inhibiting inappropriate or disadvantageous responses”. Adults are thought to rely on broad cortical areas such as supplementary motor area, frontal eye fields, anterior cingulate cortex, dorsolateral prefrontal cortex (DLPFC), ventrolateral prefrontal cortex/lateral orbitofrontal cortex, as well as temporal and parietal regions, all of which have connections with striatum in the subcortex (Leung et al., 2000; Peterson et al., 2002). According to Tau and Peterson (2010), findings from fMRI studies suggest that an improving behavioural capacity for cognitive control with ad- vancing age is associated with increasing activation of frontal and striatal circuits (Adleman et al., 2002; Bunge et al., 2002; Marsh et al., 2006). According to Tau and Peterson (2010), both self- regulating and reward systems participate in the resolution of conflicts between the intrinsic value of incentives, and their value within a broader framework of internalised social values. Thus compared with adults, adolescents are thought to be more mo-

99 100CHAPTER 3. NEURAL CIRCUITS AND BEHAVIOUR CONTROL tivated by rewards, and less averse to risks (Arnett, 1992; Stein- berg, 2007). Reduced capacity in fronto-striatal circuits is thought to limit inhibitory control in children with ADHD, who also manifest disturbances in motivational and reward processes related to disorders of ventral fronto-striatal circuits (Marsh et al., 2009). Buzsaki (2006, p30-31) points out that the basic neural circuit capable of control functions is recognisable in all vertebrate brains, small or large. “During the course of evolution, the basic circuit is not fundamentally modiﬁed, but instead multiple parallel circuits, consisting of intermediate and longer chains of neurons, are superimposed on the existing wiring”. According to Buzsaki, the brain is organised in a hierarchy of multiple parallel loops. Intermediate and long-range connections link the various loops in the cerebral cortex. Sensory information passes through the thalamus, which is under the control of neocortical feedback. The hippocampus provides a relatively random synaptic space. The strictly parallel loops in the basal ganglia and cerebellum are mainly inhibitory. The main pathways are genetically determined, but ﬁne-tuning of connections (calibration by the output- input match) is under the supervision of the body, environment, and interaction with other communicating brains. According to Buzsaki (2006, p174) collective behaviour of neurons is established by synchrony. He points out that while the temporal window for single neurons is in tens of millisecond range for single neurons, coalitions of neurons can expand the window of synchronisation from hundreds of milliseconds to many seconds.

3.1 Circuit plans of cortex, cerebellum and basal ganglia

3.1.1 Circuit plan of neocortex

Buzsaki (2006, p363) has described the anatomical design principles in which the scalable, small-world-like architecture of the neocortex can effectively deal with the statistical regularities of 3.1. CIRCUIT PLANS OF CORTEX, CEREBELLUM AND BASAL GANGLIA101 the environment, can combine or bind the various features, and can make calculated decisions on the basis of complex features, such as previous experience and the current state of the network upon which the environmental inputs impinge. “There is general agreement that the neocortex is essential for awareness and that hippocampus-supported episodic memories give rise to the feeling of individuality”. Buzsaki (2006, p62-64) has pointed out that with excitatory cortical connections only, no computation would be possible, because any input would simply recruit all neurons of the cortex into unstructured population bursts. Thus the limiting of excitatory spread and segregation of computation is solved by balanced interactions between the excitatory principal cells and inhibitory interneurons. Buzsaki describes the effect of inhibitory interneurons on excitatory neuron chains.

When an inhibitory interneuron at the beginning of the chain is activated, it will suppress the activity of its target neuron. As a result the three interneurons in the chain will be less suppressed by the second interneuron, so the activity of the third neuron may increase. [...] Networks built from both excitatory and inhibitory elements can self-organise and generate complex properties (Buzsaki, 2006, p62-64).

Buzsaki points out that in a recurrent inhibitory circuit, increased ﬁring of the principal cell elevates the interneuron’s discharge frequency, and the interneurons in turn may decrease the principal cell’s output, similar to the action of a thermostat. On the other hand, in a feed-forward inhibitory conﬁguration, increased discharge of the interneuron, as the primary event, results in decreased activity in the principal cell.

Such simple pairing of excitation and inhibition can increase the temporal precision of ﬁring substantially. This is because depolar- isation of the principal cell, initiated by the excitatory input, is reduced quickly by the repolarising effect of feed-forward inhibition, narrowing the temporal window of discharge probability. Fast coupling of the excitatory and inhibitory inﬂuences can bring about sub-millisecond precision of spike timing. Cooperation of interneurons in the same class can secure the spatio-temporal segregation of principal cells to perform a given function. [...] The co-ordinated inhibition ensures that excitatory 102CHAPTER 3. NEURAL CIRCUITS AND BEHAVIOUR CONTROL

activity recruits the right numbers of neurons in the right temporal window and that excitation spreads in the right direction. [...] The rise time and decay time of inhibitory postsynaptic potentials (IPSP’s) are much faster, and their amplitude is larger than those of excitatory postsynaptic potentials (EPSP’s) 1. This faster kinetics is the main reason why interneurons are so much more efﬁcient in timing the action potentials of pyramidal neurons than are excitatory inputs from other pyramidal cells (Buzsaki, 2006, p176-179).

Buzsaki (2006, p176-179) suggests it is difﬁcult to draw boundaries in the neocortical mantle with its myriads of cortical modules and high density local connectivity, supplied by neurons whose biophysical features do not vary across the cortex. While clustering of long-range connections allows designations of cortical systems as visual, auditory, somatosensory, motor and language-related, a further segregation of neocortical function as well as integration of information across distant regions derives from the thalamus. According to Buzsaki (2006, p205), the thalamus is a large collection of relay nuclei, which are the only source of information for neocortex about the body and surrounding world. Unlike the neocortex, where inhibitory cells are nested within the excitatory networks and adjacent to their targets, most GABAergic interneurons in the thalamus reside in a thick shell surrounding the thalamic chamber, called the reticular nucleus. Buzsaki describes the thalamus as a hub for the neocortex that provides functional shortcuts between the vast areas of the cerebral hemispheres and reduces the synaptic path lengths between the various cortical areas. “Both the excitatory thalamo-cortical and the inhibitory neurons of the reticular nucleus are endowed with various temporal spatial scales” (Buzsaki, 2006, p176-179). Buzsaki describes the inhibitory system as providing a high degree of autonomy for individual principal cells or cell groups.

1IPSP Inhibitory Post-Synaptic Potential, EPSP Excitatory Post-Synaptic Potential. 3.1. CIRCUIT PLANS OF CORTEX, CEREBELLUM AND BASAL GANGLIA103

3.1.2 Circuit plan of the cerebellum

Buzsaki (2006, p363-364) describes the cerebellum as continually monitoring the activity of skeletal muscles, and informing brain areas that regulate these muscles.

The cerebellum or ‘little brain’ has approximately the same number of cells as the rest of the brain combined, yet occupies less than 10 percent of the human skull volume. The reason for such an efﬁ- cient space savings is that the cerebellum is a truly locally organised structure. It receives its main inputs from the cerebral cortex, the basal ganglia, the reticular system and spinal pathways containing sensory inputs from muscles, spindles, tendons and joints. In short, the canonical circuit of the cerebellar cortex is two parallel feedforward inhibitory loops that exert differential and elaborate control over the output deep cerebellar neurons. This anatomical arrangement suggests that cerebellar ‘modules’, which roughly correspond with the extent of parallel ﬁbers, process the incoming inputs locally, but do not need to consult or inform the rest of the cerebellum about locally derived computation (Buzsaki, 2006, p363-364).

According to Buzsaki, the result is the execution of rapid skilled movements at speeds that are much faster than can be controlled by the “conscious” sensory systems.

3.1.3 Circuit plan of the basal ganglia

Buzsaki (2006, p365) describes the cortex-basal ganglia thalamo- cortical pathways as (similar to the cerebellar loop) as having major inhibitory steps in the loop.

The projections from one step to the next are largely toographic, providing two possible scenarios for computation. The ﬁrst possibility is a re-entrant loop, where the pathways remain segregated all the way before returning to the cortical cell groups where they originated. The second possibility is that activity spirals in the loop, so that the return message will address areas different from its origin (Buzsaki, 2006, p365).

According to Buzsaki, in both cases there is little integration between the participants of the separate loops: “another case for parallel processing”. Buzsaki describes both the cerebellar and basal ganglia architectural organisations as strikingly 104CHAPTER 3. NEURAL CIRCUITS AND BEHAVIOUR CONTROL similar, with large numbers of parallel inhibitory loops, which are funnelled back on to a relatively small excitatory hub with widespread projections (deep cerebellar nuclei and ventral thalamic nuclei, respectively). Buzsaki (2006, p369-371), suggests tha- tonly special architectures, such as the cerebral cortex can support spatially widespread oscillations at multiple temporal scales and with consequent self-organised features.

The best guess for the computational role of this arrangement is that the neurons in the loops provide the necessary calculation for precise spike timing for the numerically much smaller number of target neurons in these hubs. [...] the ability of a network to generate sustained or persistent activity is the key for the emergence of conscious experience. If no long-range connectivity exists and/or if the activity cannot persist for a sufﬁcient amount of time, locally generated activity cannot engage distant neurons: therefore integration of a large neural space cannot take place. Regenerative activity requires positive excitatory feedback, a critical ingredi- ent absent in cerebellar and basal ganglia circuits (Buzsaki, 2006, p365).

3.1.4 Basal ganglia and executive functions

Balleine and O’Doherty (2010) suggest that the striatum has a much broader role in the control of executive functions than previously suspected and “appears to be centrally involved in functions long argued to depend solely on the regions of the prefrontal cortex”. They consider evidence for two apparently independent sources of motivational control, mediated by reward and by stimuli that predict reward, and the role of cortico-ventral striatal networks in these functions. According to Balleine and O’Doherty (2010) action control may be governed by Response-Outcome (R- O) and/or Stimulus-Response (S-R) learning. In general, experienced reward determines the performance of goal-directed actions, while ‘Pavlovian’ stimuli that predict rewarding events can enhance action selection. Pavlovian ‘values’ are thought to exert their effects on actions through stimulus, as opposed to outcome control. Rodent studies have implicated prelimbic cortex and its striatal efferents on dorsomedial striatum as a key circuit respon- 3.1. CIRCUIT PLANS OF CORTEX, CEREBELLUM AND BASAL GANGLIA105 sible for goal-directed learning, with a possible homologous area of the anterior caudate nucleus involved in humans (Balleine and O’Doherty, 2010). Balleine and O’Doherty (2010) suggest that current theories indicate that outcome values are established by associating the specific sensory features of outcomes with emotional feedback (Balleine, 2001). They argue that the basolateral amygdala (BLA) has been heavily implicated in learning paradigms that have an incentive component (Balleine and Kilcross, 2006), also involving connections between the amygdala and hypothalamus. According to Balleine and O’Doherty (2010) the BLA projects to a variety of structures in the cortico-basal ganglia network implicated in the control of goal-directed action, such as the prelimbic cortex, mediodorsal thalamus, and dorsomedial striatum. Clear evidence for the presence of outcome valuation signals has been found in human ventromedial PFC (medial orbitofrontal cortex, correlating with magnitude of monetary reward (O’Doherty et al., 2001). On the other hand, Ostlund and Balleine (2007) found that the orbitofrontal cortex in rats has an important role in establishing the predictive validity of Pavlovian cues with regard to specific outcomes. Also activity in ventral striatum was found to correlate specifically with reward prediction errors, whereas activity in medial orbitofrontal cortex was found to correspond more with goal valuation. Thus both outcome itself and stimuli which predict outcome can exert a motivational influence on performance. Balleine and O’Doherty (2010) suggest that the integration of Pavlovian and instrumental learning may be accomplished through a distributed outcome “representation”, involving, on the one hand the outcome as a goal, and on the other hand, the outcome as a stimulus, with which actions become associated (DeWit et al., 2006). “This kind of finding encourages the view that in the or- dinary course of events, stimuli and goals exert complementary control over action selection and initiation respectively” (Balleine and O’Doherty, 2010). The integration of Pavlovian and outcome 106CHAPTER 3. NEURAL CIRCUITS AND BEHAVIOUR CONTROL values may be related to spiraling connections between the striatum and midbrain (Haber et al., 2000), which allow feed-forward propogation of information from limbic to associative and sensorimotor networks. Thus “Pavlovian values could influence the effective teaching signal at one level, while enhancing dopaminergic activity and hence instrumental performance at the next” (Balleine and O’Doherty, 2010; Haber et al., 2000). According to Koziol and Budding (2009, p38-42), certain thalamic loops project back on the striatum. The striatum is thought to receive thalamic projections that originate in brainstem sensorimotor structures , including the superior and inferior colliculi and pedunculopontine nucleus. The output structures of the basal ganglia, globus pallidum interna and striatal nucleus, pars reticulata (GPi and SNpr) also project back to these brainstem nuclei. Koziol and Budding (2009, p39-40) suggest the phylogenetic age of these regions reveals the long-standing evolutionary role of the basal ganglia in attention and behaviour selection. The basal ganglia are seen as an interface between subcortical and cortical systems, suggesting that basal ganglia can act independently, cooperatively, or competitively in influencing behavioural selection. Thus the basal ganglia are involved in selection processes by three distinct pathways: direct, indirect and subthalamic (hyperdirect). “ When a behaviour or response is selected, there are potentially competing aspects of responses, which may release a desired response, and inhibit competing patterns, or result in pathological motor and/or cognitive responses” (Koziol and Bud- ding, 2009).

3.1.5 Comment

Buzsaki’s description of basal ganglia modulation of the neocortex via parallel (inhibitory/excitatory) thalamocortical re-entrant CTC loops provides a neuroanatomical basis for the mechanisms which integrate sequential higher-order functions. The re- entrant basal ganglia/cortical approach outlined by Balleine and 3.1. CIRCUIT PLANS OF CORTEX, CEREBELLUM AND BASAL GANGLIA107

O’Doherty (2010); Koziol and Budding (2009) is useful from the point of view of childhood development and syndromes, in that basal ganglia involvement in automatic and learned behaviour begins early, and further cognitive development is likely to depend on optimal basal ganglia/cortical relationships. From the point of view of childhood behaviour, interpretations based on the mature brain, may not provide a full picture of brain development. Given that cortical connections are not fully mature till late adolescence or early adulthood, Buzsaki’s view would interpret consciousness, deﬁned by higher-order goal orientation, as not fully functional till that stage. Ultimately this view has implications for the question of psychopathological nosology, and whether the stages of early and adolescent development should be regarded as distinct from adult syndromes, or whether a strictly dimensional approach should be employed.

3.1.6 New Anatomy of the Basal Forebrain

The neuro-anatomical work of Heimer (2003) provides a useful understanding of cortical/basal ganglia relationships. He discusses the importance of the ‘new anatomy’ of the basal forebrain. He starts from the work of Alexander et al. (1986), who described parallel basal ganglia-thalamic-cortical circuits, in explaining the phenomenology of neuropsychiatric disorders. A classical dichotomy which reinforced a “limbic vs basal ganglia” or “limbic vs extrapyramidal” view of emotional expression was the concept that limbic forebrain structures projected to the hypothalamus, rather than to basal ganglia, whereas non-limbic cortical areas including most of the isocortex (neocortex) projected to the basal ganglia. The use of selective silver degeneration procedures combined with electron microscopic methods allowed the demonstration of a ventral-striatal-pallidal system, leading to the notion of parallel cortical-subcortical re-entrant circuits (Heimer et al., 2008, p21-28). 108CHAPTER 3. NEURAL CIRCUITS AND BEHAVIOUR CONTROL

Heimer (2003) describes cortical projections from the olfactory cortex and hippocampus (allocortex) and non-isocortical (mesocortical) areas such as the entorhinal, insular and cingulate cortex as well as posterior orbital-medial cortex and temporal pole to the striatum. The ventral striatum (nucleus accumbens, olfactory tubercle, ventral caudate-putamen) receives projections from the entire cortical mantle. According to Heimer,

[...] the realisation that the whole cortical mantle is related to the basal ganglia, has provided a new blueprint of forebrain organisation, and the ventral striatum and ventral pallidum are integral parts of a new theoretical framework for adaptive responding and neuropsychiatric disorders (Heimer, 2003).

Figure 3.1: Isocortical and greater limbic lobe projections. Reprinted with permission from Heimer et al. 2008

Heimer (2003) outlines two major neuro-anatomical systems, which play a major role in the control of behaviour. The ﬁrst are the parallel basal ganglia-thalamo-cortical circuits described above, in particular the ventral-striatal-pallidal system. The second system is the extended amygdala. As noted above the concept that the limbic forebrain system projects to the hypothalamus has been shown to be incorrect. Heimer (1972) showed that the accumbens and striatal areas of the olfactory tubercle receive cortical projections, not only from the olfactory cortex and hippocampus, but also from other parts of the greater limbic lobe. 3.1. CIRCUIT PLANS OF CORTEX, CEREBELLUM AND BASAL GANGLIA109

This led to the description by Alexander et al. (1986); Alexan- der and Crutcher (1990) of cortical-subcortical re-entrant circuits. Alexander et al. (1986) and Alexander and Crutcher (1990) described a series of five parallel segregated frontal/subcortical circuits linking specific regions of the frontal cortex to the striatum, globus pallidus, substantia nigra, and the thalamus, important in adaptive interaction with the environment. Two circuits were thought to influence motor and oculo-motor areas of the cortex, while the remaining three loops include the dorsolateral prefrontal cortex, the lateral orbitofrontal cortex, and the anterior cingulate/medial orbitofrontal cortices, thought to be involved in plannning, working memory, rule-based learning, attention and emotional regulation respectively (Middleton and Strick, 2001). Heimer (2003) characterises the ventral-striatal-pallidal system as a “motive” circuit, which is critical for the initiation and mobilisation of appropriate adaptive reward-guided behaviour (Schultz et al., 2000). Cummings (1993) postulated that a wide range of behavioural alterations, including disorders of executive function, personality changes, mood disturbances, and obsessive compulsive disorder (OCD) could be linked to dysfunction of cortical-subcortical circuits. He pointed out that lesions in circuit-related structures produced a similar behavioural disorder; second, the behavioural syndrome was not commonly seen with lesions in other brain regions; third, specific behavioural markers could be identified for the prefrontal-subcortical circuits. These markers were described as (1) executive dysfunction and motor programming deficits for the dorsolateral prefrontal circuit (2) irritability and disinhibition for the orbitofrontal circuit and (3) apathy for the anterior cingulate circuit. Weingarten and Cummings (2001) described five frontal- subcortical neuronal circuits: the motor, oculomotor, dorsolateral prefrontal, lateral orbitofrontal and anterior cingulate circuits. While the striatum is the input segment of the basal ganglia, there are two routes, direct and indirect from the striatum to 110CHAPTER 3. NEURAL CIRCUITS AND BEHAVIOUR CONTROL globus pallidus (GP), which then utilise gamma-aminobutyric acid (GABA) to connect to the thalamus and complete each fronto- subcortical circuit by connecting to the circuit specific area in frontal cortex using glutamate as its neurotransmitter. While the separation of neurobehavioural, neuropsychiatric and movement syndromes in terms of frontal-subcortical circuit area’s involvement is most specific at the cortical area, the organisation is reca- pitulated at the striatal, pallidal and thalamic levels of the circuits. Executive dysfunction such as poor motor programming, difficulty changing sets, and poor word generation is related to the dorsolateral prefrontal circuit; personality change or disinhibition is related to the lateral orbitofrontal circuit, and apathy with minimal reaction to pain is related to the anterior cingulate circuit, namely a modular circuit architecture.

Haber (2003) describes the basal ganglia as connected with frontal cortex in a series of functional modules that maintain a relatively consistent anatomical and physiological organisation, leading to the above concept of parallel processing of cortical information through segregated BG circuits (Heimer et al., 1982); (Alexander et al., 1986). However, according to Haber (2003), communication across functionally distinct circuits is required for the acquisition of new behaviour-guiding rules. She describes an emeging literature in primates, as well as rodents, which demonstrates pathways by which information from separate cortico- basal-ganglia loops can influence each other. Haber indicates three mechanisms which allow this to occur. First, the dendrites and axons within each structure often cross functional boundaries. A second mechanism is through convergence of terminals from functionally adjacent fields onto progressively smaller basal- ganglia structures (Yelnik, 2002). A third mechanism is via complex non-reciprocal loops or pathways, which provide a directional flow of information between regions. For example, a feed-forward loop by which the shell of the ventral striatum could influence the core through striato-pallido-thalamic pathways has been proposed. Also connections of the subthalamic nucleus to two pallidal 3.1. CIRCUIT PLANS OF CORTEX, CEREBELLUM AND BASAL GANGLIA111 segments are in a position to allow associative regions to influence both limbic and motor areas. Haber and McFarland (2001) have described how the basal ganglia work in concert with the thalamus to execute planned motivated behaviours. The authors describe the basic flow of information through the basal ganglia (BG) as being topographically organised from the cortex, through basal ganglia structures to the thalamus and back to the cortex. Thus within the basal ganglia, the striatum receives cortical input, and projects to the output structures, the globus pallidus (GP) and the substantia nigra reticulata (SNr). The authors describe two output routes to the thalamus: a direct pathway from the striatum to the internal segment of the GP (GPi), and the SNr to the thalamus and an indirect pathway from the external segment of the globus pallidus (Gpe), to the subthalamic nucleus to the GPi, and from the GPi to the thalamus (DeLong, 1990), with both pathways targetting the same thalamic nuclei. The striatal pathways to both direct and indirect pathways are thought to be GABAgergic and inhibitory. Thus, according to Haber and McFarland (2001), the direct pathway serves to reinforce cortically driven behaviour via positive feedback, while the indirect pathway modifies behaviour by inhibiting the positive feedback, allowing a focus on relevant behaviour. The authors describe two conceptual models of information processing from the cortex, through the basal ganglia, thalamus and back to the cortex. The first modular model describes information as “passing through the structures in separate or parallel functional circuits with little or no integration between them”. In the second model, “convergence of circuits allows information to funnel, resulting in an integration across different functional domains”. According to the authors, the role of the thalamus in each of these modules differs. The role in parallel processing is to maintain information flow from and to specific cortical areas. Where convergence occurs, information is funnelled through basal ganglia pathways to the thalamus and cortex to obtain an effective motor outcome through motor regions to the 112CHAPTER 3. NEURAL CIRCUITS AND BEHAVIOUR CONTROL cortex. Haber and McFarland (2001) point out that the frontal cortex is organised according to cytoarchitectonic features, that are also associated with specific functions. Thus, primary, supplementary, premotor and cingulate motor areas differentiate aspects of motor control. Dorsolateral prefrontal association areas are involved in executive functions, particularly working memory and the ability to switch cognitive sets (Levy and Goldman-Rakic, 1999), while the orbital and medial prefrontal areas are involved in emotional behaviour, including visceral and autonomic responses and social and emotional responses (Rolls et al., 1996). Haber and McFar- land (2001) suggest that complex behaviours require not only coordination and execution of movements, but also the motivation and cognitive function to carry them out. The association cortical area described as most closely involved in executive function is thought to be the dorsolateral prefrontal cortex (DLPFC), which projects to the head and body of the caudate nucleus (Se- lemon and Goldman-Rakic, 1985), while the orbital and medial prefrontal (OMPFC) are associated with different aspects of reward, motivational control of learning, and appropriate social and emotional responses. According to Haber and McFarland (2001), the organisation of reciprocal and non-reciprocal connections is not random, but rather form a pattern in which prefrontal areas that are involved in reward-based learning (OFPFC), have a nonreciprocal projection to thalamic areas, reciprocally associated with executive function (DLPFC). Cortico-thalamo-cortical circuits relay information from different basal ganglia loops, forming a feed- forward circuit associated with limbic, cognitive motor planning, to motor execution. Haber and McFarland (2001) suggests the notion of “parallel” information processing relies on the fact that basal ganglia output from the pallidum and substantia nigra, targets distinct thalamic relay nuclei that project to specific motor, premotor and prefrontal cortical areas (Alexander and Crutcher, 1990). However, the authors point out that thalamic relay nuclei 3.1. CIRCUIT PLANS OF CORTEX, CEREBELLUM AND BASAL GANGLIA113 have been found to be actively involved in changing the dynamics of cortical processing by setting up different oscillation patterns of frequency and synchrony, derived from inhibitory and disinhibitory effects from the globus pallidus (GP). (For example, the resting tremor of Parkinson’s disease is thought to result from changes in thalamic firing frequency due to increased inhibition of pallidal input). Not only do thalamocortical pathways project to the cortex, but cortico-thalamic projections are thought to have both reciprocal and non-reciprocal components. The non- reciprocal neurons are large rapidly conducting cells, which may affect a different part of the cortex, providing a mechanism for synchronisation of thalamic oscillations. According to Haber et al. (1995), the ventral striatum is considered an interface between limbic and motor systems. The investigators followed the orbital and medial prefrontal circuit through the monkey basal ganglia, by analysing the projection from orbitofrontal cortex to the ventral striatum, and the representation of orbitofrontal cortex in the globus pallidus and substantia nigra by utilising injections of Lucifer yellow and horseradish peroxidase. Their results indicated that the orbital prefrontal cortex projects primarily to the medial edge of the ventral striatum, and to the core of the nucleus accumbens. They were able to show that the orbital and medial prefrontal cortex is represented in a confined region of the globus pallidus, but throughout an extensive area of the dorsal substantia nigra. Haber et al. (2000) analysed a collection of retrograde and antegrade tracing studies to understand how the striato-nigrostriatal (SNS) directs information flow between ventromedial (limbic), central (associative) and dorsolateral (motor) striatal regions. They showed that when viewed as a whole, the ventromedial striatum projects to a wide range of the dopamine cells, and receives a relatively small dopamine input. In contrast, the dorsolateral striatum (DLS) receives input from a broad expanse of dopamine cells, and has a confined input to the substantia nigra (SN). The central striatum (CS) was found to receive input from 114CHAPTER 3. NEURAL CIRCUITS AND BEHAVIOUR CONTROL and project to a relatively wide range of the SN. Thus,

The results demonstrated an interface between different striatal regions via the midbrain dopamine cells, forming an ascending spiral, where the shell influenced the core, the core influenced the central striatum, and the central striatum influenced the dorsolateral striatum. The anatomical arrangement created a hierarchy of information flow, and provided an anatomical basis for the limbic/cognitive motor interface via the ventral midbrain (Haber et al., 2000).

According to Haber (2003), the premotor and motor areas mediate different aspects of motor behaviour, including planning, learning and execution, which in turn are reﬂected both anatomically and physiologically in the central and lateral caudate nucleus and in the central, dorsal and lateral putamen respectively. The dorsolateral prefrontal cortex (DLPFC) is involved in working memory, set shifting and strategic planning (executive functions), while the orbital prefrontal cortex is involved in reward- based learning and goal-directed behaviours. The lateral orbital regions project to the central and lateral parts of the ventral striatum, while medial orbital areas (OMPFC) project to the medial wall of the caudate and the nucleus accumbens consistent with the important role of the ventral striatum in the development of reward-based learning and in mental health diseases. Haber (2003) believes thalamo-striatal, and striato-pallidal substantia nigra connections are also organised topographically.

Thus the basal ganglia are connected with the frontal cortex in a series of functional modules that maintain a relatively constant anatomical and physiological organisation, leading to the concept of parallel processing of cortical information through segrated basal ganglia circuits. However the modular arrangement is not totally segregated. [...] while projections from cortex terminate in a general topography through the BG structures, the dendrites and axons within each structure often cross functional boundaries. [...] The interface between functional circuits increases with the complexity of interconnections with intrinsic BG circuitry and with the compression of pathways to successively smaller structures. [...] Thus, the thalamic output from each basal ganglia loop, although returning to the cortical area of origin is inﬂuenced by other functional areas of the cortex. [...] Information ﬂows in an ordered fashion from limbic to cognitive to motor outcome (Haber, 2003). 3.1. CIRCUIT PLANS OF CORTEX, CEREBELLUM AND BASAL GANGLIA115

Haber and McFarland (2001) point out that the complexity of the thalamo-cortico-thalamic circuit raises an “interesting” issue related to the concept of parallel versus non-parallel processing of information through basal ganglia pathways. They point out that because there are both reciprocal and non-reciprocal cortico- thalamo-cortical components to the thalamic relay nucleus from basal ganglia output structures, the information that the relay nuclei convey to the cortex is not only affected by the parallel pathway through the basal ganglia structures, but is also modiﬁed by the feed-forward component of the non-reciprocal corticothalamic pathway. Haber (2003) has described two networks, which are involved in developing new learned behaviours from emotional, cognitive and motor cortical areas. These are the striato-nigral-striatal network and the thalamo-cortical-thalamic network described above. “Within each of these sets of connected structures, there are both reciprocal connections linking up regions associated with similar functions, and non-reciprocal connections linking up regions that are associated with different basal ganglia circuits”(Haber and McFarland, 2001). Haber describes the basal ganglia as including the caudate nucleus, putamen and globus pallidus, as well as the more recent concept of the ventral striatum, including the nucleus accumbens, the medial and ventral portions of the caudate and putamen, and the striate cells of the olfactory tubercle. The striatum derives its input from the cerebral cortex, thalamus and brainstem. As described above, the striatum projects to the pallidal complex, and to the substantia nigra, via “direct” cortico-basal ganglia pathways. The above two interconnected networks contain both reciprocal connections, linking up regions connected with similar functions (parallel networks) and non-reciprocal connections, linking up regions that are associated with different cortico-basal circuits.

Parallel circuits and integrated circuits must work together so that coordinated behaviours are maintained and focussed (via parallel circuits), but can also be modiﬁed and changed according to external and internal stimuli (via integrative networks)... Indeed, both the inability to maintain and to focus in the execution of speciﬁc 116CHAPTER 3. NEURAL CIRCUITS AND BEHAVIOUR CONTROL

behaviours, as well as the inability to adapt appropriately to external and internal cues are the key deﬁcits in basal ganglia diseases, which affect aspects of motor control, cognition or motivation. When considered separately, each pathway of the system (striato- nigral or nigro-striatal pathway) creates a loose topographic organisation, demonstrating that the ventral tegmental area (VTA) and medial substantia nigra (SN) are associated with the limbic system, and the lateral and ventral SN are related to the associative and motor striatal regions (Haber and McFarland, 2001).

According to Haber and McFarland (2001), the ascending and descending limb for each functional area of the striatum differ in their proportional projections. The ventral striatum receives a limited midbrain input, but projects to a large region. In contrast, the dorsolateral striatum receives a wide input, but projects to a limited region. In addition, for each striatal region, there is one reciprocal and two non-reciprocal connections with the midbrain. The interface between different striatal regions, via dopamine (DA) cells is thus organised in an ascending spiral, interconnecting different functional regions of the striatum, and creating a feed-forward organisation. Information can thus be channelled from the shell to the core, to the central striatum, and ﬁnally to dorsolateral striatum.

Taken together, the interface between different striatal regions via the midbrain dopamine (DA) cells is organised in an ascending spiral interconnecting different functional regions of the striatum and creating a feed-forward organisation. [...] In this way information ﬂows from limbic to cognitive to motor circuit (Haber and McFar- land, 2001).

Haber (2003) describes two examples of neuronal network systems that extend beyond connecting adjacent regions, the striato- nigro-striatal network and the thalamo-cortico-thalamic interface, which allow limbic pathways to interact with cognitive pathways and which in turn interact with motor pathways. “Based on anatomical criteria, the mid-brain dopamine neurons are divided into two tiers: a dorsal tier and a ventral tier” [...] “There is an inverse dorsal-ventral topographic organization to the midbrain striatal projection. The dorsal and medial dopamine cells project 3.1. CIRCUIT PLANS OF CORTEX, CEREBELLUM AND BASAL GANGLIA117 to the ventral and medial parts of the striatum, while the ventral and lateral cells project to the dorsal and lateral parts of the striatum”. Haber (2003) describes the thalamo-cortical pathway as the last link in the circuit and is often treated as a simple “one- way relay”. However, the pathway plays a key role in regulating cortical assemblies of neurons through its projections to different cortical layers.

Thus, similar to the basal-ganglia, thalamic relay nuclei appear to mediate information flow from higher cortical “association” areas of the PFC to rostral motor areas involved in cognitive or integrative aspects of motor control to primary motor areas that direct motor execution. [...] The development and modification of goal- directed behaviours require continual processing of complex chains of events, which is reflected in the feed-forward organization of both the striato-nigral connections and the thalamo-cortical connections (Haber, 2003).

Haber maintains that parallel and integrative circuits must work together, so that coordinated behaviours are maintained and focussed (via parallel networks), but also can be modiﬁed and changed according to the appropriate external and internal stimuli (via integrative networks).

3.1.7 Comment

Haber’s anatomical work demonstrates an inverse dorsal/ventral and ventral/dorsal striatal structure, which allows coordination of motor pathways. The relationship of the cortical mantle to re- entrant circuits, through the basal ganglia and thalamus, and returning to areas of origin in the cortex, provides an anatomical basis for many of the unconscious recursive processes which modulate action. Haber’s thalamo-cortico-thalamic and striato-pallidal substantia nigra circuits differ from Edelman in being more neuroanatomically deﬁned, and providing a model of both parallel and integrative cognitive processing. Haber’s careful anatomical review updates the Alexander and Crutcher (1990), Middelton and Strick (2002), and Heimer (2003) concepts of parallel (modular) segregated circuits by allowing for lateral anatomically-based 118CHAPTER 3. NEURAL CIRCUITS AND BEHAVIOUR CONTROL subcortical levels of integration. In terms of the present thesis, and the modularity debate, the above architecture provides a basis for both focussed modular circuits and integrated behaviour, both of which, if pathological or developmentally impaired, may give rise to behavioural symptomatology. This anatomical arrangement also provides a basis for syndromal comorbidity where modular integration fails to be coordinated. Haber’s concept of parallel and integrative spiral circuits expands the notion of a Global Workspace from a primarily cortical locus, to a circuit concept which includes striato-thalamo- cortical and cortico-thalamo-striatal connections, and which integrate cognitive and motor behaviour. Failure of integration may underlie comorbid phenotypes (Levy, 2010).

3.2 Oscillation networks

Mann and Paulsen (2009, p21-48) showed that neurons can have intrinsic biophysical properties that are determined by the type and distribution of ion channels, which give them oscillatory properties.

These neurons are tuned to specific frequencies, which not only support the induction of defined frequencies, but also buffer or fil- ter out others. [...] Electrical synaptic coupling is one way to enable the transmission of high frequency oscillations. To control spike timing, specific interneurons appear to control excitatory spiking by inhibiting synaptic activity. Inhibitory and excitatory neurons can be coupled to form feedback loops, which can establish oscillations in a fast and well-controlled fashion. Preferred frequencies can be built into the system by choosing the right parameters of axonal length and synaptic transmission speed (Holscher and Munk, 2009b, p434).

Bland (2009, p287-315) utilised electrical stimulation tec- qniques to show that the origins of ascending brain stem syn- chronising pathways are in the nucleus reticularis pontis oralis (RPO) and pedunculo-pontine tegmental nucleus (PPT). Bland showed that “ purpose-built brain nuclei and projections are in- 3.2. OSCILLATION NETWORKS 119 volved in the induction and control of theta oscillations throughout the brain”. Holscher and Munk (2009c, p151-152) have outlined a “ high frequency oscillation” approach to neuronal information coding and plasticity within brain areas and neuronal populations. They start with the Hebb (1949) postulate that:

(1) Connections between neurons increase in efficiency in proportion to how successful the presynaptic neuron activates the postsynaptic neuron. When an axon of cell A is near enough to excite cell B, and repeatedly or persisistently takes part in firing it, some growth process or metabolic change takes place in one or both cells, such that A’s efficiency as one of the cells firing B is increased (Hebb, 1949, p62). (2) Groups of neurons which tend to fire together from a cell assembly, whose activity can persist after the triggering event, serve to represent information. (3) Cognitive processes are based on the sequential activation of sets of such assemblies. (Holscher and Munk, 2009c, p151-152).

According to Holscher and Munk (2009b, p437-438), Hebb not only proposed a mechanism of how synaptic weights could change in order to store information in previous active cell assemblies, but also proposed the concept of synchrony and simultaneous neuronal activity in populations that encode the same information. The investigators used cross-correlation and auto- correlation analyses of data from a delayed match-to sample task, with a working memory component, to analyse activity of neurons and neuronal ensembles, recorded in primates. Distributed neurons appeared to encode separate parts of the task, including the start, presentation of the stimulus, and presentation of the match. The linking of these parts was shown to be by synchronisation of the gamma frequency range. Thus cells that encoded similar information were synchronised by gamma, while cells that did not encode similar information were not. Holscher and Munk (2009c) concluded that cells or networks are brought together by gamma rhythms and fire either simultaneously or in a series of active assemblies, controlled in time by gamma field oscillations. Looking at hippocampal “place” cells, Holscher and Munk (2009b, p438) showed that these cells are under tight control of theta and 120CHAPTER 3. NEURAL CIRCUITS AND BEHAVIOUR CONTROL gamma activity, and the firing activity is increased when the local field potential is in the disinhibitory (depolarising) phase. A relatively new and interesting biological approach to consciousness has been derived from brain imaging studies of the precuneus (Cavanna, 2007). This approach examines the “De- fault Mode” of brain activity during conscious rest, where cortical areas associated with the highest resting metabolic rates in the conscious resting state have been shown to be located in the postero-medial parietal cortex (posterior cingulate cortex and precuneus). Raichle et al. (2001) showed that during the baseline resting state, a neural network comprising the precuneus and postero-medial parietal region, along with lateral parietal, ventromedial prefrontal parietal region, mid-dorsal prefrontal and anterior temporal cortices show high metabolic activity at rest. This decreases when subjects are engaged in goal-directed cognitive processing or perceptual tasks. Cavanna (2007) also points out that recent functional imaging studies have demonstrated that the precuneus and adjacent posteromedial cortical regions have shown a profound deactivation in pathophysiological altered states of consciousness, such as slow wave sleep and rapid eye movement sleep, the hypnotic state, general anaesthesia and veg- etative states. Positron emission tomography (PET) studies have shown that the precuneus, along with lateral parietal and prefrontal cortices were found to be significantly less active than the rest of the brain during slow wave deep sleep, providing evidence for active participation of the precuneus in conscious processing of high order self-representation. McCormick et al. (2003) have described how local cortical networks in the prefrontal cortex and visual cortex are capable of spontaneously generating sustained activity for periods of seconds or longer, maintained through recurrent excitation between pyramidal cells, and controlled by feedback inhibition. The investigators obtained intracellular and extracellular recordings from layer 5 regular spiking cells in the ferret medial prefrontal cortex, maintained in vitro in an ionic medium that mimicked in vivo 3.2. OSCILLATION NETWORKS 121 conditions. They demonstrated spontaneous “UP” states, which were typically from 0.5 to 3 seconds in duration, as well as sharp changes to “DOWN” states. The “UP” state was associated with a substantial increase in neuronal responsiveness, so that current pulses that were completely subthreshold for the generation of action potentials in the “DOWN” state could generate a strong train of action potentials during the ‘UP’ state. The facilitation of responsiveness depended in part on the amplitude of the current pulse, with larger amplitude pulses less facilitated than smaller amplitude pulses, resulting in a decrease in slope of the input- output relationship during the “UP” state. According to the authors, whether the “UP” state is the slow oscillation equivalent to the maintained membrane potential on low firing rate of the cortex in the waking, but resting state, is as yet not known. They point out that although there are similarities between persistent activity during working memory tasks in vivo, and sustained activity generated in vitro, there are limitations to the analogy.

McCormick et al. (2003) point out that cortical neurons are embedded in a richly interconnected network, and receive tens of thousands of synaptic inputs. According to the authors, many of these synaptic connections arise from local circuits, but also arise from distal cortical regions, including feedback pathways, and presumably provide a mechanism by which the responsiveness of single cortical cells can be modulated by the behavioural and cortical context. McCormick et al. (2003) point out that the spontaneous barrages of synaptic activity asociated with the “UP” state result in an increase in response probability to small inputs, with relatively little change in response to larger inputs. Thus low contrast (i.e. weak) stimuli result in an increased neural response, while high contrast (i.e.strong) stimuli are relatively unaffected. The investigators suggest that attentional mechanisms may increase neuronal responsiveness , both by depolarising some cells with barrages of synaptic activity, while increasing the ‘gain’ (slope of input-output relation) of other cells. The investigators concluded that results suggest that persistent activity associated 122CHAPTER 3. NEURAL CIRCUITS AND BEHAVIOUR CONTROL with the performance of working memory tasks may be generated largely through recurrent networks, and feedback pathways, such as those involved in selective attention, and may exert a powerful inﬂuence on neuronal responsiveness through synaptic bombard- ment.

3.2.1 Comment

The investigation of oscillatory networks represents a relatively new approach to the understanding of information processing by neuronal populations via synchronous neuronal activity, while the work of McCormick and colleagues afﬁrms the importance of recurrent networks and feedback pathways in working memory and selective attention. The modulation of these cortical networks by subcortical inﬂuences suggests increasing complexity as cortical networks develop.

3.3 Temporally structured replay of behaviour

An example of the above complexity is the function of temporally structured replay of neural activity, demonstrated by the study of hippocampal activity in rats trained to run along a circular track for food reinforcement. Following acquisition of the task, electrophysiological activity was monitored during task performance (RUN) and during periods of sleep immediately before and after behaviour. Kenway and Wilson (2001) simultaneously recorded the structure of neural activity during rapid eye movement (REM) sleep in multiple neurons in the rat hippocampus. They were able to show that temporally sequenced ensemble firing rate patterns, reflecting tens of seconds to minutes of behavioural experience are reproduced during REM episodes, at an equivalent time-scale. Within the REM episodes, behaviour- dependent modulation of the subcortically driven theta rhythm was also reproduced. The authors suggested that their results demonstrated that long temporal sequences of patterned multi- 3.4. PREFRONTAL CORTEX: COGNITIVE AND EXECUTIVE FUNCTIONS123 neuronal activity suggestive of episodic memory are reactivated during REM sleep. This work was elaborated by Lee and Wilson (2002), who repeatedly ran rats through a fixed temporal order sequence of spatial receptive place cells, and showed that these neurons fired precisely in this order in long sequences, involving four or more cells during slow wave sleep (SWS), immediately following, but not preceding the experience. Interestingly, the SWS sequences occurred intermittently in brief (approximately 100ms) bursts, each compressing the behavioural sequence in time by approximately 20-fold. Lee and Wilson (2002) suggested that the rapid encoding of sequential experience was consistent with evidence that the hippocampus is crucial for spatial learning in rodents, and possibly the formation of long-term memories of events in time in humans.

3.4 Prefrontal cortex: cognitive and executive functions

Roberts et al. (1998, p221) have described executive functions as “the optimal scheduling of the operation of different components of complex tasks that depend on more dedicated or modular mechanisms”. The authors include ‘supervisory’ functions, appropriate inhibitory mechanisms, selective attention and switching between two or more tasks under this umbrella, as well as the monitoring of processes such as retrieval from long-term memory, and performance of intended actions. While there is general agreement about the tasks involved, there is disagreement about the inter-relationship among these tasks. Goldman-Rakic (1998a, p87-90) reviewed the question of the nature and existence of a “general purpose executive” based on the premise that structure and function are inextricably related. She pointed out that a major principle of prefrontal function since mid-century has been that of a duality between dorsolateral and orbital cortices. Thus orbital lesions particularly posterior or mesopallidal areas have 124CHAPTER 3. NEURAL CIRCUITS AND BEHAVIOUR CONTROL been shown to produce selective impairments on tasks evoking appetitive or emotional resources, while the dorsolateral principal (neopallidum) has been associated with spatial delayed-response tasks. Owen et al. (1996) have advanced the idea of a two-stage hierarchical organisation of prefrontal cortex in which mid-frontal areas carry out sequential and self-monitoring functions, while the inferior convexity areas are engaged in a lower-level function, involving comparison between stimuli in short-term memory, as well as active organisation of sequences of responses based on conscious explicit retrieval of information from posterior cortical association systems. Goldman-Rakic (1987b) reviewed the development of cortical circuitry in relation to cognitive function. In particular, the area surrounding the principal sulcus was shown to be responsible for the delayed response function in human primates. The basic circuitry for this response was thought to include connections with parietal and limbic cortex, and thalamus, and projections to the caudate nucleus, superior colliculus and other pre-motor centres, being all the circuitry required for the components of the delayed response task (visual-spatial input, storage mechanisms and motor commands). Interestingly Goldman-Rakic and her colleagues found that the timing and rate of increase in the formation (and elimination) of synapses was surprisingly similar for each cortical area involved. Synaptic overproduction was highest between 2 and 4 months and was followed by a longer period of synaptic elimination. The developmental progression in the AB task, that occurs in human infants between 7 and a half and 12 months of age, was compressed in infant rhesus monkeys to between 1 and a half and 2 and a half and 4 months of age. The delayed-response task measures the capacity to utilise representational knowledge to guide behaviour. The required “object permanence” needs recognition that an object has continuity in time and space, when not in view, and depends on the capacity to form representations of the outside world, and base responses on those representations in the absence of the objects they rep- 3.4. PREFRONTAL CORTEX: COGNITIVE AND EXECUTIVE FUNCTIONS125 resent. Goldman-Rakic (1987b) emphasised that the capacity to guide behaviour by stored information is not fully mature by 4 months, and continues to grow at a slower rate into adulthood. She points out that linguistic skills depend, among other structures on Broca’s area, which lies in the human prefrontal cortex. “Analogous to delayed response behavior, spoken language requires working memory to access verbal symbols and hold them on-line until appropriate articulatory movements of the lips and mouth are executed”. Goldman-Rakic saw the use of words in children as a reflection of the maturation of Broca’s area, (Brodman’s areas 44 and 45), just as the maturation of the AB function depends in man on the maturation of the area corresponding to the principal sulcus in monkeys(area 46). According to Cohen et al. (2002) the underlying assumption in physiological models is that short-term storage of information in PFC occurs through re-circulating activity within local recurrent networks, in which DA helps to stabilise attractor states, both by making high activity states more stable, and low activity states less likely to spuriously transition to high activity states in the absence of strong afferent input. “This is accomplished by the concurrent potentiation of excitatory and inhibitory transmission, implemented as changes in ion channel properties in biologically detailed models and summarised as a change in the gain of the sigmoidal activation function in connectionist models. In simulations of DA effects, for example in the role of the PFC in working memory, enhanced stability of PFC working memory representations are made less susceptible to interference from intervening distractors, until a behavioural response is executed” (Cohen et al., 2002). Cohen et al. (2002) describe their own work as focussing on phasic vs tonic DA release. They suggest that phasic bursts of DA activity in PFC may function as a gating mechanism, by signalling when afferent inputs should be selected and stored in PFC, updating the contents of working memory. This model is supported by neurophysiological data, which suggests that 126CHAPTER 3. NEURAL CIRCUITS AND BEHAVIOUR CONTROL mid-brain DA neurons convey important reward-related information regarding external stimuli through phasic bursts of activity. Specifically these neurons are thought to show rapid, transient, stimulus-specific responses to salient environmental stimuli that reliably predict future rewards. Thus representations in PFC are updated to guide behaviour in accord with current goals, via phasic changes in DA activity. However, Cohen et al. (2002) point out that reliance on phasic DA signalling appears to contra- dict the biophysical models, which argue that tonic rather than phasic activities of DA are most relevant to PFC function. They attempt to reconcile this apparent contradiction by pointing out that most models of DA effects in PFC are based on the importance of D1 receptors (Goldman-Rakic et al., 2000). On the other hand D2 mediated effects tend to be opposite of, and antago- nistic to D1-mediated effects, such that D1 receptor activation enhances NMDA excitatory post-synaptic currrents (EPSC’s), inhibitory post-synaptic currents (IPSC’s), and interneuron excitability, while D2 receptor activation decreases NMDA EPSC’s, IPSC’s, and interneuron excitability. Also D2 receptor effects are thought to occur more rapidly and decay more quickly.

The authors believe that reciprocal tonic D1 and phasic D2 activity may explain behavioural findings such as DA agonists allowing representations in PFC to be maintained (high tonic DA), but not effectively updated, thus producing perseveration and stereotypy. Conversely, a DA antagonist would degrade PFC representations (low tonic DA) allowing intact updating and possibly over-reactive impulsive behaviour. Au-Young et al. (1999) have shown that single-pulse stimulation of the ventral tegmental area (VTA) can rapidly and transiently modulate firing patterns of PFC neurons. However, Cohen et al. (2002) point out that the tonic/phasic model of DA modulation within PFC by phasic gating may account for some task situations, but may not explain hierarchical task requirements which have a goal-subgoal structure. For such tasks, the role of basal ganglia, which have a highly segregated system of inputs to PFC subregions, may play a criti- 3.4. PREFRONTAL CORTEX: COGNITIVE AND EXECUTIVE FUNCTIONS127 cal role in reinforcement driven learning. Arnsten (2007) describes how alpha-2A adrenoreceptor stimulation, in the dorsolateral prefrontal cortex (DLPFC), is thought to result in closure of HCN channels on spines receiving inputs from neurons with similar spatial properties, thus increasing firing during delay periods for preferred directions, while moderate levels of dopamine D1 receptor stimulation leads to opening of HCN channels on spines, receiving inputs from neurons with dis-similar spatial properties, reduces delay related firing to non-preferred directions. Thus synchronous noradrenergic/dopaminergic synaptic effects combine to produce a spatial representation. Differences in dopamine D1 vs D2 receptor activity is believed important in differential prefrontal vs basal ganglia activity. The maintenance of working memory states, separate from intervening distractors, is important in understanding conditions where such stability is not maintained, such as Attention Deficit Hyperactivity Disorder (ADHD), or is over- maintained, such as Obsessive-Compulsive Disorder (OCD).

3.4.1 Comment

Goldman-Rakic’s comments on working memory and language in relation to the function of Broca’s area provides a link between cortical maturation processes and linguistic control of behaviour discussed above, in relation to recursivity as a control mechanism. The above biophysical models, suggest functions, which are very similar to those described in more abstract form by connectionist models, emphasising the value of a multi-level approach, and synaptic transmission weights, to model building. Most of the above models focus on dopaminergic modulation of synaptic gating, and/or connectivity in PFC neurons, indicating that for working memory, enhanced stability of PFC representations are made less susceptible to interference from intervening distractors until a behavioural response is executed. However the models do not entirely explain the temporal coordination of working mem- 128CHAPTER 3. NEURAL CIRCUITS AND BEHAVIOUR CONTROL ory activations. While basal ganglia are involved in control of sequential re-entry circuits during motor activity, coordination of cognitive activity is more complex, and likely involves thalamo- cortical and/or posterior cortex. Chapter 4

Development and working memory

Kohlberg and Gilligan (1971) have discussed the implications of a Piagetian stage theory of development. They state that

“the concept of stage implies an invariance of sequence in development, a regularity of stepwise progression, regardless of cultural teaching or circumstance. Cultural teaching and experience can speed up or slow down development, but it cannot change its order or sequence. Stages are the product of interactional experience between the child and the world, experience which leads to a re- structuring of the child’s own organisation, rather than to direct imposition of the culture’s pattern upon the child (Kohlberg and Gilligan, 1971).

Kohlberg and Gilligan (1971) describe a transition from logical inference as a set of concrete operations to logical inference as a set of formal operations or ”operations upon operations”. The full development of formal-operational thought is thought to occur from age 11 to adulthood, when inferences through logical operations upon propositions or “operations upon operations”, enabling an adolescent to think about thoughtand create “hypothetico- deductive” theories. Both Piaget and Kohlberg have described stages of cognitive development, but do not relate these stages to underlying neurophysiological developments, such as the onset of puberty, or the development of prefrontal cortex connectivity with subcortical modules, allowing operations upon operations (my em-

129 130 CHAPTER 4. DEVELOPMENT AND WORKING MEMORY phasis). The present thesis investigates biological and developmental aspects of this development, and behavioural syndromes resulting from developmental failures.

4.1 Structural development

Farran (2001) pointed out that the development of the cerebral cortex in humans can be divided into two major phases - the ﬁrst phase involving the birth and migration of cortical neurons, taking place early in pregnancy, and occurring only once (Hutten- locher, 1994, p233-234) . This phase is sensitive to conditions in the embryonic environment, and is thus sensitive to conditions, which slow down or prevent the timing of proper developmental processes such as poor prenatal care, malnutrition or exposure to alcohol. The second phase of development according to Huttenlocher actually has two sub-phases - the period of synaptogenesis, followed by a longer stage of loss of synapses. According to (Far- ran, 2001, 233-234), shortly before and immediately after birth, there is an explosion of synapse formation that leads to a level of synapses in the cortex at 1-2 years far above that of an adult. According to Farran, most of the overproduction takes place in asymmetric synapses, associated with excitatory signals (vision, hearing etc), while the smaller number of symmetric, inhibitory synapses are not so wildly overproduced. Thus, the major changes that occur in the human CNS between the ages of 2 and 12 years occur through a decrease in the number of synapses, as well as an increase in the complexity of their dendritic arborizations. According to Huttenlocher (1994) , the multiplicity of synaptic connections becomes re-organized through the elimination of many of them, provided the system is subjected to incoming signals that have repeating patterns. It is important to note that, while synaptic elimination in the visual cortex begins around 1 year of age, synaptic elimination in the frontal cortex does not begin until age 7 and continues through adolescence. This is an 4.1. STRUCTURAL DEVELOPMENT 131 important difference between humans and other primates, and indicates a longer period of synaptic plasticity, with greater op- portunities for learning throughout childhood.

Paus et al. (1999) pointed out that structural maturation of individual brain regions and their connecting pathways is a necessity for the successful development of cognitive, motor, and sensory functions. “The smooth flow of neural impulses throughout the brain allows for information to be integrated across the many spatially segregated brain regions. The speed of neural transmission depends not only on the synapse, but also on structural properties of the connecting fibers, including the axon diameter and the insulating myelin sheath”. According to Paus et al. (1999), postmortem studies suggest that axon diameter and myelin sheath undergo conspicuous growth during the first 2 years of life, but may not be fully mature before adolescence, or even late adulthood. Paus et al. (1999) obtained MRI scans of 111 children and adolescents aged 4-17 years. Regression analysis revealed significant age-related increases in white matter density within the left and right internal capsule, and the posterior portion of the left arcuate fasciculus. The authors suggest that the location of the changes in the posterior limb of the internal capsule suggested that the changes involved the corticospinal and possibly thalamospinal tracts, which increased linearly from age 4-17 years, is consistent with a protracted development of motor skills. “Faster conduction velocity can facilitate information flow, not only by speeding it up, but also by allowing for precise temporal coding of high-frequency bursts of neuronal activity”. Importantly for present purposes, the arcuate fasciculus contains frontal and temporal cortical regions involved in speech. Age-related white matter increases in this pathway reached significance only in the left, but not in the right hemisphere. The authors comment that the left hemisphere can be assumed to be dominant for speech in the majority of right-handed subjects. “It has been proposed that processing of speech sounds requires a neural system, capable of tracking rapid changes in accoustic input. Rapid transfer of 132 CHAPTER 4. DEVELOPMENT AND WORKING MEMORY information to the auditory cortex and beyond would require fast- conducting fiber systems” (Paus et al., 1999; Tallal et al., 2006). Paus et al. (1999) suggest that age-related increases in white matter density along the arcuate fasciculus might represent a structural correlate of the cortico-cortical pathway, mediating sensory-motor interactions between anterior and posterior speech regions. They point out that the interruption of the arcuate fasciculus in adulthood causes conduction aphasia, “perhaps as a result of the disruption of both feedforward and feedback mechanisms” (Liberman and Mattingly, 1985). The engagement of such feedback mechanisms may facilitate the late stages of speech development, requiring a fast bi-directional transfer of information between the auditory and motor cortical regions. It is also possible that the age-related increases in white matter density, both along the arcuate fasciculus and the putative corticospinal tract, reflect the effect of extensive use of these systems during the individual’s life. Pavuluri and Sweeney (2008) point out that “understanding the evolution of illness expression needs to be nested in a solid understanding of the normal trajectory in the maturation of brain functions”. Thus a comprehensive understanding of the “normal developmental trajectory in youths without disorder” is essential to map the pattern of deviance in at-risk people”. The authors point out that functionally, the prefrontal cortex is one of the last regions of the human brain to attain an adult level of development, typically by 15 to 16 years. “From a behavioural perspective, subjects continue to show improvement in response inhibition and working memory abilities through at least mid- adolescence” (Luna et al., 2001). Thus changes in the participation of the frontal cortex and the ability to recruit widely distributed brain circuitry support increases in the efficiency of the cognitive control of behaviour during adolescence (Pavuluri and Sweeney, 2008). According to Giedd et al. (1996) developmental curves for frontal and parietal gray matter peak at around age 12 years, 4.1. STRUCTURAL DEVELOPMENT 133 temporal lobe around age 16 years, while cortical white matter re- finements continue until nearly 20 years. Their results suggested that efficient top-down modulation of reflexive acts may not be fully developed until adulthood, and provide evidence that maturation of function across widely distributed brain regions, lays the groundwork for enhanced voluntary control of behavior during cognitive development. Giedd et al. (1999) in a large-scale longitudinal paediatric neuroimaging study, confirmed linear increases in white matter, and nonlinear changes in cortical grey matter, with a pre-adolescent increase, followed by a post-adolescent decrease. The changes in cortical gray matter were regionally specific, with frontal and parietal lobes peaking around age 12 years, temporal lobe around 16 years and occipital lobe through age 20.

Shaw et al. (2009) investigated the development of cortical asymmetry in children with and without ADHD. Asymmetry is thought to be an evolving property of the childhood brain. The authors describe variations in structural asymmetries as linked with left or both-handedness. Atypical hemispheric specialisation for language is found to occur in approximately 30% of non-right- handed people, but fewer than 10% of right-handed individuals (Woods et al., 1988). Shaw and colleagues utilised magnetic resonance images (MRI’s) to delineate typical and atypical patterns of development in 358 participants, of whom 22 were predominantly left-handed, and 20 were ambidextrous. In typically developing children, they found that in early childhood, the left orbitofrontal and inferior frontal gyrus was relatively thicker than the right, but by late adolescence resembled the adult pattern. A similar increase with age in relative right hemispheric thick- ness was found in the medial occipital region. Thus in childhood, the posterior cortex was thicker, but by adulthood the asymmetry had reversed, and was relatively thicker on the left. On the other hand, ADHD-diagnosed children, who were right-handed showed typical posterior, but atypical anterior evolving asymmetry, while the non-right-handed ADHD group also showed an atypical right- ward temporal asymmetry with age. Thus in typical development, 134 CHAPTER 4. DEVELOPMENT AND WORKING MEMORY the increased dimensions of the right frontal and left cortical regions emerge in adulthood from the reversed pattern of childhood cortical asymmetries, while “loss of the prefrontal component of this evolving asymmetry in ADHD is compatible with disruption of prefrontal function in this disorder” (Shaw et al., 2009). While the control of early PFC development is extremely complex, Rakic (1988) described how the “immense population of neurons that constitute the human cerebral neocortex is generated from progenitors lining the cerebral ventricle and then distributed to appropriate layers of distinctive cytoarchitectonic areas can be explained by the radial unit hypothesis. According to this hypothesis, the ependymal layer of the embryonic cerebral ventricle consists of proliferative units that “provide a proto-map of prospective cyto-architectonic areas”. According to Rakic, the output of the proliferative units is translated via glial guides to the expanding cortex in the form of ontogenetic columns, whose final number for each area can be modified through interaction with afferent input. His radial unit model provides a framework for understanding cerebral evolution, epigenetic regulation of the parcellation of cytoarchitectonic areas, and insight into the pathogenesis of certain cortical disorders in humans. For the present investigation, the significance of early cellular and receptor events is illustrated by a series of experiments by Kellendonk et al. (2006) who investigated the effects on working memory, of generating increased levels of D2 receptors (D2Rs) in transgenic mice. They found that selective over-expression of D2 receptors in the striatum led to impairment in tasks requiring working memory and behavioural flexibility. The deficits persisted even after the transgene was switched off, indicating that the deficit resulted from developmental, rather than concurrent functioning of the upregulated D2 receptors. The investigators first showed that delayed non-match to sample (DNMTS) maze tasks require mice to integrate information held on-line, with a learned rule. Kellendonk et al. (2006) lesioned the medial PFC (mPFC) of wild-type mice, specifically affecting the infralimbic, 4.1. STRUCTURAL DEVELOPMENT 135 prelimbic and anterior cingulate cortices. They showed that the mPFC is required in mice performing a DNMTS T-maze task. Because the lateral and medial ganglionic eminences give rise to both neostriatal neurons and GABAergic neurons in the cortex, the investigators analysed the density of GABA interneurons in the mPFC, but detected no differences in the number of GAD positive neurons in the prelimbic cortex between genotypes. Kel- lendonk et al. (2006) then measured the ex-vivo tissue content of dopamine and its metabolites in the mPFC. They found an increase in dopamine levels and a decrease in dopamine turnover. They also determined D1 receptor activation in the PFC of transgenic mice by measuring the induction of the early gene c-fos by the selective D1 agonist chloro-APB. After injection of the D1/D5 selective agonist, a significant increase in the number of c-fos positive cells was observed in D2 transgenic mice. However transgenic mice in which over-expression of D2 receptors was switched off, showed a significant decrease in the number of c-fos positive neurons, suggesting that they developed compensatory processes for D1 receptor function. Since working memory requires an optimal level of D1 receptor activation in the mPFC, the increase in D1 receptors in non-treated, and decrease in D1 receptors in doxycycline-treated transgenic animals, was thought in both cases to be possibly responsible for deficits in working memory in both conditions (indicating the importance of D1 receptors). According to Kellendonk et al. (2006), working memory and behavioural flexibility have, based on imaging and lesion studies, been generally associated with the dorso-lateral PFC in humans and non-human primates (Goldman-Rakic, 1994b; Fuster, 1997). They suggested that the medial PFC is the homologous structure of the dorso-lateral PFC in primates, and that the mPFC is also required for acquisition of a working memory task in mice. Because the striatum participates in a cortico-striatal pallido- thalamo-cortical associative loop (Alexander et al., 1986), which includes the DLPFC, perturbing the circuit at any point may affect its function. Thus the cognitive impairments in the striatal 136 CHAPTER 4. DEVELOPMENT AND WORKING MEMORY

D2 over-expressing mice may be due to effects of excess D2 receptors on cortico-striatal synapses, or indirect effects through midbrain dopaminergic neurons. The investigators found that in D2 receptor over-expressing mice, activation of c-fos by a D1 agonist is increased in the prelimbic area of the PFC, indicating a higher D1 receptor sensitivity or density. However, when the D2 receptor binding was normalised by switching off the transgene, mutant mice continued to demonstrate the working memory task deﬁcits. The investigators pointed out that opposing regulation of D1 receptor activation can potentially affect working memory because of the U-shaped effects on dopamine transmission in the PFC. They suggest that during development the dopaminergic inner- vation of the cortex is set up in a homeostatic manner. “Increased D2 receptors in the striatum may alter the physiology of the dopaminergic mid-brain neurons during developments, leading to a new equilibrium, that may be sensitive to adult changes and affect PFC function”. While Kellendonk et al. (2006) relate these effects to schizophrenia, they may equally apply to ADHD, possibly in opposite directions, such that upregulation of D1 receptors may occur in schizophrenia, and down-regulation in ADHD. In both cases, these compensations in the adult may occur as a result of embryonic changes in D2 receptors in embryonic or perinatal stages of development, with later effects on working memory.

4.2 Working memory

In the mature brain, working memory is thought to depend on an intact dorsolateral prefrontal cortex (DLPFC) (Petrides et al., 1993; Goldman-Rakic, 1994a). According to Tau and Peterson (2010), rudimentary working memory capacities have been observed in infants as young as 6 months of age, but performance on Piaget’s A not B task (retrieval of a hidden object after a delay) is not in place till 9 months, and is not solidly in place for difﬁcult tasks till middle childhood. It is believed that fronto-parietal networks are recruited with increasing task difﬁculty between child- 4.2. WORKING MEMORY 137 hood and adolescence (Durston et al., 2006; Geier and Luna, 2009; Scherf et al., 2006). Funahashi (2007, p213-214) describes working memory (WM) as “mechanisms or processes that are involved in the control, regulation, and active maintenance of task-relevant information in the service of complex cognition, including novel as well as familiar skilled tasks” (Miyake and Shah, 1999). He outlines a revised model of working memory proposed by Baddeley (2000), which includes one master component (the central executive) and three slave components (the viuospatial sketchpad, the phonological loop and the episodic buffer). Baddeley and Hitch (1974) originally described a three-component model, which assumed an attentional controller or the central executive, aided by two sub- sidiary systems, the phonological loop, holding speech-based information, and the visuospatial sketchpad, which performed a similar function for visual information. In the Baddeley (2000) review, the episodic buffer was assumed to be capable of storing information in a multidimensional code, thus providing a temporary interface between the slave systems (the phonological loop and the visuospatial sketchpad) and long-term memory (LTM). The central executive was seen as responsible for binding information from a number of sources into coherent episodes, assessed to be retrievable consciously. According to Baddeley (2000) frontal areas are very likely to be important for both the central executive and the episodic buffer. He describes the episodic buffer as emphasising the importance of coordination and confronting the need to relate WM and LTM. Thus, according to Fu- nahashi (2007), the episodic buffer is a temporary storage buffer with a limited capacity to integrate information from a variety of sources, including long-term memory as “chunks”. The central executive is an attentional “master component” which selects appropriate control processes or strategies for performing current tasks, and supervises the performance of the three slave systems. Funahashi (2007, p214-216) suggests that all the slave components include neural mechanisms for maintaining and process- 138 CHAPTER 4. DEVELOPMENT AND WORKING MEMORY ing information. He proposes a “process-based” model of working memory, including a selection process, a temporary information process, an output process and modulatory inputs.

[...] the DLPFC receives a variety of information including sensory, motor, motivational, and emotional information from other cortical or subcortical areas. [...] Close relationships between working memory and the dorsolateral prefrontal cortex could provide important clues for common neural mechanisms of working memory. [...] In fact, neurophysiological studies have shown that many neurons in the DLPFC exhibit tonic sustained activation (delay-period activity) during the delay period of spatial working memory tasks (Funahashi et al., 1989; Sawaguchi, 1998).

Funahashi (2007) suggests that information processing in the DLPFC can be seen as a temporal change of the information represented by a population of neural activities. “[...] in order for the DLPFC to perform such higher cognitive functions as thinking, reasoning, decision-making and language comprehension, the DLPFC not only receives a variety of information from cortical and subcortical areas, but also sends a variety of information to these areas”. He describes the temporary information process as an essential component of the neural mechanism of working memory in the DLPFC. “[...] by interactions among temporary information storage processes, maintained information would be processed, replaced, and updated”. Funahashi also describes an output process, which sends maintained information to brain areas, where it is utilised, such as motor areas, and areas related to LTM, while modulatory input includes feedback information from motor centers, and motivational or emotional information from limbic areas (Barbas, 1992), or modulation signals by catecholaminergic inputs (Arnsten et al., 1998; Sawaguchi, 1998; Wang et al., 2004b). Thus information processing in working memory can be explained as a change of the information represented in the temporary information-storage processes, caused by dynamic interactions among complex processes. Funahashi (2007, p214-216) describes modality-specific working memory systems as mechanisms for temporary maintenance 4.2. WORKING MEMORY 139 and processing of certain types of information, such as sensory, linguistic, or motor information. “Modality specific working memory will be present in cortical and subcortical areas. On the other hand, a general-purpose working memory system is a mechanism to maintain or process a variety of information, including the behavioural context to perform various cognitive tasks or control signals (e.g. top-down signals) to monitor or regulate operations of other cortical and subcortical areas”. The contents of a general- purpose working memory system, according to Funahashi, would effect or regulate the contents of modality-specific working memory systems in necessary directions.

Funahashi (2007, p222-223) describes the functions of the general-purpose working memory systems as similar to proposed executive functions such as Cohen et al. (1996), who posited that accomplishment of a goal requires maintenance of an internal representation, suppression of unnecessary behaviours and temporal coordination of behaviours. Funahashi also describes cortico-subcortical connections with the DLPFC, which is described as having reciprocal connections with the posterior parietal cortex, the inferior temporal cortex, superior temporal poly- sensory areas, the anterior cingulate, the retrosplenial cortex and the parahippocampal gyrus, as well as strong reciprocal connections with the mediodorsal nucleus of thalamus (Fuster, 1997; Petrides, 1994; Petrides and Pandya, 2002) and frontal eye fields, motor areas and caudate nucleus (Morris et al., 1999). Thus the DLPFC has reciprocal anatomical connections with cortical and subcortical areas, especially sensory and motor association areas and limbic brain areas, allowing dynamic and flexible interactions between the general-purpose working memory system, and the modality specific working memory systems. Funahashi (2007) discusses the importance of top-down modularity effects by the DLPFC, which affect a wide variety of cognitive activities, including sensory perceptions, directed attention, episodic memory encoding and retrieval through dynamic and flexible functional interactions as the typical function of the DLPFC as the general- 140 CHAPTER 4. DEVELOPMENT AND WORKING MEMORY purpose working memory system. He describes the dynamic and flexible nature of these interactions as an essential feature of the general-purpose working memory system. According to Rolls (2007), the ventral visual system is closely related to consciousness when explicit or conscious planning is required, because we normally apply multi-step planning processes to objects represented in the ventral visual system. Rolls (2007, p433) describes two coupled networks (one in the inferior temporal visual cortex for perceptual functions, and another in the prefrontal cortex) for maintaining short-term working memory during intervening stimuli in pursuit of a goal. The selection between reward and punishment signals is made, according to Rolls, by Pavlovian learning processes, which are able to elicit autonomic and endocrine responses to be elicited by conditioned stimuli, and by action oriented learning to obtain particular goals. Lewis (1997) points out that adult levels of performance on some cognitive tasks mediated by the DLPFC are not achieved until after puberty in both monkeys and humans. In particular, the delayed-response task first appears about one year of age in humans, but does not reach adult levels until after puberty. During post-natal development, the percentage of DLPFC neurons that exhibit delay-period activity doubles between 12 and 36 months of age, suggesting that developmental changes in DLPFC circuitry facilitate the recruitment of neurons to this functional role. According to Lewis, recent studies have shown that in DLPFC layer 3 pyramidal neurons, certain classes of local circuit GABA neurons and dopamine afferents are (1) synaptically linked; (2) undergo substantial temporally associated maturational changes during adolescence and (3) may be critical components of the neural substrate for delayed-response tasks. (Thus working memory cannot be considered fully mature until late adolescence or early adulthood). Miller and Cohen (2001) have pointed out that while the PFC is not critical for performing simple automatic behaviours such as orientation to an unexpected sound, the PFC is important 4.2. WORKING MEMORY 141 when top-down processing is needed to guide behaviour by internal states or intentions. That is, when selective attention, behavioural inhibition, working memory and rule-based or goal- directed behaviour is required, even when in competition with habitual or stronger responses. Miller and Cohen describe the PFC as having the properties required to achieve top-down behavioural control. These include the ability to maintain its activity robustly until a goal is achieved, and second to have interconnections with all sensory systems, cortical and subcortical motor systems, and with limbic and midbrain structures involved in affect, memory and reward. Thus the lateral and mid-dorsal PFC receives visual, somatosensory and auditory information from the occipital, temporal and parietal cortices. The dorso-lateral area 46 is connected with pre-motor areas that send connections to primary motor areas and the spinal cord, as well as cerebellum and superior colliculus. There are also dense interconnections with basal ganglia.

The orbital and medial PFC are closely associated with medial temporal limbic structures, critical for long-term memory and the processing of internal states, such as affect and motivation. This includes direct and indirect (via the medial dorsal thalamus) connections with the hippocampus and associated neocortex, the amygdala, and the hypothalamus. Also most PFC regions are interconnected with most other PFC regions. [....] Thus the PFC provides a venue by which information from wide-ranging brain systems can interact through relatively local circuitry (Miller and Co- hen, 2001).

Miller and Cohen also point out that the PFC neurons are both individually selective and others bimodally selective for sensory cues, but in addition PFC neural activity is able to represent rules required to perform a particular task. The Miller and Cohen model requires feedback signals from the PFC to reach throughout the brain. Miller et al. (1996) were able to show that monkeys were able to maintain a working memory of a rewarded stimulus over time, and that target-speciﬁc activity appeared simultaneously in the PFC and parietal cortex. While other brain areas can sustain activity up to several seconds, the PFC is distinguished 142 CHAPTER 4. DEVELOPMENT AND WORKING MEMORY by the ability to sustain such activity in the face of intervening distractions.

Thus the PFC exhibits sustained activity that is robust to interference: multimodal convergence and integration of behaviourally relevant information; feedback pathways that can exert biasing influences on other stuctures throughout the brain; and ongoing plasticity that is adaptive to the demands of new tasks. The authors believe this specialisation is optimal for a role in the brain-wide control and coordination of processing. They point out that the mechanisms responsible for updating representations in the PFC must be responsive to changes in the environment, as well as resistant to updating irrelevant changes. They hypothesise that dopamine (DA) might play an important role in this gating function. They suggest that dual concurrent influences on midbrain DA allow the system to learn while it gates, and where a DA-mediated gating signal leads to a successful behaviour, its concurrent reinforcing effects will strengthen the association of the signal with cues representing the pattern of activity that produced the behaviour. Thus this self-organising bootstrapping mechanism averts the invocation of a “homunculus” to control behavioural selection. This still leaves open the question of “sub-goaling” or hierarchical updating, and the proper sequencing of actions. The authors suggest this may rely on interactions between PFC and basal ganglia. Miller and Cohen (2001) point out that capacity for active maintenance may involve the bistabil- ity of neurons, but also circuit-based models which propose that recirculation of activity through closed or recurrent loops of interconnected neurons or attractor networks (Hopfield, 1982), likely to support sustained activity. These loops could be intrinsic to the PFC, or they might involve other structures such as cortex- striatal-globus pallidus-thalamus-cortex loops (Alexander et al., 1986). 4.2. WORKING MEMORY 143

4.2.1 Higher Order Cognition and Working Memory

Goldman-Rakic (1987b) proposed that a working memory process is the fundamental specialization of the prefrontal cortex and the mechanism for directing responses by internal representations:

which can be considered the basis for memory-guided responding. Further, this process distinguishes the prefrontal contribution to behavior from those systems of the brain. [...] The products of learning and past experience are accessed by prefrontal neurons, which process them and amalgate them with the ongoing stream of information, currently experienced (Goldman-Rakic, 1987b).

Thus, Goldman-Rakic differentiates prefrontal (working memory) processes based on reverbatory cortico-cortical and thalamo- cortical circuits, from associative and inhibitory cortico-striatal networks. She includes language as a purely representational process, guided by an on-line processor. Robbins (1998, p116) has described four important forms of neural interaction involving the prefrontal cortex. These are (1) dedicated processing modules of the posterior cortex such as parietal and temporal lobes. (2) limbic structures such as the amygdala and hippocampus (3) the output of the striatum, which targets the frontal lobe and (4) the ascending monoaminergic and cholinergic systems of subcortical origin, which exert potentially diverse effects on forebrain functioning. He reviewed investigations utilising a number of ‘frontal’ tasks, including extra- dimensional shift, spatial span, spatial working memory (errors and strategy) and Tower of London. Robbins (1998, p127-128) suggests that at a theoretical level the results show that identical compound stimuli can be processed at more than one site in the prefrontal cortex, perhaps simultaneously. The results also suggest that distinct processes of response inhibition are recruited to control processes of extra-dimensional shift and associated reversal learning. However these data are thought not incompatible with hierarchical models of prefrontal cortex, emphasising serial ﬂow of information to superordinate regions, such as dorsolateral prefrontal cortex, but are more consistent with parallel process- 144 CHAPTER 4. DEVELOPMENT AND WORKING MEMORY ing by relatively independant sectors of the prefrontal cortex. Goldman-Rakic (1998b, p89) points out that despite a widespread belief that the prefrontal cortex is a composite of functionally distinct or hierarchically arranged areas, engaged respectively with processes of attention, affect, emotion, memory and motor aspects of behaviour, she suggests:

(1) the dorsolateral prefrontal cortex has a ‘generic’ function, namely ‘on-line’ processing of information (working memory) in the service of a wide range of cognitive functions. (2) this process is iteratively represented throughout several and possibly many subdivisions of the prefrontal neopallidum. (3) each autonomous subdivision integrates attentional, memorial, motor and possibly affective dimensions of behaviour by virtue of network connectivity with relevant sensory, motor and limbic areas of the brain, and do not need to be allocated to separate architectonic regions (Goldman-Rakic, 1998b).

Goldman-Rakic (1998b) ascribes a pre-eminent role in working memory to the prefrontal cortex, but describes functional integration with posterior parietal, and inferior prefrontal with temporal lobes. According to Goldman-Rakic, these networks are reciprocal.

Our view is that the central executive may be composed of multiple segregated special purpose processing domains, rather than one central processor served by slave systems, converging to a central processor; and that each specialised domain consists of local and extrinsic networks with sensory, mnemonic, motor and motivational control elements. [...] It is possible to view the co- activation of working memory domains and their associated cortical networks, as a well-designed parallel processing architecture for the brain’s highest level cognition (Goldman-Rakic, 1998b).

Goldman-Rakic et al. (1990) have described the anatomical overlap of different mono-aminergic receptors in the same cortical strata, suggesting that there may be families of receptors linked by localization on common targets. This would provide the anatomical basis for subcortical inﬂuences on prefrontal/parietal systems. “On the other hand, the complimentary laminar distribution of D1 vs D2, 5-HT1 vs. 5-HT2, and alpha vs. beta adrenergic receptors, raises the possibility that different subtypes within 4.2. WORKING MEMORY 145 a given class, may have distinctive actions in the cortex, by virtue of their localisation on different cells, or possibly on different portions of the same cell”. Here, Goldman-Rakic points to mechanisms of reciprocal receptor actions in response to the same neurotransmitter. This would appear to be a useful biological mechanism, so that a transmitter such as dopamine and/or noradrenaline, may have reciprocal balancing effects at different cellular and/or circuit locations.

According to Roberts et al. (1998), the Goldman-Rakic hypothesis that “on-line” and inhibitory control processes are emergent processes throughout much of the prefrontal cortex. Regions differ primarily with respect to informational domain, in which spatial processing takes place mainly in the dorsolateral prefrontal cortex and feature processing takes place in the ventrolateral prefrontal cortex, which is involved in the processing of reinforcement or affective stimuli. This differs from the differential functional hypothesis of Petrides. Petrides (1998, p106-112) suggested that the ventrolateral prefrontal cortex is involved in ﬁrst-order executive processes such as active selection and comparison of stimuli held in short-term memory, while the dorsolateral prefrontal cortex is involved in higher-order executive compononents of working memory. Petrides distinguishes between automatic and active retrieval processes. According to his hypothesis, automatic retrieval is the by-product of the triggering of stored representations in posterior cortical association regions, either by incoming sensory input that matches pre-existing representations, or by recalled events that trigger stored representations of related information on the basis of strong pre-existing associations. On the other hand active retrieval is thought to imply conscious (i.e. willed) effort to retrieve a speciﬁc piece of information, guided by the subject’s intentions, plans, or instructions. Thus Goldman- Rakic postulates a continuation of modularity at cortical levels, while Petrides suggests an active hierarchical integration process in the cortex. 146 CHAPTER 4. DEVELOPMENT AND WORKING MEMORY

Working memory is deﬁned as the ability to hold information cognitively online for a brief period of time but long enough for task completion. Koziol and Budding (2009, p46-52) point out that while higher order cognition has traditionally been considered mediated by the cortex, the mechanisms that maintain multiple representations online, manipulate those representations, prevent intrusions by distractions, and update the contents of working memory are thought to be mediated by interactions between cortex and basal ganglia.

The basal ganglia perform various selection operations, by interacting with prefrontal cortex in a way analogous to the role of a doorman or bouncer in a nightclub. [...] Prefrontal-cortical connections maintain information online, while basal ganglia ‘gate’ manipulations of information and prevent the intrusion of distractions. [...] Output from the GPi, because of its inﬂuence over the thalamus, appears to be a key player in updating functions. [...] Therefore working memory is characterised by a division of labour between maintaining information online, which is a function of the cortex, and updating information, which is gated by the basal ganglia (Koziol and Budding, 2009, p48).

. Koziol and Budding (2009, p48) point out that while working memory’s dependence upon cortical processing is well docu- mented, the ability or capacity is easily disrupted by any condition that affects the fronto-striatal system, including neuropsychiatric disorders such as Tourette’s Syndrome, Attention Deﬁcit Disorder, and schizophrenia. Koziol and Budding (2009, p50) believe that the cognitive control system is likely an evolutionary extension of the frontal-basal ganglia motor control system.

[...] we can think of motor plans or behaviours as represented or maintained in cortex. [...] Serial order processing is made possible by re-entrant projections that ‘loop back’ to the same areas of cortex from which they originated. The basal ganglia, by interacting with the cortex in a proper sequence release the proper order of behaviour. As the behaviour is released, the sequence needs to be updated, while remaining elements of the motor sequence need to be maintained as representations ‘online’, until the program is completed. We believe that the basal ganglia interact in the same way with the representations of plans and goals that reside in the 4.2. WORKING MEMORY 147

prefrontal cortex and in this way, the processes that guide motor functioning have much in common cognition of working memory that ultimately guides behaviour (Koziol and Budding, 2009, p50).

According to [p59]Koziol and Budding (2009), refrontal- cortical connections are excitatory, so that vast amounts of processed information is potentially in competion to be considered in working memory, or to intrude upon its functions. However, the same PFC regions that have reciprocal connections with cortical areas also activate the head of the caudate through the direct pathway. This activation of the striatum selectively reduces GPi inhibition in the thalamus, so that the PFC-posterior cortical working memory loops can be selectively identiﬁed, chosen and maintained. This “selection” function performed by the striatum (through its interactions with the GPi) allows the information in prefrontal-cortical circuitry to be kept “online” for the purpose of thinking about the information or representations (Ponzi, 2008).

In a well-learned routine or highly familiar situation, one would expect the striatum to be highly activated [...]. In a novel or new situation striatal neurons associated with a particular behaviour would be only weakly activated, because the context of the situation is not associated with any behavioural pattern [...]. Therefore the PFC stimulates the head of the caudate to maintain activation of PFC-cortical working memory loops, so that a new behaviour can be devised or programmed, since there is no automatic or stimulus- based behaviour that works under those novel conditions. This allows for alternating episodes of stimulus-based behaviour with appropriate adjustments from PFC supervisory higher-order control mechanisms. In particular, the anterior cingulate (medial prefrontal) circuitry plays a critical role in the monitoring of performance, in detecting performance deviations, and in signalling the need for behavioural adjustments (Koziol and Budding, 2009, p59).

According to Koziol and Budding (2009, p59), the prefrontal cortex is biased by two sets of inputs. One set is biased by selection from the striatum, which is based upon the pattern of cortical activity that it reads. The other set of inputs comes from the cortical cell assemblies that are speciﬁcally activated by the current situation. Koziol and Budding (2009, p61) point out that even when the basal ganglia are not recruited to execute a routine 148 CHAPTER 4. DEVELOPMENT AND WORKING MEMORY behavioural pattern, the basal ganglia continue to play a role in higher-order behavioural control by participating in the division of labour by actively gating cognitive activity through direct and indirect pathways. Thus “at best we may be considered creatures of habit, and at worst we may be creatures of impulse”. Arnsten (2007) points out that although Goldman-Rakic (1987b) used spatial working memory as a model system for examining functional circuitry, she proposed that these principles applied to other sensory and affective domains, and described the process as “representational knowledge within parallel processing streams”. According to Arnsten, Goldman-Rakic (1991) spoke of prefrontal cortical (PFC) network activity as a fundamental contribution to mind, and the disruption of this process as a primary contribution to thought disorder in mental illness. “She used the term working memory to describe a building block of cognition: the ability to represent information no longer in the environment through recurrent excitation of pyramidal cells with shared stimulus properties” (Goldman-Rakic, 1995).

4.2.2 Comment

Thus adequate PFC functioning appears critical for not only mature reasoning, but also involves behavioural fuctions, including inhibition of task-irrelevant behaviours, processing of affect, motivation, and reward attainment by virtue of connections with wide-ranging cortical centers. It could thus be argued that the process of development is closely dependant on adequate PFC development, and many if not most behavioural syndromes of childhood reflect deficits in PFC development. A consequence of such deficits in PFC development is an incapacity for sequential reasoning, lack of affect regulation, a lack of capacity for working on sustained goal achievement, and a tendency for impulsive and repetitive behaviours, under either environmental, or subcortical control. This proposition will be further examined in relation to a number of DSM-IV defined childhood syndromes. 4.2. WORKING MEMORY 149

Barkley (1997)’s concept of the fundamental importance of an inhibitory deficit in ADHD, with consequent sub-deficits in working memory, regulation of affect, internalisation of speech, and reconstitution of behaviour, could be conceptualised as a primary working memory deficit, based on failure of adequate PFC connectivity. The present WM model based on PFC-related circuits, differs from the more traditional psychometric WM measures, (such as Tower of London, which may involve a number of abilities), and is rather based on the capacity to maintain representational activity until a goal is achieved. 150 CHAPTER 4. DEVELOPMENT AND WORKING MEMORY Chapter 5

The emotional brain

Papez (1937) proposed the “hypothalamus, anterior thalamic nuclei, gyrus cinguli, hippocampus, and their connections” as a system, which elaborated the functions of emotion and partici- pated in emotional expression. This concept was further elaborated by MacLean (1949, 1952). According to Heimer et al. (2008); MacLean (1949) based his concept of the limbic system on his own studies of temporal lobe epilepsy and the Papez proposal. Heimer et al. (2008, p69-100) describe the cortical limbic lobe as derived from Broca (1878)’s description of a ring or limbus of each hemisphere encircling the brainstem and forming its medial-most edge or border. According to Heimer, when Papez formulated his famous circuit for emotion, involving major portions of the limbic lobe, little was known about sensory input to the circuit other than olfaction. However, in the last four decades, it has been learned that sensory inputs from all modalities inﬂuence the limbic lobe, and sensory information coverges into multimodal association areas of frontal and temporal lobes, sending axons to the limbic lobe, allowing complexity, and integration . The rise of contemporary neuroanatomical and histochemical methods in the 1960’s, 70’s, and 80’s, and the combination of these methods with molecular biological approaches in the 1990’s and 2000’s have fueled an acquisition of exponentially increasing amounts of detailed information about neural pathways and circuits, which may expose the limitations of the limbic system

151 152 CHAPTER 5. THE EMOTIONAL BRAIN

(Heimer et al., 2008, p12). Importantly, Papez and MacLean failed to realise that the major output of their limbic systems was to the basal forebrain, rather than the hypothalamus. Hypothala- mic output is believed by Heimer and colleagues to yield deterministic and stereotypic outputs, be they neural, autonomic or endocrine. Heimer et al. (2008) describe a greater portion of our behaviour as non-deterministic and governed by genetic predis- position, learning, reward, and social norms (integrated at prefrontal cortical levels) This concept differs fundamentally from the Papez/MacLean concept of a limbic subcortical/hypothalamic emotional system. Rolls (1998, p78-81) has postulated that the orbitofrontal cortex is involved in the execution of behavioural responses when these are computed by reward or punishment. Association learning is a function for which the orbitofrontal cortex is specialised in terms of representations of primary (unlearned) reinforcers, and in rapidly learning and readjusting associations of stimuli with their primary reinforcers. The orbitofrontal cortex is thought to encode reward associations of visual, face and non-reward stimuli. This function is similar to, but more effective than amygdala learning, as indicated by the greater difficulty in eliciting reversal from amygdala neurons. According to Rolls, the orbitofrontal cortex links to output systems, which control behaviour, via the striatum depending on how strongly each part of the cerebral cortex is calling for output. Rolls (1998, p75-76) has described an important function implemented by the orbitofrontal cortex as rapid stimulus- reinforcement association learning and the correction of these associations when reinforcement contingencies in the environment change. Rolls defined a positive reinforcer as a stimulus, which the animal will work to obtain, and a negative reinforcer, which an animal will work to escape. According to Rolls, the amygdala is concerned with some of the same functions as the orbitofrontal cortex, and receives similar outputs, but functions less effectively in the very rapid learning and reversal of 153 stimulus-reinforcement associations, and by the greater effect of orbitofrontal lesions in leading to continuing behavioural responses to previously rewarded stimuli. “In primates the necessity for very rapid stimulus-reinforcement re-evaluation, and the development of powerful cortical learning systems, may result in the orbitofrontal cortex effectively taking over this aspect of amygdala functions” (Rolls, 1998, p76). This places greater emphasis on the cortex “taking over” the regulation of subcortical functions. According to Rolls the above theory is very different to that of Damasio (1994), who tried to resurrect a weakened version of the James-Lange theory of emotion from the last century, “by arguing with his somatic marker hypothesis, that after reinforcers have been evaluated, a bodily response (’somatic marker’) normally occurs, which then leads to a bodily feeling, that in turn is appreciated by the organism, to then make a contribution to the decision-making process”. Rolls points out that the James-Lange theory was weakened by evidence that inactivation of peripheral feedback did little to abolish feelings or behaviour to emotion- provoking (reinforcing) stimuli. Rolls (1998) points out that the functions performed by the orbitofrontal cortex need not be performed with explicit (conscious) processing, but can be performed with implicit processing. “This differs from routes involving language, which enable long-term planning, where the plan involves many syntactic arrangements of symbols (e.g. many if [...] then statements)” [...] “It is suggested that this latter process is closely related to explicit, conscious processing” (Rolls, 1998, p79). Schultz et al. (1997) have reported on the mechanism of dopaminergic reward prediction, utilising conditioned experiments. The investigators postulated that reward-dependent learning is based on the unpredictability of the reward, and that no further learning takes place when the reward is entirely predicted by a sensory cue. An analogy from engineered systems is the temporal difference (TD) algorithm, in which artificial systems learn to predict. Similarly dopaminergic activity was 154 CHAPTER 5. THE EMOTIONAL BRAIN thought to encode expectations about external stimuli or reward, when there is an error between the actual reward received in predictions of the time and magnitude of reward. Neurons were activated only if the time of the reward is uncertain or unpredicted by any preceding cues. Thus dopamine neurons were thought to be excellent detectors of the “goodness” of environmental events relative to learned predictions about these events. According to Schultz et al. (1997) they emit a positive signal (increased spike production) if an appetitive event is better than predicted, no signal (no change in spike production), if an appetitive event occurs as predicted, and a negative signal (decreased spike production) if an appetive event is worse than predicted. Schultz et al. (1997) believes the TD algorithm assumes a con- sistency of prediction error through time. If the time of reward delivery is changed relative to cue onset, the same cue will come to predict the new time of reward delivery. Fluctuations in dopaminergic activity are able to represent an important “economic evaluation”, which is broadcast to target structures. Thus the authors conclude that dopamine neurons in the ventral tegmental area (VTA) and substantia nigra report ongoing prediction errors for reward. Delivery of the signal to target structures influences the choice of reward-maximising actions. For example, in the striatum, scalar evaluation may have a direct effect on action choice. The authors believe the striatum may be a site where the dopamine signal influences behavioural choices by modulating the level of competition in the dorsal striatum. In the PFC, dopamine delivery is thought to exert a dramatic influence on working memory, a central PFC ability in maintaining on-line monitoring of behaviour. Hollerman et al. (2000) define rewards as appetitive environmental events that serve as goals for voluntary, intentional behavioural actions. The authors point out that whereas sensory systems utilise specialised receptors for detecting specific types of physical stimuli in the environment, the neural systems involved in the processing of reward must extract the reward component 5.1. EMOTION CIRCUITS 155 for a variety of different environmental stimuli. Three brain regions have, according to the authors, been shown to be of particular importance in goal-directed behaviour. These are, according to Hollerman et al. (2000), the ventral mesencephalon, the ventral striatum and the orbitofrontal cortex. While these regions are in- terlinked, there appears to be a hierarchical progression in the degree of abstraction of reward processing, from mesencephalon to orbitofrontal cortex. Primate experiments, in which the activity of neural discharges were monitored continuously during rewarded discrimination tasks, in which contingencies and timing were varied. Thus Mirenowicz and Schultz (1997) and Hollerman et al. (1998) were able to show that the response of dopamine neurons in the ventral mesencephalon were specific to rewards, but also sensitive to the predictability of the occurrence of a reward. Thus response magnitude was significantly greater during learning of novel stimuli before criterion was met, than when reward occurrence became predictable. The dopamine neurons were also sensitive to the timing of the occurrence of a predicted reward. The authors suggest that the dopamine response at this level might play the role of a “teaching signal” or an appetitive “tag” for conditioned stimuli reflecting the initial representation of a goal as appetitive.

5.1 Emotion circuits

Mayberg et al. (1999) pointed out that theories of human behaviour from Plato to Freud have repeatedly emphasised links between emotion and reason attributed to pathways connecting phylogenetically “old” and “new” brain regions. The authors utilised positron emission tomography (PET) to clarify these mechanisms. They carried out two experiments - the first with healthy subjects, who were asked to recount scripts of two recent sad personal experiences - and the second in a selected group of unipolar depressed patients treated with fluoxetine. Regional cerebral blood flow was measured in sad and neutral states, and 156 CHAPTER 5. THE EMOTIONAL BRAIN in depressed vs remitted dysphoria and other syndromal features. The results demonstrated increases in limbic-para-limbic (subgenual cingulate and anterior insula) and decreases in neocortical (right dorsolateral prefrontal and inferior parietal) regions). With recovery from depression, the reverse pattern involving the same regions was seen - limbic metabolic decreases and neocortical increases. The investigators concluded that the presence and maintenance of functional reciprocity between these regions with shift in mood in either direction suggested that these interactions were obligatory, with the possibility of both “top-down” and “bottom- up” effects on interventions such as cognitive behaviour therapy (CBT), surgical interventions, or pharmacotherapy. According to Grace (2001), the amygdala is involved in emotional or affective properties of stimuli, enabling the subject to respond to events that are emotionally charged, and therefore of immediate survival value. In pathological states, the amygdala input may be overdriven to the extent that the maintenance of focus is overly disrupted by minor events. As described by Arnsten (2000), too high levels of dopamine and norepinephrine may have additive effects on information processing in PFC, reducing signals and increasing noise. Arnsten points out that although PFC functions are often essential for successful organization of higher- order behaviour, there may be some conditions, when it may be adaptive to ‘shut down’ these complex, reflective operations and to allow more automatic or habitual responses, dependent on posterior cortical or subcortical structures to control behaviour. Phillips et al. (2003b) have reviewed the neurobiological basis of emotion perception in terms of three related processes. These are described as (1) the identification of emotionally salient information, (2) the production of affective states in response, and (3) the regulation of the emotional state. Based on animal and human studies, the authors describe a ventral system, including the amygdala, insula, ventral striatum and ventral regions of the anterior cingulate gyrus and prefrontal cortex, predominantly important for processes (1) and (2), and automatic regulation of emo- 5.1. EMOTION CIRCUITS 157 tional responses; and a dorsal system, including the hippocampus and dorsal regions of anterior cingulate gyrus and prefrontal cortex, predominantly important for process (3). Thus the ventral system is important for the identification of the emotional significance of environmental stimuli, production of affective states, and automatic regulation of autonomic responses to emotional stimuli, while the dorsal system is important for executive functions, including selective attention, planning, and effortful rather than automatic regulation of affective states. Herba and Phillips (2004) discussed developmental studies of amygdala activation to fearful faces, and pointed out that with age, there is increased prefrontal and decreased subcortical activity. Thus early in development, children may have difficulty labelling neutral faces, which may be interpreted as ambiguous, giving rise to amygdala activation, whereas adults show greater amygdala activation in response to fearful facial expressions. Skuse (2003) has described an important evolutionary connection between emotions, feelings and ability to interact appropriately in social situations. He points out that the amygdala responds specifically to eye contact in adults and is maximally activated by exaggerated wide-open eyes, associated with fearful expressions. Thus direct eye contact in primates can elicit an instinctive ‘fear- response’ of fight or flight. However, in humans, response to a stimulus that could be a threat is normally determined by a full evaluation of that stimulus by means of complex neocortical connections, allowing an appropriate social response. According to Skuse, a crucial component of the modulating circuitry in humans is the recruitment of language centers, and the conscious processing of a ‘feeling’ response. Thus neural circuitry that evolved for the purpose of fear detection in others’ faces are now associated with the development of social skills. LeDoux (2002, p156-157) studied how fear responses are coupled to specific stimuli, namely fear conditioning. He found that when a sound is presented to a naive animal, it reaches the lateral nucleus of the amygdyla (transmitted from sensory areas in 158 CHAPTER 5. THE EMOTIONAL BRAIN the thalamus and cortex) and mildly activates neurons there.

Inhibition, by GABA, prevents much from happening in response, and if the sound is repeated without consequence, the cells quickly stop responding. But if the sound is followed by a shock, the weak presenting pre-existing response is greatly ampliﬁed, because the shock activates the postsynaptic cell, while the sound is causing the presynaptic cell terminals to release glutamate. During this activation, a relation between the presynaptic and postsynaptic neurons is stabilised, and the sound acquires the ability to elicit strong activation in the amygdala. (LeDoux, 2002, p156-157).

According to LeDoux (2002), fear conditioning by the amygdala is implicit (unconscious). However, during any experience in which we are awake, working memory is aware, if what is going on is significant, then the executive directs storage of information about the situation in the explicit memory system, allowing later conscious recall into working memory. LeDoux et al. (1989) suggested that emotional memories established via the thalamo- amygdala pathways may be relatively indelible. Contextual fear conditioning, which involves more complex stimuli from multiple sensory modalities, may require projections to the amygdala from higher-order cortical areas that integrate inputs from many sources and the hippocampus. Williams et al. (2004) have used a novel technique for simultaneous neuroimaging (fMRI) and skin conductance response (SCR) recording, in which both brain responses and SCR’s are time- locked to individual stimuli, providing a reliable measure of phasic increases in autonomic responses. They were able to examine the time course of cortico-amygdala and autonomic arousal integration in response to implicit perception of fearful faces. Over the experimental time course, changes in autonomic arousal (in- dexed by SCR’s) followed a “U-shaped” profile in which they were most frequent in the early phase of stimulus processing, declined during the middle phase and persisted during the final phase. The corresponding time course of brain activity highlighted three regions: relatively greater engagement of somatosensory-related cortices with early SCR’s, the medial prefrontal cortex with 5.1. EMOTION CIRCUITS 159 middle-phase SCR’s, and the amygdala with persistent SCR’s during the final phase. The authors interpret their findings as suggesting that sustained cortico-amygdala and autonomic responses may serve to prime the emotional content of fear signals, and differentiate them from the initial novelty reaction. Furthermore, Williams et al. (2001) utilised the above techniques to measure phasic arousal while subjects viewed fearful versus neutral faces. Interestingly, amygdala/medial frontal activity was observed only with SCR’s, whereas hippocampal- lateral frontal activity occurred only in the absence of SCR’s providing evidence for a dissociation between human visceral-emotional (amygdala) versus declarative fact (hippocampal/cortical) processing of fear. Thus the authors differentiate between an immediate fast visceral response to fear stimuli (necessary for a survival response) and more detailed and involved processing necessary in order to consciously see and process the stimulus. This approach is consistent with the above concepts of a differentiation between visceral subconscious and conscious higher level processes. An interesting application of ‘affective neural circuit’ theory is found in the work of Schore (1994) in his extensive text on ‘Affect Regulation and the Origin of the Self’. Schore’s hypothesis represents an attempt to align psychoanalytic theory, particularly attachment theory with contemporary neuroscientific research. In so doing he examines the child’s early years and in particular “the orbitofrontal generation of affect-regulating internal representations” (Schore, 1994, p177-182). Schore describes a major maturational change in the prefrontal cortex at 10-12 months. He draws on the work of Goldman-Rakic (1987c) and states:

The orbitofrontal cortex is known to subserve cognitive and memory functions and to generate mental images and faces [...] A major function of prefrontal circuits is to access and hold on line representational short-term knowledge of the outside world (Goldman- Rakic, 1987a). These fundamental capacities enable it to be fundamentally involved in the generation of internal representations of the self, interacting with the outside world, especially the pri- 160 CHAPTER 5. THE EMOTIONAL BRAIN

mordial other, who mediates the social world. [...] The concept of internal representation of external object relations of models of the nascent self, interacting with the early social environment, is now being utilised by a number of different disciplines” (Schore, 1994, p179).

Schore (2003, p59-63) has discussed the contributions from neuroscience to attachment theory. According to Schore:

The limbic system has been suggested to be the site of developmental changes associated with the rise of attachment behaviours. [...] face-to-face transactions of affect synchrony between caregiver and infant directly influence the imprinting, the circuit wiring of the orbital prefrontal cortex, a cortico-limbic area that is known to begin a major or maturational change at 10-12 months, and to complete a critical period of growth in the middle to end of the second year. [...] This fronto-limbic system provides a high-level coding that flexibly coordinates exteroceptive and interoceptive domains and functions to correct responses as conditions change [...] the integrity of the orbitofrontal cortex is necessary for acquiring very specific forms of knowledge for regulating interpersonal and social behaviour. [...] Due to its location at the ventral and medial hemisphere surfaces, it acts as a convergence zone, where cortex and subcortex meet [...] But the orbital frontal system is also deeply connnected into the autonomic nervous system and the arousal- generating reticular formation, and due to the fact that it is the only cortical structure with such direct connections, it can regulate autonomic responses to social stimuli, and modulate instinc- tual behaviour (Schore, 2003, p61).

Schore (2003, p61) describes the orbital prefrontal region as especially expanded in the right hemisphere, which is specialised for “inhibitory control”, and it comes to act as an executive control function for the entire right brain. This hemisphere, which is dominant for unconscious processes, computes on a moment-to- moment basis, the affective salience of external stimuli. “It also contains a ‘nonverbal affect lexicon’, a vocabulary for nonvebal affective signals, such as facial expressions, gestures and vocal tone or prosody. The right hemisphere is thus faster than the left in performing valence-dependent, automatic, pre-attentive appraisals of emotional facial expressions” (Schore, 2003, p61). 5.1. EMOTION CIRCUITS 161

5.1.1 Comment

Schore’s hypothesis is of present interest for a number of reasons. First, Schore mistakenly cites Goldman-Rakic as basing her conclusions in relation to working memory on the orbitofrontal cortex, whereas her work is based on the dorsolateral prefrontal cortex, that is, the dorsal rather than ventral system. This is important as Goldman-Rakic’s work is more related to cognitive, rather than emotional development, and Schore’s assumptions violate the modularity of these circuits, unless an interactive mechanism is provided. Secondly, the Rolls criticism of Damasio, could also be made in relation to Schore, namely that his theory represents a weakened version of the James-Lange theory of emotional development, which assumes a limbic/emotional circuit, based on integration of unconscious autonomic responses, at early stages of development. Schore musters diverse studies of the “right brain” to support a basically psychodynamic theory of child and adult development based on a “right brain” explanation of Bowlby’s attachment theory. Nonetheless his emphasis on brain circuit pathology has drawn attention to the need to integrate neurobiological and clinical ﬁndings.

5.1.2 Basal forebrain organisation

Mogenson et al. (1980) described the role of the nucleus accumbens in functioning as an interface between limbic and motor systems (from motivation to action). Mogenson quoted Graybiel’s findings that the nucleus accumbens is a key structure in linking motivation and action at the interface of the limbic system with motor mechanisms, receiving direct connections from amygdala, hippocampus and other limbic forebrain structures, as well as indirect connections via mesolimbic dopaminergic projections from the ventral tegmental area. “The nucleus accumbens has direct motor connections to the globus pallidus and indirect connections via the substantia nigra and nigrostriatal dopaminergic system. Mesolimbic dopamine projections to the nucleus accum- 162 CHAPTER 5. THE EMOTIONAL BRAIN bens in rats were also implicated by abolition of the hyperactivity effect of systemically administered amphetamine by damage to ventral tegmental projections to the nucleus accumbens by injections of 6- hydroxydopamine” (Mogenson et al., 1980). Heimer et al. (2008, p15-16) reviewed developments leading to the “eroding relevance of the limbic system”. He describes how MacLean (1990) in his text on ‘The Triune Brain in Evolu- tion’ suggested that a paleo-mammalian brain, represented by the Limbic System, (important in emotional behaviour, feeding, repro- duction and parenting) was added in early mammals to the already existing reptilian brain (basal ganglia, important for daily routines, ritual displays, aggression, territoriality and courtship). This was thought to be followed in modern mammals by the development of the mammalian brain (neocortex related to problem solving and cognition). While the concept of a hierarchical brain structure was and remains influential, Heimer believes the concept to be fundamentally incorrect, as demonstrated by later improved histotechnical advances, which showed that sensory systems in non-mammalian vertebrates have telencephalic (cortical) representations. A further “incorrect” thesis is the suggested close relationship of the limbic system to the hypothalamus, leading to a distinction between the limbic system and the basal ganglia. Importantly, Heimer et al. (2008, p21) point out that the realisation that the whole cortical mantle projects to subcortical basal ganglia structures allows for a closer integration of neurological and psychopathological thinking, than in the past. The concept that the subcortical telencephalic nuclei are recipients of massive inputs from not only isocortex (neocortex), but also the non- isocortical (limbic lobe) parts of the cerebral cortex provides a better way of understanding, in neuroanatomical terms, clinical phenomena from epilepsy to schizophrenia and from “laugh to cry”. Heimer et al. (2008, p24) describe the ‘ventral striatopallidum’ and ‘extended amygdala’ as essential concepts, which presaged new thinking about forebrain organisation. The demonstration that the ventral parts of the striatal cortex, including the 5.1. EMOTION CIRCUITS 163 accumbens and striatal territories in the olfactory tubercle, extended to the ventral surface of the brain and received allocortical and mesocortical (i.e. “limbic” cortical) inputs, just as the dorsal parts of the striatal complex receive massive projections from the neocortex (isocortex) was central in the 1970’s, to an understanding of the ventral and dorsal topographically organised cortico- striatopallidal-thalamocortical re-entrant circuits (Santini, 1975) . Thus the concept of the accumbens as a “limbic-motor” interface (Mogenson et al., 1980) was replaced by viewing the accumbens as an integral component within the large system of cortico- basal ganglia-thalamocortical re-entrant circuits, which also include the caudate, putamen (’motor’ centers) and olfactory tubercle. Heimer points out that if it is considered that most, if not all, motor disorders have psychiatric symptoms, and all psychiatric disorders reveal to the careful observer, some accompanying motor component, this allows for greater merging of neuropsychiatric and neuromotor understandings. A second important discovery of the 1970’s was, according to Heimer et al. (2008), the identification of a prominent neuroanatomical continuum, which included the centromedial nucleus of the amygdala and bed nucleus of the stria terminalis, establishing a second “emotional-motor” interface (Heimer et al., 2008, p62). “The concept of the ‘extended amygdala’ views the bed nucleus of the stria terminalis and centromedial amygdala as part of one and the same macrostructure, and the outputs from both are thought to represent an effector structure, strategically placed to coordinate multiple cortical limbic lobe regions for the development of behavioural responses through its output channels”. The third fundamental concept outlined by Heimer et al. (2008, p69-100) is that of the cortical ‘greater limbic lobe’, consisting of cingulate gyrus dorsally, and parahippocampal gyrus ventrally, and bridged by smaller areas of cortex, to complete the cortical limbus (Broca, 1878). Sensory information to the limbic lobe is derived from visual modalities, via temporal association 164 CHAPTER 5. THE EMOTIONAL BRAIN cortices before relay to perirhinal and posterior parahippocampal cortices and basaolateral amygdala. According to Heimer, these pathways are critical for many forms of behaviour, including fear, face recognition, declarative memory and object recognition. The hippocampus and amygdala are described by Heimer as key players in the limbic lobe. Other important connections to the greater limbic lobe from association areas are described from the somatosensory, parietal, insula and precuneus. Heimer et al. (2008, p86) also describes limbic lobe reciprocity with all of the projections it receives, including reciprocity with prefrontal and multimodal association areas. A major output of the limbic lobe is described by Heimer as directed to the basal forebrain (rather than the hypothalamus). Heimer (2003) was able to show that the hypothalamus was only one of a number of effector organs. Thus while hypothalamic output is described by Heimer as deterministic and giving rise to focussed and stereotypic behaviours, the limbic lobe and basal forebrain are largely concerned with non- deterministic predispositions, governed by societal norms, learning and reward, providing a cognitive base for feelings and emotions, and a link with memory mechanisms.

Heimer and Van Hoesen (2006) discussed the implications for emotional functions and adaptive behaviour of the cortical limbic lobe and its output channels, which have been defined as a result of histochemical advances. He describes two major systems, the ventral striatopallidal system and extended amygdala, which together with the basal nucleus of Meynert and the septum diagonal band system serve as output channels for an expanded version of the “classic limbic lobe of Broca”. He points out that thus defined, the limbic lobe contains all the major cortical, namely orbitofrontal, cingulate and insular cortices, in addition to hippocampal formation and “cortical-like” structures (laterobasal cortical amygdala), known to be important for emotional and motivational functions. Heimer points out that the concept of the limbic system posited by MacLean (1952), while heuristic in the past, has been outdated by neuroanatomical and functional 5.1. EMOTION CIRCUITS 165 advances, but the restriction of the greater limbic lobe to the cortical mantle continues to be important. Heimer and Van Hoe- sen (2006) distinguishes the ventral striatum and ventral striatopallidal system as being continuous with the accumbens. The ventral striatum was originally was originally defined, based on its input from allocortex (olfactory cortex and hippocampus) and amygdala, but also receives input from anterior cingulate and insula, and is thus largely related to allocortex. Heimer suggests that the ventral striato-pallidal (emotional-motivation functions) system cannot strictly be separated from dorsal prefrontal projections, serving cognitive-executive functions, while dorsolateral prefrontal projections to striatum, overlap with projections from motor cortical areas.

cortex

d. striatopallidum v. striatopallidum extended amygdala

reticular formation

motor effectors

Figure 5.1: Major basal forebrain macrosystems, Adapted from Zahm, 2006, Heimer et al. 2008

Heimer and Van Hoesen (2006), described the discovery that the ventral pallidal neurons project to the medio-dorsal thalamic nucleus, rather than the ventrolateral thalamic ‘motor’ nucleus, as leading to the concept of parallel cortical, basal ganglia, thalamo-cortical re-entrant circuits. As described above, this notion was further developed by Alexander et al. (1986), so that six major circuits are now recognised, two of which are motor and oculo-motor (frontal eye field) and four relate to prefrontal cortical areas (an executive circuit originating in the dorsolateral pre- 166 CHAPTER 5. THE EMOTIONAL BRAIN frontal cortex, an anterior cingulate circuit, originating in the anterior cingulate, and medial and lateral orbitofrontal circuits, all of which return to their area of origin in the cortex. Thus the ventral striatopallidal system (limbic loop) is thought to choose and mobilise ‘reward-guided’ behaviour (Schultz et al., 2000), while the executive circuit (and its subcircuits) is implicated in planning and working memory. Heimer and Van Hoesen (2006) have also raised the question of the closed vs open transfer of information between circuits, also suggested by Haber (2003). The notion of spiralling pathways between the ventral striatum and the rest of the striatal complex, with the mesencephalic dopamine system, provides a suggestive mechanism for cross-modular information transfer. In addition to the ventral striatum, Heimer and Van Hoesen (2006) describe the ‘extended amygdala’, consisting of a ring-like structure, which includes the central and medial amygdaloid nuclei and their extensions along the arc of stria terminalis, linking to the bed nucleus of stria terminalis. The extended amygdala is thought to receive significant inputs from the latero-basal amygdala and other parts of the limbic lobe. The system is characterised by long associative connections and has prominant projections to autonomic and somato-motor centers in the hypothalamus, brainstem, and to endocrine-related hypothalamus. While the concept of extended amygdala has been questioned (Swan- son, 2003), Heimer and Van Hoesen (2006) point out that the system is rich in neuropeptides, including opiods, cholecystokinin, neurotensin, somatostatin, corticotrophin releasing factors and receptors for oxytocin/vasopressin and androgens, important in emotion-related functions. Heimer and Van Hoesen (2006) describe the basolateral amygdala as associated with the hippocampal para-hippocampal part of the limbic lobe, and thus of importance in memory consolidation. “Massive projections from the basolateral amygdala to the anterior hippocampus (Saunders et al., 1988), and to the striatum, especially the ventral striatum, are well positioned to trans- 5.1. EMOTION CIRCUITS 167 mit the modulating effects from the basolateral amygdala to the hippocampal and striatal-dependant memory processes”. “In the primate brain, and especially in humans large bundles of axons leave the cingulum and turn dorsally into the parietal association cortex”, possibly carrying limbic lobe output. “In this sense, feedback connections from the limbic lobe could provide part of the anatomical underpinning for well-known persistence of emotion- related memories and their resistance to decay over a lifetime”. Finally, Heimer and Van Hoesen (2006) discuss the relationship of the cortical limbic lobe to emotion, and conclude that within the cortical mantle, it is the basolateral amygdala, anterior amygdala, anterior insula, caudal orbitomedial prefrontal cortex, and anterior cingulate gyrus, which are of special significance for emotional functions. The importance of subcortical structures and circuits for emotional functions is thought “unde- niable”.

Together with the basal nucleus of Mynert and the septum- diagonal band system, the ventral striatopalidal system and extended amygdala serve as output channels for the limbic lobe, of which the large cortical-like laterobasal cortical amygdala is an important component. [...] The central role of the large baso-lateral amygdaloid complex in appraising and identifying the emotional signiﬁcance of sensory stimuli, underscores the heuristic value of this expanded version of Broca’s great limbic lobe (Heimer and Van Hoesen, 2006).

5.1.3 Comment

Heimer’s concept of a greater cortical limbic lobe in addition to neocortex, projecting to striato-pallidum both simpliﬁes and allows for integration of cognitive and emotional circuits. The concept of the Limbic System, as developed by MacLean (1949) was thought central to emotional development. MacLean believed that the limbic system output was to the hypothalamus, but modern staining techniques by Heimer (2003) and others have shown that this was incorrect, and that the limbic forebrain projects to the ventral striatum (accumbens) and extended amygdala. 168 CHAPTER 5. THE EMOTIONAL BRAIN

Heimer has shown that this system allows the amygdala to in- ﬂuence motor and somato-motor output. Zahm (2006) has pointed out that an appreciation of the neuroanatomical composition of basal forebrain is fundamental to understanding its function. Alheid and Heimer (1988) proposed that the ventral striato-pallidum, extended amygdala, septum (with associated structures) and magnocellar corticopedal system represent separate processing units that receive speciﬁc information derived largely from the cortex and provide “either local [...] or distal feedback to cortex” and output to brainstem motor effectors. He notes that the terms ‘functional-anatomical system’ and the synonymous term ‘macrosystem’ emerged in subsequent papers (Heimer, 1991; Heimer et al., 1991). Despite differences in neurochemical, connectional and functional details Alheid and Heimer (1988) pointed out an underlying organisational framework or “template” that is shared by all macrosystems. They proposed an important model, according to which, inputs bearing discrete information, largely from cortex terminate within input structures comprising predominantly medium-sized, densely spiny inhibitory (i.e. GABAergic) neurons. Among such input structures would be included, e.g. the dorsal and ventral striatum, extended amygdala and lateral septum. These in turn, project prominantly to structures comprising mostly medium to large sparsely-spined GABAergic neurons dispersed in the pallidum, extended amygdala and preoptic area, that project to the cortex, via thalamo- cortical re-entrant pathways, and to autonomic and somatic motor effectors in the hypothalamus and brainstem. In addition, the authors proposed that the macrosystems establish connectional relationships with forebrain and mesencephalic cholinergic and dopaminergic neuromodularory cell groups, providing additional thalamic and extra-thalamic feedback loops to cortex and telencephalic nuclei. Heimer et al. (2008, p136-137) point out that the hypothesis that neural correlates of affect are generated subcortically, is consistent with a wealth of experimental data. “From this point of 5.1. EMOTION CIRCUITS 169

cortex cortex

dorsal and ventral extended amygdala striatum

cholinergics cholinergics

dorsal and ventral extended amygdala pallidum

dopaminergics dopaminergics

thalamus and thalamus and brainstem brainstem

Figure 5.2: Striatopallidum and Extended amygdala. Adapted from Zahm 2006, Alheid and Heimer, 1988, Heimer et al. 2008

view, subcortical information processing systems are substrates, within which outputs from all parts of the cortex are processed prior to engaging multisynaptic pathways to (1) hypothalamic and brainstem somatomotor and visceromotor effectors and (2) back to the cortical cognitive apparatus via “re-entrant trajectories”. Heimer and colleagues describe the striato-pallidal complex, extended amygdala and septum-preoptic systems, as having organisational characteristics, that while fundamentally striatopallidal are nonetheless highly distinct. “It is well established that each of these functional-anatomical constructs is innervated by the same high-level associational areas, comprising the greater limbic lobe, including the prefrontal cortex, basolateral amygdala, and ventral hippocampus, as well as dopaminergic and “nonspe- 170 CHAPTER 5. THE EMOTIONAL BRAIN cific” thalamic projections” (Heimer et al., 2008, p136-137). As described by Zahm (2006) each macrosystem relates preferentially to specific parts of the cholinergic and dopaminergic ascending neuromodulatory projection systems. Lieberman and Eisenberger (2006) have postulated that dACC activity is related to the experience of social and physical pain. The authors point out that while clinical and psychopathology researchers might regard the function of dACC as producing attention-getting affective-maturational states, such as pain, anxiety and distress, cognitive researchers are likely to focus on the monitoring of conflict and detection of error, or a discrepancy between one’s goal and one’s prepotent response. The dACC is postulated as “notifying” lateral prefrontal cortex that top-down control processing is necessary to promote contextually appropriate responding (Botvinick et al., 2001). A review by Bush et al. (2000) concluded that cognitive tasks tended to activate dACC, and deactivate rostral ACC (rACC). Lieberman and Eisenberger (2006) also postulate that dACC and rACC may differ in regard to the extent to which a conflict is represented symbolically or non-symbolically. The authors suggest that symbolic processes conscious awareness of and attention to a specific symbolic representation. They point out that the resources of awareness and attention are limited, such that only a handful of symbolic representations can be attended to and processed at any one time, and are typically thought to be processed serially. “Thus, although many conflicts may be non- symbolically processed, only a single symbolic conflict can be processed at any one time”. Lieberman and Eisenberger (2006) propose that symbolic conflict is processed primarily in the rACC, whereas non-symbolically represented conflict is processed primarily in dACC.The authors suggest that cognitive, affective, and pain processes are each distributed across dACC and rACC, and are organised such that more symbolic forms of cognition, emotion and pain are processed in rACC, with less symbolic forms processed in dACC. They conclude that because of the role of so- 5.1. EMOTION CIRCUITS 171 cial attachment in mammals, the social pain system may have piggy-backed onto the physical pain system during evolution, and that the dACCmay have been one of the primary sites in which overlap evolved, and today the region produces similar experience of distress in response to both physical and social injuries. Posner et al. (2009) have postulated a dual “circumflex model” of affect, in which all emotions are described as a linear combination of two underlying, largely independent neurophysiological systems, valence and arousal. “The valence system determines the degree to which an emotion is pleasant or unpleasant, and the arousal system determines the degree to which it is behaviourally activating. Yang and Seamans (1996) described the effects of ventral tegmental area (VTA) activation on prefrontal (PFC) neurons and networks. In control conditions where there is little dopaminergic (DA) tone, VTA neurons fire at a low basal rate, until a better than expected rewarding stimulus is encountered, giving rise to bursting in VTA neurons, and release of glutamate in the PFC. This gives rise to an “up” state in the PFC, which terminates in a few hundred milliseconds, but if large levels of extra-synaptic DA are present, supra-optimal D1 receptor stimulation may push the PFC into a “locked state” with stereotypic thoughts or actions. Psycho-pathological symptoms can thus occur either as a result of either too high or too low neurotransmission (arousal) at synaptic gates situated in both affective and motor circuits (inverted-U phenomenon).

5.1.4 Discussion

While the above studies were mostly conducted in adults, they demonstrate cogent examples of the importance of cortical- subcortical functional coupling, as well as dual ‘immediate’ vs ‘executive’ systems of behavioural and emotional control. As Heimer et al. (2008, p53) contends, a consensus seems to have emerged that the cortico-subcortical re-entrant circuit (and its sub-circuits) through the ventral striatopallidal system, or “lim- 172 CHAPTER 5. THE EMOTIONAL BRAIN bic loop” is critical for selecting and mobilising appropriate behaviour, and in the same vein, the “executive” circuit (and its subcircuits) is implicated in action planning and working memory’. Thus an understanding of the neuroanatomical composition of basal forebrain and related circuits is fundamental to the understanding of both emotional and cognitive control processes. The suggested distinction by MacLean between the limbic system and the basal ganglia is replaced by a number of cortical-subcortical circuits which include dorsal and ventral cortico-striatal-thalamic and cortico-amygdala-thalamic circuits. Thus the ventral systems are thought to choose and mobilise ‘reward-guided’ behaviour (Schultz et al., 2000), while the dorsal executive circuit (and its subcircuits) is implicated in planning and working memory. The macrosystem morphology allows for both modular separation and lateral integration of cognitive and emotional processes, as suggested by the work of Haber (2003), as well as providing a better understanding of the pharmacological transmitters involved, in each system. The macrosystem architecture, which utilises direct excitatory and indirect inhibitory circuit structures and a number of processing units that collectively receive information from the entire cortical mantle, and project either to somatic and autonomic effectors in the lateral hypothalamus and brainstem or, via re-entrant pathways, back to the forebrain, could allow the classification of childhood behavioral syndromes discussed below in terms of dorsal and ventral striato-pallidal and extended amygdala, (utilising dopaminergic, serotonergic and noradrenergic neurotransmission respectively). However, the important feature of chilhood symptomatology is the consideration of developmental factors influencing the establishment of flexible macrosystem functioning.

5.2 Social cognition and neural circuits

While the above macrosystems underlie individual cognition, development occurs in a social context. As discussed in Chapter 2 5.2. SOCIAL COGNITION AND NEURAL CIRCUITS 173 in relation to language, Gallese et al. (2004) have outlined a “unifying view of the basis of social cognition”. They point out that the conventional conceptual approach for understanding actions performed by others, is to consider observed actions in a similar way to all other visual stimuli. However, the authors outline a different approach, based initially on the discovery of mirror neurons, in primates, located in the ventral premotor cortex (F5). They describe these neurons as responding when a monkey performs a particular goal-directed action. Their “core proposal” is that the observation of an action leads to the activation of parts of the same cortical neural network that is active during its execution. Thus the observer “understands the action” because of his/her motor knowledge of the action and its potential outcome (Gallese et al., 119); (Rizzolatti et al., 1996). The authors additionally tested monkeys in two conditions, in one of which the monkey could see an entire action (e.g. a hand grasping action), but subsequently the hand-object interaction was hidden by a screen. The mirror neurons were shown to also respond in the hidden condition (Umilta et al., 2001). A similar result was recorded when monkeys heard an action, after observing it (Kohler et al., 2002).

Gallese (2006) has discussed the implications of his theory of social cognition for a neurophysiological perspective on “intentional attunement” in Autism. He points out that in primates, the capacity to understand conspecifics’ behaviours as goal-related provides considerable benefits to individuals, as they can then predict, influence, and manipulate the behaviour of conspecifics. Similarly the traditional view in the cognitive sciences holds that human beings are able to understand the behaviour of others in terms of their mental states. “The capacity for attributing mental states, beliefs and desires to others has been defined as Theory of Mind (TOM) (Premack and Woodruff, 1978). “Accord- ing to this perspective, social cognition becomes almost synonymous with mind reading abilities”. Gallese (2006) argues that social cognition is not only “social metacognition” (thinking about the contents of someone else’s mind by means of abstract repre- 174 CHAPTER 5. THE EMOTIONAL BRAIN sentations), but there is also an experiential dimension of interpersonal relationships, which enables a direct grasping of the sense, emotions and sensations experienced by others. Gallese describes this mechanism as “embodied simulation”. Koziol and Budding (2009, p175) point out that while the role of mirror neurons in the ventral and rostral posterior parietal cortices has received extensive study in relation to empathy and theory of mind (Rizzolatti et al., 2006), dopamine-mediated reward systems in the ventromedial prefrontal lobe and ventral striatum play a significant role in social valuation (Fliessbach et al., 2007). Koziol and Budding (2009) believe that social cognition can be broadly organised around two systems.

There is a system of automatic social skills that develops through instrumental conditioning and procedural learning. This system governs the stimulus-based social skills that are necessary for successful routine interactions. There is also a system of higher-order social processing that allows us to think about what to say or do, to consider the thoughts, ideas and feelings of others, to manipulate or comply with situations, and to reﬂect upon ourselves. These two systems can operate independently, but they mostly interact, since most social circumstances require alternating implementation of these two systems [...] Failures in social behaviour can result from impairment of the “automatic” stimulus-based processing system or because of failure in higher-order social processing (Koziol and Budding, 2009, p175-177).

Koziol and Budding (2009, p171) point out that in addition to gesture and language, “non-verbal” communication such as body language and intuition are equally important to adaptive socialisation. “Categorisations that generate positive or negative reinforcement, or survival “rewards” routinely recruit the basal ganglia. This type of learning need not be conscious. In fact the data reviewed imply that the basal ganglia are essential in reward- related learning, and that this is independant of higher-order control systems”. According to Gallese (2006), one important difference between humans and monkeys could be the higher level of recursivity attained by the mirror neuron system for actions in our species. “A quantitative difference in computational power and degree of 5.2. SOCIAL COGNITION AND NEURAL CIRCUITS 175 recursivity could produce a qualitative leap in social cognition” (Hauser and Fitch, 2004; Hauser et al., 2002). Intentional attunement is thought by Gallese (2006) to enable a direct grasping of the sense of the actions and emotions experienced by others. Thus empathy “ [...] entails the capacity to experience what others do experience, while being able to attribute those shared experiences to others and not the self. This is achieved by “intentional attunement”. Gallese describes “embodied simulation” as a speciﬁc mechanism by means of which our brain/body system models its interactions with the world. He suggests that embodied simulation mechanisms may be crucial in the course of the long learning process required to become fully competent in how to use propositional attitudes, like during the repeated exposure of children to the narration of stories. He suggests “embodied simulation” is also at play during language processing (Gallese, 2005).

Gallese (2006) also suggests that the defects observed in Autis- tic Spectrum Disorders (ASD) may be ascribed to a deficit or malfunctioning of “intentional attunement”, because of a malfunctioning of embodied simulation mechanisms, produced in turn by a dysfunction of the mirror neuron system. Gallese describes an fMRI studies by (Dapretto et al., 2006; Wang et al., 2004a), which demonstrated that during observation and imitation of facial expressions, high functioning ASD individuals did not show activation of the mirror neuron system in the pars opercularis of the inferior frontal gyrus (F5) during observation of facial expressions. The investigators found that children with ASD actually showed greater activity than did typically developing children in right visual and left anterior parietal areas (modulated by visual and motor attention, respectively, whereas typically developing children were able to rely on a right hemisphere - mirroring neural mechanism-interfacing with the limbic system via the insula-whereby the meaning of the imitated (or observed) emotion was directly felt or understood. In contrast, Dapretto et al. (2006) point out that when the mirroring mechanism is seemingly not engaged in children with ASD, they must then adopt 176 CHAPTER 5. THE EMOTIONAL BRAIN an alternative strategy of increased visual and motor attention, whereby the internally felt emotional significance of the imitated facial expression is probably not experienced, while typically developing subjects are able to read others’ emotional states from a mere glance. Gallese (2006) suggests that a theory of “intentional attunement”, while parsimonious, provides a neurophysiological mechanism, at the basis of the activation of the cortical regions involved in social cognition. However, he adds that cognitive neuroscience “must seek to know why, how, and because of which functional mechanism, a particular brain area or cortical network happens to be activated in a particular social cognitive task”. Fogassi et al. (2005) has discussed findings which suggest that the posterior parietal cortex integrates multisensory inputs, and also codes motor actions. A number of monkey experiments in which neural activity was recorded, indicated that Inferior Pari- etal Lobule (IPL) neurons code the same motor act differently, depending on whether the goal was eating or placing. Similarly, when the monkeys observed a similar motor act, the same motor neurons were triggered. The authors suggest that in addition to recognising the goal of an observed motor act, the IPL mirror neurons ‘recognise’ the goal of the observed motor act, because it is recognised as part of a chain leading to the final goal of the actions, allowing the intention to be “read”. The authors also note rich connections of the IPL with areas coding biological action (areas of the superior temporal sulcus), and object semantics (inferotemporal lobe). Thus at a basic level, the linkage of action and cognition in a neural circuit “may determine the emergence of complex cognitive functions”. Carr et al. (2003) postulated that if action representation mediation is critical for empathy, and the understanding of the emotions of others, then even the mere observation of emotional facial expression should activate the same brain regions of motor significance that are activated during the imitation of the emotional facial expressions. Moreover activity should be greater during im- 5.2. SOCIAL COGNITION AND NEURAL CIRCUITS 177 itation, compared with observation of emotion, throughout the above network, because of the simultaneous encoding of sensory input and planning of motor output. “Within mirror areas, the inferior frontal cortex seems particularly important here, given that understanding goals is an important component of empathy. The superior temporal cortex would be more active during imitation than observation, as it receives efferent copies of motor commands for mirror areas. The insula would be more active during imitation because its relay role would become more importamt during imitation. Finally, limbic areas would also increase their activity because of the modulating role of motor areas with increased activity”. Carr et al. (2003) tested these hypotheses, utilising functional MRI (fMRI) while subjects were either imitating or simply observing emotional facial expressions. They found that imitation and observation of emotions activated a largely similar network of brain areas, and that within the network, there was greater activity during imitation compared with observation of emotions, in premotor areas, including the inferior frontal cortex, as well as in the superior temporal cortex, insula and amygdala.

Thus the authors describe a “moderately recursive circuit”, connected to the limbic system, via an insula relay. They point out that the anterior insula receives slow-conducting unmyelinated fibers that respond to light caress-like touch, and may be important for emotional and affiliative behaviour between individuals. The activity of the amygdala during imitation, compared with observation of the action representation circuit onto limbic activity was thought to be a critical structure in emotional behaviours and in the recognition of facial emotional expressions of others. “Taken together, these data suggest that we understand the feelings of others, via a mechanism of action representation, shaping emotional content, such that we ground our empathic resonance in the experience of our acting body and the emotions associated with specific movements”. “To empathise , we need to invoke the representation of the actions associated with the emotions we are witnessing. In the human brain, this empathic resonance occurs 178 CHAPTER 5. THE EMOTIONAL BRAIN via communication between action representation networks and limbic areas, provided by the insula. Lesions in this circuit may determine impairment in understanding the emotions of others and the inability to empathise with them”.

Leslie et al. (2004) have expanded the interpretation of mirror activity to include Broca’s area. The investigators utilised fMRI to examine passive and imititative face viewing, as well as hand viewing and imitative movements, in 15 right-handed subjects. They also utilised a verb generation task, to identify language processing areas. The authors questioned whether there is a common imitation circuit, that is independent of response modality, and also whether there would be a dissociation between viewing and imitation of faces. They reasoned that a common imitation circuit would be active during both face and hand imitation, and found common activation in several areas, including left pars opercularis (Broca’s area), bilateral premotor area, right superior temporal gyrus, and medial wall of the superior frontal gyrus. They did not find left pars opercularis activity during passive viewing of either faces or hands, but there was significant activation in the right premotor area, whereas imitation produced bilateral activation. The authors interpreted their data as consistent with evidence for right hemisphere dominance for emotional processing, suggesting a right hemisphere mirroring system, that could provide a neural substrate for empathy. They also suggested that a possible interpretation of their results might be that the left hemisphere pars opercularis (classic Broca’s area) is involved in conscious goal-directed movements, whereas the mirroring involved in unconscious mimicry and empathy is mediated by the right hemisphere and ventral premotor cortex. The goal-directed imitation of faces involved strong bilateral activation, with an important role for the left pars opercularis, which might involve “putting on a face”, or even masking one’s intentions, while passive viewing might involve more unconscious empathic processes. They also suggest the possibility of a “motor theory of empathy”, similar to the motor theory of speech which was thought to derive 5.2. SOCIAL COGNITION AND NEURAL CIRCUITS 179 from representations of motor activity (Liberman and Mattingly, 1985; Liberman et al., 1967; Liberman and Wahlen, 2000). Gray et al. (2002) utilised functional MRI (fMRI) to test the hypothesis that emotional states can selectively influence cognition- related neural activity in the lateral prefrontal cortex. They showed six short videos for induction of emotional states comedy (amusement, pleasant, approach related), documentary (neutral, calm), or horror (anxiety, unpleasant, withdrawal related) films. After each video, the participants were scanned while performing a 3-back working memory task, with either words or faces as stimuli. They showed that word 3-back was enhanced by a pleasant state and impaired by an unpleasant state, whereas face 3-back showed the reverse, indicating a crossover interaction. The neuroimaging data showed no hemispheric specialisation, nor main effects of the stimulus or emotion factors (when treated either as two separate regions or as a single combined region). The pattern of activity in the integration-sensitive ROI across six conditions was a crossover interaction, indicating that the emotion induction selectively influenced task-related brain activity. Neural activity was greatest in the word-unpleasant and face-pleasant conditions, intermediate in the neutral condition and lowest in the word-pleasant and face-unpleasant conditions, meeting all formal criteria for integration. Finally, there were specialised effects of stimulus type in the lateral PFC. Word stimuli were found to preferentially activate the left hemisphere, and face stimuli preferentially activated the right hemisphere. They also found greater right-hemisphere task related activation during the pleasant emotion conditions and greater left-hemisphere task-related activation during the unpleasant emotion condition. The investigators concluded from the emotion x stimulus crossover interaction, with no main effects, and with activity pre- dicting task performance, indicating that “emotion and higher cognition can be truly integrated, i.e., at some point of processing, functional specialisation is lost, and emotion and cognition conjointly and equally contribute to the control of thought and be- 180 CHAPTER 5. THE EMOTIONAL BRAIN havior”. On the other hand, “Other regions in lateral PFC showed hemispheric specialisation for emotion and for stimuli separately, consistent with a hierarchical and hemisphere-based mechanism of integration”.

5.2.1 Comment

Put more simply unconscious ‘mirror’ processes may have relevance for both language and emotional development. Unconscious parallel modules, which “mirror” biologically relevant perceptions such as facial expression, appear to be associated with affective reactions such as fear, disgust or sadness, and are thus fundamental to social competance. It is likely that the concept of unconscious mirroring of facial expressions described above is fundamental to emotional and social interaction, and may provide a model for deﬁcits in emotional development. However the model also requires a cortical component for integration of emotional and cognitive regulation. Koziol and Budding (2009, p182-183) point out that although language functioning and social cognition have traditionally been understood as solely under cortical mediation, they propose a dual-system model, based on rule-based stimulus processing and higher-order control. The mirror neuron view appears to support a ‘theatre in the mind’, based however on observations of actions and emotions of others, presumably encoded cortically, but retrieved or ‘reawakened’ by active participation of subcortical motor and/or visceral centers. It is thus a motor theory and implies both a participation and activity requirement from cortical-subcortical circuits. This ﬁts well with the parallel limbic/basal forebrain macrosystems described above. It also implies a social dimension in addition to cognitive and motivational functions of cortical-subcortical circuits. While mechanisms of mirror processing are not fully understood, the capacity of primates to reproduce and integrate observed emotional and behavioral states appears central to successful social functioning, discussed below in relation to Autism 5.2. SOCIAL COGNITION AND NEURAL CIRCUITS 181 and conduct disorders.

5.2.2 Reappraisal theory

Oschner et al. (2002) have reviewed studies of the human ability to cognitively regulate emotional responses to aversive events. The authors define the term “reappraisal” as “ the cognitive trans- formation of emotional experience”. 1 They suggest that reappraising an aversive event in unemotional terms reduces negative affect with few of the physiological, cognitive or social costs associated with other emotion-regulating strategies, such as the suppression of emotion-expressive behaviour (Jackson et al., 2000). The investigators cite the general findings (Miller and Cohen, 2001; Smith and Jonides, 1999) that cognitive control is thought to involve interactions between regions of lateral (LPFC) and medial (MPFC) and subcortical and posterior cortical regions that encode and represent specific kinds of information “by increasing or decreasing activation of particular representations, prefrontal regions enable one to selectively attend to and maintain goal- relevant information in mind and resist interference (Miller and Cohen, 2001). Oschner et al. (2002) hypothesised that cognitive reappraisal systems would involve three processes implemented by LPFC and medial PFC (MPFC) frontal cortices. 1/ Active generation of a strategy for cognitively reframing an emotional event in unemotional terms, and keeping that strategy in mind as long as the eliciting conditions endured (associated with working memory processes localised in the LPFC) 2/ the second process monitored interference between top-down reappraisals that neutralised affect, and bottom-up evaluations that continued to generate an affective response, signaling the need for reappraisal to continue. These functions were thought to be associated with the dorsal anterior cingulate cortex. 3/ The third process involved reevalu- ating the relationship between internal (experiential or physio-

1Reappraisal is defined as a technique by which individuals reframe the meaning of an event, and thereby change its emotional significance and impact. 182 CHAPTER 5. THE EMOTIONAL BRAIN logical) states, and external stimuli, used to monitor changes in one’s emotional state during reappraisal. The dorsal regions of the MPFC were thought to be activated when making attribu- tions about one’s own or another person’s emotional state. Activa- tion of the MPFC when anticipating a painful shock was thought to be inversely correlated with the experience of anxiety, suggesting regulatory control. Oschner et al. (2002) hypothesised that reappraisal would modulate the processes involved in evaluating a stimulus as affectively significant, via two processes (Scherer et al., 2001). The first type was relatively automatic, and determined whether a stimulus was affectively relevant, whereas the second type was important for evaluating contextual meaning, and the appropriateness of possible responses. Oschsner and colleagues suggest that “two highly interconnected brain structures are associated with these two types of emotion processing. Thus the amygdala was thought to be important for the pre-attentive detection and recognition of affectively salient stimuli, while the medial orbitofrontal cortex (MOFC) was thought important for representing the pleasant or unpleasant affective value of a stimulus (Os- chner et al., 2002). The amygdala and MOFC were postulated to differentially encode and represent the affective properties of stimuli. Oschner et al. (2002) tested the hypothesis experimentally via 4-second presentations of a negative or neutral photo, viewed by subjects on a screen. Participants were instructed to either “attend”, or “reappraise” (interpret photos so that they no longer felt negative in response to them). Those regions more active on fMRI when reappraising than attending were thought to reflect processes used in emotional control, while regions more active for Attend than Reappraise trials were thought to be deactivated by reappraisal. The results showed that, consistent with the predictions, significantly activated regions included the dorsal and ventral regions of the left LPFC, as well as dorsal PFC. Additional activa- 5.2. SOCIAL COGNITION AND NEURAL CIRCUITS 183 tions were found in the left posterior insula, right medial occipital cortex and right inferior parietal cortex. The amygdala showed significant activation at a liberal ( less stringent) threshold. The amplitude of response was greater for the right amygdala on At- tend trials, while the response to most negative photos on the Reappraise trials was not significantly different for the Attend trials. Interestingly, contrasts carried out for a set of functionally defined ROI’s indicated that for MOFC, activation to most negative photos on Attend trials was greater than on Reappraise trials, whereas activated regins if LPFC exhibited the precisely opposite pattern. The cognitive processes supporting reappraisal were discussed by Oschner et al. (2002) as illustrated by the above study. He pointed out that the neural correlates of effective reappraisal were 1/ activation in the regions of the LPFC and MPFC essential for working memory and cognitive control and self-monitoring and 2/ decreased activation in two regions involved in emotion processing, the MOFC and the amygdala (Adolphs, 2001). “Taken together, these findings provide the first evidence that reappraisal may modulate emotion processes implemented in the amygdala and MOFC, that are involved in the evaluating of the affective salience and contextual relevance of a stimulus” (Os- chner et al., 2002). Oschner et al. (2002) pointed out that further work was needed to determine which specific aspects of amygdala and OFC functioning were modulated by reappraisal. Thus while the appraisal function of the amygdala was characterised as automatic, reappraisal did not modulate the early amygdala response, which was thought to depend on subcortical inputs from the senses.

Reappraisal may have inﬂuenced a more sustained response, dependant on cortical inputs, more amenable to control by cognitive processes [...] Whereas the evaluation process supported by OFC may support the selection of appropriate, and the transient suppression of inappropriate affective responses, the reappraisal processes by lateral and medial prefrontal regions may be important for modulating these evaluation processes (Oschner et al., 2002). 184 CHAPTER 5. THE EMOTIONAL BRAIN

Oschner et al. (2002) also discusssed the observed inverse correlation between lateral prefrontal and amygdala activation during reappraisal, although these two structures shared few direct connections. Two possible connection routes were suggested: 1/ Indirect modulation from LPFC to MOFC and 2/ Prefrontal modulation of posterior perceptual and semantic inputs to the amygdala from the occipital and parietal regions. “Reappraising the affective significance of images in working memory may reorganise these input so that the amygdala and MOFC no longer register the presence of aversive stimuli”. The regulation of different aspects of emotion is also discussed by Oschsner and colleagues in relation to the lateralisation of activations and deactivations related to reappraisal and emotion processing respectively. According to Oschner et al. (2002) left lateralisation of reappraisal-related prefrontal activations might reflect a common verbal component of reappraisal strategies employed by those participants who typically reported mentally talking to themselves through their reappraisals. On the other hand, mechanisms involving attentional deployment, or suppression of expressive emotional behaviour might activate right prefrontal systems. A further aspect of lateralised activation was thought to relate to findings associating negative affect with the right hemisphere and positive affect with left hemisphere, where right amygdala deactivation could reflect down regulation of systems generating negative appraisals, whereas left PFC activation could reflect engagement of systems supporting positive reappraisals. Greater activity of left vs right PFC is thought to be associated with resistant depression, while individuals with depression or obsessive-compulsive didorder have shown hypoactivation of the prefrontal regions, coupled with hyperactivity of the amygdala and/or orbito-frontal cortex (Davidson et al., 2000; Saxena et al., 1999). Lieberman et al. (2007) carried out an experiment utilising fMRI to demonstrate how putting feelings into words disrupted amygdala activity in response to affective stimuli. Subjects were 5.2. SOCIAL COGNITION AND NEURAL CIRCUITS 185 asked to view target faces displaying emotional expressions and to label the expressions. A control gender-labeling task was also used. Lieberman and colleagues were able to demonstrate that affect-labeling, relative to other forms of encoding, diminished the response of the amygdala to negative emotional images. Affect labeling was shown to produce increased activity in the right ventrolateral prefrontal cortex (RVLPFC), which was inversely correlated with amygdala activity. In addition the directly amygdala connected medial prefrontal cortex (MPFC) was found to statis- tically mediate the relationship between RVLPFC and amygdala activity. The authors suggest this is a neuroanatomically plausible route for inhibitory effects on the amygdala, possibly similar to reappraisal mechanisms.

Everitt and Robbins (2005) have described the basolateral amygdala, nucleus accumbens core system as being involved in Pavlovian associative conditioning. “ Basolateral amygdala lesions, like nucleus accumbens lesions increase the choice of small immediate rewards over larger delayed rewards - indicating greater impulsivity” [...] “Selective lesions of the orbital prefrontal cortex (OFC) also impair the acquisition of cocaine seeking and responding with conditioned reinforcement”. According to Everitt and Robbins (2005), there is general consensus on the function of the amygdala, nucleus accumbens core and OFC, in the control over goal-directed behavior, by discrete conditioned stimuli acting as conditioned reinforcers. However they report that stimulation experiments suggest that the hippocampal formation underlies conditioning to contextual or spatial stimuli, and therefore the motivational impact of conditioned stimuli on behaviour such as drug seeking. Hippocampal contextual information and amygdala-dependant discrete conditioned stimuli may compete for control over goal-directed behaviours. “Thus amygdala lesions not only impair appetitive behavioral responses under the control of discrete conditioned stimuli, but also result in enhanced control by contextual cues; similarly hippocampal lesions impair contextual conditioning to discrete conditioned stimuli. The neural basis 186 CHAPTER 5. THE EMOTIONAL BRAIN of such competition between associative inﬂuences on behaviour is unclear, but may depend upon the projections of the basolateral amygdala and hippocampus to the nucleus accumbens shell” (Everitt and Robbins, 2005).

5.2.3 Discussion

The verbal component of reappraisal described above by Os- chner and colleagues and by Lieberman and colleagues provides a ‘higher order’ mechanism for the control of emotion, and may explain the stimulus-bound emotional reactions seen in a number of childhood syndromes, where language development is delayed or deﬁcient. For such children their behaviour does not follow the normal developmental progression from immediate stimulus- based reactivity via the amygdala, to emotional control and planning capacity. This important aspect of emotional control tends to be overlooked in most childhood diagnostic systems, with the possible exception of autism. The DSM-IV allows for a categorical Developmental Language Disorder diagnosis, but does not provide a language axis, which could provide a useful background to behaviour disorder. The neurophysiological basis of fear and emotion, based on the limbic/hypothalamic system of MacLean, has been replaced by an updated striato-nigral-striatal and thalamo- cortical-thalamic motivational networks, and an emotional circuit consisting of amygdala, hippocampus and other limbic forebrain structures (orbitofrontal cortex), as well as indirect connections via mesolimbic dopaminergic projections from the ventral tegmental area. At a therapeutic level, both the LeDoux and Zahm models describe reciprocal cortical-amygdala connections, providing an underlying circuit basis for both cognitive behavioural or “top-down” interventions, as well as pharmacological or “bottom-up” interventions. Immediate fast visceral response to fear stimuli (necessary for a survival response), and more detailed and involved processing necessary in order to consciously see and process the stimulus, appears to reﬂect immediate subconscious 5.2. SOCIAL COGNITION AND NEURAL CIRCUITS 187

“bottom-up” and conscious higher level “top-down”processes respectively. Top-down ’higher-order’ approaches will thus require maturation of cortical macrosystem connections. 188 CHAPTER 5. THE EMOTIONAL BRAIN Chapter 6

Attention Deﬁcit/Hyperactivity Disorder

The American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, (DSM-IV), describes Attention Deﬁcit/Hyperactivity Disorder (ADHD) as:

Either (1) or (2): (1) six (or more) of the following symptoms of inattention have persisted for at least 6 months to a degree that is maladaptive and inconsistent with developmental level: Inattention (a) often fails to give close attention to details or makes careless mistakes in schoolwork, work, or other activities (b) often has difficulty sustaining attention in tasks or play activities (c) often does not seem to listen when spoken to directly (d) often does not follow through on instructions and fails to finish schoolwork, chores, or duties in the workplace (not due to oppositional behaviour or failure to understand instructions (e) often has difficulty organising tasks and activities (f) often avoids, dislikes, or is reluctant to engage in tasks that require sustained mental effort (such as schoolwork or homework) (g) often loses things necessary for tasks or activities (e.g., toys, school assignments, pencils, books or tools) (h) is often easily distracted by extraneous stimuli (i) is often forgetful in daily activities (2) six or more of the following symptoms of hyperactivity- impulsivity have persisted for at least 6 months to a degree that is maladaptive and inconsistent with developmental level: Hyperactivity (a) often figits with hands or feet or squirms in seat

189 190CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER

(b) often leaves seat in classroom or in other situations in which remaining seated is expected (c) often runs about or climbs excessively in situations in which it is inappropriate (in adolescents or adults may be limited to subjective feelings of restlessness) (d) often has difficulty playing or engaging in leisure activities quietly (e) is often “on the go” or often acts as if “driven by a motor” (f) often talks excessively Impulsivity (g) often blurts out answers before questions have been completed (h) often has difficulty awaiting turn (i) often interrupts or intrudes on others (e.g., butts into conversations or games) B. Some hyperactive-impulsive or inattentive symptoms that caused impairment were present before age 7 years C. Some impairment from the symptoms is present in two or more settings (e.g., at school [or work] and at home). D. There must be clear evidence of clinically significant impairment in social, academic, or occupational functioning. E. The symptoms do not occur exclusively during the course of a Pervasive Developmental Disorder and are not better accounted for by another mental disorder (e.g., Mood Disorder, Anxiety Dis- order, Dissociative Disorder, or a Personality Disorder).

6.1 Theories of ADHD

ADHD is often described as the most commonly presenting syndrome at Child Psychiatry clinics. The DSM-IV describes a Pre- dominantly Inattentive, Predominantly Hyperactive/Impulsive and Combined subtypes. The former, based on factor analytic field studies emphasises difficulty sustaining attention in tasks or play activities and symptoms of distractibility, while Hyperactiv- ity is described as often fidgeting with hands or feet and running about or climbing excessively, while Impulsivity is described difficulty awaiting turn, and often interrupting or intuding on others. The dual nature of ADHD has been investigated in twin studies. McLoughlin et al. (2007) have reported a twin study of 6,222 approximately 8-year-old twins from the Twins Early Develop- ment Study. They found that both subscales were highly heritable (hyperactive- impulsive, 88 percent and inattentive,79 percent. However bivariate genetic modeling indicated substantial overlap between the two components. 6.1. THEORIES OF ADHD 191

Castellanos and Tannock (2002) pointed out that partly in response to controversy about the validity of ADHD and concern about an apparent increase in its prevalence in the 1990’s, investigators have unsuccessfully attempted to formulate a single theory of ADHD, appealing to psychological constructs such as response inhibition, regulation of arousal, activation and delay aversion (avoidance of delay). Despite this the authors believed that relatively little progress had been made towards elucidating its pathophysiology, despite a burgeoning literature. They suggest that nearly two decades of unsuccessful efforts in psychiatric genetics have led to the conclusion that symptom-based diagnostic classiﬁcation systems do not facilitate mapping between susceptibility genes and behavioural outcomes. However a “loosely formulated dopamine hypothesis motivated candidate-gene studies that have been surprisingly productive. Anatomical MRI studies have found reduced volumes, mainly supporting the idea that a distributed circuit that includes the right prefrontal cortex, the caudate nucleus, the cerebellar hemispheres and a subregion of the cerebellar vermis, underlies ADHD” (Castellanos and Tan- nock, 2002).

Bush (2010) has pointed out that ADHD is a developmental syndrome with multiple forms, encompassing an inattentive form, a fairly rare purely hyperactive form, and a combined type with features of inattention and hyperactivity. Conceptualisa- tions have varied from Still (1902) - a deﬁcit of moral regulation, to Barkley (1990) - a deﬁcit in executive function, to Sonuga- Barke (2003) - dual pathways, incuding motivation and executive function, and Nigg and Casey (2005) who also include frontocere- bellar functions (Bush, 2010). Leiner et al. (1993) reviewed neuroanatomical, neuroimaging and behavioural reports of cerebellar involvement in cognitive and language functions. For example, the authors point to the enormous enlargement of the neo-dentate part of the cerebellum. “Although it is generally recognised that the projection from the cerebellum reaches the motor areas of the frontal lobe (areas 4 and 6 of Brodman), it is not widely recog- 192CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER nised, as yet, that the cerebellar projection also reaches some prefrontal areas of the lobe”. Leiner et al. (1993) describe a striking activation in the inferior lateral part of the cerebellum, which was anatomically distinct from the activation of the paramedian part of the cerebellum, during motor tasks including speech. The sequential organisation of both language and motor tasks is likely to be a fundamental aspect of executive functions, with impairment occurring in both ADHD and autism.

6.1.1 ADHD circuits

Castellanos and Tannock (2002) suggest that indices of disease liability or risk, termed endophenotypes may predict the risk for ADHD in the same way that serum cholesterol predicts the risk of cardiovascular disease. They nominated measures such as dopamine metabolism, response inhibition, or slow and variable responses in a stop-signal task, shortened delay gradient, temporal processing and working memory as possible endophenotypes. Sagvolden et al. (2005) described mesocortical, mesolimbic, and nigrostriatal dopaminergic pathways (modules), which give rise to specific symptom classes in ADHD, relating to executive function, emotional regulation and motor symptomatology. The authors postulated that the behaviour and symptoms of ADHD derive from altered dopaminergic function with a consequent failure to modulate non-dopaminergic (predominantly glutamate and GABA) signal transmission. They postulate that a hypo-functioning mesolimbic dopamine branch produces altered reinforcement of behaviour and deficient extinction of previously reinforced behaviour, giving rise to delay aversion, impulsivity and failure to ‘inhibit’ responses or disinhibition. A hypofunctioning mesocortical dopamine branch is postulated to cause attentional deficits, while a hypo-functioning nigrostriatal branch gives rise to impaired motor functions. Sagvolden and colleagues have argued that the main component of altered reinforcement processes in ADHD children is a steeper 6.1. THEORIES OF ADHD 193 delay-of-reinforcement gradient, or shorter time interval between response and effective reinforcer, resulting in less effective reinforcement and also less effective extinction of previously established, but no longer reinforced responses.

Sonuga-Barke (2003, 0002, 2005) described a dual-pathway model of ADHD, which integrated executive function and motivational theories of ADHD. Executive function refers to dysregulation of cognitive controls, resulting in failure to modify actions, thoughts, and feelings so that they conform to the social and intellectual requirements of a situation. On the other hand, motivational models focus on the inability of ADHD children to delay gratification for later reward (delay-aversion). Sonuga-Barke has pointed out that previous theories of psychopathology in ADHD had attempted a unitary construct at the level of symptom expression, but given that the ADHD diagnosis was made on the basis of behavioural symptoms, there was no one-to-one match necessary between clinical and neuro-biopsychological characteristics of ADHD. He pointed to the growing body of evidence from neuroanatomical, psychopharmacological, imaging and clinical studies to support a model of cortical-subcortical interaction, based on parallel but functionally segregated distributed brain circuits connecting specific cortical foci with basal ganglia and thalamic nuclei. The executive circuit, was thought to involve glutamergic excitatory inputs down from the prefrontal cortex to the dorsal portion of the neo-striatum, with reciprocal excitatory connections back up to cortical regions. These connections were mediated through GABA-based inhibitory inputs to dorso-medial sec- tions of the thalamus via dedicated direct and indirect pathways within subcortical relays (globus pallidus, substantia nigra, and the subthalamic nucleus). On the other hand, activity within the reward circuit was centered on the ventral striatum (in particular the nucleus accumbens). In this case the excitatory connections from frontal regions (Anterior Cingulate and Orbito-frontal) were reciprocated via the ventral pallidum and related structures through the thalamus. In both the above cases, dopamine was 194CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER thought to be a key neuro-modulator of both the executive and reward circuits. Activity within the executive circuit was modulated by the mesocortical and nigrostriatal dopamine branches, while on the other hand, mesolimbic branches of the dopamine system from the VTA, modulated motivational circuits. Thus the inhibitory based executive dysfunction was related to dorsal fronto-striatal dysfunction (I-EDF), and motivational dysfunction was underpinned by ventral frontostriatal circuits, and linked to signalling of delayed rewards and delay aversion DAv). A recent neuropsychological study of ADHD by Sonuga- Barke et al. (2010) has gone “beyond” the dual pathway model. The investigators utilised I-EDF, Dav and temporal processing TD) tasks to further investigate the dual pathway model. Using a principal components factor analysis, they described four factors with eigenvalues greater than 1 as follows: Component 1: Stop signal reactiontime (SSRT) and Go/no-go (GNG and modified Stroop(MStroop); Component2: Temporal processing deficits (TPD) and Working memory(WM); Component 3: Negative effect of imposed delay and Component4: Delay positive. These components were relatively independent in ADHD subjects, but proband sibling correlations were significant only for Inhibition and Timing. Multiple regression with proband scores in the four domains as predictor and sibling scores as the outcome showed the sibling domain scores were specifically predicted by proband scores for Inhibition, and Timing. The investigators interpreted their data in terms of showing a third dissociable neuropsychological component of ADHD, namely temporal processing. Interestingly, Delay deficits were associated with IQ and literacy, whereas Timing was significantly associatedwith literacy only. When IQ was added as a predictor, the effects of Delay, but not Timing on literacy were significantly reduced. While the dual pathway model is generally consistent with the present dual phenoypic behavioual model, the demonstration of a further neuropsychological construct, related to temporal processing, might suggest undue simplicity in the present model. However, the ad- 6.1. THEORIES OF ADHD 195 dition of a third TPD construct which according to Sonuga-Barke et al. (2010) might share neural components with I-EDF (ie basal ganglia) and (DAv), but remaining distinct in other ways, possibly related to cerebellum, serves to underline the necessity for further studies of neural circuit phenomena in children. Addition- ally the importance of the relation between literacy and Timing is yet to be determined, but is consistent with the present emphasis on literacy. According to Floresco et al. (2003), the mesolimbic dopamine system plays a central role in reward and goal-directed behaviour, and has been implicated in multiple psychiatric disorders. Under- standing the mechanism by which dopamine participates in these activities requires comprehension of the dynamics of dopamine release. Floresco et al. (2004) describe the DA system as com- partmentalized consisting of a synaptic (and potentially peri- synaptic) compartment and a tonically maintained extrasynaptic compartment, each of which is differentially affected by uptake processes. The description of separate synaptic and peri- synaptic compartments suggests separate cortical and subcortical genetic influences, and separate pharmacological influences at these levels. For example, DA metabolism is predominantly via the dopamine transporter (DAT) at subcortical levels, but via COMT at the PFC (Bilder et al., 2004). Bilder and colleagues described a tonic/phasic theory of dopaminergic actions where subcortical phasic DA is subject to fast reuptake by DAT, whereas most DA in the PFC is removed via reuptake into noradrenergic terminals, requiring DA to diffuse long distances before inactivation via this route. Floresco et al. (2003) describe the synaptic or phasic levels of dopamine as being mediated by bursting events at the level of the cell body, restricted by high affinity and rapid uptake systems, and associated with reward- conditioned prediction. On the other hand extrasynaptic or tonic dopamine levels are modulated by presynaptic limbic and cortical glutamergic inputs. Alterations in tonic levels of dopamine efflux occur on a much slower time-scale 196CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER and allow a wide variety of motor, cognitive and motivational functions. “The findings provide insight into multiple regulatory systems that modulate dopamine system function; burst firing inducing massive synaptic dopamine release, rapidly removed by reuptake before escaping the synaptic cleft, whereas increased population activity modulates tonic extra-synaptic dopamine levels that are less influenced by reuptake, and presumably affect long-term disposition” (Floresco et al., 2003). In ADHD children, impairment of tonic/phasic relationships may influence reinforcement gradients described by Sagvolden et al. (2005), as a result of lowered availability of tonic DA levels in the mesolimbic and mesocortical systems, resulting in the stimulus-bound, impulsive “fearless” behavior of ADHD children. On the other hand impaired synaptic gating by PFC at the accumbens level allows greater access to conditioned amygdala reactions and the anxiety (or aggression) described in some ADHD children (Levy, 2004). Arnsten (2006) has reviewed neuropsychological and imaging studies that have shown that attention-deficit/hyperactivity disorder (ADHD) is associated with alterations in prefrontal cortex (PFC) and its connections to striatum and cerebellum. She points out that the PFC is important for sustaining attention over a delay, inhibiting distraction, and dividing attention, while more posterior cortical areas are essential for perception and the allocation of attentional resources. Lesions to the PFC produce a profile of distractibility, forgetfulness, impulsivity, poor planning, and locomotor hyperactivity and optimal levels of norepinephrine and dopamine are needed for proper PFC control of behavior and attention. “Recent electrophysiologic studies in animals suggest that norepinephrine enhances signals through postsynaptic alpha2A-adrenoceptors in PFC, while dopamine decreases noise through modest levels of D1-receptor stimulation. Blockade of alpha2-receptors in the monkey PFC re-creates the symptoms of ADHD, resulting in impaired working memory, increased impulsivity, and locomotor hyperactivity” (Arnsten, 2006). Noradrener- 6.1. THEORIES OF ADHD 197 gic transmission in the prefrontal cortex may, according to Arn- sten, also act via the D1 receptor. Arnsten (2006) has described the functional contributions of the higher association cortices in attention. She points out that these cortical areas are intricately interconnected, creating both feed-forward and feed-back loops that work together to provide a unified emotional experience. The inferior temporal cortex is thought to process visual features such as colour and shape. In- terference from ‘top down’ PFC or parietal association areas can direct visual feature processing by suppression from nearby stimuli in the same visual field. The parietal association cortex is thought specialised for analysis of movement and spatial relationships, creating ‘world-reference’ maps. The PFC is thought to use representational knowledge (working memory) to guide overt responses (movement) as well as covert responses (attention), allowing inhibition of inappropriate behaviours and gating of irrelevant stimuli. Thus PFC lesions “impair the ability to sustain attention, particularly over a delay, and reduce the ability to gate sensory input”. According to Arnsten (2006) the association cortices project down to both the basal ganglia and cerebellum in a series of parallel, closed-loop circuits. The PFC and parietal and temporal association cortices project to the caudate nucleus as part of the ‘cognitive circuit’, through the basal ganglia, projecting back to the PFC and premotor cortices. According to Arnsten, the PFC and parietal association cortices project to the cerebellum by way of the pontine nuclei and the cerebellum in turn projects back to association cortices by way of dentate projections to the thalamus, suggesting a role in higher cerebellar functions. Arnsten suggests that the PFC appears to thrive under conditions of moderate catecholamine release, when norepinephrine alpha2a receptor stimulation “signals” and optimal dopamine D1 receptor stimulation decreases “noise”. In contrast, PFC working memory functions are impaired under conditions of high catecholamine release that engage alpha1, and beta receptors and excessive D1- 198CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER receptor stimulation. These catecholamines may act as a chemical switch, turning on the PFC during normal waking and turning it off during drowsiness or stress. In contrast, high levels of catecholamines may turn on more primitive brain structures such as the amygdala for more automatic control of behaviour under conditions of danger. Solanto (2002) discussed the circuits involved in both motor overactivity and cognitive dysfunction. While the dorsal striatum is involved in the selection, initiation and execution of voluntary motor responses, two parallel prefrontal-striatal and thalamic- cortical circuits, direct and indirect are involved.

The ‘direct pathway’ extends from the PFC through the internal segment of the globus pallidus, through the thalamus, feeding back to exert a net amplification (via disinhibition) of the original cortical output. [...] The other ‘indirect’ pathway projects through the external segment of globus pallidus, and synapses on inhibitory projections from the subthalamic nucleus to the internal globus pallidus, producing a net inhibition of cortical output. Insufficient dopaminergic activity in this pathway will result in excessive motor activity (Solanto, 2002). Solanto (2002) also describes the PFC as essential for attentional control, organisation and planning. “With respect to dopamine, it has been amply demonstrated in primates that D1 agonists enhance and D1 antagonists impair working memory”. Solanto points out that since the PFC projects to many subcortical areas, including dorsal and ventral striatum, thalamus, amygdala, substantia nigra and ventral tegmental area, PFC deficits may lead to disinhibition in these regions. Nigg and Casey (2005) have described an ‘integrative theory of ADHD’, based on three related neural circuits. They suggest that fronto-striatal circuits are involved in response output and suppression, as well as working memory, while fronto-cerebellar circuits are involved in temporal information processing. They point out that both the cerebellum and basal ganglia project to the prefrontal cortex via the thalamus, and both utilise gamma-amino butyric acid (GABA) as an inhibitory neurotransmitter. Function- ally the detection of unpredicted or novel events has been linked 6.1. THEORIES OF ADHD 199 to fronto-striatal functioning (Schultz et al., 1997), whereas alert- ing, monitoring and detecting violations in the timing of events has been linked to fronto-cerebellar function (Spencer et al., 2003). The authors suggest that in ADHD there is diminished ability to alter behaviour when predictions about the environment are violated, possibly emerging from altered catecholamine modulation of the above circuitry. In this view, ADHD represents a failure of adaptive pre-frontally guided behaviour when salient or unexpected events occur. Thus disruption of the prefrontal cortex results in weakened ability to carry out actions represented in working memory, and which require monitoring of new information or holding multiple goals in mind. Nigg and Casey (2005) also describe an affect-related fronto- amygdala circuit which is important in ADHD. They suggest that activation of the amygdala is associated with avoidance of potentially unpleasant events, while activation of the nucleus accumbens is related to approaching a potentially positive event. Hare et al. (2005) have suggested an inverse correlation between ventral striatum and amygdala such that the activity of one inhibits or suppresses the other. The authors suggest that affective systems may be involved in the impulsive features of ADHD, possibly by overvaluation of reward and failure of avoidance learning. The authors suggest that developmental pathways, particularly those involved in prefrontal projections may be a critical aspect of the cognitive and affective control, involved in ADHD, with crosstalk among neural circuits. Thus ADHD may represent a failure of both mature cognitive control and well-regulated affective response systems. They point out that it is important that these systems support one another, or fail to do so during early development. They suggest that early temperamental predispositions may represent precursors to difficulty with cognitive control and/or affective (reactive) intensity. According to Goldman-Rakic (1987c, 1998b), prefrontal neurons have sensory receptive fields and can be driven by both auditory and visual stimuli. Also principal sulcus neurons show ac- 200CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER tivity time-locked to the delay period of a delayed response. Thus the principal sulcus participates in an essential mechanism for holding spatial information “on-line” for the time necessary to guide a response. Goldman-Rakic describes the caudal two-thirds of the principal sulcus as receiving the major outflow of posterior parietal projections, with a precise topographically organised network of connections between particular sectors of the parietal cortex and arial subdivisions of the principal sulcal cortex, with at least a rough organization of different modalities in the inferior parietal lobule corresponding to cytoarchitectonic distinctions. In sensory systems, it was suggested that parietal-prefrontal projections have the properties of feed-forward pathways, whereas prefrontal-parietal projections exhibit characteristics of feedback pathways. Goldman-Rakic (1998b) suggested that the reciprocal circuitry between the principal sulcus and parietal cortex provided a regulatory mechanism for selecting, adjusting and maintaining a flow of relevant information from the parietal to the prefrontal cortex, providing visuo-spatial coordinates to guide behaviour in the absence of external cues. In addition, reciprocal hippocampal connections were thought to provide access to long-term memories. Finally Goldman-Rakic (1998b) described a larger system of connections between circuitry linking prefrontal, parietal, limbic and temporal cortices, allowing for “multiple functional nuances”.

Castellanos et al. (2008) have investigated an alternate (to executive function) model of networks involved in ADHD, to the more traditional fronto-striatal circuitry. The investigators utilised functional magnetic resonance imaging (fMRI) scans to image attentional lapses in 20 adults with ADHD, and 20 age- and sex-matched healthy volunteers. They hypothesised abnormalities in functional connectivity between frontal foci involved in cognitive control, and the non-goal-directed ’default-mode’. They developed a study-specific template to optimise automated spatial normalisation, and were able to verify the presence of an antiphasic or negative relationship between activity in dor- 6.1. THEORIES OF ADHD 201 sal anterior cingulate cortex, and in default-mode network components. Group analysis revealed compromises in this relationship, while secondary analyses revealed an extensive pattern of ADHD-related decreases in connectivity between precuneus and other default-mode network components, including ventromedial prefrontal cortex and portion of the posterior cingulate. The investigators commented that the demonstration of a negative relationship between control regions and the default-mode network suggests that these interactions are implicated in preventing attentional lapses found in ADHD. An additive regression analysis revealed ADHD-related decreases in connectivity among components of the default-mode network between precuneus and ventromedial prefrontal cortex. This could, according to the authors suggest a novel locus of dysfunction for working memory deficits commonly observed in ADHD. Better performance in a working memory task has been reported to be positively related to the strength of functional connectivity between anterior and posterior default-mode components (ventromedial PFC and posterior cingulate cortex) (Hampson et al., 2006). Vijayraghavan et al. (2007) showed that dopamine D1 receptor stimulation in PFC produced an ‘inverted-U’ dose-response, whereby either too little or too much D1 receptor stimulation impaired spatial working memory. This response has been observed across species, including genetic linkages with human cognitive abilities and PFC activation states, and DA synthesis. Ac- cording to the authors the cellular basis for the inverted-U has long been sought, with in vitro intracellular recordings supporting a variety of potential mechanisms. Their study demonstrated that the D1 receptor agonist inverted-U response was observed in PFC neurons of behaving monkeys: low levels of D1 receptor stimulation enhanced spatial tuning by suppressing responses to non-preferred directions, whereas excessively high levels reduced delay-related firing for all directions, eroding tuning. These actions of D1 receptor stimulation were mediated in monkeys and rats by cyclic AMP intracellular signaling. The evidence for an 202CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER inverted-U at the cellular level in behaving animals promises to bridge in vitro molecular analyses with human cognitive experience.

Imaging studies

Recent imaging studies have drawn attention to the dorsal anterior midcingulate cortex (daMCC) (previously called dorsal anterior cingulate), dorsolateral prefrontal cortex (DLPFC), ventrolateral prefrontal cortex (VLPFC) and parietal cortex (Bush, 2010). According to Bush, the daMCC has critical roles in attention, cognitive processing, target detection, novelty detection, response selection, response inhibition, error detection, and motivation. The daMCC is believed to be a key modulator of moment-to-moment adjustments in behaviour, through its primary role in feedback- based decision making. The daMCC’s role in attention and cognition is believed to integrate goal and feedback-related information from various sources and then to use this information to modulate activity in executive brain regions that direct attention and produce motor responses. “The daMCC thus acts within cognitive- reward-motor networks to increase the efficiency of decision making and execution” (Bush, 2010). The striatum is also described by Bush (2010) to have multiple roles relevant to ADHD (Haber, 2003). Castellanos et al. (2006) postulated that executive function deficits will be linked with anterior striatum, DLPFC and daMCC dysfunction, whereas delay aversion symptoms will be tied to dysfunction of motivational /reward areas, including the ventral striatum and orbitomedial prefrontal cortex, suggesting bi-directional influences (Bush, 2010).

6.1.2 Comment

Arnsten’s work suggests that ADHD may represent a disorder of executive function based on deﬁcits in the above executive circuits. Until the advent of positive effects of the noradrenergic drug Atomoxetine, the striking results of stimulant medica- 6.2. AGE EFFECTS 203 tions were thought to have primarily dopaminergic effects in the basal ganglia. It now appears likely that both cortical executive and subcortical effects can occur depending on the dose levels of stimulant (dopamine releasing) medications, and/or noradrenergic postsynaptic effects.

6.2 Age effects

Castellanos et al. (2002) utilised automated MRI, to measure initial volumes and prospective age-related volume changes of total cerebrum, cerebellum, gray and white matter and caudate nucleus in ADHD children (age range 5-18 years) and age and sex-matched controls (age 4.5-19 years). Their results demonstrated that on initial scan, patients with ADHD had significantly smaller brain volumes in all regions. However, while caudate nucleus volumes were initially abnormal for patients with ADHD, as caudate volumes decreased for both patients and controls, diagnostic differences disappeared. This developmental effect is consistent with the present hypothesis in relation to activity vs attention phenotypes with increasing age. Taerk et al. (2004) have examined PFC and COMT (catechol- O-methyltransferase) associations in 118 children aged 6-12 years, meeting DSM-IV criteria for ADHD, utilising the Wiscon- sin Card Sorting Test (WCST), which measures set-shifting ability and has differentiated ADHD children from controls, Tower of London, which measures planning ability, and the Self Ordered Pointing Task, a measure of working memory, also found capable of differentiating ADHD children from controls. The investigators reported no association of the catechol-O-methyltransferase (COMT) Val 108/158 Met polymorphism with these tasks. Taerk et al. (2004) argued that the absence of an association between the COMT polymorphism and their measures of executive function could be related to the relatively young age of their sample, and that the functional importance of COMT in the PFC may only be observable in adults. They quote animal studies indicating a 204CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER positive relationship between aging and COMT. Certainly in humans COMT is likely to become more important with age, as the PFC matures. The age effect is important in understanding developmental factors involved in ADHD, where PFC functioning is thought to be impaired or delayed in ADHD children, as suggested by Levy (2007b,c). On the other hand, Diamond et al. (2004) in a study of healthy children examined the Dots-Mixed task, which requires that participants remember two rules and inhibit the tendency to respond on the same side as the stimulus on one-half of the trials. It is poorly performed by primates and by children with phenylke- tonuria, who have low prefrontal DA, and showed that approximately 26 percent of the variance was shared with COMT genotypes, while the Self-Ordered Pointing task, where children are shown six blocks of trials with line drawings and abstract de- signs presented in a rectangular grid, and asked to point only once to each stimulus, (thought insensitive to DA levels) was not associated. The authors claimed a high level of specificity, in that the COMT polymorphism was linked only to tasks sensitive to the level of dopamine. Forssberg et al. (2006) used a 3D positon emission tomography (PET) study to measure alterations in brain dopamine transmission in a group of 8 male adolescents with ADHD and 6 age-matched controls. They found lower dopamine metabolism indices in subcortical areas, associated with the ADHD diagnosis, and with more severe attention deficits within the ADHD group. Cortical variables had a positive correlation with attention within the ADHD group. Wilens et al. (2004) reviewed the features of Adult ADHD. They point out that longitudinal studies of ADHD youths show that symptoms of hyperactivity and impulsivity may decay, but inattention tends to persist (Achenbach et al., 1998). Studies of clinically referred adults with ADHD show that while up to half may retain clinically important levels of hyperactivity and impulsivity, up to 90 percent have prominent attentional symptoms. Functional imaging studies implicate fronto-subcortical systems 6.2. AGE EFFECTS 205 in the pathology of ADHD. Zametkin et al. (1990) utilised positron emission tomography to show reduced global and regional glucose metabolism in the premotor cortex and superior prefrontal cortex, while neuroimaging studies suggest that 3 subcortical structures - caudate, putamen, and globus pallidus are part of the neural circuitry that underlies motor control, executive function, inhibition of behaviour and modulation of reward. According to Wilens et al. (2004) these fronto-striatal-pallidal circuits provide feedback to the cortex for the regulation of behaviour.

6.2.1 ADHD and early onset bipolar disorder

The issue of the covariance and co-occurrence early-onset bipolar disorder (PBD) with ADHD has been subject to recent debate. According to Zepf (2009), the question of whether these two disorders share common underlying neurobiological processes which produce the same phenomenology is controversial because PBD symptoms have frequently been shown to overlap with ADHD symptoms. Also the question of drug-induced manic symptoms by treatment with antidepressants and/or stimulants has been raised by differences in prevalence rates between Europe and the USA (Zepf, 2009). Biederman et al. (1995) utilised a proﬁle based on the Child Behavior Checklist (CBCL), with scales related to Anxiousness/Depressiveness, Attention problems and Aggression to discriminate children with PBD and ADHD. However, the longitudinal outcome of individuals diagnosed in this way appear to predict diverse personality disorders (Wingo and Ghaemi, 2007), suggesting a lack of predictive speciﬁcity.

6.2.2 Working memory and ADHD

Goldman-Rakic (1996, 1987b) described the role of the primate pre-frontal cortex in spatial cognition, from the point of view of connectivity with major neurological centers. The pre-frontal cortex in primate studies was found to be integral to delayed response tasks, which required behaviour to be guided by rep- 206CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER resentations of discrimination stimuli, rather than those stimuli. Thus, the work of Goldman-Rakic in primates suggest that prefrontal-parietal reverberating reciprocal circuits maintain the visuo-spatial representation required for delayed response and for CPT performance. The activation of cortical circuits is maintained by subcortical tone acting directly and indirectly (Glowin- ski et al., 1984). Arnsten and Li (2005) have pointed out that the prefrontal cortex (PFC) guides behaviours, thoughts and feelings using representational knowledge, allowing the ability to inhibit inappropriate behaviours and thoughts, regulate attention, monitor actions and future planning. Animal research has shown that optimal levels of norepinephrine acting at post-synaptic alpha-2A- adrenoreceptors and dopamine acting at D1 receptors are essential to prefrontal cortex function. While animal and neuroimaging studies (Sagvolden et al., 2005; Solanto, 2002) have implicated dysfunction within fronto-striatal circuits as a putative mechanism, associated with impulsivity in ADHD. Levy (2007b) has described a dissociation or differential dopaminergic effects on the D1 receptor in the PFC, and D2 receptors in the striatum. While striatal effects largely involve motor systems (Sagvolden et al., 2005), PFC effects on working memory and inhibition are more likely to involve D1 receptors. Denny and Rapport (2001) have described a model, which posits that working memory plays a pivotal role in the organisation of behaviour. Organised responding is functionally dependant on the capacity of working memory to (a) generate and maintain representations of input stimuli (b) search memory traces for matches and (c) access and maintain representations of behavioural responses appropriate to input stimuli. Failure of working memory leads not only to disorganised behaviour but also mo- tivates children to redirect their attention to other stimuli in the environment (stimulation seeking), giving rise to frequent rapid shifts in activity. Working memory is viewed as a causal cognitive process that is the direct consequence of one or more neu- 6.2. AGE EFFECTS 207 robiological substrates. Thus prefrontal-parietal activation is an important component of working memory in tasks, shown to be impaired in ADHD children. While the inhibitory role of prefrontal/striatal circuits have been described by Castellanos (1997) and others, the importance of prefrontal/parietal pathways for working memory had not been emphasised. A further gating mechanism in the PFC is described by Arn- sten (Medscape Psychiatry and Mental Health, 2007). She points out that recent studies of alpha-2 noradrenergic agonists in the treatment of ADHD have shown that D1 activation is not the only receptor important for PFC working memory. Noradrener- gic drugs such as clonidine and guanfacine (Arnsten et al., 1996; Avery et al., 2000) have shown that monkeys performing a spatial working memory task show increased regional cerebral blood flow in the dorsolateral prefrontal cortex after administration of guanfacine, as well as a delay-related improvement in working memory. At the cellular level, delay-related cell firing is thought to be the electro-physiological signature of working memory, needed to overcome distraction and inhibit inappropriate responding. According to Arnsten, the mechanism of guanfacine’s ability to strengthen PFC cognitive function is now known to occur at the level of the synaptic ion channel (Wang et al., 2007). Alpha-2A receptors are co-localised with hyperpolarisation-activated cyclic nucleotide-gated (HCN) channels on prefrontal dendritic spines. HCN channels are opened by cyclic adenosine monophosphate (cAMP), and when opened can allow passage of both Na+ and K+ ions. This diminishes membrane resistance, and voltage shifts from synaptic inputs are unable to pass through the spine into the cell, essentially disconnecting network connections. Guan- facine apparently reverses this process by inhibiting production of cAMP, opening the HCN channels. Thus Arnsten’s work indicates a synaptic gating mechanism at the cellular level, suggesting that PFC working memory depends on a number of such op- ponent processing mechanisms. Cohen et al. (1998, p209) have utilised a connectionist ap- 208CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER proach to capture the central features of computation such as simple processing units, graded flow of activation, and modifiable connection weights. They postulate that in order for a system to be able to maintain information in the face of intervening and potentially interfering stimuli, a mechanism is required to stabilise these representations. They propose that dopaminergic neuromodulation of the prefrontal cortex implements a gating function, that governs which representations gain access to the PFC. These in turn provide representations that enable the flexible biasing of representations in posterior neocortex, over intervals required to guide task performance in a temporally extended manner. The prefrontal parietal feedback system should correctly be termed a micro-circuit, which helps to maintain optimal activity in the PFC during working memory and vigilance tasks. 1

6.2.3 Imaging studies

Event Related Potential (ERP) studies of ADHD have also been used to examine activation processes in ADHD children. The most commonly examined component is the P300, that is a late positive wave with a latency of 300 to 800 ms (Klorman et al., 1991). The amplitude of the P300 is influenced by stimulus probability or relevance, and the latency is influenced by cognitive, perceptual or memory load. It is thought to reflect allocation of attention and stimulus evaluation processes, as well as updating of internal representations and working memory. Most studies in ADHD children have found the amplitude of the P300 to be reduced in children with ADHD, suggesting deficits in activation processes. Silberstein et al. (1998) utilized a Steady-State Visually Evoked Potential (SSVEP), which measures cortical electrical activity over the course of an A-X Continuous Performance Task (children are asked to press a button when they see an X pro- ceeded by an A). The technique, also known as Steady-State Probe Tomography (SSPT), enables examination of disturbances in the

1A basic microcircuit requires direct positive and indirect inhibitory feedback. 6.2. AGE EFFECTS 209 spatial distribution and dynamics of brain electrical activity. (Re- ductions in SSVEP amplitude have previously been associated with increased cortical activation). The investigators found that compared to the mean amplitude during a reference task, control subjects demonstrated SSVEP amplitude reductions, during the A-X interval. Transient amplitude reductions occurred in frontal regions, while right parietal and occipital amplitude reductions were sustained throughout the 3.5 second A-X interval. In contrast, ADHD subjects demonstrated much smaller frontal amplitude reductions and increased parieto-occipital amplitude (i.e. less activation) suggesting they failed to increase cortical activation according to task demands. The largest group differences occurred at the disappearance of the A, when controls showed extensive activation, particularly in the parieto-occipital region, while ADHD subjects showed reduced activation in the parieto- occipital region. The disappearance of the A also coincided with large latency reductions in controls at frontal and temporal sites, while in ADHD subjects there were latency increases.

A further unpublished study by the same group examined changes in SSVEP following methylphenidate administration in 60 boys with ADHD (mean age 10 years 1 month), recently diagnosed according to DSM-IV criteria (Diagnostic and Statisti- cal Manual of the American Psychiatric Association - 4th edition). SSVEP amplitude and latency during the A-X interval, before methylphenidate, and 90 minutes after administration of 0.2 mg/kg of methylphenidate was examined. After methylphenidate, there was increased activation predominantly in frontal and occipital regions, and reduced parieto-occipital latency, at the disappearance of A, in the A-X interval. The most sustained and signiﬁcant changes occurred in the right parietal region. There were also transient reductions in frontal latency, particularly after the disappearance of the A. The results are thought to suggest that right prefrontal and right parietal processes, which are involved in performance of the CPT-AX task in healthy control children, are deﬁcient in ADHD children, and are enhanced 210CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER by methylphenidate in ADHD children. In particular, these processes may be involved in the maintenance of activation in the interval between A and X, allowing a ‘working’ retention of the parameters necessary for correct task performance.

6.2.4 Resting/default state and vigilance variability

Fassbender et al. (2009) have utilised fMRI to examine the relationship intra-individual variability in task RT , and suppression of response in regions of the “default network”, using a working memory task and two levels of control tasks in 12 ADHD and 13 healthy controls. The investigators demonstrated that children, diagnosed with ADHD, demonstrated more RT variability, than controls, while neural measures showed that although both groups displayed a pattern of increasing deactivation of the medial PFC, with increasing task diffeculty, the ADHD group was significantly less deactivated than controls in medial prefrontal areas. The authors discussed the “default-mode interference” hypothesis of Sonuga-Barke and Castellanos (2007), who suggested that the characteristic increase in RT variability in ADHD might be due to intrusions by the default attention network into goal- directed activity. Namely, the normally adaptive state of periodically attending to potential novel events in the environment (subserved by the default attention network) would become maladaptive when this interfered with ongoing task processes. Castel- lanos et al. (2008) proposed this effect as a model for ADHD, with inattention resulting from an imbalance between suppresion of the default network, and activation of task-appropriate regions. According to Bush (2010), recent studies on “resting brain” activity have provided complementary information to data produced using cognitive activation paradigms. A few studies have focussed on “deactivations” or decreases in regional cerebral blood flow (rCBF) or fMRI signal during a task of interest relative to a control task. Raichle et al. (2001) have suggested that such de- 6.2. AGE EFFECTS 211 creases provide evidence of a ’default’ mode or homeostatic brain state maintained during rest or visual fixation. According to Bush (2010) three separable “networks” of brain support (1) cognitive or focussed, goal-directed behaviour; (2) internal state monitoring, involving the regulation of emotions, motivational state and endogenous stimuli; and (3) vigilance for salient external stimuli. Gusnard and Raichle (2001) noted that, compared with a fixation/rest “default state”, cognitive tasks are thought to activate brain regions such as the MCC, DLPFC and posterior parietal cortex. Conversely, cognitive tasks deactivate the perigenual ACC, medial PFC, portions of the VLPFC, amygdala and postero- medial areas such as the posterior cingulate cortex (PCC), retrosplenial cortex and precuneus, and other posterolateral parietal areas near the angular gyrus (Gusnard et al., 2001). The authors suggested that these latter areas are “tonically active” during unstructured rest periods, to support environmental vigilance and monitoring of the internal milieu. Bush (2010) points out that “resting state studies” point to a need for taking individual variability differences into account, in addition to differences in means between groups.

6.2.5 Continuous Performance Task (CPT)

Koziol and Budding (2009, p246-248) describe the CPT as assessing capacity to attend and capacity to inhibit. “ [...] the subject is required to detect the proper stimuli, while inhibiting response to competing and distracting inﬂuences”. The task measures errors of commission, errors of omission, and in some versions, mean reaction time to a nominated stimulus, presented over time among randomly distributed distracting stimuli. Levy (1980) demonstrated a clear developmental progression in capacity to perform the CPT. Errors of omission and comission demonstrated learning curves with numerous early errors of both decreasing to minimal errors in normal preschool children around age 41/2 to 5 years, and mean reaction time stabilising around 212CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER

500 msec by 5-6 years. Levy and Hobbes (1981) were able to show a relationship between age-normalised CPT performance and clinician diagnosis of ADHD. Subsequent researchers have investigated reaction time (RT) variability in ADHD (Fassbender et al., 2009).

Figure 6.1: Age vs CPT Omission Errors: From Levy, F. 1980

Figure 6.2: Age vs Mean Reaction Time: From Levy, F. 1980 6.2. AGE EFFECTS 213

6.2.6 Discussion

The developmental progression from primarily motor hyperactivity in early childhood, to primarily attentional dysfunction in adulthood is likely to reflect the importance of prefrontal and cingulate cortical areas in the development of working memory, control over motor impulsivity and ability to sustain attention. While the above phenotypes have been frequently described (Achenbach et al., 1998; Barkley, 1997), the developmental progression from primarily ‘motor’ to primarily ‘inattentive’ phenotypes is not well understood, but is postulated as relating to the optimal development of prefrontal and cingulate cortical-subcortical circuits, the former associated with working memory, and the latter with motivational aspects of attention. It is of interest in the current context to reflect on the behavioural “deficit” indicated by poor CPT performance in either normal young preschool children and/or ADHD school-aged and adolescent children, in light of the Koziol and Budding (2009) hypothesis of dual cognitive control systems, namely stimulus- bound and higher-order systems. In this context, a ‘normal’ CPT performance with minimal omission and commission errors, and a regular consistent reaction time to a nominated stimulus, can be postulated to reflect not only a normal capacity for inhibition, but also implies a capacity for higher-order control, (utilising a verbally nominated stimulus). On the other hand a pre 4-5-year old or pathological school-aged performance, with age- inappropriate errors of commission, errors of omission and a ir- regular, inconsistent and usually slower than normal mean reaction time, is indicative of a behavioural deficit. In the present context it can be argued that this deficit is a manifestation of a deficit in the development of higher-order cognitive functioning, leaving the child (and/or adolescent) subject to stimulus-bound behaviour, with deficient or absent capacity for both routinisation of responses and higher-order planning processes. By formulat- ing the observed distractible and impulsive behaviour observed 214CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER in ADHD children in this way, current neuroscientific findings of basal ganglia and prefrontal deficits in ADHD are more un- derstandable, as resulting from deficits in the cortico-thalamic- striatal-cortical (CSTC) system. The relationship of ’resting state’ models to vigilance models is an important topic for future research.

ADHD symptomatology may represent the opposite of a recurrent syndrome. That is, distractible behaviour, which is controlled by environmental stimuli, suggests an inability to maintain a sustained behaviour pattern, leaving the child (or adult) vulnerable to random sensory stimuli, which will transiently control behaviour, but be rapidly replaced by other salient stimuli. This gives rise to the distractible behaviour pattern observed in hyperactive children. The recurrent circuit architecture may still apply in ADHD, but in this case there is insufﬁcient neurotransmitter at subcortical levels to maintain a recurrent feedback/reinforcement loop at the PFC level.

While the Sonuga-Barke and Castellanos (2007) and Castel- lanos et al. (2008) default mechanism models are heuristic and consistent with previous demonstrations of RT variability in ADHD children, the model does not explain how the normal integration or balance between task-related and default processing is achieved. Here the cortical/subcortical models postulated by Koziol and Budding (2009), Balleine10 provide mechanisms, whereby appropriate modulation of task-related (higher- order) processing and stimulus-bound (default) processing may be achieved. This modulation may require both striatal and cerebellar inputs to achieve appropriate timing of prefrontal activity during task-related activity. The incorporation of default or stimulus- bound reactivity into the ADHD model helps to explain the overreactivity of young and of ADHD children to external stimuli and distracting internal concerns, in terms of circuit deﬁcits in task related and default networks. 6.3. TREATMENT MODELS 215

6.3 Treatment models

Arnsten (2006) has pointed out that it is likely that both dopamine D1 and norepinephrine actions contribute to the therapeutic effects of stimulants in patients with ADHD. In contrast excessive doses of stimulant medication may produce cognitive inflexibility. A recent approach to the treatment of ADHD ex- emplifies the utility of a detailed understanding of the architecture and pharmacology of the ‘executive’ network. The finding of beneficial effects from alpha-noradrenergic agents such as atomoxetine has recently modified the uniform use of stimulant medications to treat ADHD. According to Arnsten (1997), high doses of alpha-2 agonists appear to have unilaterally beneficial effects on cognitive function, though these effects may be eroded by emerging sedative and hypotensive effects. However, Arnsten et al. (1998) have shown that the ability of alpha-2 agonists to improve PFC function without side effects was found to correspond with selectivity for the alpha-2A receptor site. Arnsten and Dudley (2005) suggest that noradrenergic transmission has a vi- tal beneficial effect on PFC function, particularly via postsynaptic alpha-2-adrenoreceptors. The investigators point out that the PFC guides behaviour, thought and affect using working memory, and that these processes are the basis of executive functions, including regulation of attention, planning and impulse control. PFC function is thought to be fundamental in ADHD. The role of DA in PFC function has been extensively researched (Sawaguchi and Goldman-Rakic, 1994; Murphy et al., 1996). Stimulation of the D1 receptor within optimal levels improves working memory performance, while higher levels may erode performance. Interestingly, while it has been accepted in the past that the beneficial effect effects of stimulant medication were mediated by DA transmission, (Arnsten and Dudley, 2005) suggest that performance on PFC tasks in rats and mice on low oral doses of methylphenidate may be mediated in part by the alpha-2 adrenoreceptor. Thus both stimulants and atomoxetine may exert 216CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER therapeutic actions via D1 and alpha-2A receptor mechanisms, but under conditions of stress, guanfacine may be more protective given that high doses of stimulants may impair cognition and in- duce obsessive and perseverative or restricted thinking. The authors interpreted their data as suggesting that guanfacine acts on the prefrontal cortex (probably post-synaptically at alpha2 - receptors) to increase cognitive and associated functions, known to be dysfunctional in ADHD sufferers, and also helps in the regulation of locomotor activity via inhibitory control of subcortical brain regions, particularly the caudate putamen and nucleus accumbens. Thus guanfacine appears to have the ability to ‘turn down’ striatal activity, possibly of benefit in the treatment of mo- toric hyperactivity. This increased selectivity is important where side effects of stimulant medications are of concern, because of motor side effects of DA at striatal levels. It is still somewhat unclear whether DA effects in the PFC are exerted primarily through the DAD1 receptor, or alpha-2A receptor. Cornil et al. (2002) have shown that dopamine activates noradrenergic receptors in the quail preoptic area. They found that DA-induced inhibitions/excitations were not blocked by selective dopaminergic receptor antagonists, but were suppressed by selective alpha noradrenergic antagonists. While the mechanism of the cross-talk between DA and NE receptors was unclear, the relatively similar structure of DA and NE could potentially explain the binding of both amines to the two receptor subtypes, as could co-localisation of cAMP, HCN channels and alpha- 2A adrenoreceptor (Wang et al., 2007). Studies of transgenic mice demonstrated that the transporters for NE and DA lacked specificity, so that in the prefrontal cortex, DA was mainly taken up by the NE transporter (Moron et al., 2002). Thus, the mechanisms and specificity in human therapeutics remains an intrigu- ing problem whose solution offers more targeted understanding of future treatments. The actions of dopaminergic vs. noradrenergic agents, currently available for the treatment of ADHD have overlapping, 6.3. TREATMENT MODELS 217 but different actions in the prefrontal cortex (PFC) and subcortical centers. While stimulants act on D1 receptors in the dorsolateral prefrontal cortex, they also have effects on D2 receptors in the corpus striatum and may also have serotonergic effects at orbitofrontal areas. At therapeutic levels, dopamine (DA) stimulation (through DAT transporter inhibition) decreases noise level acting on subcortical D2 receptors, while NE stimulation (through alpha-2A agonists) increases signal by acting preferentially in the PFC possibly on DAD1 receptors. On the other hand, alpha-2A noradrenergic transmission is more limited to the prefrontal cortex (PFC), and is thus less likely to have motor or stereotypic side effects, while alpha-2B and alpha-2C agonists may have wider cortical effects. The data suggest a possible hierarchy of specificity in the current medications used in the treatment of ADHD, with guanfacine likely to be most specific for the treatment of prefrontal attentional and working memory deficits. Stimulants may have broader effects on both vigilance and motor impulsivity, depending on dose levels, while atomoxetine may have effects on attention, anxiety, social affect, and sedation via noradrenergic transmission (Levy, 2007b).

At a theoretical level, the advent of successful noradrenergic therapies poses the question of the role of working memory in ADHD. From the present architectural point of view, the need for maintenance of an optimal level of neuronal activity in the PFC via feedforward and feedback relationships with modular cortical and subcortical centers (via striato-thalamo-cortical loops) may underlie the basis of distractible symptomatology. While the aetiology of ADHD is recognised as being heterogeneous both genetically (Faraone et al., 2005) and environmentally (Swanson et al., 2007), Levy (2007b) has pointed out that as distinct from aetiology, the executive functional effects of stimulant and other medications appear to depend on prefrontal dopamine and/or noradrenergic stimulation at the D1 receptor. This allows goal-related control of sequential motor (and language) actions, whereas lack of this control results in random sensory control of behaviour via 218CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER subcortical effects. Interestingly, adrenoreceptor stimulation is thought to result in closure of HCN channels on PFC dendritic spines receiving inputs from neurons with similar spatial properties, thus increasing firing during delay periods for preferred directions, while moderate levels of D1 receptor stimulation leads to opening of HCN channels on spines receiving inputs from neurons with dissimilar spatial properties, reducing delay related firing to non- preferred directions (Arnsten, 2007). This reciprocal process between D1 and alpha-2A activity at prefrontal levels allows representation by ‘preferred’ spatial stimuli to be maintained in working memory, and has important implications for mechanisms of representation in the PFC (Levy, 2008). It also suggests that the alpha-2A agonist guanfacine may have specific effects on working memory in ADHD children. Weinberger (2003, p129-154) and colleagues, who have shown that dopamine (DA) modulation of the prefrontal cortex is an important factor in signal to noise (STN) processing of information. Their data have implicated the gene for catechol-O- methyltransferase (COMT) as playing an important role in prefrontal dopamine (DA) metabolism in the cortex, but not in the striatum. In humans, the COMT gene contains a highly functional and common variation in its coding sequence, (guanine to adenine substitution), which translates into a valine to methionine (val/met) change in peptide sequence. This single amino acid substitution affects the temperature lability of the enzyme ; at body temperature the met allele has one-fourth the enzyme activity of the val allele, suggesting that individuals with the val alleles would have greater inactivation of prefrontal DA and therefore relatively poorer prefrontal function. Weinberger discusses studies which show that prefrontal neurons project mono- synaptically to DA neurons that return to the prefrontal cortex. He suggests that the prefrontal cortex tonically inhibits striatal DA activity (Carlsson, 2001). Egan et al. (2001) have demonstrated poorer performance (ie more perseverative errors) on the 6.3. TREATMENT MODELS 219

Wisconsin Card Sorting Test (WCST) in a large sample of subjects (n=457), with the val allele being associated with the relatively poorer performance. Egan et al. (2001) also studied the cortical physiological response during a working memory task with fMRI and found the predicted effect of the genotype. Work- ing memory deficits are found in a number of behavioural syndromes such as Attention Deficit Hyperactivity Disorder (ADHD), and schizophrenia. The example illustrates how an apparently small genetic variation can have a major impact on medication response, mediated via neurotransmitter relationships between the midbrain and cortex. Interestingly, a study by (Mattay et al., 2003) on the effects of this genotype on individual variation in the brain response to amphetamine (AMP), reported significant main effects of COMT genotype and of AMP on the distributed cortical activation patterns associated with a working memory (WM) task, with prominent locales of activation in prefrontal and parietal cortices, but a significant genotype by drug interaction was restricted to the prefrontal cortex (PFC). At all levels of WM load, subjects with the val/val genotype had a more efficient prefrontal activation response. However, when above an optimal AMP threshold, subjects with the low activity met/met genotype, who were initially superior in performance showed a deterioration, suggesting an inverted-U functional response to increased dopamine signalling in the prefrontal cortex. Thus molecular genetic studies of genes involved in working memory were able to explain previously puzzling behavioural observations of an inverted-U response to stimulant medications. Because symptoms of PFC dysfunction are characteristic of many neuropsychiatric disorders, the above examples indicate the importance of empirically-based biological models of neuropsychiatric disorders, which are made possible, as neuroscientific and genetic methods increase in so- phistication. 220CHAPTER 6. ATTENTION DEFICIT/HYPERACTIVITY DISORDER Chapter 7

Conduct Disorder and neural circuits

7.1 Diagnostic issues

Both Conduct Disorder (CD) and Antisocial Personality Disorder (APD) are DSM-IV diagnoses, and some aspects of CD are considered predictive of APD. It is necessary to manifest at least three antisocial behaviours, for at least 6 months, to be diagnosed with CD, while APD must present with a pervasive pattern of antisocial behaviour that begins in childhood and continues into adulthood for diagnosis to be achieved. Blair (2001) describes “acquired sociopathy” as characterising persons who, after acquired lesions of the orbitofrontal cortex, fulfil the DSM-III diagnostic criteria for “sociopathic disorder”, an antecedent of APD. Blair also distinguishes Psychopathy in both childhood and adulthood as being defined by high scores on clinically-based rating scales. He points out that the above diagnosed populations may be heterogeneous, and that different risk factors are associated with the life course of persistent CD, appearing before or after 10 years. He also distinguishes “reactive aggression”, elicited in response to frustra- tion/threat, from “instrumental aggression”, which is more pur- poseful and goal-directed. However in both cases there is a deficit in the individual’s capacity to regulate emotion and to manifest

221 222 CHAPTER 7. CONDUCT DISORDER AND NEURAL CIRCUITS empathy. Blair and Frith (2000) point out that a number of studies have explored the performance of individuals with CD, APD and psychopathy on measures of executive function, including planning, monitoring and inhibition of prepotent responses. The results indicated that neither CPD, APD nor psychopathy were specifically associated with poorer performance on general measures of executive functioning (Kandel and Freed, 1989; Pennington and Ozonoff, 1996). The authors suggest that while executive dysfunction is a risk factor for violence in some special populations, such as ADHD comorbid with CD, it has not been shown to be directly implicated in the development of antisocial personality disorder. Blair and Frith (2000) define the term ‘executive emotion processing’ as referring to cognitive structures that modify responses to emotion-laden stimuli. These mechanisms are those that modify behavior when the reinforcement associated with a particular stimulus changes from positive to negative (e.g. changes from reward to punishment), allowing reversal learning and extinction and inhibition of a primary learned response to occur. Blair and Frith describe the ventral and orbitofrontal cortex as implicated in executive emotion processing. They point out that only one model of executive emotion processing has been proposed: Damasio’s somatic marker hypothesis, referring to bodily responses to stimuli that guide reasoning and decision making. However, patients with APD’s and psychopathy do not manifest typical somatic marker responses, though they do show impairments in tests of extinction and reversal learning (Rolls et al., 1994). Thus while there may be commonality between patients with orbitofrontal lesions and adult psychopaths, Blair and Frith believe there is no unified theory of cognitive functions that require intact function of the orbitofrontal cortex. Blair (2001) describes both the “Somatic Marker Model” (Damasio, 1994), which postulates a “body loop”, that is activated in guiding or controlling aggressive responses via somatosensory responses, and a “Social Response Reversal Model” (SRR), which 7.1. DIAGNOSTIC ISSUES 223 emphasises the role of social behaviour (Rolls, 1997; Dias et al., 1996). The latter system is activated by another’s angry expressions, and negative valence expressions or representations of situations previously associated with anger. Thus the main difference between the two models is thought to be the stimuli activating each system. In the case of the SRR, this was thought to be angry expressions, while in the case of the Somatic Marker model, a broad range of stimuli are involved. However, Blair (2001) argues that neither the SRR nor the Somatic Marker model provide an explanation for instrumental aggression associated with developmental psychopathy. He points out that no satisfactory explanation has been offered for why psychopathic persons show intact autonomic responses to basic threat stimuli, such as bared teeth displays, but not to threat stimuli produced through visual imagery (Blair et al., 1997). Blair suggests that according to a “Violence Inhibition Mechanism” (VIM), where moral socialisation occurs through the pairing of the activation of distress cues in relation to sad and fearful expressions, with representations of acts, which caused the distress. Thus an appropriately developing child “[...] initially finds the pain of others aversive and then, through socialisation, thoughts of acts that cause pain to others also aversive”. Blair et al. (2001) point out that the VIM model was prompted by work suggesting that most social animals possess mechanisms for the control of aggression (Eibl-Eibesfeldt, 1970; Lorenz, 1981), who note that submission cues displayed to a conspecific aggressor terminate attacks. In this sense, sad facial expressions may function as a human submission response. Blair (2001) proposes that psychopathic persons have had a disruption to the violence inhibition (VIM) system, such that it fails to be triggered by representations of acts that cause harm to others.

This is thought to occur because the signal to the learning system concerning emotionally aversive stimuli is muted. That is, the unconditioned stimulus (US) signal is diminished, thus impairing the formation of unconditioned-conditioned stimulus (US-CS) associations. [...] At the neural level, this has been attributed to dysfunc- 224 CHAPTER 7. CONDUCT DISORDER AND NEURAL CIRCUITS

tion within the amygdala (Blair et al., 1999; Blair, 1995).

Thus Blair distinguishes between “acquired sociopathy” after orbito-frontal cortex lesions, and psychopathy after possible early amygdala dysfunction, developmental psychopathy. The latter is associated with an impairment in capacity to form associations between distress cues and representations of transgres- sions. Blair (2003) points out that while classifications of psychopathy are not synonymous with diagnoses of DSM-IV Con- duct Disorder or Antisocial Personality disorder, they represent an extension of the diagnoses. He suggests that the psychiatric diagnoses are poorly specified and concentrate on antisocial behaviour, whereas psychopathy is defined not only by antisocial behaviour, but also by emotional impairment such as lack of guilt. Also much of the antisocial behaviour of individuals with psychopathy is instrumental in nature, directed towards achieving money, sexual gratification or status. Blair (2003) has outlined the “neural basis of psychopathy”. He points out that the amygdala is involved in all the processes that, when impaired, give rise to the functional impairments shown by individuals with psychopathy (Patrick, 1994; Blair et al., 1999). Blair describes two fMRI studies (Tiihonen et al., 2000; Kiehl et al., 2001), which found that high levels of psychopathy were associated with reduced amygdyloid volume in the former, and a reduced amygdala response in an emotional memory task in the latter. Additionally, the medial orbitofrontal cortex (OFC) is, according to Blair (2003) involved in instrumental learning and response reversal, both of which functions are impaired in individuals with psychopathy. Blair suggests that the neural structures implicated in psychopathic pathology include the amygdala and OFC. While the causes of this pathology are unclear, administration of the beta-adrenergic blocker propranalol, as well as amygdala damage, block episodic memory for emotionally arous- ing events, as well as the processing of sad facial expressions (Harmer et al., 2001). 7.2. BASIC EMOTION THEORIES 225

Blair and Frith (2000) reviewed the neurocognitive impairments underlying the antisocial personality disorder. They consider the possibility of three neuropsychological and cognitive ﬁndings:

(1) An impairment in executive functioning, including planning and inhibition, primarily implicating (1) the dorsolateral prefrontal cortex. (2) An impairment in executive emotion processing involving somatic marking and primarily implicating orbitofrontal cortex and (3) An impairment in emotion processing of fear and empathy, primarily implicating the amygdala and septo-hippocampal system. (Blair and Frith, 2000). Blair and Frith (2000) define executive functioning as the ability to maintain an appropriate problem-solving set for attainment of a future goal. “The executive functions are considered to be involved in the mediation of controlled and non-automatic behaviour, and are thought to be mediated by the dorsolateral prefrontal cortex. Specific functions, which have been classified as ‘executive’, are inhibition, planning and monitoring (Shallice, 1988)”. The authors point out that most models of executive function predict decreased levels of inhibition, and hence increased aggression, if the functioning of the executive system is disrupted.

7.2 Basic emotion theories

Pavuluri and Sweeney (2008) point out that “new models” are emerging in the understanding of social cognition in children with empathic failures, such as psychopathy and autism, where amygdala and related noradrenergic connectivity responsible for making stimulus-punishment associations necessary for successful socialisation, are thought to be disrupted in psychopathy, where there is a failure to learn to avoid actions that will harm others, and ultimately elicit punishment for oneself. According to Blair and Frith (2000), the theory of a basic emotion deﬁcit has been proposed by (Hare, 1970; Patrick, 1994); (Quay, 1993), suggested a basic impairment of fear. Eysenck (1964) and (Raine, 226 CHAPTER 7. CONDUCT DISORDER AND NEURAL CIRCUITS

1997) have suggested chronic under-arousal or dysfunctional ‘behavioural inhibition system’ (Gray, 1971; Patterson and New- man, 1993; Quay, 1993). Blair and Frith describe three main observations that support the suggestion that there is impairment in fear processing in individuals with APD. First antisocial populations show impoverished classical conditioning (Hare, 1978). Second psychopaths fail to show an augmented startle re- ﬂex (Patrick et al., 1993). Third, antisocial populations show poor performance on a card-playing task, in which winning cards provide monetary gain, while losing cards involve monetary loss. Over time, the ratio of punished responses increases, but adults and children continue to view signiﬁcantly more cards. How- ever, according to Blair and Frith, the extent to which antisocial behaviour is inhibited in normally developing individuals by punishment has been questioned (Blackburn, 1988; Blair, 1995). On the other hand, studies have shown that moral socialisation is better achieved through empathy induction (Baumrind, 1971; Hoffman and Saltzstein, 1967).

7.2.1 Fear circuits

Blair and Frith (2000) assert that the importance of empathy for moral socialisation was one of the reasons for the development of the violence inhibition mechanism model of psychopathy (Blair, 1995; Blair et al., 1997). “According to the model, moral sociali- sition occurs through the pairing of the activation of the mechanism by distress cues, with representations of the acts which caused those distress cues”. The authors point out that while the VIM empathy account is consistent with the literature on socialisation in normally developing children, it cannot clearly account for findings motivated by fear-impairment theories. How- ever, they believe the fear-based and empathy-based accounts may not be incompatible, and that they can be integrated at the anatomical level. While early accounts of the dysfunctional fear system position implicated the septo-hippocampal system as a 7.2. BASIC EMOTION THEORIES 227 crucial component of a fear system (Fowles, 1988; Gray, 1987; Quay, 1993), Blair and Frith believe that more recent data indicate that the amygdala is the locus for fear conditioning and that the hippocampal system provides contextual data for this conditioning (LeDoux, 1995; O’Keefe, 1991) . For example, PET functional imaging studies in normal adults have confirmed that the amygdala is crucially involved in processing the facial affect of fear (Morris et al., 1996; Blair et al., 1999) have shown that there is a specific neural response in the left amygdala that is proportional to the intensity of sad facial affect. Blair and Frith thus suggest that at least one locus of dysfunction in the psychopath is within the amygdaloid body. They conclude that both the executive function and basic emotion models predict that for an individual to present with the full-blown syndrome of psychopathy, he/she needs to have a socially disadvantaged childhood environment. From the executive function position, the social environment determines whether aggression is a prepotent response to social frustrations. From the basic emotion position, the social environment determines the motivation to offend, via a lack of empathy. Blair et al. (1999) used PET to study 13 male volunteers who viewed static grey-scale images of emotionally expressive faces, taken from a standard set of pictures of facial affect, depicting either sad or angry expressions. They found that increasing intensity of sad facial expression was associated with enhanced activity in the left amygdala and right temporal pole. They also found that increasing intensity of angry facial expression was associated with enhanced activity in the orbitofrontal and anterior cingulate cortex. They found no support for the suggestion that angry expressions generated a signal in the amygdala, suggesting “dissociable but interlocking” systems for the processing of negative social affect. Blair (2006) has developed a neurocognitive model of psychopathy, based on three systems. He suggests that sensory representations are primarily implemented by the temporal cortex. 228 CHAPTER 7. CONDUCT DISORDER AND NEURAL CIRCUITS

Second, valence representations are crucially implemented via the amygdala, which is involved in the formation of stimulus- reward and stimulus-punishment associations, whose connections with further systems are implemented through Hebbian learning (Hebb, 1949). Thus a “distress cue” activated by a valence representation will bias the processing of sensory representations, increasing attention to the sensory stimulus. According to Blair (2006) the third system mediates the production of motor responses, implemented by basal ganglia and motor cortex. Blair argues that instrumental psychopathy results from an impairment of amygdala responses of fear and empathy. He describes a relatively specific deficit in processing stimulus-punishment associations, rather than stimulus-reward associations. Thus individuals with psychopathy differ from patients with amygdala lesions, who also require the amygdala to form stimulus-reward associations. This suggests a relatively specific module, possibly genetically determined, for stimulus fear (or punishment) associations.

Blair (2006) has described an important difference between children and adults with psychopathy in their response to contingency reversal (reversal of reward vs punishment). While children are as impaired as adults in responding to diminished reward (or punishment), children with psychopathy show a less marked response in detecting change in stimulus-reward associations, thought to depend on the orbital/ventrolateral cortex. Blair suggests that the difference between children and adults could reflect a lack of afferent input from the amygdala to orbital frontal cortex as development proceeds. Thus while the neural architecture for emotional processing, i.e. the amygdala, hypothalamus and periaqueductal gray is present from birth, there may be a developmental progression to orbital frontal cortex effects found in adults. Thus cortical/subcortical relationships may be influenced by developmental and/or environmental factors such as stress. Blair suggests that while the responsiveness of basic stress systems may be influenced by the environment, the neural architecture is not constructed out of environmental experi- 7.2. BASIC EMOTION THEORIES 229 ence. He contrasts this approach with that of Karmilloff-Smith et al. (2003) who posits important environmental influences on cortical-subcortical relationships. An explanation for the differences may be found by examination of the development of cortical-subcortical functional relationships. Herba and Phillips (2004) have pointed out that the functional relationships between subcortical and prefrontal cortical regions in the response to emotionally salient stimuli has been relatively unexamined. Thus with the first two years of life there is a large increase in myelination, improving the efficiency of the cortical and subcortical pathways (Herschkowitz, 2000). Ac- cording to Herba and Phillips (2004), Herschkowitz has pointed out that the connections between the amygdala and the developing hippocampus (with its role in memory) may help explain the emergence of the amygdala response to fearful stimuli. “These findings, taken together with the knowledge that the prefrontal cortex continues to develop into adolescence (Stuss, 1992) suggest that the functional connections between subcortical and prefrontal cortical regions may continue to develop into childhood and adolescence (Herba and Phillips, 2004). Adolphs et al. (1996) have indicated that the timing of amygdala damage impacts on the subsequent ability to recognise emotional expressions. They concluded that the association between the recognition of fearful facial expressions and the knowledge of the meaning of fear is acquired over development. The authors also concluded from these and other findings that the amygdala is crucial during development for establishing the networks necessary for emotion expression recognition. However once the networks are established they may function independently of the amygdala. Thus with age, Herba and Phillips (2004) describe increased prefrontal and decreased subcortical activity. In adults, Phillips et al. (2003b) and Phillips et al. (2003a) have proposed that specific patterns of structural abnormalities in parallel neural systems, important for the response to emotional stimuli and regulation of emotional behaviour may be associated 230 CHAPTER 7. CONDUCT DISORDER AND NEURAL CIRCUITS with the generation of different symptoms of psychiatric disorders, including schizophrenia, bipolar and major depressive disorders. In children, numerous studies have demonstrated deficits in emotional functioning and inability to recognise facial expression in autism (Baron-Cohen et al., 2000b; Dyck et al., 2001). While some studies have been criticised for not controlling for mental age, the association of autism and early amygdala dysfunction is an important area of study. Similarly, Herba and Phillips (2004) point out that it has been shown that physically abused children label expressions of anger more frequently, compared with non- abused children (Pollack and Kistler, 2002). Herba and Phillips (2004) have also described variations in brain development in relation to sex. Thus the amygdala, which is smaller in females has been linked to affective disorder (Rubia et al., 1999, 2000).

7.2.2 Reward circuits

Hollerman et al. (2000) point out that the striatum can be subdi- vided into three territories based on cortical and subcortical inputs:

(1) a sensorimotor region (principally dorsal putamen) receiving inputs from primary and secondary motor cortices. (2) ventral or limbic striatum (including but not restricted to the nucleus accumbens) receiving inputs from the amygdala and orbitofrontal cortex and (3) associative striatum (dorsal and anterior caudate) receiving inputs from dorso-lateral prefrontal cortex. The authors focus on the ventral striatum and associative striatum. (Hollerman et al., 2000).

Hollerman et al. (2000) suggest that a fundamental difference between mesencephalic dopamine and striatal neurons is that reward-related activity in the striatum is divided into subpopulations processing different aspects of reward. While one population responds to prediction of reward, other neurons exhibited activation following reward (integrating motivational and motor responses), and other striatal neurons responded to conditioned stimuli associated with rewards. Thus the differential striatal subpopulations provide a more complex analysis of aspects 7.3. GENETIC EFFECTS 231 of reward-related information. Some neurons encoded reward- related information in the environmental stimuli, while others reflected the immediate anticipation of the outcome of a behavioural act, and still others encoded the occurrence of that outcome. In- formation regarding behavioural contingency and outcome was most present with conditioned stimuli. Activation reflecting reward exclusively might serve as representations of reward as a goal, while activation reflecting both reward and behaviour could serve as representations of both the goal of reward and the behaviour required to obtain that goal. Hollerman et al. (2000) describe neuronal activity in the orbitofrontal cortex as also heterogeneous. While the classes of activation are similar to those in the striatum above: (1) responses to primary rewards (2) activity preceding a predicted reward and (3) responses to conditioned stimuli associated with rewards, activity in the orbito-frontal cortex largely isolates reward information from motor information, such that rewards are processed largely in relation to their relative motivational significance. The authors point out that the ability to process the relative motivational value of competing goals is essential for goal selection when various goals are available.

7.3 Genetic effects

Forbes et al. (2009) have reported an ‘imaging-genetics’ study of genetic variation in components of dopamine neurotransmission, impacting on ventral striatal (VS) reactivity associated with impulsivity. The authors point out that the ventral striatum has been widely implicated in reward processing, and that individual differences in its function are linked to traits involving high reward sensivity. Because dopamine (DA) plays a critical role in reward processing, it is thought to be a potent modulator of VS reactivity. Thus functional polymorphisms in dopamine-related genes may be associated with impulsivity. The authors investigated the association of a number of DA related polymorphisms and self- 232 CHAPTER 7. CONDUCT DISORDER AND NEURAL CIRCUITS reported impulsivity. They found that genetic variants associated with relatively increased striatal DA release (DRD2-141C deletion) and availibility (DAT1 9-repeat), as well as diminished inhibitory postsynaptic DA effects (DRD2-141C deletion and DRD4 7-repeat), predicted 9-12 percent of the interindividual variability in reward-related VS reactivity. By contrast, genetic variation directly affecting DA signalling only in the prefrontal cortex (COMT Val158Met) was not associated with variability in VS reactivity. The authors suggested that altered VS reactivity might represent a key neurobiological pathway through which these polymorphisms contribute to behavioural impulsivity.

Some important indicators for the present study that derive from Caspi et al. (2008) have shown that the COMT valine/methionine polymorphism at condon 158 is associated with phenotypic variation among children with ADHD. They found that across three large samples, valine/valine homozygotes had more symptoms of conduct disorder, were more aggressive, and were more likely to be convicted of criminal offences, illustrating that genetic information can provide evidence suggesting clinical subtypes. The authors discussed emerging data (in adults) regarding the importance of COMT variation in prefrontal cortex (PFC) function. One speculation from the above data, was that deficits in PFC function might be modulated by the COMT gene, which is important for dopamine metabolism in the PFC. Such deficits were thought to be important in conduct disordered children’s ability to control their behaviour or consider the implications of their behaviour. On the other hand, initial reports from the Cardiff ADHD Genetic study did not find an association between the COMT genotype (Mills et al., 2004), (Taerk et al., 2004), and several tests of executive function. However, this findng could be explained by factors which influence the stage of development when the activity of COMT becomes important in the PFC, as the PFC develops. A further suggestion draws from studies indicating that COMT variants may be associated with emotional regulation, and hypotheses that COMT Val alleles are related to 7.3. GENETIC EFFECTS 233 reduced responsiveness to unpleasant stimuli (as occurs in conduct disorder) (Smolka et al., 2005). Another example by Young and Wang (2004) described a neurobiological model for pair-bond formation, based on studies of monogamous rodents. They describe prairie voles as forming en- during pair bonds, while meadow voles are non-monogamous and typically do not display biparental care. Using partner preference tests in the laboratory, two neuropeptides emerged as critical mediators of partner preference formation: oxytocin and vasopressin. Infusion of oxytocin into the cerebral ventricles of female prairie voles accelerated pair bonding. Similarly, vasopressin facilitated pair-bonding in male prairie voles without mating. Administration of selective oxytocin receptor and vasopressin receptor antagonists prevented pair-bond formation in female and male prairie voles respectively. Results from anatomical and pharmacological studies indicate that the prefrontal cortex, nucleus accumbens and ventral pallidum are all critical brain regions in pair-bond formation. These regions are also thought to be involved in the mesolimbic dopamine reward system, suggesting that pair-bond formation uses the same neural circuitry as reward. According to the authors, reward processing is believed to depend on the meso-cortico-limbic dopaminergic system, consisting of dopamine neurons in the ventral tegmental area, and their projections to the nucleus accumbens, prefrontal cortex and other brain areas. The ventral pallidum is thought to be a major target of the nucleus accumbens, to mediate locomotor reponses to rewarding stimuli. Dopamine release within this circuit has been shown to be critically involved in natural reward such as food intake, as well as maladaptive reward such as drugs (DiChiara, 2002). Thus, consistent with the hypothesis that pair-bonding involves conditioned learning, dopamine neurons within the nucleus accumbens, are critical for partner preference in prairie voles. Mating has been shown to increase dopamine turnover in the nucleus accumbens of males, and results in an increase in ex- 234 CHAPTER 7. CONDUCT DISORDER AND NEURAL CIRCUITS tracellular dopamine in the nucleus accumbens of females (Arag- ona et al., 2003a); (Aragona et al., 2003b). The dopaminergic regulation of pair-bond formation is described by Young and Wang (2004) as receptor subtype-specific for activation of the dopamine D2 receptor. However, activation of the dopamine D1 receptor in the nucleus accumbens prevents pair-bonding in males, which according to Young and Wang (2004), may serve to prevent the formation of new pair bonds. Young and Wang (2004) point out that if their model is correct, pair bonding could potentially be induced in a non-monogamous species by expressing the oxytocin, or vasopressin receptor in the nucleus accumbens or ventral pallidum. They tested this prediction using viral vector-mediated gene transfer to overexpress Avpr1a, the gene responsible for encoding the vasopressin receptor in the ventral pallidum of the non-monogamous male meadow vole. After co-habitation with a receptive female, during which copulation occurred, the transgenic animals showed enhanced partner preference compared to controls. Pre-treating virus-treated voles with a D2 receptor antagonist prevented partner preferences. According to the authors, the study had ‘’remarkable implications” for the evolution of complex behaviour, suggesting that mutations altering the expression pattern of a single gene can have a profound impact on complex social behaviours.

7.3.1 Comment

Emerging understandings of the PFC/subcortical circuits involved in executive function vs those involved in reward and emotional control will potentially contribute to the understanding of a number of childhood syndromes. The theme of small variations in numbers of important genes inﬂuencing important behavioural connectivity, is both directly and indirectly relevant to the present thesis, in suggesting that variation in neural circuits, is ultimately genetic in origin, and thus subject to biological understanding and targeted interventions. 7.3. GENETIC EFFECTS 235

A futher implication of the Forbes et al. (2009) study relates to the development of executive control over impulsivity, via prefrontal-striatal circuits dependant on cortical-subcortical connectivity and representational processes. The processing of evaluating reward-related information is carried out by a hierarchical system, signaling predominantly different aspects of reward related information at different levels of reward processing, though at each level all aspects can be processed. The process of successively greater abstraction via successive specialisations of parallel neuronal systems, culminating at an abstract cortical level is a possible blueprint for information processing in a number of the syndromes under discussion, including ADHD and Conduct Dis- order. While studies examining the neural correlates of emotion in children and adolescents have tended to focus mainly on the response to fearful facial expressions and amygdala development, Herba and Phillips (2004) point out that little is understood about the functional development throughout childhood and adolescence of neural systems important for the response to displays of other emotions, “Evidence to date from studies of emerging emotion processing skills and aberrant patterns of development suggest distinct developmental trajectories for the recognition of different emotional expressions”. The establishment of functional cortical-subcortical relationships over the course of development is thus important for both cognitve and emotional functions, depending on the particular circuit involved. The above ‘empathy related’ theories of conduct-related problems, are based on fear conditioning, but can be integrated with executive models, via a modular theory of development, which postulates the development of an emotional circuit which includes the amygdala, cingulate and orbitofrontal cortex and hippocampus, and an ‘executive’ cognitive prefrontal, striatal, thalamo- cortical circuit. The possibility of a “motor theory of empathy”, involving ‘mirror circuits’ (Arbib, 2005), similar to the motor theory of speech, (Liberman and Wahlen, 2000) may also be useful in un- 236 CHAPTER 7. CONDUCT DISORDER AND NEURAL CIRCUITS derstanding language and affective deﬁcits observed in psychopathy. The absence of empathy in some conduct disordered children has been associated with language and reading disability in some children, suggesting both cognitive and affective deﬁcits (Levy, 1989).

A young boy, C was seen at 6.8 years of age, referred by his speech pathologist, due to concerns about his slow progress with literacy, and his ‘violent’ behaviour at home and at school. He had punched a teacher when in kindergarden, and recently kicked a teacher and hit his mother. His early development was normal, but by the age of 2 1/2 years, he was noticed to have articulation and expressive language difficulties and commenced speech therapy. He had at this time also developed oppositional and aggressive behaviour with temper outbursts at child-care. His articulation remained almost unintelligible even when assessed at age 5. A subsequent speech pathology assessment at age 7 years indicated receptive and expressive language skills in the average range, but weaknesses in phonological awareness, and articulation difficulty. His oral reading was stilted and often inaccurate, which affected comprehension. His working memory at that stage was in the 19th percentile. C was diagnosed around this time by a paediatrician as having At- tention Deficit Hyperactivity Disorder (ADHD), but the most striking aspect of C’s behaviour was the suddenness and rapidity of his outbursts, including running on to the road, in response to minor frustrations, as well as a lack of guilt or empathy.

7.3.2 Discussion

While a single case, C’s development illustrates a progression from articulation problems to literacy and behaviour problems. The extreme reactivity of his oppositional behaviour suggested direct engagement of noradrenergic amygdala and posterior cortical systems in control of his behaviour. Levy et al. (1987) showed that in a clinical sample, children with attention deﬁcit disorder with hyperactivity (ADDH) overlapped with children with conduct disorder (CD), on tests of vigilance, while children with severe ADDH performed signiﬁcantly worse on tests of vigilance and reading age, illustrating the close relationship between early reading problems and disorders of conduct and attention. Levy 7.3. GENETIC EFFECTS 237

(1989) demonstrated that phonetic spelling was signiﬁcantly inversely related to a measure of vigilance or sustained attention in a group of 51 ADHD boys. While these associations may or may not be causal, there does appear to be an important relationship between language, reading and attention in a number of externalising syndromes, where comorbidity (as discussed in Chapter 7) is often manifest. 238 CHAPTER 7. CONDUCT DISORDER AND NEURAL CIRCUITS Chapter 8

Autistic Disorder/Asperger’s Syndrome

8.1 Autism

The American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) describes Autistic Disorder as stipulating a total of six or more items from A, B and C with at least two from A, and one from B and C.

A. Qualititative impairment in social interaction, as manifested by at least two of the following: (1) marked impairment in the use of multiple nonverbal behaviours such as eye-to eye gaze, facial expression, body postures, and gestures to regulate social interaction. (2) failure to develop peer relationships appropriate to developmental level (3) a lack of spontaneous seeking to share enjoyment, interests, or achieve- ments with other people. (4) lack of social or emotional reciprocity. Qualitative impairments in communication as manifested by at least one of the following: (1) delay in, or total lack of, the development of spoken language, not accompanied by an attempt to compensate through alternative modes of communication such as gesture or mime. (2) in individuals with adequate speech, marked impairment in the ability to initiate or sustain a conversation with others. (3) stereotyped and repetitive use of language or idiosyncratic language. (4) lack of varied, spontaneous make-believe play or social imitative play

239 240 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME

appropriate to developmental level. Restricted repetitive and stereotyped patterns of behavior, interests and activities, as manifested by at least one of the following: (1) encompassing preoccupation with one or more stereotyped and restricted patterns of interest that is abnormal either in intensity or focus. (2) apparently inflexible adherence to specific, non- functional routines or rituals (3) stereotyped and repetitive motor mannerisms (e.g. hand or finger flapping or twisting, or complex whole-body movements). (4) persistent preoccupation with parts of objects. B. Delays or abnormal functioning in at least one of the following areas, with onset prior to age 3 years: (1) social interaction, (2) language as used in social communication, or (3) symbolic or imaginative play.

The essential features of Autistic Disorder are the presence of markedly abnormal or impaired development in social interaction and communication and a markedly restricted repertoire of activity and interests. The impairment in social interaction is gross and sustained and there may be marked impairment in the use of multiple nonverbal behaviours (e.g., eye-to eye gaze, facial expression, body postures and gestures). There may also be a lack of varied, spontaneous make-believe play or social imitative play appropriate to developmental level. Speech development is delayed and the pitch, intonation, rate, rhythm or stress may be abnormal, with abnormal grammar and repetitive use of language or idiosyncratic language. Behavioural symptoms include odd responses to sensory stimuli (e.g. oversensitivity to sounds or being touched, and fascination with certain repetitive stimuli and a lack of fear in response to real danger), as well as overactivity, short attention span, impulsivity, aggression, and self-injurious behaviour.

8.1.1 Language in Autism

Walenski et al. (2006) reviewed language deficits in Autism, Autistic Spectrum Disorder (ASD) and Aspergers Syndrome. They pointed out that while 20 percent of children with autism are essentially non-verbal, others acquire functional language to 8.1. AUTISM 241 varying extents, and that ASD may be associated with a particular pattern of both relatively spared and impaired language functions. The authors described two exploratory theories of language in ASD, which are complementary in that they focus on different sets of language functions: The “theory of mind” (TOM) hypothesis seeks to explain pragmatic impairments of language and communication in terms of social deficits with some specific accompanying neurocognitive deficits, while the “procedural deficit hypothesis” (PDH) posits that grammatical impairments in the disorder, including syntax, morphology, and phonology are explained by neurocognitve deficits in the procedural memory system, whereas lexical knowledge, which depends on the declarative memory system remains relatively spared.

The authors describe pragmatics as the use of language appropriately for the social and real-world contexts in which utterances are made, including interpreting a speaker’s intended meaning across different social contexts. Widespread pragmatic impairments have been described in Autism and ASD. Prosody is described as the timing, rhythm and intonation of speech. The pragmatic functions relate to non-grammatical pauses, and the use of stress in language. Prosodic deficits are common in ASD, as are difficulties with metaphor, irony and jokes. According to Walenski et al. (2006) , ASD involves pervasive impairments in the pragmatic aspects of language usage. The theory hypothesises “that people with ASD are fundamentally impaired at causally linking their own and other people’s behaviour to mental states”. This is thought to be evidenced by deficits in the understanding of false belief and emotion (Baron-Cohen et al., 1985, 2000b). The authors note that functional neuro-imaging studies in ASD suggest abnormalities in the neural structures thought to underlie the processing of theory of mind. These include the medial frontal cortex (BA 8/9) bordering on the cingulate gyrus, lateral inferior frontal cortex (primarily Broca’s area BA 44/45), and posterior superior temporal/temporo-parietal cortex. While neuroimaging studies of ASD suggest a tendency for decreased activation (rela- 242 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME tive to controls) in both frontal areas, they show activation (relative to controls) in the posterior superior temporo-parietal cortex (Allen et al., 2003). The PDH is, according to Walenski et al. (2006) implicated in the learning of new tasks, and in the control of long-established motor and cognitive skills, habits and other procedures such as typing, or game playing (Graybiel, 1995). The authors describe the procedural system as being composed of a network of neural interconnected brain structures, particularly in the left cerebral hemisphere, encompassing frontal lobes, and basal ganglia (neostriatum). Within the frontal cortex, the supplementary motor area (SMA) and Broca’s area are thought to be important. The authors describe the anticipated effects in ASD, depending on this system as deficits of grammar, syntax, morphology and phonology. However, because the circuits are composed of parallel and functionally segregated loops related to a particular set of cortical/subcortical structures, deficits may be heterogenous. The heterogeneity relates to different pathways and also “direct” disinhibitory vs “indirect” inhibitory influences from the basal ganglia. In contrast to the procedural memory system, the declarative memory system is described as subserving long-term learning representation and the use of knowledge about facts (somatic memory). Medial temporal structures-hippocampus and parahippocampal gyrus consolidate new memories, which eventually depend on neocortical areas, such as the temporal lobes. The authors believe that declarative and procedural memory systems interact both co-operatively and competitively. However, dysfunction of one system can lead to enhancement of the other. Thus Walenski et al. (2006) posit that in ASD, the declarative memory system will tend to take over certain grammatical functions from the dysfunctional procedural memory system. Complex structures which are able to be composed by the procedural system (walk plus ‘ed’) are simply stored as chunks in lexical/declarative memory in individuals with ASD, particularly shorter higher fre- 8.1. AUTISM 243 quency and less complex forms. Where declarative memory is dysfunctional, such compensation is less available. The authors suggest that declarative memory is often but not always spared in ASD, resulting in relative sparing of lexical knowledge. However rule-governed compositional aspects of grammar are largely abnormal in ASD. This results in dependance on the use of memorised complex representations. Baron-Cohen (1988) reviewed the literature describing language and pragmatics (social use of language) in autism. He described an Affective theory and a Cognitive theory. The Affec- tive theory proposes that the social and communicative deficits in autism are primarily affective. The theory articulated by Hobson (1986) starts from the assumption that normal infants are pre- wired to be sensitive to and comprehend another person’s emotions, which are perceived “directly” in their bodily expressions. He then proposes that the development of a symbolic capacity and a conceptual role-taking ability are both directly derived from the infant’s affective relationships with others. The infant thus comes to appreciate another person’s way of conceiving and seeing an object, and it is this that provides the infant with the notion of symbolic interpretation and other people’s conceptual viewpoints. Baron-Cohen (1988) points out that while Hobson’s model may account for his 1986 experiment showing that autistic children have difficulty in matching facial, vocal, and gestural emotional expressions, it does not necessarily imply difficulty in understanding beliefs. A second axiom of Hobson’s theory is that a non-functional ability to perceive people’s emotional states is an inability to abstract and symbolise, accounting for autistic children’s deficits in pretend play. However Baron-Cohen (1988) believes that the mechanism by which the development of a symbolic capacity occurs requires more clarification and empirical evidence. The Cognitive theory (Boucher, 1981; Rutter, 1983) also considers as central the autistic child’s difficulty in understanding other people’s mental states. According to Baron-Cohen (1988), 244 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME this view starts from the premise that mental states are not directly observable, but have to be inferred, requiring a complex cognitive mechanism. Thus the Cognitive theory places more emphasis on the ability to infer mental states, such as “beliefs about something” implying intentionality (Dennett, 1978), rather than emotions. Baron-Cohen (1988) describes this as the ability to attribute mental states with content to others, as a “theory of mind” (Premack and Woodruff, 1978), because it involves the person pos- tulating the existence of mental states and using them to explain and predict another person’s behaviour.

Our beliefs about or concepts of the physical world may be called ’primary representations’. However our beliefs about other people’s mental states (such as their beliefs and desires) are representations of other representations, which may be called ‘second- order representations’ (Dennett, 1978), or ‘meta-representations’ (Pylyshyn, 1978; Leslie, 1987). The Cognitive theory points out that in autism the capacity for meta-representation is impaired. Thus according to Baron-Cohen (1988), the Cognitive theory proposes that the observed pragmatic deﬁcits in autism are those that would be expected if autistic children are using language without a theory of mind (Baron-Cohen, 1988).

Numerous studies have shown motor impairments in ASD (Smith and Bryson, 1994; Leary and Hill, 1996). The aquisition of both verbal and non-verbal sequences has been reported to be abnormal in ASD. Thus abnormalities in the production or recall of hierarchically structured sequences are reported with relative sparing of repetitive sequences (Boucher, 2001). Thus Walenski et al. (2006) describe lexical memory for single-word production as spared in ASD, whereas retrieval involving frontal and cerebellar regions is deficient. Episodic memory which depends on frontal structures is also impaired. ASD is thus associated with impairments of the compositional aspects of grammar across linguistic domains related to procedural memory (Broca’s area). The authors point out that theory of mind and grammatical/procedural functions may both be related to the “dorsal stream” (Frith and Frith, 1999). However, Walenski et al. (2006) point out that whether or how the neurocognitive abnormali- 8.1. AUTISM 245 ties underlying theory of mind and procedural memory are related to each other, or to other abnormalities in ASD (eg underconnectivity, weak central coherence, or impaired executive functions) remains to be examined. According to Tager-Flusberg (2005) , language is integrally linked to the social-cognitive component of theory of mind. Ast- ington and Jenkins (1999) were able to show that language, and in particular, syntactic knowledge predicted later TOM in a longitudinal study of preschoolers. Tager-Flusberg describes studies by de Villiers and Pyers (2002), which postulate that sentential or tensed compliments are a pre-requisite of children’s acquisition of a representational TOM. Sentential complements allow for the embedding of tensed propositions under a main verb. Thus while a main clause may be true, an embedded clause may be false. Tager-Flusberg (2000) compared autistic with age-, IQ- and language- matched mentally retarded adolescents on three experiments that tested knowledge of the syntactic and semantic properties of sentential compliment constructions, including both communication and mental state verbs. Performance on the complementation tasks was significantly related to whether they passed or failed false belief tasks. This ability was the single best predictor over and above IQ and language. Tager-Flusberg (2005) concluded that her claim for the distinction between perceptual and cognitive levels for representing mental states was consistent with other accounts of the hierarchical nature of representational systems (Sperber, 1997), including the significant role played by language at the meta-representational level (Sperber, 2000b) . Pulvermuller et al. (1995) have shown that processing of individual words is associated with increased gamma EEG activity, compared with processing of non-words. Furthermore, the location of this activity depends on the semantic nature of the words. Concrete nouns are associated with impaired gamma activity over the visual and visual association cortex, whereas action verbs lead to increased activity overlying the motor cortex. Thus gamma activity corresponding to words processed in differ- 246 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME ent local regions must be bound together into a coherent global pattern to represent the meaning of the sentence. “A deficit in this binding process would lead to individual words being processed out of context. However processing the meaning of individual words would be unimpaired because it involves local activity and well-learned associations between semantic and phonological representations” (Pulvermuller et al., 1995). Thus the authors have proposed that in autism, local specialised networks process information in increasing isolation, caused by deficits in temporal binding between local networks. Also abnormalities of visual perception and language processing are associated with weak central coherence in terms of reduced coherence of oscillatory activity in the gamma frequency band.

8.1.2 Theory of Mind and Autism

Boucher (1989) reviewed the hypothesis that autistic people have impaired meta- representative ability and, as a result, lack a theory of mind (TOM). According to Boucher, the term TOM derives from a paper by Premack and Woodruff (1978) which was entitled “Does a chimpanzee have a theory of mind?”. Premack described experiments in which chimpanzees exhibited rudimentary understanding of the mind of others, though the question has remained a matter of debate. In the same volume, Dennett (1978) called beliefs about beliefs “second-order” representations. He also allowed for “third-order” representations, such as “my belief that John thinks that Mary imagines that snow is black”. Boucher (1989) suggests that Perner (1988) and Leslie and Frith (1988) both used the term ‘theory of mind’ in a narrow sense. Perner (1988) suggested three levels of representational abilities. At the first level, a child encodes perception (knowledge base). At the second level, the child is able to manipulate his/her knowledge base to engage in pretend play and distinguish between things real and imaginary, while at the third level the child becomes able to reflect on his/her own representations. Perner calls the ability 8.1. AUTISM 247 to reflect on alternative models or representations of the world “meta-representation”, and maintains that only at this stage, can a child be said to have a ‘theory of mind’. Leslie (1987) also maintained that pretence, like TOM involves second-order representations and that pretence is a manifestation of a primitive theory of mind. Leslie and Frith (1988) defined pretend play as requiring at least one of: (1) One object is substituted for another. (2) Non-existent properties are attributed to an object. (3) Absent objects are imagined. Leslie (1987) suggested that pretend play was dependent on second-order representations, requiring first a primary representation and then a secondary representation, which is a copy of the primary representation, but can be manipulated in play, without distorting the original primary representation, utilising a decoupling mechanism. A broad perspective on the TOM hypothesis relates the children’s cognitive modelling of other people to the relationship between this modelling and communication skills. Boucher (1989) has also discussed the development of the connection between the TOM hypothesis and autism, first postulated by Baron-Cohen et al. (1985); Baron-Cohen (1987). Baron- Cohen et al. (1986) utilised false-belief pictures to demonstrate that autistic children performed significantly worse than controls. In Baron-Cohen (1987), he investigated symbolic (pretend) play to support the hypothesis of a dissociation between functional and pretend/symbolic play in autism, citing this as evidence of impaired meta-representational ability. However, according to Boucher (1989), one of the difficulties in assessing the theory of mind hypothesis of autism is that there has been no clear statement of what is meant by key terms such as ‘theory of mind’ and ‘meta-representation’. She outlined a definition which might be acceptable to its proponents as:

(1) That autistic individuals have impaired ability to attribute mental states to others. (2) That this is caused by some speciﬁc impairment of higher-order representational capacity. (3) That this speciﬁc impairment, when fully understood will be seen to be primary in the sense of (a) not itself being in need of further be- 248 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME

havioural explanation and (b) being able to explain all the criterial features of autistic behaviour. (Boucher, 1989)

Thus impairment of higher-order representational capacity is consistent with a higher-order domain-general incapacity. Boucher (1989) questions whether the fact that autistic children appear not to attribute mental states to others means that they are unable to attribute mental states to others. She quotes her own work which suggests that autistic children can pretend play as adequately as language-matched controls, but do not spontaneously do so. For example, Lewis and Boucher (1988) reported a study in which autistic children spent less time than controls in both pretend and non-pretend (functional) play. Boucher (1989) suggests that autistic children’s paucity of spontaneous imaginative play (both functional and pretend) is at least partly the result of a problem with motivation or initiative. She also points out that the TOM hypothesis is a circular argument in that autistic children’s impaired ability to attribute mental states to others is caused by impaired meta-representative ability. Thus other tests than pretend play are required to break this circu- larity. She points out that DeGelder (1987) argues that having a theory of mind is not a homogeneous ability, but involves biological functions of interaction, linguistic and conversational skills, as well as high level conceptual and representational abilities. DeGelder (1987) argues that it is likely that lower order biological functions are impaired in autism, since autism originates in infancy, long before there is evidence of an impaired theory of mind. Boucher (1989) concludes that impaired meta-representational ability in autism is likely to be secondary to some impairment of lower level and earlier developing capacities. While Baron-Cohen (1988) and Frith (1989b) have argued that impaired ability to attribute cognitive/volitional states to others can explain all features of autism, Boucher (1989) argues that social and communicative impairment are unlikely to be completely explained by impaired meta-representation. For example “failure to use available language and repetitive use of language are 8.1. AUTISM 249 not so easily explained in this way”. Importantly, Frith (1989b) states that repetitive behaviour cannot be explained in terms of impaired meta-representation. Wing (1988) described a triad of impairments in autism consisting of: 1. Social impairment 2. Communicative impairment 3. Impairment of imaginative activity with substitution of repetitive activity. The triad is similar to the DSM-IV, but signiﬁcantly theory of mind explanations do not explain the restricted repetitive and stereotyped patterns of behaviour in Wing’s triad (which may suggest a domain-general deﬁcit to explain the control of behaviour by lower-level stereotyped routines, as discussed by Belmonte et al. (2004a) below).

8.1.3 Central coherence theory

Frith (1989a) proposed that autism is characterised by a “specific imbalance in integration of information at different levels”. She points out that a characteristic of normal information processing is the tendency to draw together diverse information to construct higher-level meaning in context, “central coherence”. Frith proposed that central coherence is disturbed in autism, providing a parsimonious explanation of the above deficits and assets. Frith (1989a) cite data from the embedded figures test (Shah and Frith, 1983) which demonstrated that autistic children picked out hidden figures more rapidly than controls. The authors suggested there was preliminary evidence to suggest the central coherence hypothesis as a good candidate for explaining ‘idiot savant’ phenomena. They also suggested that the central coherence hypothesis might explain executive function deficits in autism, where inhibition of pre-potent reactions was dependent on recognition of context-appropriate responses. They also pointed out that an area for future definition was the level at which coherence is weak in autism. “While Block Design and Embedded Figures tests appear to tap processing characteristics at a fairly low or perceptual level, work on memory and verbal comprehension suggests higher-level coherence deficits” (Frith, 1989a). The authors sug- 250 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME gest that a way forward might be to contrast local coherence within modular systems and global coherence across these systems in central processing. For example, they suggest that within text there may be the word-to-word effect of local association, the effect of sentence context, and the longer effect of story structure. Similarly, Norris (1990) found that building, a connectionist model of an “idiot savant date calculator” only succeeded when forced to take a modular approach. Frith and Frith (1999) have pointed out that the success of human social interactions depends on the development of a ‘social intelligence’, whose components include knowing one’s place in society, learning from others and teaching novel skills to others. While the awareness of others and social relationships are highly developed in apes and monkeys, the use of deception in social situations is thought to be based on learned behaviour rather than mental state attribution (Byrne and Whiten, 1990). How- ever, according to Frith and Frith (1999) mental state attribution is demonstrated by age four for children, where deliberate deception is commonplace. Children orient towards persons and imitate observed actions by eighteen months. However autistic children appear not to orient towards other people’s attention focus and do not engage in pretend play, or understand false belief (for example in the Sally-Anne Task most 4 year olds recognise that Sally, who has not seen Anne transfer a ball from one hiding place to another will think the ball is in its original hiding place, but this ability is lacking in autistic children). It is suggested by Frith and Frith (1999) that empirical studies of mentalising in normal development and autism find that this ability is largely independent of other abilities. The abilities relevant to mentalising are described as (i) the ability to distinguish between animate and inanimate entities (ii) the ability to share by following the gaze of another agent (iii) the ability to represent goal-directed actions and (iv) ability to distinguish between actions of the self and others. The authors describe the location of regions in the superior temporal sulcus (STS) activated 8.1. AUTISM 251 in studies of mentalising, while adjacent cells respond to particular directions of gaze. They describe evidence that activity in the STS relates to observations of monkey’s goal-directed movements (Emery et al., 1997). Such mirror neurons could provide the basis for abstract representation of goals. However it appeared that cells which were activated by sights and sounds generated by others were separate from those representing the self. According to Frith and Frith (1999), evidence from brain imaging studies in humans suggested a role for medial frontal areas in reporting mental states. Shima et al. (1991) reported cells in the posterior part of the anterior cingulate (anterior to the motor cingulate) where activity was observed before the production of self-initiated movements.

Frith and Frith (1999) suggested that a human mentalising system might include (i) STS for the detection of the behaviour of agents and analysis of the goals and outcomes of this behaviour. (ii) inferior frontal regions for representations of actions and goals and (iii) anterior cingulate cortex ACC/medial prefrontal regions for representations of mental states of the self. The authors propose that the analysis of another agent’s behaviour in conjunc- tion with the representation of one’s own mental state allows us to make inferences about the intentions of that agent. They also point out that a further striking implication of the data is that ability to mentalise appears to have evolved from the dorsal action system, rather than the ventral object identification system (Milner and Goodale, 1995). According to the authors, the components of social intelligence that developed in the monkey before the emergence of mentalising abilities include recognition of subtle differences in emotional expression, recognition of other individuals, and recognition of their status and relationships. These abilities are thought to depend on complex and sophisticated object recognition of the kind supported by the ventral system. In contrast Frith and Frith (1999) believe that the emergence of mentalising required the development of capabilities relating to the representation of actions, the goals implicit in actions and the 252 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME intentions behind them. (This is consistent with the inheritance of the dorsolateral PFC in human goal-related activity). Tager-Flusberg (2005) has pointed out that the assumptions that underlie studies of adult patients with acquired lesions do not hold for neuro-developmental disorders, where disruption of early brain development affects cognitive systems. Thus adult studies of modularity may not apply in a developmental context, where brain plasticity is of greater significance (Courchesne et al., 2003). In addition, intellectual level and language variation may obscure syndrome-specific patterns. Tager-Flusberg (2005) has approached these issues by comparing studies of William’s Syn- drome (WMS) and autism. The TOM hypothesis has been thought to explain the failure of children autism on tasks investigating TOM, and to also explain deficits in pretense, social functioning and communication. These deficits were taken as evidence in support of the modularity of the mind (Baron-Cohen, 1995; Leslie and Roth, 1993) . In contrast, WMS is thought to be characterised by sparing in the TOM domain, because of relatively good language skills, excellent face processing ability, strong social interest and attention to faces and people (Mervis et al., 2003), suggesting a double dissociation, where in autism TOM is impaired while visual-spatial skills are spared, while in WMS, TOM is spared, while visual spatial skills are impaired. However, Tager-Flusberg (2005) also points out that the TOM hypothesis of autism has been criticised on the basis that mentally retarded children and adolescents also show TOM deficits (Zelazo et al., 1996). Also autistic symptoms emerge very early in infancy, before TOM is measurable or developed in normal children. Finally some autistic children pass TOM and false belief tasks (Baron-Cohen, 1995). Tager-Flusberg (2000) has shown that adolescents with WMS performed no better than IQ-matched controls on TOM, false belief and explanation of action tests. Tager- Flusberg (2005) has proposed a two-component model to help explain the above contradictory findings. The model describes a primary social-perceptual and a higher-order social-cognitive ca- 8.1. AUTISM 253 pacity. The latter refers to the meta-representational capacity to make more complex cognitive inferences about the content of mental states, requiring information across time and events. Thus, Tager-Flusberg (2005) describes a social-perceptual component of TOM, which builds on the innate preferences of infants to attend to human social stimuli, especially faces and voices (Mehler and Dupoux, 1994). This component develops in the first year of life, allowing infants to utilise perceptual information from faces, voices and gestures to interpret the intentions and emotional states of other people. The social-cognitive component of TOM builds on the earlier emerging perceptual component. It is involved in integrating information, not only from perceptual cues, but from sequences of events over time, and is linked to working memory and language. This capacity is developed by four years of age (Hale and Tager-Flusberg, 2003). Inter- estingly, neurophysiological evidence on the ‘social brain’ (Broth- ers, 1990) suggests that the amygdala and associated regions of the medial temporal cortex are involved in the facial recognition of emotions and other mental states (Baron-Cohen et al., 1999) and intentional motion (Bonda et al., 1996). On the other hand, the areas subserving the social-cognitive component of TOM are thought to involve the orbito-frontal cortex (which monitors social appropriateness of action), and medial frontal cortex. Tager-Flusberg (2005) describes the early social-perceptual deficits in autism as being associated with failure to perceive others as intentional and joint attention deficits. The ability to read mental states from the eye region of the face is deficient even in older high functioning autistic or Asperger people. In addition, Tager-Flusberg describes examples where autistics are able to pass TOM tasks using a language strategy, while in the case of WMS, it was found that young WMS children showed greater empathy and more appropriate affect than a matched group of Prader-Willi children. However in some cases WMS subjects performed worse than normal controls, which might implicate some domain-general processing deficits such as attention 254 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME and response speed.

8.1.4 Face processing in autism

Pierce et al. (2001) have investigated the development of face processing in autism. They point out that processing of the human face is the focal point of most social interactions, and that autism is one of the only disorders where affected individuals spend reduced amounts of time engaged in face processing from birth. The investigators utilised functional MRI haemodynamic responses during a face processing task to investigate four regions of interest (ROI’s): the fusiform gyrus (lateral to the parahippocampal gyrus in the temporal lobe), the inferior temporal gyrus, the medial temporal gyrus, and the amygdala in adults with autism and controls. They point out that unlike autistic subjects, 100 percent of normal subjects showed maximal neural responsiveness to faces in the fusiform gyrus, and that two opposing explanations have been put forth to explain the invariance of the fusiform face area (FFA) activation in normals. The first interpretation suggests that this reflects an innately determined face module that is specific to and required for face processing - namely a ‘domain-specific’ view. The second view posits that the FFA is an experience-dependent neural region, evolved to process subordinate levels of extremely familiar classes of objects, a ‘domain general’ view. The investigators found that while autistic subjects could perform the face perception task, none of the regions supporting face processing in normals were found to be significantly active in the subjects investigated. They found that in every autistic subject, faces maximally activated aberrant and individual- specific neural sites (e.g. frontal cortex, primary visual cortex and cerebellum). Pierce et al. (2001) also found decreased structural amygdala volume in adult autistics. They point out that this finding is consistent with the idea that the amygdala is abnormal in autism, and that the amygdala plays a key role in establishing the social 8.1. AUTISM 255 significance of a face, interpreting it as threatening or fearful, monitoring eye gaze, and assigning hedonic values to stimuli in general. The authors suggest that an absence of normal amygdala functioning would thus prevent many of the normal social perceptual activities of a newborn and young child, preventing activity-dependent development and refinement of the amygdala. Thus malfunction of the amygdala could represent an essential neural insult that initiates a cascade of social maldevelopments found in the disorder. They posit a “critical period” for the development of the FFA, not just for experience with faces per se, but between exemplars of a particular class of objects. The authors finally suggest that abnormal neurofunctional responding to faces in autism is probably the result of inefficient or faulty networks, extending beyond the FFA and amygdala, relating to top-down processes such as fronto-parietal networks involved in attention. Baron-Cohen et al. (2000a) have discussed the “amygdala theory of autism”. They cite Brothers (1990), who proposed a network of neural regions comprising the “social brain”. They describe the amygdala as a collection of nuclei lying beneath the uncus of the temporal lobe at the anterior end of the hippocampal formation and the anterior horn of the lateral ventricle. “Evi- dence over the last two decades has revealed that the amygdala is intricately connected with many brain regions, including neocortex, basal forebrain, the limbic striatum (nucleus accumbens and ventral pallidum), the neostriatal structures (the caudate nucleus and the putamen), the hippocampal formation and the claustrum (Baron-Cohen et al., 2000a)”. According to the authors, the amygdala receives considerable visceral inputs from olfactory input and the hypothalamus, as well as inputs from temporal and anterior cingulate cortex, while fibers leave the amygdala to reach many of the same areas that send efferents to it. Adolphs (2002) has reviewed neural systems for recognising emotions. He focusses on the “so-called” basic emotions, most reliably recognisd from facial expression: happiness, surprise, fear, anger, disgust and sadness. Adolphs believes that these brain pro- 256 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME cesses involve perceptual processing (identifying the geometric configuration of facial features in order among different stimuli on the basis of their appearance, and recognition of the emotional meaning of a stimulus). Thus recognition may rely on different strategies, including, for example, knowledge of the concept of fear, from both the lexical label, perception of the response of fear triggered in the subject (or a cortical representation of fear), or knowledge of the motor representations required to produce the expression shown in the stimulus. Adolphs (2002) describes a large number of different structures participating in the recognition of facial emotion, including the occipito-temporal cortices, basal ganglia and right parietal cortices. Adolphs (2002) points out that single-unit studies in monkeys’ intracranial field potential, and studies in neurosurgical human patients have all provided evidence that cortical areas in the lateral parts of the inferior occipital gyrus, fusiform gyrus and superior temporal gyrus are disproportionately important in face processing. While the fusiform gyrus is more activated by faces than other objects, it is also activated by subordinate-level categorisation. Adolphs suggests that the fusiform gyrus is especially involved in representing the static features of faces, and conse- quently contributing to encoding identity, whereas the superior temporal gyrus is especially involved in representing the dynamic changeable features of faces, and therefore to encoding facial expression and direction of gaze. According to Adolphs (2002) the construction of a detailed perceptual representation of a face appears to require 170 secs, but some rapid, coarse categorization of gender and emotion can occur with substantially shorter latencies, indicating the existence of cruder perceptual routes that occur in parallel to routes for the full encoding of the stimulus. Thus the earliest activity that dis- criminates between emotional facial expressions is seen in midline occipital cortex as early as 80 msec to 110 msec. While information sufficient to distinguish faces from other objects is encoded in 120 msec, responses encoding fine-grained subordinate 8.1. AUTISM 257 information sufficient to distinguish different emotional expressions only appear in 170 ms. According to Adolphs (2002) the findings suggest the possibility that responses to emotional stimuli in visual cortices are modulated by feedback, perhaps from structures such as amygdala and orbitofrontal cortex.

8.1.5 Neural abnormalities in autism

The above psychological theories of autism suggest a ‘theory of affect’ in autism, related to a deficit in shared attention and use of context, giving rise to impairments in social relatedness. Thus an incapacity to process facial expression suggests an incapacity to recognise and process emotion in others. Morris et al. (1996) utilised positron-emission-tomography (PET) to measure response to fearful and happy facial expressions in five normal volunteers. They found that the neural response in the left amygdala was significantly greater in response to fearful as opposed to happy expressions. The response showed an increase with intensity to fearful and decrease with happy expressions. On the other hand, a contrast of happy with fearful expressions was associated with activations in the right medial temporal gyrus, right putamen, left superior frontal lobe and left calcarine fissure. The authors comment that the primate amygdala receives substantial inputs from temporal visual-association areas, and also has a strong anatomical link with the autonomic system. Thus damage to the amygdala impairs the formation of conditioned automomic responses to aversive stimuli. Perception of an expression of fear in a conspecific may trigger an autonomic response to potential danger, accounting for the observed amygdala response to fearful faces. It has been shown that amygdala-lesioned monkeys become socially isolated (Kling and Brothers, 1992). For example, Klu- ver and Bucy (1939) described a syndrome of ‘psychic blindness’ in which lesions of the temporal lobe (including amygdala, hippocampus, and temporal cortex) produced a syndrome of over- 258 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME reaction to all objects, hypoemotionality, loss of fear, hypersexual- ity and hyperorality. Baron-Cohen et al. (2000a) describe the only animal model of autism as involving the ablation of the amygdala in rhesus monkeys (Bachevalier, 1991), and the Kluver-Bucy syndrome as a fairly good animal model of autism (Hezler and Griffin, 1981). They also describe the orbito-frontal and medial frontal cortex as important for social intelligence, suggesting a social network as described by Brothers (1990). Ohnishi et al. (2000) have investigated brain abnormality in autistic patients using single photon emission tomography (SPECT), to show the relationship between autistic symptomatology and regional cerebral blood flow. They selected nine items from the Childhood Autism Rating scale (Schopler et al., 1980): namely echolalia; abnormal prosody; abnormal grammatical construction; treat people as interchangeable; no symbolic play; no use of finger pointing; fascination with certain objects; insistence on sameness; and stereotyped behaviour. Interestingly, a factor analysis (based on mild, moderate, and severe scores in 23 subjects) generated two factors with eigenvalues greater than unity, accounting for 83 percent of the variance. The first factor loaded heavily on: treating people as interchangeable; abnormal prosody, echolalia; and no symbolic play and was designated as ‘impairments in commnication and social interaction’. The second factor loaded heavily onto: stereotyped behaviour; insistance on sameness; and fascination for certain objects, designated as ‘obsessive desire for sameness’. The investigators found that the autistic children had abnormal regional cerebral blood flow rCBF in the bilateral insula (close to the claustrum), superior temporal gyri and left prefrontal cortices compared with non-autistic children. The authors point out that disorders of perception and modulation of sensory experience have previously been related to disorders of insular cortex/paralimbic connections (Mesulam and Mufson, 1984; Augus- tine, 1996). They also found that impairments in communication and social interaction (thought to be related to TOM) were corre- 8.1. AUTISM 259 lated with altered perfusion in the medial prefrontal cortex and anterior cingulate cortex, and the obsessive desire for sameness and with altered perfusion in the right medial temporal lobe. The authors discuss the relevance of TOM as an explanation of cognitive disorder in autistic children (Baron-Cohen et al., 1985; Frith and Happe, 1994), referring to the ability to explain and predict the behaviour of others in terms of their own mental states. The authors suggest that their findings of association between impairments in communication and social interaction and left anterior cingulate (BA 32) could be a crucial component of a brain system that underlies TOM. According to the investigators, the area BA 32 receives afferents from BA 9, the temporal pole, and orbito frontal cortex, a brain system which has been implicated in higher cognitive function, and in the expression and recognition of affect, functions that are compromised in autism. The investigators did not find rCBF changes in the amygdala in factor 1, but did find an unexpected correlation between the score for factor 2 and altered rCBF in the right medial temporal lobe (the ‘repetetive behaviour’ factor). Frith and Happe (1994) have pointed out that the mentalising deficit theory of autism does not explain all features of autism, including:

Restricted repertoire of interests (DSM-III-R,1987). Obsessive desire for sameness (Kanner and Eisenberg, 1956). Islets of ability (Kanner, 1943). Idiot savant abilities (Rimland and Hill, 1984). Excellent rote memory (Kanner, 1943). Preoccupation with parts of objects (DSM-IV) (Frith and Happe, 1994).

The important omission of restricted repetitive behaviour patterns tends to strengthen Boucher’s argument that the fundamental features of Autistic Disorder are not completely accounted for by impaired meta-representation or TOM. Turner (1997, p62) pointed out that while the importance of repetitive behaviour to the autistic syndrome has been emphasised since Kanner (1943), 260 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME though it was sometimes regarded as a marker of non-specific impairment. A second reason for the apparent lack of interest in repetitive behaviour is the assumption that the behaviours function as a coping mechanism, and thus as a secondary feature of the disorder. However, such accounts fail to explain why repetitive behaviours are so ubiquitous, pervasive and endur- ing in autism. Turner (1997, p60-61) systematically investigated the various theories of repetitive behaviours in autism. She first described a taxonomy of repetitive behaviours, including tics, stereotypic movements, self-injury, stereotyped manipulation of objects, abnormal object attachments and preoccupations, insistence on sameness of environment, rigid adherence to routines and rituals, repetitive use of language, circumscribed interests and obsessions and compulsions. These behaviours are characterised by 1. high frequency of repetition, 2. the invariant way in which the behaviour or activity is pursued and 3. is inappropriate or odd in its manifestation or display. Turner (1997) compared 22 high-functioning autistic children and adults with a verbal IQ above 75 (HFA) and 22 learning disabled individuals with autism and IQ below 75 (LDA), as well as two groups of age-, sex-, and ability matched controls. She found that 98 percent of the autistic subjects displayed repetitive behaviours in three or more of the eleven classes of the above behaviours in comparison with 17 percent of controls, but showed little effect of age or ability on the repetitive behaviour. Turner suggested that the findings supported Kanner (1943)’s original assertion that repetitive behaviour is a core feature of autism: “if we want to provide a full explanation of the pervasive tendency to repetitive behaviour that characterises autism, and sets it apart from other clinical disorders, then we must look towards those hypotheses which provide a mechanism of autistic repetitions”. Turner (1997, p66-68) also investigated whether the theory of mind deficit is able to explain repetitive behaviour. The theory predicts that levels of repetitive behaviour will be highest when the individual is in a novel or unpredictable social situation, and 8.1. AUTISM 261 lowest in a highly familiar environment. However, most studies report that rates of stereotypic behaviour are lowest during periods of social interaction (Clark and Rutter, 1981; Dadds et al., 1988). The theory of central coherence (Frith, 1989b) explains the performance of autistic subjects in terms of inability to integrate perceptual information, leading to a focus on seemingly insignifi- cant details of the environment. According to Turner (1997) , this account is consistent with the common insistence that even mi- nor features of the environment remain unchanged, including the narrowness of repetitive behaviour. However, when Turner investigated the relationship between performance on the Children’s Embedded Figures task and four domains of repetitive behaviour, she found no difference between high and low scorers on this task. Thus Turner (1997, p69) is critical of the ability of the central coherence theory to explain the high degree of repetition and invariance characteristic of autistic behaviour. On the other hand it may be that because autistic performance on the Embedded Fig- ures Test is superior at all levels it may not discriminate.

Turner (1997, p72) suggested executive function or the supervisory attentional system (SAS) is crucial to normal, flexible and adaptive regulation of behaviour (Norman and Shallice, 1980). Inability to inhibit actions is thought to result in perseverative behaviour. However response perseveration does not account for inappropriate responding, though it may account for “stuck in set”. Turner reports that both HFA and LDA suspects who produced the lowest number of novel responses and highest number of immediate repetitions on a Sequence task also showed the highest number of repetitive movements and extreme circumscribed interests. However her argument appears somewhat circular in terms of low novel responses being associated with circumscribed interests. Finally Turner (1997, p89) found that autistic subjects produced significantly fewer responses than learning disabled controls despite a mean IQ 40 points higher. She concluded that repetitive behaviour is associated with general disruption of a system such as the SAS, which is responsible for con- 262 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME trolling volitional activity, but this does not explain the insistence on sameness of environment, or why repetitive behaviours are so constant across different individuals and different cultures. Belmonte et al. (2004a) have pointed out that while people with autism have been described as suffering from a lack of ‘central coherence’, the field of autism itself suffers from a lack of integration of differing analytical and theoretical concepts, including executive function, complex information processing, theory of mind and empathy. They point out that autism, defined and diagnosed by purely behavioural criteria was first described and investigated, using the tools of behavioural psychology, but more recent results from brain anatomy, physiology, genetics and biochemistry have not been fully integrated. The triad of deficits comprising impaired social interaction, impaired communication, restricted interests, and repetitive behaviours are believed to be the extreme of a spectrum of abnormalities including Asperger’s syndrome and the “broader autistic phenotype”. Belmonte et al. (2004a) have attempted to integrate some of the above observations and theories in terms of abnormal neural connectivity. They differentiate local connectivity within neural assemblies from long-range connectivity between functional brain regions and distinguish physical connectivity associated with synapses and tracts from the computational connectivity associated with information transfer. The authors posit that in the autistic brain high local connectivity may develop in tandem with low long-range connectivity, perhaps as a consequence of widespread alterations in synapse elimination and/or formation. Furthermore, indiscriminately high physical connectivity and low computational connectivity may reinforce each other by failing to differentiate signal from noise. “The model is consistent not only with impairments in higher order cognition described by the diagnostic triad, but also impairments of motor coordination (Teit- elbaum et al., 1998), perceptual abnormalities such as high visual motion coherence thresholds (Milne et al., 2002) and broad tuning of auditory filters (Plaisted et al., 2003) and abnormal growth 8.1. AUTISM 263 within regions of local, but not long-range white matter projections (Herbert et al., 2004), and the substantial comorbidity with epilepsy (Ballaban-Gil and Tuchman, 2000)”. Belmonte et al. (2004a) have utilised functional Magnetic Res- onance Imaging (fMRI) to investigate the above concepts in autistic subjects. They postulated that in an over-connected network, sensory inputs should evoke abnormally large activation for attended and unattended stimuli alike, giving rise to an overall increase in activation but reduction in selectivity within sensory regions. Conversely “brain regions subserving integrative functions will be cut off from their normal inputs and should therefore manifest reductions in activation and in functional correlations with sensory regions” (Belmonte et al., 2004a). Thus a combination of EEG (Belmonte, 2000) and fMRI (Belmonte and Yurgelun-Todd, 2003) measures in a task of visual spatial attention demonstrated exactly this pattern in autistic subjects. Belmonte and Baron- Cohen (2004) were able to show abnormally strong activation in parietal cortex during suppression of distractors at the same time as integrative regions in prefrontal and medial temporal cortices were abnormally quiescent. A further example of the above principles is described by Bel- monte et al. (2004a) in relation to cerebellar functions in autism. Allen and Courchesne (2003) demonstrated that in autism, cerebellar activation is abnormally low during a task of selective attention and abnormally high during a simple motor task. At the macroscopic level, there is a reduction of Purkinje cell numbers, releasing deep cerebellar nuclei from inhibition, producing abnormally strong physical connectivity and potentially abnormally weak computational connectivity along the cerebello- thalamocortical circuit. They also describe abnormal regional activation patterns during face processing (Keil et al., 1999) and describe Fragile X as a single gene disorder with proximal disruption of complex patterns of gene expression, in which dendritic spines in specific cortical regions are present in high density and are abnormally long and thin, suggesting an immature morphol- 264 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME ogy that may produce overconnectivity. The work of Courchesne et al. (2003) on head circumference measurements suggests over- growth and exuberent synaptogenesis and dendritic arborization within the first 6-14 months in the frontal lobes (essential for attention, social behaviour and language), but lesser effects in the precentral gyrus and orbital cortex. Thus “the cortical areas most affected are precisely those broadly projecting, phylogenetically and ontogenetically late-developing regions that are essential to complex cognitive functions such as attention, social behavior, and language” (Belmonte et al., 2004a). The authors conclude that abnormal neural connectivity qualifies as an explana- tory framework within which genetic and neuropathological findings in autism may be unified with neuroanatomy, neurophysiology and behaviour. Additionally the approach promises greater understanding of the mechanisms involved in both normal and pathological development of neural and cognitive systems, important for the understanding of childhood behavioural syndromes. Belmonte et al. (2004b) point out that because autism is developmental disorder, primary dysfunction can be masked by the evolution of compensatory processing strategies, and by the induction of activity-dependant secondary dysfunctions which dis- rupt behaviour in new ways. They note that a decrease in signal- to-noise can arise from abnormalities of neural connectivity in either direction, over-connection allows so much noise to pass that it swamps the signal, whereas an under-connected network passes so little signal that it becomes lost in noise.

In either case large segments of the network are constrained to either an all-on or an all-off state, and the network’s information capacity is thereby reduced. [...] Thus a failure to delimit activation can give rise to hyper-arousal in response to sensory input, a decreased ability to select among competing sensory inputs. [...] Physiologically, functional imaging has demonstrated heightened activity in autism in brain regions associated with stimulus- driven, sensory process and decreased activity in regions that normally subserve higher-order processing. Studies have shown heightened activity during face processing in peri-striate cortex (Critchley et al., 2000), inferior temporal gyrus (Schultz et al., 2000), while fusiform activity is abnormally low. Similarly height- 8.1. AUTISM 265

ened activity in the superior temporal gyrus during inference of mental state from pictures of eyes, and decreased connectivity between extrastriate visual areas and prefrontal and temporal areas associated with inference of mental state, while prefrontal and temporal activations are again abnormally low (Belmonte et al., 2004b).

In modular terms these deficits can be seen as excess activity in specific modules and diminished activity in higher order domain-general activities (areas which integrate activity in higher brain centers). Thus in children with autism, the visual evoked N2 potential to novel stimuli is augmented during task performance, even when these stimuli are not relevant to the task. However when a response is required to an auditory stimulus, the P3 in these same children with autism is abnormally generalised to occipital sites overlying visual processing areas. Thus perceptual filtering occurs in an all-or-none manner with little specificity for location of the stimulus, behavioral relevance of the stimulus, or even for the sensory modality in which the stimulus appears. Thus in the absence of a normally functional mechanism to bias sensory processing towards attended stimuli, all stimuli receive much the same degree of sensory evaluation, and the irrelevant stimuli must then be actively discarded in a manner that creates a bottleneck. Selective attention is thus impaired. Belmonte et al. (2004b) have outlined a developmental theory of abnormalities of perceptual processing, dependent on the idea of a developmental chain of abnormal function.

When a developing brain is confronted with an abnormal constraint on information processing, it will evolve an abnormal organisation in order to accomodate that constraint, resulting in a succession of autistic behavioural abnormalities extending into sensory, motor and later developing cognitive functions. [...] From the earliest months of infancy, the flood of input generated by over-aroused, under-selective primary procesing would overload nascent higher-order cognitive processes (Belmonte and Yurgelun- Todd, 2003). Faced with the bottleneck, Belmonte et al. (2004b) believe the brain would likely evolve a cognitive style that avoids reliance 266 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME on high-level integrative processing and instead emphasises low- level features. This pattern of autistic perception has been characterised by Frith (1989a) as weak central coherence, which results in a loss of identity of fragments once they are assembled into a single object, but can confer advantages in tasks that demand attention to detail. Thus people with autism have an advantage in the Embedded Figures Test and process faces as collections of individual features rather than centrally coherent gestalts, while their rote memory is superior to normal. According to Belmonte et al. (2004b) weakened central coherence may be a secondary property emerging from the interaction of normal cognitive development with abnormal neural information processing, encouraging unusual cognitive dependence on low-level processing of individual details. They suggest this may impair the use of contextual information on complex perceptual and executive tasks and may impact on capacity to form a model of another person’s mental state (theory of mind), impairing the development of joint attention and shared affect. Also failure to use context may, according to the authors, base learning on statistical associations, leading to a preference for ritualised, scripted repeatable interactions. Brock et al. (2002) have proposed a ‘temporal binding’ theory of central coherence in autism. They suggest that whereas typical brain development involves the emergence of functionally specialised but nevertheless integrated brain regions, brain development in autism involves the emergence of functionally specialised brain regions that become increasingly isolated from each other over time. Thus while localised activity will proceed as normal, cognitive activity requiring integration will be impaired. This differs from Belmonte et al. (2004b) who propose excess activity at the local level. However both emphasise the relationship between local modular and central integrative activity. Brock et al. (2002) have suggested that temporal vs combinatoral binding of features is an important deficit in autism. Temporal binding implying that neurons responding to the same object are tagged by temporal 8.1. AUTISM 267 correlation of their firing patterns. Thus local groups of cells representing individual features may be bound together into a larger global group by their synchronous firing. The authors have applied the temporal binding theory to an explanation of the reduced influence of context on language in language processing. Levy (2007c) has described the case of an autistic girl who was able to sing memorised tunes at preschool but unable to produce sentences and/or communicative language:

C (Age 13 years) was first seen following a 6-month period of ‘elec- tive mutism’ at age 3 years, despite initially ‘normal’ speech development, where she could sing learned preschool songs. At age 5 she was seen at a developmental clinic after her class teacher reported communication problems with difficulty following directions, picking up cues and answering questions. She was found to have a significant delay in both language and speech and had difficulty in forming peer relationships. At age 6 she repeated the year to assist with social skills development and received speech therapy. Psychological assessments indicated a mild global delay in general ability, verbal skills and verbal comprehension. C was intermittently prescribed dexamphetamine, when this was ceased following concerns regarding slow growth and depressed mood. At the beginning of 2003 (age 11), C was noticed to have become anxious, unreactive and flat in affect and unmotivated. She was increasingly socially withdrawn, with depressed mood and affect, minimal verbal responses, repetitive behaviours and obsessive slowness, and was diagnosed as having a Depressive Disorder. She was later noticed to be smiling and giggling to herself and com- plaining of voices in her head. She was minimally communicative with mostly yes/no answers and episodes of shouting “be quiet”. C was treated for a psychotic disorder, but her underlying condition was subsequently diagnosed as autism. (Levy, 2007c)

Of interest in the present context is the complex behavioural presentation following C’s early communication problems. After her psychotic illness was treated, her communication deﬁcit was clearly autistic in that she was able to repeat words and short sentences, but was unable to construct sentences. Her early capacity to sing may also have represented a capacity to repeat songs, but she lacked capacity to integrate words into sentences, illustrating a localised modular form of language as described by Walenski et al. (2006) above. 268 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME

8.2 Asperger’s Disorder

Is defined by similar criteria to Autistic Disorder, namely qualitative impairment in social interaction and restricted, repetitive and stereotyped patterns of behaviour, interests and activities, but specifies that there is no clinically significant general delay in language (e.g. single words used by 2 years and communicative phrases used by age 3 years). Autistic Spectrum Disorders include Autistic Disorder and Asperger’s Disorder (DSM-IV). Ac- cording to Allen et al. (2003) these disorders exhibit significant repetitive behaviours and restricted interests, including ritualistic behaviours such as counting, tapping, flicking, or repeatedly restating information, and compulsive behaviours such as lining up objects requiring a rigid adherence to routine and a marked resistance to change. While the onset of these disorders is believed to be prior to or at birth, they are poorly understood. According to Allen et al. (2003), recent advances in neuroimaging have shown delayed frontal lobe maturation as well as bilateral temporal hy- poperfusion.

8.2.1 Discussion

From the point of view of the present investigation, the concept of local overconnectivity and long-range underconnectivity in modular cortical-subcortical circuits, responsible for integration and control of behaviour provides a useful basis for the understanding of a number of childhood syndromes, where domain-specific routines may be repeated in the absence of domain-general executive integration at higher-order levels. The Belmonte model of a cognitive style that avoids reliance on high-level integrative processing and instead emphasises low-level features is particularly useful in integrating the psychological, language and rigid/repetitive behaviours characteristic of autism. Also the model of abnormal brain organisation as a result of early dysfunction is relevant not just for autism, but also for later developing deficits in the syndromes discussed below. The thesis postulates that some child- 8.2. ASPERGER’S DISORDER 269 hood syndromes may result from a deficit in the development of long-range communication between parallel specialised subcortical systems and a cortical global workspace, in which goal accomplishment is made possible by the updating of working memory and serial testing of outcomes. In the case of autism, the early absence of normal language development, accompanied by impaired capacity for social interaction and attachment, all suggest a profound absence of integration of a number of specialised subcircuits. Thus an initial domain-specific (modular) deficit in the amygdala-FFA circuit results in later domain-general deficits and a secondary lack of normal social experience, with dependence on local domain specific networks, giving rise to cognitive rigidity and problems such as repetitive behaviour (Levy and Krebs, 2006; Levy, 2007a). The relationship of the modular deficits to mirror phenomena is also of interest, given the tendency of autistic children to engage in miror-like echolalic behaviour.

8.2.2 Therapeutic implications

A recent study by Dadds and colleagues, Dadds et al. (2008) has shown that children with high psychopathic traits showed decreased eye ﬁxations during an emotional fear-recognition task. They postulated that amygdala dysfunction might contribute to failure of fear recognition in this population. Conversely, autistic subjects show hyper-reactivity to sad or fearful faces. Thus while the latter are not able to “look one in the eye”, the former are encouraged to look at the eyes in sad or fearful pictures. The authors have applied behavioural approaches to train facial expression recognition in both conduct-disordered and autistic children, where “opposite” mechanisms at the amygdala-FFA circuit may be involved. 270 CHAPTER 8. AUTISTIC DISORDER/ASPERGER’S SYNDROME Chapter 9

OCD and Tourette’s Disorder

The American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV), describes Obsessive Compulsive Disorder as:

A. Either obsessions or compulsions: Obsessions as deﬁned by (1), (2), (3), and (4): (1) recurrent and persistent thoughts, impulses, or images that are experienced, at some time during the disturbance, as intrusive and inappropriate and that cause marked anxiety or distress. (2) the thoughts, impulses, or images are not simply excessive worries about real-life problems (3) the person attempts to ignore or suppress such thoughts, impulses, or to neutralise them with some other thought or action (4) the person recognises that the obses- sional thoughts, impulses, or images are a product of his or her own mind (not imposed from without as in thought insertion) Compulsions as deﬁned by (1) and (2) (1) repetitive behaviours (e.g., hand washing, ordering, checking) or mental acts (e.g. praying, counting, repeating words silently) that the person feels driven to perform in response to an obsession, or according to rules that must be applied rigidly (2) the behaviors or mental acts are aimed at preventing some dreaded event or situation; however these behaviours or mental acts are not connected in a realistic way with what they are designed to neutralise or prevent or are clearly excessive B. At some point during the course of the disorder, the person has recognised that the obsessions or compulsions are excessive or unreasonable. Note: This does not apply to children

271 272 CHAPTER 9. OCD AND TOURETTE’S DISORDER

C. The obsessions or compulsions cause marked distress, are time consuming (take more than 1 hour a day), or signiﬁcantly interfere with the person’s occupational (or academic) functioning, or usual social activities or relationships. D. If another Axis I disorder is present, the content of the obsessions or compulsions is not restricted to it (e.g., preoccupation with food in the presence of an Eating Disorder; hair pulling in the presence of trichotillomania; preoccupation with drugs in the presence of a Substance Use Disorder; preoccupation with having a serious illness in the presence of Hypochondriasis; preoccupation with sexual urges or fantasies in the presence of a Paraphilia; or guilty ruminations in the presence of Major Depressive Disorder). E. The disturbance is not due to the direct physiological effects of a substance (e.g., a drug of abuse, a medication) or a general medical condition. Specify if: With Poor Insight: if, for most of the time during the current episode, the person does not recognise that the obsessions and compulsions are excessive or unreasonable.

9.1 Neural circuits in OCD

According to Saxena and Rauch (2000) various frontal-subcortical circuits subserve different behavioural functions and mediate the symptomatic expression of different psychiatric syndromes. The circuits originate in nearly every part of the cerebral cortex and project through different sub-compartments of the basal ganglia and thalamus, back to the cortex. Saxena and Rauch (2000) describe the lateral prefrontal subcortical circuits as being associated with major depression, whereas orbito-frontal-subcortical circuits are thought to be involved in obsessive-compulsive disorder (OCD), while sensori-motor circuits are thought to mediate symptoms of movement disorders. Saxena and Rauch (2000) outline the classical route of cortical subcortical circuits as having two loops (1) a “direct” pathway and (2) an “indirect” pathway. In primates the direct pathway is thought to project from cerebral cortex to striatum to the internal segment of the globus pallidus, then to thalamus and back to cortex. On the other hand, the indirect pathway has a similar 9.1. NEURAL CIRCUITS IN OCD 273 origin from cortex to striatum, but then projects from striatum to the external segment of the globus pallidus and then to subthalamic nucleus before returning to globus pallidus/substantia nigra, where it rejoins the common pathway to the thalamus and cortex (this is similar to the basic frontal-striato-cortical (FSC) structure described above). “Impulses along the direct pathway with two excitatory and two inhibitory projections act as a net positive feedback loop, whereas activity along the indirect basal ganglia control system, with three inhibitory connections provides negative feedback and decreases thalamo-cortical drive” (Saxena and Rauch, 2000). OCD symptoms were thought to be due to an imbalance in tone between direct and indirect striato-pallidal pathways. Prefrontal cortex and thalamus also reciprocally activate each other. Impulses along the direct pathway with two inhibitory connections tend to “disinhibit” the thalamus and activate the system, whereas activity along the indirect pathway with three inhibitory connections provides negative feedback and disinhibits the thalamus. Sachdev (2005) have described evidence implicating the orbito-frontal (OF) and anterior cingulate (AC), as well as the striatum and amygdala.

Cortex

indirect

GPe

Striatum direct

Amygdala

GPi

Thalamus

Figure 9.1: Recurrent Neural Network

According to Saxena and Rauch (2000), the different regions 274 CHAPTER 9. OCD AND TOURETTE’S DISORDER of the striatum receive input from different cortical regions. The orbitofrontal cortex, a paralimbic isocortical area which projects to the ventromedial caudate nucleus, while the dorsolateral prefrontal cortex (an associative neocortical area) projects to the dorsolateral caudate, and the anterior cingulate gyrus and hippocampal formation (limbic areas) project to the nucleus accumbens. Thus the loops consist of motor, associative, and limbic loops. Naturally occurring activity along the direct pathway would tend to direct behaviour to the execution of appropriate behavioural ‘macros’ until the need is judged by the organism to have passed. Conversely, activity of the indirect pathway may have, as part of its function, the suppression of direct pathway driven behaviours, when it is time to switch to another behaviour. With the advent of magnetic resonance imaging (MRI) studies of OCD Saxena et al. (1998) reviewed the basic science and brain imaging studies of OCD, and found that although not all studies agreed, the structural and functional brain imaging data suggested abnormalities in orbitofrontal cortex, anterior cingulate cortex, and elements of the basal ganglia circuits (described by (Alexander et al., 1986)). Thus in a “classical frontal-subcortical direct loop there are two inhibitory connections which disinhibit the thalamus and activate the system in a self-perpetuating positive feedback loop, while activity along the indirect pathway (with three inhibitory connections) provides negative feedback, inhibiting the thalamus” (Saxena et al., 1998). The “working model” of the pathophysiology of OCD suggested that in persons with OCD, there was a neurologically-mediated response-bias toward stimuli relating to socio-territorial concerns about violence, hygiene, order, and sex. The authors described the experimental and clinical evidence as suggesting that the orbitofrontal cortex is involved in the mediation of emotional responses to biologically signiﬁcant stimuli (Zald and Kim, 1996), while animal studies suggested that hoarding, another OCD symptom was mediated by the ventromedial striatum (Mogenson and Wu, 1988). Modell et al. (1989) have described a neurophysiologic dysfunc- 9.1. NEURAL CIRCUITS IN OCD 275 tion in basal ganglia/limbic striatal and thalamocortical circuits as a pathogenic mechanism of obsessive-compulsive disorder. Acc- cording to the authors, “The basal ganglia/limbic striatum appear to serve an integrative role for the limbic system which allows for the production of a coherent and goal-oriented stream of behavioural and emotional output and suppression of unwanted or inappropriate responses”. Thus lesions that disturb the caudate nuclei (whether degenerative, clinical or surgical) often result in repetitive and compulsive behaviours, whereas interventions or pathology that lead to a relative increase in neural output from the caudate may, according to the authors, cause arrest, overin- hibition, or perseveration of behaviour and blunted emotional responsiveness. They describe the basal ganglia/limbic striatum as intimately connected with the orbitofrontal cortex and thalamus.

Orbitofrontal cortex appears to mediate the ability to alter behaviours and cognitive strategies with changing tasks and reinforcement contingencies, such that ablation of the region results in perseverative behaviours and difﬁculty executing newly learned instructions or associations (Modell et al., 1989).

The authors describe the “critical circuitry” implicated in the pathogenesis of OCD, as a neuronal loop from posterior portions of the orbitofrontal cortex, and running sequentially through to the ventral striatum (ventral or ventromedial caudate and accumbens nuclei), ventrommedial pallidum, and certain medial thalamic nuclei, and back to orbitofrontal cortex. Based on neuroanatomic, pharmacologic and pathological con- siderations, Modell et al. (1989) proposed that the primary pathogenic mechanism of OCD lies in a dysregulation of the basal gangla/limbic striatal circuits that modulate neural activity in and between the posterior portion of the orbitofrontal cortex and the associated medial thalamic nuclei. “The verbal nature of obsessions and the forced intellectualisation of compulsions suggest that left hemisphere structures may be particularly involved, although the strong affective components to the illness might implicate right-sided dysfunction as well”. More speciﬁcally, the au- 276 CHAPTER 9. OCD AND TOURETTE’S DISORDER thors proposed the OC symptoms occur when an aberrant positive feedback loop develops in the reciprocally excitatory fronto- thalamic neural interchange, which is inadequately integrated or inhibited by the ventromedial (limbic) portion of the striatum. “We postulate that the net result of ventral striatal activation is increased inhibitory pallidothalamic output: this would occur through the effects of the excitatory intrastriatal neurotransmitters and GABA-mediated striato-nigral inhibition” (Modell et al., 1989). Saxena et al. (1998) suggested that patients with OCD might have a low threshold for “system capture” by stimuli involving the above concerns, possibly due to excess “tone” in the direct, relative to the indirect orbitofrontal-subcortical pathway, allowing concerns about violence, hygiene, order and sex, and atten- dant compulsive behaviours to predominate. They concluded that while they did not know which brain center hosted the “primary dysfunction”, they believed that concerns triggering activity in the orbitofrontal cortex had “undue advantage” through pathways from the basal ganglia to the thalamus, in the competition among various stimuli for dominance in the modulation of thalamic outputs. The authors also propose that drugs which strongly inhibit serotonin (5-HT) re-uptake change the relative balance of tone through the indirect vs. direct orbitofrontal-subcortical pathways, thereby decreasing activity in the overall circuit.

9.1.1 The glutamate hypothesis

Rosenberg et al. (1998) hypothesised a role for glutamate in OCD. McMaster et al. (2008) have suggested that two astrocyte glutamate transporters inﬂuence the pathophysiology of OCD and its pharmacological regulation by regulating regional glutamate levels, in the conversion of glutamate to non-toxic glutamine. En- hanced expression of SLC1A2 and SLC1A3 is thought to increase the local residence time of glutamate and glutamine in a neuron- astrocyte cycle, resulting in higher glutamergic concentrations 9.2. OBSESSIVE COMPULSIVE SPECTRUM DISORDERS 277

(glutamate/glutamine [Glx]), as observed in the caudate in paediatric OCD. Underexpresion of SLC1A2 and SLC1A3 would by contrast result in diversion of incoming synaptic glutamate to neurons, causing faster neuron ﬁring, greater remote synaptic ex- port of glutamate, and/or consumption in the Krebs cycle to sustain the higher metabolic rate and lower levels of tissue GLx as observed in the anterior cingulate in paediatric OCD (McMaster et al., 2008).

9.2 Obsessive Compulsive Spectrum Disorders

Boyer and Lienard (2006) have outlined an ethological approach to ritualised behaviour. They describe rituals as being intuitively recognisable by their stereotypy, rigidity, repetition and apparent lack of rational motivation as being found in both cultural rituals, children’s complicated routines, and in the pathology of obsessive- compulsive disorders. He attributes rituals to an evolved “Precau- tion System” geared to inferred threats to fitness. This system is believed by them to be distinct from fear-systems geared to manifest danger, but includes a repertoire of clues for potential danger. The authors attribute OCD pathology to a failure of the system to supply appropriate feedback in relation to the appraisal of potential threats, resulting in doubts about the proper performance of precautions and repetition of action. They assert that anxiety focuses attention on low-level gestural units of behaviour, rather than the goal-related higher level units normally used in parsing the action-flow. Thus actions which are normally automatised continue to be submitted to cognitive control, swamping working memory. The resulting compulsions provide temporary relief from intrusions, but also result in long-term strengthening of rituals. Allen et al. (2003) have described a spectrum of compulsive versus impulsive disorders. Compulsive disorders such as Obses- sive Compulsive Disorder (OCD), hypochondriasis, and anorexia nervosa are characterised by avoidance of risk and an exaggerated sense of harm. Compulsive behaviours serve to avoid harm 278 CHAPTER 9. OCD AND TOURETTE’S DISORDER and reduce anxiety. Impulsive disorders are characterised by pathological gambling (PG) and sexual compulsivity (SC). Impul- sive risk takers who underestimate the likelihood or severity of possible harm seek pleasure, arousal, or gratification, and their repetitive behaviours involve risk-taking, rather than risk avoidance. Baxter et al. (1990) have postulated that OC spectrum disorders may involve cortico-striatal dysfunction with specific disorders having different areas of dysfunction within the system. They found that compulsive disorders showed increased sensitivity of specific serotonin receptor subsystems, while impulsive disorders were characterised by decreased frontal lobe activity and decreased presynaptic serotonergic function. Obsessions are defined by Allen et al. (2003) as recurrent thoughts, impulses or images, which are intrusive and ego- dystonic; they are thought to be related to basic fears or urges that are distressing to the individual such as contamination, aggression, sex, religion/scrupulosity, order/symmetry, hoarding or pathological doubt. Compulsions on the other hand are repetitive behaviours, including mental acts that the individual feels com- pelled to perform to reduce the anxiety created by the obsessions. As both obsessions and compulsions are intrusive, preoccupying, and distressing and interfere with attention and cognitive tasks, they are socially disabling, as well as time-consuming. OC spectrum disorders are thought by Allen et al. (2003) to involve the orbitofrontal cortex and the caudate nucleus. Studies have shown that successful treatment of OCD with either a specific serotonin re-uptake inhibitor (SSRI) or cognitive behaviour therapy (CBT) results in normalisation of orbito-frontal activity. Pathological gambling is an example of a disorder of impulse control characterised by recurrent gambling behaviour that is maladaptive, while sexual compulsivity is on the impulsive side of the compulsive-impulsive spectrum. While these conditions do not occur in childhood, they are of interest in indicating a compulsive-impulsive spectrum of behaviour. In fact, paraphilia- related disorders (PRD’s) are thought to begin in childhood, ado- 9.2. OBSESSIVE COMPULSIVE SPECTRUM DISORDERS 279 lescence or early adulthood (Bremer, 1992), with a preponderance of males. Yin and Knowlton (2006) have reviewed the role of the basal ganglia in habit formation. They differentiate intentional goal- directed actions from habitual responses. They define instrumental behaviour as an outcome or goal, which is contingent on a response. In action-outcome contingencies (A-O) performance varies according to how the contingency is manipulated. Instrumental contingency is viewed as the probability of reward given a particular action, relative to the probability or reward given no action. If these probabilities are the same, the contingency is said to be completely degraded and if degrading the contingency has no effect on work, the behaviour is habitual. For habit formation, behaviour is not guided by outcome expectancy, rather it is strengthened or weakened by reinforcement (S-R). According to the authors, the number of rewarded responses appears to be a crucial factor in determining the shift from the A-O system to the S-R system or habit formation. Both overtraining and the schedule of re-inforcement (interval schedule), tend to produce habitual responding, whereas ratio schedules do not. In ratio schedules a response results in a certain probability of reward, whereas in interval schedules a response is only rewarded after a certain time interval has elapsed. The authors believe that changes in neural activity in habit formation occur at too rapid a rate to be explained by the slow and gradual changes posited by traditional S-R reinforcement theory. They describe initially rapid A-O learning or association and subsequent S-R or habit acquisition. They conducted a series of training experiments to test the hypothesis that the dorsolateral striatum (DLS) is involved in S-R learning, whereas the dorsomedial striatum (DMS) is involved in A-O learning. They have proposed that involvement of the hippocampus may be required for A-O learning of spatial and/or temporal configurations, and that the distributed cortico-basal ganglia network is a functional associative network, involving the medial prefrontal cortex (PFC) and 280 CHAPTER 9. OCD AND TOURETTE’S DISORDER

DMS. With habit formation, control of behaviour shifts from a higher level of functional integration to a lower one, more specifically from the associative cortico-basal network to the sensorimotor cortico-basal ganglia network, which is more effector specific (possibly owing to its more lateralised cortico-striatal projections). However extensive damage to either network results in the other network assuming control over instrumental behaviour. Van Ameringen et al. (2001) have pointed out that the clearest distinction to be made among subtypes of OCD is the distinction between tic-related and non-tic related OCD. While Tourette’s Syndrome (TS) is characterised by motor tics and one or more vocal tics, beginning before the age of 18 years, the tics of TS can share some phenomenological similarities with the compulsions of OCD “ People with TS report that while they can delay their tics, they find them irresistible, experience relief when they perform them until they are just right”. Miguel et al. (1997) compared 3 groups of patients - OCD without tics, TS without OCD, and a group with both OCD and TS. They found that for OCD without comorbid tics or TS, repetitive behaviour was preceded by cognitive phenomena (obsessive thoughts or images) and anxiety, but largely not by sensory behaviour. By contrast, the TS group and the OCD with comorbid tics group reported more sensory phenomena, and fewer cognitions, than did the OCD group. It has also been found that OCD responds well to antidepressants that target the serotonergic system, while neuroleptics are ineffective. On the other hand, TS responds to neuroleptics but not to serotonergic drugs. Also OCD with comorbid tics is less responsive to serotonin reuptake inhibiting drugs (SRI’s) alone, but responds better to a combination of SRI’s and neuroleptic medication. These findings suggest the phenomenology of impulses and tics may be related to dopaminergic systems, while obsessions and compulsions may be related to serotonergic systems. As described above, Van Ameringen et al. (2001) outline four well-defined dopaminergic pathways in the brain. The nigrostriatal dopamine pathway is part of the extrapyramidal motor sys- 9.2. OBSESSIVE COMPULSIVE SPECTRUM DISORDERS 281 tem, and controls motor movements. It projects from dopaminergic cell bodies in the substantia nigra of the brainstem, via axons terminating in the basal ganglia (composed of caudate and putamen parts of the striatum), globus pallidus, substantia nigra and subthalamic nucleus. The cortical areas projecting to the striatum are roughly divided into ‘motor’ areas (somatosensory, motor and premotor cortices) and ‘limbic associative’ areas including amygdala, hippocampus, orbital, entorhinal, temporal, prefrontal, parietal, cingulate and association cortex. According to Van Ameringen et al. (2001), TS is most often associated with dysfunction of the putamen, while OCD is most often associated with dysfunction of the caudate. “ [...] it seems likely that involvement of the ventromedial caudate nucleus, receiving projections from the anterior orbitofrontal cortex, leads to affectively tinged cognitive symptoms of OCD, whereas involvement of the putamen, receiving projections from sensori-motor cortices leads to the somatosensory premonitory symptoms and tics of TS” (Van Amerin- gen et al., 2001). The mesolimbic dopamine system projects from dopaminergic cell bodies in the ventral tegmental area (VTA) of the brainstem to axon terminals in the limbic area of the brain, such as the nucleus accumbens. Stimulation of the VTA leads to immediate and sustained increase in extracellular dopamine in the nucleus accumbens, thought to correlate with reward. This may be the final common pathway for positive reinforcement of survival behaviours such as eating, drinking and copulation, as well as addictive drugs. Rauch and Savage (1997) have proposed a model of the functional anatomy and organisation of cortico-striatal pathways, in which they distinguish sensorimotor, oculomotor, dorsal cognitive, ventral cognitive, and affective motivational cortico-striatal circuits. According to Van Ameringen et al. (2001), the dorsal cognitive circuit projects primarily from the dorsal, anterior and lateral regions of the prefrontal cortex, via the dorso-lateral portion of the head of the caudate nucleus to the ventral anterior and 282 CHAPTER 9. OCD AND TOURETTE’S DISORDER medial dorsal nuclei of the thalamus. This circuit is thought to have a role in working memory, and ability to shift cognitive sets. The “ventral cognitive circuit” projects from the anterior and lateral orbitofrontal cortex, via the ventro-medial portion of the caudate nucleus as well as to the ventral anterior and medial dorsal nuclei of the thalamus, and is thought important for response inhibition, in relation to social and emotional functioning. The “affective-motivational circuit” projects from paralimbic cortical territories (i.e. the posterior orbitofrontal cortex and the anterior cingulate, via the nucleus accumbens (i.e. ventral striatum) to the medial dorsal nucleus within the thalamus, and is heavily influenced by limbic structures such as the amygdala, and has a role in emotional or reward-based information processing (Rauch and Savage, 1997). Van Ameringen et al. (2001) postulate that sensori-motor, oculomotor, and affective motivational circuits are more vulnerable to dopaminergic dysfunction, while dorsal and ventral cognitive circuits may be more dependant on serotonergic dysfunction, and may respond to specific serotonin re-uptake inhibitors. Thus symptoms of obsessions and compulsions are probably associated with functional changes in the caudate and orbitofrontal cortex, with dysregulation of serotonergic function. In contrast, Tourette’s Syndrome may be associated with functional changes in the putamen and dysregulation of dopaminergic function, while impulsive obsessive compulsive spectrum disorders, such as trichotillomania (hair picking) and skin picking are thought more related to TS than OCD, and may respond to dopaminergic agents.

9.2.1 Developmental perspective

According to Leckman et al. (2009) typically developing children engage in a significant amount of ritualistic, repetitive and compulsive-like activity, reaching a peak at about 24 months of age. The content of these behaviours is described as closely re- 9.2. OBSESSIVE COMPULSIVE SPECTRUM DISORDERS 283 sembling obsessive compulsive (OC) symptom dimensions, including arranging objects, repeating actions, and concern with dirt or cleanliness, as well as hoarding behaviours. While direct evidence linking these behaviours to OCD is lacking, subclinical obsessions and compulsions are encountered in childre at a rate as high as 8% (Fullana et al., 2009). Leckman et al. (2009) also describe a tic- related subtype in as many as 10-40% of paediatric OCD cases. Children with tic-related OCD are described as having higher rates of disruptive behaviour disorders, including ADHD, oppositional defiant disorder, and trichotillomania. Using data collected by the Toutette Syndrome Association International Consortium for Genetics Affected Sibling Peer Study, Leckman et al. (2003) selected all available affected TS pairs and their parents for whom OC symptom dimensions (factor scores) could be generated. They report that 50% of the siblings with TS were found to have comorbid tic-related OCD and greater tha 30% of mothers and 10% of fathers also had a diagnosis of OCD. Based on complex segregation analyses, significant evidence for genetic transmission was obtained for all factors (Leckman et al., 2009). Leckman and colleagues believe that in addition to the existence of subtypes of OCD, a strong case can be made to support the use of a dimensional approach to OC symptoms, allowing integration of scien- tific findings from genetics and neurobiology.

9.2.2 Therapeutic implications

Schwartz (1998) has argued that an understanding of basal ganglia physiology, including the micro-circuitry of the striatum helps in the understanding of OCD and its treatment. In particular, the striatum contains ‘modules’ comprised of specialised zones called striasomes, dispersed with a larger component called the matrix. According to Schwartz (1998) the striasomes primarily receive inputs from the orbital cortex and anterior cingulate gyrus, while the matrix receives projections from the rest of the prefrontal cortex. Schwartz suggests that specialised cells called 284 CHAPTER 9. OCD AND TOURETTE’S DISORDER tonically active neurons (TANS) may serve to integrate or gate information ﬂow and generate new striatal activity patterns in response to the integration of behaviourally signiﬁc information. He describes a ‘neurogenic’ approach to cognitive behavioural treatment of OCD - in which patients are advised to construct behavioural hierarchies, to structure self-exposure to alternative or avoided behaviours, and re-focus away from ritualised habitual OCD behaviours. Exposure and response prevention are thus hypothesised to activate alternate cortico-striatal circuits, allowing habituation and extinction of anxiety, and escape from ritualised behaviour patterns. Kopell and Greenberg (2008) have discussed the concept of cortico-striatal-pallido-thalamo cortical loops, in terms of possible applications of deep brain stimulation, particularly for OCD and major depression. While the use of this technology is not established in children, it promises a possible non-invasive technique with potential for OCD, and possibly Tourette’s disorder, when better established in terms of safety and ethical issues.

9.2.3 Comment

For present purposes, the basal ganglia, limbic striatal, thalamocortical circuits described by Saxena and Rauch (2000) with a direct pathway from the cortex to striatum to internal segment of the globus pallidus, to thalamus and back to the cortex, and an indirect inhibitory pathway, via the external segment of the globus pallidus, which rejoins the common pathway to the thalamus and cortex, provides a fundamental circuit structure, in which disequilibrium gives rise to repetitive behaviour. Imbalance between inhibitory and disinhibitory inﬂuences at thalamic level will result in a failure of appropriate decay of activity at cortical levels, experienced as repetitive thoughts and concerns. Thus OCD can be characterised as a primary circuit disorder, rather than a local modular disorder, in which anxiety effects feedback systems, so that a ’habit circuit’ is maintained. 9.2. OBSESSIVE COMPULSIVE SPECTRUM DISORDERS 285

However, it can be argued that the imbalance between inhibitory and dis-inhibitory activity at thalamic levels prevents workspace integration at cortical levels, leading to repetitive emotional and cognitive concerns. From a develomental perspective, OCD shows a lack of integration of affective and cognitive modules. The understanding of circuit phenomenology is useful , both for pharmacological and behavioural interventions. For the latter, cognitive and response prevention approaches help to alter the dysregulated balance in malfunctioning circuits. Regulation of glutamate/glutamine concentration in the cingulate vs caudate may be an important physiological mechanism in OCD. Faster neuronal ﬁring in the anterior cingulate may be involved in response suppression deﬁcits described in OCD, but further fMRI studies are needed to elucidate this mechanism. The Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) describes Tourette’s Disorder as:

A Both multiple motor and one or more vocal tics have been present at some time during the illness, although not necessarily concurrently. (A tic is a sudden, rapid, recurrent, nonrhythmic, stereotyped motor movement or vocalisation.) B. The tics occur many times a day (usually in bouts) nearly every day or intermittently throughout a period of more than 1 year, and during this period there was never a tic-free period of more than 3 consecutive months. C. The disturbance causes marked distress or signiﬁcant impairment in social, occupational, or other important areas of functioning. D. The onset is before age 18 years. E. The disturbance is not due to the direct physiological effects of a substance (e.g., Huntingdon’s disease or post-viral encephalitis) (e.g. stimulants or a general medical condition).

Leckman (2002) describes the cardinal features of Tourette’s Syndrome as motor and phonic tics that wax and wane in severity. He describes motor tics as usually beginning between the ages of 3 and 8 years, with transient periods of intense eye blinking or some other facial tic. Phonic tics, such as repetitive bouts of sniff- ing or throat clearing can begin as early as 3 years, but typically 286 CHAPTER 9. OCD AND TOURETTE’S DISORDER they follow the onset of motor tics by several years. According to Leckman, the severity of uncomplicated cases, peaks early in the second decade of life, but extreme cases can arise in adulthood, and involve forceful bouts of self-injurious behaviour, such as hitting or biting, and socially unacceptable coprolalic utterances, e.g., shouting obscenities, racial slurs and gestures. The syndrome can wax and wane over time, including days, weeks or months. According to Leckman (2002), episodes are usually characterised by stable intra-tic intervals, and self-similarity of patterns across different times. A further feature is the associated sensory symptoms, including premonitory urges, as well as a generalised inner tension, relieved only by the performance of a tic. Tics, especially loud phonic tics can be very socially debilitating, and require psychological and/or pharmacological treatment. Co- morbidity of Tourette’s Disorder with hyperkinesis and obsessive- compulsive disorder has often been described. Towbin et al. (1999) point out that leaving aside the chance that a diagnostic error has occurred, there are three possible relationships that may exist between two disorders when they are observed in one person. The first is that they are entirely separate entities, running independant courses (true comorbidity). The second is when one disorder ‘sets the stage’ for another such as an opportunistic fungal infection, but both are aetiologically independent. The third case is when two disorders are aetiologically related, but are differentially expressed (such as peripheral neuropathy and retinal haemorrhages in diabetes). Leckman and Cohen (1999) describe human studies supporting the importance of basal ganglia in Tourette’s Syndrome pathophysiology (Haber and Wolfer, 1992). As described above, Leck- man and Cohen (1999) point out that complex tics can be extremely difficult to distinguish from compulsions, and that 30 to 60 percent of Tourettes’ patients are also diagnosed as having Obsessive Compulsive Disorder (OCD), with some studies suggesting genetic relationships (Pauls et al., 1991). Similarly Leck- 9.2. OBSESSIVE COMPULSIVE SPECTRUM DISORDERS 287 man and Cohen (1999) claim that approximately 50% of all clinically identified Tourette’s syndrome patients also have ADHD. While ascertainment bias may account for some of these relationships, Leckman attributes the underlying pathology of these conditions to parallel cortico-striato-thalamo cortical (CSTC) circuit pathology (Alexander and Crutcher, 1990; Goldman-Rakic et al., 1990; Parent and Hazrati, 1995). According to Leckman and Cohen (1999), current consensus held that CSTC circuitry had at least three components, those initiating from and projecting back to sensory-motor cortex (SM); orbitofrontal cortex; or association cortices. Further components were also thought to include the limbic system. Leckman and Cohen (1999) describe cortical projections to the caudate and putamen (striatum) appear oriented in somatotopically organised domains. Information leaves the basal ganglia principally through the internal segment of the globus pallidus (GPi) and its brainstem counterpart (substantia nigra, SNr) before ascending to the thalamus and cortex. Striatal projections to the external segment of the globus pallidus (GPe) then project to the reticular thalamic nucleus and internal segment (GPi), exerting inhibitory influences modulated by GABA. Projections from striatum to pallidum are also GABAergic, and increased striatal activity (such as associated with movement) will phasically silence the tonically active GPi and SNr neurons, reducing GABAergic transmission to the thalamus, thereby activating thalamic target neurons.

However, the CTSC are also implicated in obsessive compulsive disorder (as described above) (Baxter, 1992). Finally, abnormalities in striatal function and related association cortex circuitry are thought to contribute to ADHD symptomatology. Castellanos et al. (1996); Leckman and Cohen (1999) also suggest that limbic system circuitry, including the temporal lobe may be involved in the sexual and aggressive content of many complex motor and vocal tics, while the sexual and aggressive content of many obsessions and compulsions suggests involvement of amygdala and related circuitry. Leckman and Cohen (1999) describe 288 CHAPTER 9. OCD AND TOURETTE’S DISORDER the limbic system as consisting of the amygdala and hippocampus, the temporal lobe, the cingulum, the caudate nucleus and other basal ganglia structures in the subcortex, the hypothalamus and periaqueductal grey matter in the brain stem and connections with associative frontal cortex. The involvement of the cingulate cortex in both Tourette’s and Obsessive Compulsive Dis- order is described as inhibitory. Decreases in severity of OCD symptoms have been correlated with decreases in right anterior cingulate metabolic activity during successful medication therapy (Baxter, 1992). Towbin et al. (1999) examined the relationships between ADHD, Obsessive Compulsive Disorder and Tourette’s syndrome. In the case of ADHD and Tourette’s syndrome, the authors believe that aetiology may be independent, but in some cases symptoms of ADHD may precede Tourette’s Disorder. On the other hand, Obsessive-Compulsive disorder may in some cases be the product of the same aetiology (Pauls et al., 1986). The authors point out that a categorical classification would view the three disorders as separate, whereas a concept of overlapping mechanisms, evolving from impairment at related anatomical sites or systems (such as the cortico-striato-thalamo-cortical system) provides a better explanation of comorbidity. Leckman (2002) has described cortical-subcortical circuits, which project to the striatum and back to specific regions of the cortex. Leckman describes two structurally similar, but neuro- chemically distinct compartments in the striatum - striasomes and matrix, which differ by their cortical inputs. The striasomal medium spiny projection neurons mainly receive convergent limbic and prelimbic inputs, while neurons in the matrix receive convergent input from ipsilateral primary motor and sensory motor cortices, and contralateral motor cortices. According to Leckman (2002), the response of particlar medium spiny projection neurons in the striatum is partly dependant on perceptual cues that are judged salient, so rewarding and aversive stimuli can both serve as cues. “Tonically active neurons are very sensitive to salient 9.2. OBSESSIVE COMPULSIVE SPECTRUM DISORDERS 289 perceptual cues, because they signal the networks.” [...] “The fast spiking interneurons of the striatum are electronically coupled via gap junctions that connect adjacent dendrites. Once activated, these fast spiking neurons can inhibit many striatal projection neurons synchronously.” (Leckman, 2002). Tics are thought to reflect a lack of balance (striasome vs. matrix). According to Leckman (2002), alterations in the structure of the basal ganglia, including loss of right-left symmetry, and cortico-thalamic inputs may be important in tic and hyperki- netic disorders. Recent neurosurgical procedures have, according to Leckman (2002), reinforced the functional importance of thalamic regions, that are part of the above cortical-subcortical loops. Additionally, ascending dopaminergic pathways are thought to have a role in the consolidation and performance of tics. Leck- man (2002) points out that early results of in-vivo neuroimaging studies have shown that voluntary tic suppression involves activation of regions of the prefrontal cortex and caudate nucleus and bilateral deactivation of the putamen and globus pallidus. According to Leckman, the findings accord with the well-known finding that chemical or electrical stimulation of inputs into the putamen provokes motor and phonic responses resembling tics. Additionally, Leckman describes a case study, which showed that high frequency stimulation of the median and rostral intralaminar thalamic nuclei produced a reduction of tics. This effect was thought to be possibly caused by the effect of the midline thalamic nuclei on tonically active neurons, or on a broadly distributed cortical system, suggesting the possibility of circuit-based therapies.

9.2.4 Discussion

While the above syndrome descriptions appear to reﬂect a modular, motor syndrome, the association of premonitory sensory urges, the presence of motor and phonic tics, as well as coprolalia, and associations with hyperkinesis and obsessive-compulsive disorder suggest greater complexity. The division of tics into those 290 CHAPTER 9. OCD AND TOURETTE’S DISORDER with and without cognitive phenomena suggest that TD may be characterised in this way. Haber (2003), in relation to these phenomena, describes the mediation of cognitive and motor planning, and drive behaviours by the parallel cortical-subcortical neural circuits originally described by Alexander et al. (1986). However, the Alexander model fails to address how information ﬂows between circuits. Haber points out that recent anatomical evidence from primates demonstrates that neuro-networks within the basal ganglia are in a position to move information across functional circuits. She describes two networks: the striato-nigral, and the thalamo-cortical-thalamic network. Impor- tantly, within each of these nets of connected structures, Haber describes both reciprocal connections, linking up regions associated with similar functions, and non-reciprocal connections linking up regions that are associated with different cortico-basal ganglia connections. Thus each component of information (from limbic to motor outcome sends feedback and feed-forward connections), allowing the transfer of information, which may help explain comorbid symptomatology in Tourette’s Disorder. Thus Tourette’s Disorder, Obsessive Compulsive Disorder, and ADHD could be conceptualised as manifesting respectively more or less cognition, as the involvement of orbitofrontal cortex in the case of OCD increases, with more or less affective reactivity as the involvement of limbic and temporal and cingulate lobes increases. This concept also applies to ‘dorsal’ (cognitive) and ‘ventral’ (affective) systems and symptomatology. Chapter 10

Posttraumatic Stress Disorder

The American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) describes Posttraumatic Stress Disorder as follows:

A. The person has been exposed to a traumatic event in which both of the following were present: (1) the person experienced, witnessed or was confronted with an event or events that involved actual or threatened death or serious injury, or a threat to the physical integrity of self or others (2) the person’s response involved intense fear, helplessness, or horror. Note: In children, this may be expressed instead by disorganised or agitated behaviour B. The traumatic event is persistently reexperienced in one (or more) of the following ways: (1) recurrent and intrusive distressing recollections of the event, including images, thoughts or perceptions. Note: In young children, repetitive play may occur in which themes or aspects of the trauma are expressed. (2) recurrent distressing dreams of the event. Note: In children, there may be frightening dreams without recognisable content. (3) acting or feeling as if the traumatic event were recurring (includes a sense of reliving the experience, illusions, hallucinations, and dissociative ﬂashback episodes, including those that occur on awakening or when intoxicated). Note: In young children, trauma-speciﬁc reenactment may occur. (4) intense psychological distress at exposure to internal or external cues that symbolise or resemble an aspect of the traumatic event C. Persistent avoidance of stimuli associated with the trauma and numbing of general responsiveness (not present before the trauma)

291 292 CHAPTER 10. POSTTRAUMATIC STRESS DISORDER

as indicated by three or more of the following: (1) efforts to avoid thoughts, feelings, or conversations associated with the trauma (2) efforts to avoid activities, places, or people that arouse recollections of the trauma (3) inability to recall an important aspect of the trauma (4) markedly diminished interest or participation significant activities (5) feeling of detachment or estrangement from others (6) restricted range of affect (e.g., unable to have loving feelings) (7) sense of foreshortened future (e.g., does not expect to have a career, marriage, children, or a normal life span) D. Persistent symptoms of increased arousal (not present before the trauma), as indicated by two (or more) of the following: (1) difficulty falling or staying asleep (2) irritability or outbursts of anger (3) difficulty concentrating (4) hypervigilance (5) exaggerated startle response E. Duration of the disturbance (symptoms in Criteria B, C and D) is more than 1 month. F. The disturbance causes clinically significant distress or impairment in social, occupational, or other important areas of functioning. Specify if: Acute: if duration of symptoms is less than 3 months Chronic: if duration of symptoms is 3 months or more Specify if: With Delayed Onset: if onset of symptoms is at least 6 months after the stressor

10.1 Neural circuits in PTSD

Childhood amnesia was thought by Freud (1895) to be due to overwhelming traumatic experiences, which were blocked or re- pressed. However, developmental psychologists suggest that immaturity in domains such as sense of self and time, verbal ability and narrative capacity impair recall for the period before 2-3 years (Rovee-Collier, 1993). The incomplete development of the hippocampal/medial temporal lobe and prefrontal regions during the first years of life may be possible causal factors in childhood amnesia (Siegel, 2001). Hippocampal development is required for conscious memory. This means that determinist childhood trauma theories rely on memories from the implicit stage of 10.1. NEURAL CIRCUITS IN PTSD 293 memory development. However, it is interesting that while determinist theory is based on events occurring early in infancy, there has been little attention paid to neurophysiological processes during this stage of development. Thus, the stage of childhood amnesia has remained the subject of much theorising and little understanding. According to Krystal et al. (1995) the most predominant feature of Postraumatic Stress Disorder (PTSD) is that memories of traumatic experiences remain indelible for decades and are easily reawakened by all sorts of stimuli and stressors. They describe a number of neural mechanisms of learning and memory, of relevance to the re-experiencing symptoms of PTSD, and also discuss failure of extinction, behavioural sensitisation, and dissociation, as neural mechanisms involved in PTSD. The continued ability of conditioned stimuli to elicit traumatic memories and flashbacks may represent a failure of extinction. Extinction may not erase the original aversive memory, but may involve the learning of a new memory, which masks or inhibits the original one. Behavioural sensitisation refers to the increase in behavioural or physiological responsiveness that occurs following repeated exposure to a stimulus. It may be context-dependent or conditioned. Single or repeated exposure to a stressor has been shown to increase dopamine function in the forebrain (Kalivas and Duffy, 1989) as well as norepinephrine release in the hippocampus and hypothalamus. Krystal et al. (1995) believe that traumatisation appears to alter patterns of memory encoding, leading to the formation of memories with reduced contextual information. Extreme environmental stimulation has been assumed to affect the ability to process and synthesise information by overloading cognitive capacity and attention resources. According to Krystal, stress and dissociative states may also alter the predominant mode of memory encoding. Memories may be encoded using detail-focused strategies in which components of memory are encoded and rehearsed separately, or holistic strategies where memories are rehearsed 294 CHAPTER 10. POSTTRAUMATIC STRESS DISORDER as composite images. “It is possible that during traumatisation, there is a shift away from verbal encoding toward encoding in the emotional, pictorial, auditory and other sensory-based memory systems. This shift would help to explain the common experience that verbal descriptions fail to convey an accurate representation of the horror and distress of traumatic experience” (Krystal et al., 1995). Schulkin et al. (1994) have provided a framework that integrates animal studies on the regulation of chronic anxiety, fear and helplessness with human fear, anxiety and post-traumatic stress (PTSD), all of which may result from anticipation of negative events (allostasis). They discuss the role of central nucleus of amygdala (CEA) in expectations of adversity giving rise to elevated glucocorticoid secretion. They argue that increases in corticotropin-releasing hormone (CRH) caused by stress or glucocorticoids in the amygdala may affect allostatic load. On the other hand, CRH in the paraventricular nucleus of the hypothalamus is to some extent increased by glucocorticoids. They conclude that the ‘hormones of stress’ CRH and glucocorticoids arouse a sense of adversity, possibly via the amygdyla, and that expectations of adversity may result in pathological arousal and allostatic load. According to Bremner (2006) brain imaging studies have shown alterations in a circuit, including medial prefrontal cortex, anterior cingulate, hippocampus and amygdala, in PTSD. A variety of methods have been used to trigger PTSD symptoms. For example, exposure to traumatic reminders in the form of slides or sounds, giving rise to an increase in PTSD symptoms, resulted in decreased blood flow and/or failure of activation in the medial PFC/anterior cingulate, while other findings showed decreased function in the hippocampus, visual association cortex, parietal cortex and inferior frontal cortex, but increased function in amygdala and posterior cingulate. According to Bremner (2006) similar studies in children are more variable, possibly related to chronicity of illness, or interaction between trauma and development. Bremner (2006) describes findings of smaller hippocampal vol- 10.1. NEURAL CIRCUITS IN PTSD 295 ume in a number of trauma-related psychiatric disorders, including depression with early abuse, borderline personality disorder with early abuse and Dissociative Identity Disorder with early abuse. Bremner suggests that traumatic stress occurring at different stages of the life cycle interacts with the developing brain. “In the first 5 years of life an overall expansion of brain volume is related to development of both gray matter and white matter structures. From 7-17 years of age, there is a progressive increase in white matter (myelination) and decrease in gray matter (pruning). Basal ganglia decrease in size, while corpus callosum, hippocampus and amygdala appear to increase in size. During middle life there are further gradual decreases in caudate, dien- cephalon and gray matter” (Bremner, 2006). The regulation of stress is carried out by a hormonal neural system consisting of the hypothalamic-pituitary adrenal (HPA) axis. Corticotropin-releasing factor (CRF) is released from the hypothalamus, stimulating adreno-corticotropic hormone (ACTH) release from pituitary, resulting in release of cortisone, which feeds back to the hypothalamus, and also to the hippocampus. Noradrenaline is also released from the locus coeruleus during stress. Animal studies have shown that the hippocampus, which is involved in verbal declarative memory is very sensitive to the effects of stress. In addition the amygdala is involved in memory for the emotional valence of events, and is involved in the acquisition of fear responses. In addition, the medial prefrontal cortex (including anterior cingulate, subcallosal gyrus and orbitofrontal cortex modulate emotional responsiveness through inhibition of amygdala responsiveness. Thus according to Brem- ner (2006) “studies in patients with PTSD show alterations in brain areas implicated in animal studies, including amygdala, hippocampus and prefrontal cortex, as well as in neurochemical stress response systems, including cortisol and norepinephrine”. Mason et al. (2001) examined intra-psychic correlates of individual differences in cortisol levels in male Vietnam combat veterans with PTSD. They suggest that cortisol levels reflect the 296 CHAPTER 10. POSTTRAUMATIC STRESS DISORDER ongoing balance between the undifferentiated emotional arousal state of engagement (associated with higher cortisol levels) and opposing anti-arousal disengagement, associated with lower cortisol levels. Thus, the low cortisol levels often seen in patients with PTSD were thought to reflect a dominating effect of disengagement coping strategies, representing compensatory secondary adaptations during chronic PTSD to counteract primary arousal symptoms. This phenomenon is consistent with the “massive emotional numbing” widely recognised as a clinical sign of PTSD. McFarlane et al. (2002) have pointed out that the emergence of PTSD with intrusive and distressing recollections is a progressive rather than immediate process, in which a traumatic event is a necessary but not sufficient explanation. The authors have developed a model, which incorporates neural network models, as well as modeling the manner in which complex associative networks are trained and modified (similar models can also be applied in OCD, schizophrenia and dissociative disorders and possibly autism). The disruption of cortical/subcortical connections is “critical” to the development of the model. The authors suggest that an individual’s cognitive and affective adaptability becomes impaired by the dominance of internal memories. This occurs as a consequence of “primary” and “top-down activation”, which in turn interferes with “the development of semantic networks to explain and integrate the trauma”. The neural network model of “parallel distributing processing attractor networks” implies that “memory and cognition reflect the cooperative interaction of multiple cortical regions rather than hierarchical control” (McFarlane et al., 2002). The authors cite the work of Goldman-Rakic (1988) on functional reciprocity of frontal and posterior association cortex, with complimentary connection to up to fifteen other cortical and subcortical areas. McFarlane et al. (2002) describe three processes characteristic of neural networks of interest in the study of PTSD. These are (1) interactive learning (2) pruning of neural intercon- 10.1. NEURAL CIRCUITS IN PTSD 297 nections, in which some connections die as a result of competition, and (3) top-down activation in which dominant or inflexible networks prime or bias activation toward stimuli relevant to certain memories. Thus a consequence of iterative learning is the facilitation of some pathways at the expense of destroying or inhibiting others by preferential adjustment of the synaptic strength between neurons. The authors suggest that the repeated replaying and contemplation of the traumatic memories modify associated fear networks. “For some, replaying the memory involves an aug- mentation of the related affect and distress. Thus the ‘trauma’ is related to both the initial experience, and the subsequent replaying of the inactive visual and linguistic representation of the memory, where a host of affective associations embue ongoing meaning”. McFarlane et al. (2002) suggest that the loss of too many dendrites by pruning can result in the development of autonomous foci, whereby they can become activated in the absence of environmental primes, as well as leading to the inappropriate fusion of memory networks. This may explain several phenomena including the emergence of flashbacks and the development of dissociative reactions in response to traumatic triggers.

Pruning can also explain how a particular traumatic event could serve as an activator of previously traumatic memories and create a domino effect in the processing of similar traumata. [...] Fi- nally, pruning has a secondary effect on general working memory structures and systems. For example, parasitic foci may have the ability to assert their inﬂuence over other concurrent inputs with the secondary effect of inducing systemic inactivity (with loss of ﬂexibility) (McFarlane et al., 2002).

The authors also suggest that “top-down processing” can be used to explain the increasing generalization of traumatic triggers, disturbed concentration and general emotional numbing in which an individual’s attention stance becomes biased to particular classes or categories of stimuli, so that they are more likely to be attended to than others. Thus “the specific content of the memories in PTSD may be less important to its phenomenology than 298 CHAPTER 10. POSTTRAUMATIC STRESS DISORDER the effect of the iterative processing of these memories and their effects on the underlying structures of memory and perception” (McFarlane et al., 2002). McFarlane et al. (2002) also discussed the role of the noradrenaline (NA) system in PTSD. They point out that the central noradrenaline system is one of the few systems in which the afferents have widely dispersed projections. Furthermore the function of the NA system has been conceptualised as manifesting tonic/phasic activity. According to McFarlane et al. (2002) the phasic mode of activity promotes focused selective attention, whereas the tonic mode leads to attentional scanning, with associated effects on behavioural flexibility. Thus the interaction between the two states effects signal to noise ratio. (This tonic/phasic model is similar to tonic/phasic effects in the dopamine system described above by Floresco et al. (2003)). As described by McFarlane et al. (2002), the wide receptive domains of the NA system, originating from the locus ceruleus (LC) domains of the NA system carry non-specific information to many areas, thus integrating and modulating other transmitters and brain regions. The NA projections in the PFC are mediated by alpha 1 and alpha 2 receptors. According to Friedman et al. (1999), the alpha 2 receptors improve prefrontal functions by inhibiting irrelevant and distracting sensory processing through the modification of prefrontal connections with the sensory association cortices; which in turn improves the cognitive processing of relevant information. On the other hand, as described by Arnsten et al. (1999), the alpha 1 receptors in the PFC inhibit cognitive performance in the PFC, possibly to allow rapid use of habitual subcortical routines in the face of danger. (These contrasting NA effects are similar to the contrasting effects of D1 and D2 receptors in the PFC and corpus striatum). In the present PTSD context, a hyper-catecholaminergic state in PTSD “releases the inhibition typically maintained by the PFC, thereby releasing traumatic memories”. As in the DA system, Aston-Jones et al. (1999) suggest an inverted-U relationship (Grossberg, 1999), between tonic 10.1. NEURAL CIRCUITS IN PTSD 299

LC activity and performance on discrimination tasks, where moderate tonic activity in association with phasic responses to targets leads to optimal performance. Thus in PTSD, the NA system is overactivated, resulting in suboptimal performance. This overac- tivation is thought to underlie symptoms such as flashbacks and intrusive thoughts (Southwick et al., 1993). According to McFarlane et al. (2002) the neural network model helps explain how the same neurotransmitter can produce both symptoms related to overactivity and to underactivity. For example, the loss of the normal flexible modulation of parallel distributed processing by cortical networks may result in overactivity in phasic states and under-responsiveness to non-specific sensory stimuli, creating an environment of sensory understimu- lation and sensory deprivation. McFarlane also describes a “kin- dling” hypothesis in relation to hippocampal functioning in PTSD, in which the storage of memories by the hippocampus is disrupted by a biological cascade that modifies the individual’s sensitivity to a range of stressors, with alterations in the hypothalamic- pituitary-adrenal axis functioning, with evidence of a decreased hippocampal volume after child sexual abuse (Bremner et al., 1997). Phelps (2004) describes the amygdala and hippocampal com- plexes as two independent memory systems, which interact in subtle but important ways. The amygdala is described as more or less specialised for the processing of emotion.

The hallmark of this memory system is that it is crucial for the acquisition and expression of fear conditioning, in which a neutral stimulus acquires aversive properties by virtue of being paired with an aversive event. [...] On the other hand the hippocampal system is thought to govern the recollection of events at will (Phelps, 2004). According to Phelps, most research has focused on how the amygdala can influence hippocampal-dependent episodic memory for emotional stimuli. “It has been shown that there is an enhanced response in the amygdala to emotional stimuli (e.g. fearful faces) and this response is correlated with a similar response 300 CHAPTER 10. POSTTRAUMATIC STRESS DISORDER in the visual cortex” (Phelps, 2004). Phelps (2004) suggests that the amygdala can alter the encoding of hippocampal-dependent episodic memory, such that emotional events receive priority. The second stage is thought to be hippocampal storage or slow consolidation. Phelps suggests that one reason for the slow consolidation process is to allow an emotional reaction to an event to influence storage by arousal, and the release of stress hormones. Stress hormones are thought to activate adrenergic receptors in the basolateral amygdala, modulating their effect on hippocampal consolidation. While amygdala influences on hippocampus have been well researched, Phelps (2004) also discusses evidence of episodic memory influences on the amygdala. Thus for humans, it is possible to learn about the emotional significance of stimuli in the environment through verbal communication. This type of learning is thought to require the hippocampal complex for acquisition, and possibly for retrieval. Phelps (2004) illustrates this process in an fMRI study in which subjects were told they would receive one or more mild shocks when a blue square was presented. An arousal response to the blue square was demonstrated in the absence of the blue square. It was suggested that the relationship between the amygdala and hippocampus might be bi-directional during the encoding of emotional events (Richardson et al., 2004). Phelps (2004) points out that there is currently a relatively poor understanding of the precise mechanisms of storage for hippocampal- dependent memories in general, although it is likely that working memory plays an important part when an episodic memory is retrieved. Olsson and Phelps (2007) have pointed out that insight into the workings of social fear learning and neural mechanisms allows integration of biological principles of learning with cultural evolution. They point out that research in the neurobiology of fear conditioning has focused on the amygdala in the medial temporal lobe as a key structure in the brain’s circuitry. This is thought to involve sensory input and motor output systems, as well as re- 10.1. NEURAL CIRCUITS IN PTSD 301 gions that contribute to explicit and conscious aspects of learning and expression of fear. The basolateral complex in the primate amygdala has strong connections to visual cortex, in particular to inferotemporal cortex region that responds to face identity and to facial expression. The authors also suggest the hippocampus, another medial temporal lobe structure adjacent to the amygdala is critical for coding contextual information about the fear learning situation, such as relationships between different features and the timing of events. Thus while the amygdala is responsible for forming associations between somatosensory states and representation of individual stimuli, the hippocampus is important for encoding relations between the various cues that comprise the learning context.

Olsson and Phelps (2007) suggest two interacting mechanisms in social learning of fear. While the amygdala is believed to process and store representations of conditioned-unconditioned contingencies, via connections with the medial prefrontal cortex (PFC), observational fear learning is believed to be supported by shared affective representations in the anterior insula, as well as more explicit hippocampal representations about context and relevant social information, such as social status and familiarity. They also suggest that for fears that are acquired through verbal communication, the CS-US is represented in a distributed cortical network, which is left-lateralised, reflecting the verbal nature of the US. The authors also point out that animal studies of fear conditioning have not emphasised the striatum, beyond its role in avoidance learning and active coping (LeDoux and Gorman, 2001). While unidirectional projections have been emphasised, most of the regions involved have bidirectional connections with the amygdala. Thus while fear learning is first expressed after social and nonsocial means of acquisition, further learning may result, which could change the nature of the representation further. “For instance in instructed fear, co-occurrence of the CS and autonomic arousal may cause the CS to act as a secondary re-inforcer, which projects its emotional salience to the lateral nucleus to fa- 302 CHAPTER 10. POSTTRAUMATIC STRESS DISORDER cilitate an amygdala dependent representation of the CS-threat association that was not present after initial verbal instruction”. This model of changing representations of verbally commu- nicated fears, shows how they may, according to Olsson and Phelps (2007) change over time with “recirculation” of emotionally salient fears, and may provide a “neural mechanism supporting socially transmitted fears and socioemotional impairments in many psychological disorders, such as phobias and anxiety disorders”. The model is also relevant in the present context where re-entry and recirculation are hypothesised as fundamental aspects of symptomatology. An interesting study by Rauch et al. (1996) examined trauma-related symptoms in PTSD and demonstrated decreased activity of Broca’s (language) area in the left hemisphere and increased activation in a number of right hemisphere regions, particularly in the temporal lobe, suggesting that in PTSD subjects, the images were processed less in the language domain, and more in the limbic structures involved in the affective representation of memory. A 10-year series of studies of PTSD in Vietnam combat veterans, rape victims and Holocaust survivors and their children, has been reported by Yehuda and colleagues (Yehuda, 2002; Ma- son et al., 2001; Yehuda et al., 2000). Yehuda (2001) reported that over the previous 10 years, biological observations unexpectedly showed that, in PTSD, urinary and plasma cortisol levels were considerably lower than in non-PTSD trauma survivors and normal controls. She believes that reduction in cortisol levels results from an enhanced negative feedback by cortisol, which is secondary to an increased sensitivity of glucocorticoid receptors in target tissues. Sensitisation of the HPA axis is consistent with the clinical picture of hyper-reactivity and hyper-responsiveness. Interestingly, Yehuda et al. (2000) found that low cortisol levels were significantly associated with PTSD in parents, and lifetime PTSD in their adult offspring, whereas having a current psychiatric diagnosis other than PTSD was relatively but non- significantly associated with higher cortisol levels. The authors 10.1. NEURAL CIRCUITS IN PTSD 303 point out that studies of World War I (5-7) and World War II (8) veterans and civilians exposed to disaster (9) all described a significantly higher rate of familial mental illness in trauma survivors with PTSD, compared to non-PTSD survivors. Community- based studies have confirmed that respondents with PTSD were three times more likely than trauma survivors without PTSD to report family mental illness. Also the risk for PTSD after trauma exposure was significantly greater for MZ than for DZ twins, suggesting a genetic susceptibility factor. However, non-genomic transmission of the stress response may also have been a factor. The authors comment that this effect of maternal behaviour has been shown to persist across multiple generations, and to be associated with increased hippocampal glucocorticoid receptor expression.

10.1.1 Neurocircuitry models

Alter and Hen (2009) have pointed out that physiologically primitive regions, such as the amygdala and hippocampus develop earlier than the phylogenetically more advanced regions, such as the frontal cortex. In the rat hippocampus, a dramatic rise and fall of dendritic growth and spine formation, integral to the formation of synapses and circuits is described in the first three weeks of life. In humans, a similar process is thought to occur early in the 3rd trimester, and to continue throughout much of childhood. This period is thought to represent a period of increased vulnerability to the effects of both experience and genetic variation. Shin and Liberzon (2010) describe neurocircuit models of PTSD as implicating the amygdala, medial PFC and hippocampus (Rauch et al., 2006). According to the authors, some models describe the amygdala as hyperresponsive in PTSD, while portions of the ventromedial PFC are hyporesponsive and fail to inhibit the amygdala. These deficits potentially lead to deficits in extinction, emotion regulation, attention, and contextual processing (Liberzon and Sripada, 2008). Abnormal hippocampal func- 304 CHAPTER 10. POSTTRAUMATIC STRESS DISORDER tion (with decreased volume) is thought to contribute to deficits in contextual processing, as well as impairments in memory and neuroendocrine dysregulation. According to Shin and Liberzon (2010), recent studies have suggested that the dorsal anterior cingulate cortex (dACC) is hyperresponsive in PTSD, perhaps underlying fear learning. In general, functional neuroimaging findings in PTSD support the hypothesis that the amygdala is hyperresponsive, and ventral portions of medial prefrontal cortex are hyporesponsive in at least some groups of PTSD patients (Etkin and Wager, 2007). McGowan et al. (2009), have demonstrated for the first time, a mechanism by which childhood adversity may be linked to altered stress responses in humans, which are associated with an increased risk for multiple forms of psychopathology. Previous studies had demonstrated that maternal care influenced hypothalamic-pituitary-adrenal (HPA) function in the rat, through programming of glucocorticoid receptor expression. Mc- Gowan et al. (2009) examined differences in a neuron-specific glucocorticoid receptor (NR3C1) promotor in postmortem hippocampus obtained from suicide victims with a history of childhood abuse versus that obtained from either suicide victims with no childhood abuse or controls. Epigenetic regulation of the glucocorticoid receptor gene, NR3C1, observed in humans, who had been abused as children, was consistent with predictions derived from a rodent model in which early postnatal experience influences adult response to stress. Early maternal interactions with rat pups were shown to produce persistent effects on the responsiveness of the HPA axis. Thus pups of mothers who showed high levels of licking and grooming, have substantially smaller HPA axis responses to stress, and release lower levels of adrenocor- ticotrophic hormone (ACTH) than offspring of mothers with low licking and grooming. The mechanism of an early “stress” mechanism, transmitted through environmental modification of a genetic promotor in rat pups, and an apparent similar early mechanism in suicide vic- 10.1. NEURAL CIRCUITS IN PTSD 305 tims, influencing the HPA circuitry indicates both the value of a detailed understanding of neural circuits, (and provides a possible mechanism for early intervention).

10.1.2 Nucleus Accumbens/Ventral striatum and Anxiety Syndromes

Levy (2004) has discussed the role of the nucleus accumbens in gating anxiety levels. Mogenson et al. (1980) described the role of the nucleus accumbens in functioning as an interface between limbic and motor systems (’from motivation to action’). Mogen- son described the nucleus accumbens is a key structure in linking motivation and action at the interface of the limbic system with motor mechanisms, receiving direct connections from amygdala, hippocampus and other limbic forebrain structures, as well as indirect connections via mesolimbic dopaminergic projections from the ventral tegmental area. The nucleus accumbens has direct motor connections to the globus pallidus and indirect connections via the substantia nigra and nigrostriatal dopaminergic system. Mesolimbic dopamine projections to the nucleus accumbens in rats were also implicated by abolition of the hyperactivity effect of systemically- administered amphetamine by damage to ventral tegmental projections to the nucleus accumbens by injections of 6- hydroxydopamine (Graybiel, 1995). The ventral striatum including the nucleus accumbens (NAcc) receives input from cortical areas other than the motor and sensory areas of the frontal and parietal lobes. It receives input mainly from areas that do not project to the dorsal striatum notably the temporal (including the hippocampal formation), limbic, and orbitofrontal cortical areas as well as the basolateral amygdala. The ventral striatum in turn projects to the ventral pallidum. Information is then relayed via the dorsomedial and ventral anterior nuclei of the thalamus to the cingulate, orbitofrontal and prefrontal cortex and to the ventral tegmental area of the brainstem. The ventral circuitry parallels the dorsal circuitry and is thought to be involved with 306 CHAPTER 10. POSTTRAUMATIC STRESS DISORDER behavioral phenomena, reward and punishment and in integrating cognitive with emotional responses. Thus the ventral striatum, including the accumbens is thought critical in integrating cognitive and emotional responses. Heimer (2003) has described the ventral striatal (including accumbens), pallidal system and extended amygdala as the major components of the “new anatomy” of the basal forebrain. He describes three re-entrant circuits-the anterior cingulate, the lateral orbito-frontal, and the medial orbito-frontal, which are closed in that they originate and terminate in the same area of the frontal lobe, but suggests they can also interact. He suggests that the entire cerebral cortex including the hippocampus and major parts of the amygdala project to the basal ganglia and are involved in the regulation of specific emotional functions and adaptive behaviours ranging from fear-anxiety, and addictive-reward and appetitive behavior. The ventral striatum integrates various cortical and subcortical inputs to adapt ‘motivational’ behavior in a similar way to that in which the motor loop through the dorsal parts of the basal ganglia is important for movement control. Grace (1995) suggested that when DA is released into the striatal synaptic cleft in response to action potentials, it is rapidly removed from the synapse by a highly efficient reuptake system (Grace, 2000). On the other hand, tonic DA level is thought mediated by stimulation of pre-synaptic heteroreceptors on DA terminals by corticostriatal glutamergic projections. Tonic DA exerts a suppressive influence on subcortical DA systems and is regulated, in part by frontal and cortical afferents to the accumbens. “The nucleus accumbens receives input from a number of limbic- related cortical structures, including the prefrontal cortex, hippocampus and amygdala . In particular, the hippocampus and amygdala strongly influence the ability of the prefrontal cortex to activate accumbens cell firing. As a result these systems alter- nately modulate the flow of information from the prefrontal cortex through the nucleus accumbens, where it can ultimately influence thalamo-cortical function via projections through the ventral 10.1. NEURAL CIRCUITS IN PTSD 307 pallidum” (Grace, 2001). According to Grace (2001), the nucleus accumbens is the striatal region in which the limbic system has overlapping inputs. Accumbens neurons exist in a bistable state, with their membrane potential alternating between a hyperpo- larized non-firing state and a depolarised plateau lasting several hundred milliseconds, during which spike activity is generated. Grace (2001) also suggests that the amygdala gates events based on their affective valence. “The amygdala also influences the accumbens by facilitating prefrontal stimuli but only within a very narrow single-event related time interval”. Thus this bistable accumbens state allows the operation of a synaptic gating mechanism between prefrontal and limbic influences on behaviour.

”Studies have suggested that the subiculum of the hippocampus is involved in a type of context dependent gating inﬂuence over the ability of the prefrontal cortex to control thalamocortical information processing, enabling the organism to stay on-task and focused. This mesocortical DA system provides a potent direct in- ﬂuence over activity within the prefrontal cortex. Among the many functions of the prefrontal cortex is the generation of a set of potential responses that the subject can utilise to react to a stimulus. Thus, while the hippocampus generates long duration activity in the accumbens neurons, keeping the subject focused on the current task, the amygdala provides a more brief, event-related gating of prefrontal throughput in the accumbens” (Grace, 2001).

According to Grace (2001) the amygdala is involved in emotional or affective properties of stimuli, enabling the subject to respond to events that are emotionally charged, and therefore of immediate survival value. In pathological states, the amygdala input may be overdriven to the extent that the maintenance of focus is overly disrupted by minor events. As described by Arn- sten (2000) high levels of dopamine and norepinephrine may have additive effects on information processing in PFC, reducing signals and increasing noise. Arnsten points out that although PFC functions are often essential for successful organization of higher- order behaviour, there may be some conditions, when it may be adaptive to ‘shut down’ these complex, reflective operations and 308 CHAPTER 10. POSTTRAUMATIC STRESS DISORDER to allow more automatic or habitual responses, dependent on posterior cortical or subcortical structures to control behaviour. Goto and O’ Donnell (2002) have reported timing-dependent limbic-motor synaptic integration in the nucleus accumbens. In the current context the demonstration of differing effects of coincident and asynchronous PFC and limbic activity suggests a possible gating mechanism, whereby PFC activity can gate limbic input depending on the state of the neurons and the timing of inputs. According to Floresco et al. (2003), the mesolimbic dopamine system plays a central role in reward and goal-directed behavior, and has been implicated in multiple psychiatric disorders. Under- standing the mechanism by which dopamine participates in these activities requires comprehension of the dynamics of dopamine release. They report dissociable regulation of dopamine neuron discharge by two separate afferent systems in rats; inhibition of pallidal afferents selectively increases the population activity of dopamine neurons, whereas activation of pedunculo-pontine inputs increases burst firing. Synaptic or phasic levels of dopamine are mediated by bursting events at the level of the cell body, restricted by high affinity and rapid uptake systems, and associated with reward-conditioned prediction. On the other hand extrasynaptic or tonic dopamine levels are modulated by presynaptic limbic and cortical glutamergic inputs. Alterations in tonic levels of dopamine efflux occur on a much slower time-scale and allow a wide variety of motor, cognitive and motivational functions. Ramos and Arnsten (2007) have pointed out that norepinephrine (NE) has widespread projections throughout the brain, and is thus ideally positioned to orchestrate neural functions based on arousal state. Thus NE can increase signal/noise ratio in the processing of sensory stimuli, and enhance long- term memory consolidation in the amygdala and hippocampus, through actions at alpha-1 and beta adrenoreceptors. In addition, NE has been shown to play a powerful role in regulating the working memory and attention functions of the PFC. The modest levels of NE released under controlled conditions strengthen 10.1. NEURAL CIRCUITS IN PTSD 309

PFC functions, via actions at post-synaptic alpha-2 adrenoreceptors, while high levels released during stress impair PFC cortical functions via alpha-1 and possibly beta-1 receptors. Thus levels of NE can rapidly determine whether prefrontal cortical or cortical systems control behaviour and thought. Thus while tonic/phasic NE mechanisms are able to rapidly switch behavior to fast-acting posterior cortical/amygdala control, slower acting dopaminergic tonic/phasic mechanisms are involved in prefrontal control of working memory and sequential behaviour. According to Friedman and Karam (2009, p4), animal research, fortified by brain imaging studies has identified neurocircuitry that mediates processing of threatening or fearful circuitry. “Such stimuli activate the amygdala, which, in turn, produces outputs to the hippocampus (to mediate memory consolidation and spatial learning); orbital frontal cortex (to process memory of emotional events and choice behaviours); locus coeruleus/thalamus/hypothalamus (to mediate autonomic and fear reactions; and dorsal/ventral striatum (to instigate instu- mental approach or avoidance behaviour) (Davis and Whalen, 2001). According to Friedman and Karam (2009) in PTSD, the normal restraining influence of the medial prefrontal cortex, especially the anterior cingulate gyrus and orbitofrontal cortex, has been severely disrupted. “The resulting disinhibition of the amygdala is thought to increase the likelihood of recurrent fear conditioning because stimuli are more likely to be misinterpreted as threatening; normal counterbalancing inhibitory prefrontal cortex restraint is nullified; and sensitisation of key limbic nuclei may occur, thereby lowering the threshold for fearful reactivity (Charney, 2004). Friedman and Karam (2009, p13-14) also discuss the factor structure of PTSD. They point out that few studies have supported the DSM-IV three-factor model of reexperiencing, avoidence/numbing and arousal. On the other hand, three studies have supported a two-factor model including intrusion/hyperarousal and avoidance (Creamer et al., 2003), while 310 CHAPTER 10. POSTTRAUMATIC STRESS DISORDER other studies have supported four factors. The authors point out that fear conditioning has been proposed as a component of a two- factor theoretical model, and within a cognitive context of activated fear networks (Foa and Kozak, 1986). Rauch and Drevets (2009, p229-230) have reviewed neuroimaging studies of PTSD and point out that taken together, imaging data support the current neurocircuitry model of PTSD that emphasises the functional relationship between the amygdala, medial prefrontal cortex (mPFC), and hippocampus. “When exposed to reminders of traumatic events, subjects appear to recruit anterior paralimbic areas and the amygdala, while exhibiting decreased activity within other heteromedial cortical areas. In comparison with control subjects, patients with PTSD exhibit diminished pregenual anterior cingulate cortex (pACC) activation, but exaggerated activation within other limbic (amygdala) and paralimbic regions (including the insula), along with exaggerated neurophysiological deactivations within widespread areas associated with higher cognitive functions.

10.1.3 Sensitive period: gene environment interaction

According to Alter and Hen (2009) gene-knockout studies and phenotypic characteristics of individuals with functional polymorphisms have highlighted several genes of the serotonin system that affect the developmental course of circuits involved in anxiety. The serotonin transporter (SERT) and serotonin 5-HTRB genes are thought to be imprtant in this process. The normal functioning of the (l/l) allele of the SERT polymorphism is thought to be protective against many of the behavioural consequences of early life stress in humans (Caspi et al., 2003; Jorm et al., 2000). Alter and Hen (2009) comment that deﬁning circuits on which to focus research “opens the door for a careful dissection of genetic and environmental inﬂuences on neural function, that is not available when using the gross outcome measures of behaviour and disease”. 10.1. NEURAL CIRCUITS IN PTSD 311

10.1.4 Comment

While the above data provide evidence of how neural network structures undergo modification in PTSD, they also have relevance in a developmental context, such as the functioning and adaptation of neural networks in autism, where local networks predominate over long-range connections. The McFarlane et al. (2002) model of PTSD is particularly cogent for the present thesis where the importance of iterative information, providing access to linguistic representations and to memory has been central to the understanding of symptomatology in conditions such as autism, where this is absent. In terms of the present investigation, the stress response system demonstrates the principle that when homeostasis of an iterative system is disrupted (in this case by an exaggerated amygdala response and diminished PFC control), a repetitive behavioural syndrome (which can go on for many years in terms of flashbacks and anxiety) results, possibly as a continuing attempt to restore homeostasis. The process of circuit formation in PTSD appears to be in an opposite direction to the process of normal development. In the latter, there is a progression from modular to more domain general or integrated information processing, while in PTSD there is a progression to more modular and fragmented processing. Also the Friedman and Karam (2009) and Rauch and Drevets (2009) descriptions of diminished restraining influence of the medial prefrontal cortex, especially the anterior cingulate gyrus and orbitofrontal cortex with resulting disinhibition of the amygdala, lowering the threshold for fearful reactivity, fits with the basic architecture described for OCD in terms of fronto-striatal disinhibition in a parallel circuit. In this case the circuit consists of mPFC, orbital frontal cortex, amygdala and hippocampus (possibly incorporating the extended amygdala described by Heimer et al. (2008)). 312 CHAPTER 10. POSTTRAUMATIC STRESS DISORDER Chapter 11

Comorbidity

11.1 Diagnostic issues

Comorbidity between classes of Child and Adolescent psychiatry disorders has been frequently described. Angold et al. (1999) reviewed the literature on the overlap between these conditions. He pointed out that the way in which comorbidity is treated has changed over time. For example, the ICD-10 allows a mixed category of conduct and emotions, whereas proponents of a factor analytic approach to diagnosis developed the idea that “patterns of psychopathology were a matter of continuous variation along a series of scales” suggesting a failure of categorical diagnosis to appropriately describe separable syndromes on the one hand, and on the other hand, an artifact of the failure to recognise the co- variation between naturally occurring scalar syndromes. For example, between 50 percent and 80 percent of children with ADHD also meet diagnostic criteria for other disruptive behaviour disorders, namely Oppositional Defiant Disorder (ODD) (Waldman and Lilienfeld, 1991) and Conduct Disorder (Thapar et al., 2001) or for Learning Disorders and Communication Disorders (Tannock et al., 1995). There are also high rates of comorbidity for internalising problems such as anxiety (Goodyear and Hynd, 1992). Krueger et al. (2005) proposed that classification in the fifth edition of the Diagnostic and Statistical Manual (DSM) should re-

313 314 CHAPTER 11. COMORBIDITY

flect the etiological and clinical commonalities of the externalising disorders. They have pointed to connections among externalising disorders and personality traits associated with aggression and impulsivity. A twin study by Young et al. (2009) has linked liability for externalising spectrum disorders with psychometric measures of response inhibition at a genetic level. The investigators assessed behavioural (dis)inhibition using measures tapping substance use, conduct disorder, attention deficit hyperactivity disorder (ADHD), and novelty seeking in a cohort of 293 same-sex and twin pairs from the Colorado Longitudinal Twin Study at ages 12 and 17 years. Executive functions were assessed using a number of psychometric measures, including response inhibition, working memory updating and task-set-shifting. The results indicated that at age 12, behavioural disinhibition was demonstrated by ADHD and conduct problems, and was highly heritable, while at age 17, the proportion of variance at- tributable to genetic factors was somewhat smaller, with additional variance due to shared genetic factors. At both ages, behavioural disinhibition was more closely related to response inhibition than working memory updating and task-set-shifting. Response inhibition utilised an antisaccade task, where subjects were required to override their automatic tendency to move their eyes towards a cue, when briefly flashed (for 150msec) on the left or right computer screen, in the opposite direction of a small target, which appeared immediately after the cue for 175 msec, pointing in the opposite direction. Additionally a stop-signal task, in which participants had to periodically stop a prepotent categorisation response, in response to an auditory signal, as well as a Stroop task, requiring inhibition of a prepotent word reading response in favour of a colour categorisation response were employed. The investigators characterised behavioural disinhibition as a common latent factor, representing a general vulnerability to various externalising disorders. In contrast, they describe an alternate model, which conceptualises behavioural (dis)inhibition, 11.1. DIAGNOSTIC ISSUES 315 as a heterogenous dual process, associated with two distinct, interacting neural circuits (Nigg, 2003, 2006). Nigg (2003) argued that conduct disorder involves a primary failure of reactive or motivational control processes (i.e. detecting and responding to immediate contextual cues, related to incentives, punishment, or social anxiety, with secondary breakdowns in effortful control of attention or executive control processes. According to Nigg, ADHD develops in part from a failure of neural systems of executive control, with secondary deficits in reactive or motivational controls. Young et al. (2009) investigated which separate components of ADHD (attention problems and hyperactivity/impulsivity) were differentially related to other externalising disorders, as well response inhibition. They found that conduct problems were correlated less highly with attention problems than with hyperactivity/impulsivity, while the correlations with substance use and novelty seeking were nearly identical for the two ADHD components. Attention problems were more highly correlated with response inhibition, than hyperactivity/impulsivity problems, but both components showed highly significant genetic correlation with response inhibition. The investigators concluded that their findings partially mirrored previous research, but suggested that conduct problems and associated motivational problems (e.g. high risk taking, low fear of punishment) were strongly associated with deficits in response inhibotion, and the relationship was largely genetic. The authors comment that their common pathway model takes a different perspective to Nigg’s dual process model, but both approaches examine both the unity (common underlying risks) and diversity (differential underlying risks) of externalising problems and associated cognitive deficits. Lahey et al. (2008) utilised confirmatory factor analysis (CFA) of twin data in a large representative sample of 4,069 six to sev- enteen year old twins, in 28 urban, suburban and rural coun- ties in Tennessee, to investigate a dimensional model of DSM- IV symptomatology. They utilised adult caretaker and youth in- 316 CHAPTER 11. COMORBIDITY terviews on the Child and Adolescent Psychopathology scale. The symptom dimensions were organised hierarchically within higher-order “externalising” and “internalising” dimensions. Of present interest, the investigators found that for major depressive disorder (MDD)/generalised anxiety disorder (GAD), the higher- order CFA model, which fit best, loaded on both “internalising” and “externalising” dimensions, indicating an overlap between the traditional separation of these dimensions. Angold et al. (1999) explored the ways in which research findings on pairs of comorbid diagnoses have been analysed in attempts to understand the causes of psychopathology and re- fine nosology. Angold and colleagues distinguish homotypic comorbidity, where there is continuity of phenomena over time, from heterotypic comorbidity where a continuing process is differently manifested over time. The authors point out that a large body of literature, based on parent questionaires has established at least seven highly replicable factor-analytically derived syndromes that can be applied to boys and girls at all ages, including aggressive, anxious/depressed, attention problems, delin- quent, schizoid, somatic complaints and withdrawn. The presence of “mixed disorder” was established by the Isle of Wight study (Graham and Rutter, 1973) where mixed disorder was diagnosed 14 times more often than would be expected from the prevalence of separate disorders at age 10-11, and 8 times more often at age 14 to 15. Angold et al. (1999) conducted a meta-analysis of general population estimates of the strengths of associations between disorders. The authors found that the association between conduct disorder (CD) and oppositional defiant disorder (ODD) was weaker than that between CD and depression. This argued against the existence of a unitary association between “internalising” and “externalising” disorders. While it has been argued that these comorbidity findings point to flaws in “medical models” of psychopathology, Angold et al. (1999) point out that the diagnostic literature has shown that associations between symptoms of dif- 11.2. CIRCUIT MODELS 317 ferent types occur at the extremes of psychopathology, and not just at the level of mild factor scores. Thus “the issue for future research is to explore why symptoms group together in ways that they do, and why there is overlap between syndromes, whether defined by diagnostic criteria or factor scores”. While the present investigation does not aim to answer these questions, it is useful to examine the connections between modular systems referred to above, which may provide an underlying basis for the comorbidity observed in epidemiological and clinical studies. Here the work of Haber (2003) above, who has demonstrated that within each area of cortico-basal connections, linking up areas with similar functions (maintaining parallel networks), and in addition, non-reciprocal connections associated with differing cortico-basal ganglia circuits. Haber points out that the development and modification of goal-directed behaviours require continual processing of complex chains of events, reflected in the feed-forward organisation of both striato-nigral connections and thalamo-cortical connections. Where parallel and integrative circuits are not coordinated, maladaptive syndromes may result.

11.2 Circuit models

Haber and McFarland (2001) has discussed direct and indirect output routes from the thalamus (the direct pathway from the striatum to the internal segment of the globus pallidus, GPi and substantia nigra pars reticulata, SNr to the thalamus), and indirect route (from the external segment of the globus pallidus, GPe to the subthalamic nucleus to the GPi, and from the GPi to the thalamus (discussed above in relation to obsessive compulsive disorder). She points out that the complexity of the thalamo- cortico-thalamic circuit raises an issue related to the concept of parallel versus nonparallel processing of information through basal ganglia pathways. Because there are both reciprocal and non-reciprocal cortico-thalamic relay nucleus from basal ganglia output structures, the information that the relay nuclei convey to 318 CHAPTER 11. COMORBIDITY the cortex is not only affected by the parallel path through the basal ganglia structures, but is also modified by the feed-forward component of the non-reciprocal corticothalamic pathway, allowing influence from other functional regions of the cortex. Because these functional cortical areas are also, according to Haber involved in their own basal ganglia loop, the transfer of information occurs between different functional basal ganglia loops in the final pathway to the cortex. Haber and McFarland (2001) describe a tight anatomical and functional triad between the cortex, basal ganglia, and thalamus, with the same regions of the thalamus that convey basal ganglia output to the cortex, and projecting back to the striatum. “These thalamic nuclei therefore play a dual role in basal ganglia circuitry; both relay basal ganglia output to frontal cortical areas, and provide direct feedback to the striatum”. The authors speculate that thalamic input might sustain disinhibition of specific thalamocortical circuits, which would release responses and keep a behaviour focused on a task until its goal is achieved, or alternatively it may further reinforce or facilitate selection of basal ganglia circuits to enable desired and suppress unwanted behaviours. The control of adaptive motor behaviour thus results from a complex interplay of excitatory and inhibitory reciprocal and non-reciprocal cortico-thalamo-striatal circuits. Comorbidity between motor and attention syndromes may be partially explained by connections between parallel modular pathways, comorbidity between internalising and externalising syndromes suggests connections with motivational circuits, chal- lenging the traditional modular separation of these syndromes. Heimer et al. (2008, p121) has discussed the degree of segregation of cortico-basal ganglia-thalamocortical circuits or loops. He points out that while it may be tempting to overstate their closed and segrated nature, the work of Haber (2003) has suggested substantial feed-forward routing of impulse conduction from limbic lobe circuits, via executive circuits to cortical motor circuits. Acccording to Heimer, similar neuroanatomical consider- 11.2. CIRCUIT MODELS 319 ations (Zahm and Brog, 1992; Joel, 1994) suggest a general flow of information from ventromedial to dorsolateral parts of the striatal complex “consistent with the idea of a passage of information from “motivational” to “motor” parts of the cortico-basal ganglia- thalamocortical circuitry. While cortical-subcortical functional areas are broadly thought to relate to motor, prefrontal and limbic cortical areas, the separation of functional domains is described by Heimer et al. (2008, p51) as blurred because the corticostriatal projections from adjacent functional domains within the striatum overlap extensively with each other. “Limbic lobe projections (related to emotional-motivational functions) overlap with dorsal prefrontal projections (serving cognitive executive functions). Dorsolateral prefrontal projections to striatum, likewise overlap with projections from motor cortical areas (Heimer et al., 2008, p51).

11.2.1 Dual system model and comorbidity

The investigation of the neuropathophysiology of Attention Deﬁcit Disorder and comorbid conditions is providing new insights into both normal and problematic cognitive development and comorbidity. Bush (2010) has reviewed the major components of neural systems, potentially relevant to ADHD, including a ’cingulo-fronto-parietal’ (CFP) attention network. This is suggested to consist of the dorsal anterior mid-cingulate cortex (daMCC), dorsolateral prefrontal cortex (PFC), ventrolateral prefrontal cortex (VLPFC) and parietal cortex. While the DLPFC and VLPFC are believed to support vigilance, selective and divided attention, planning and executive control, as well as working memory, the daMCC is believed to be a key modulator of moment- to-moment adjustments in behaviour through its primary role in feedback-based decision making. On the other hand the striatum is believed to contain separable components of parallel distributed circuits that support executive and motor functions. 320 CHAPTER 11. COMORBIDITY

Sesack and Grace (2010) have described the nucleus accumbens (NAc) as part of the ventral striatum, which serves as a critical region where motivations derived from limbic regions, interface with motor control circuitry to regulate appropriate goal- directed behaviour. “The ventral subiculum of the hippocampus provides contextual spatial information, while the basolateral amygdala conveys affective inﬂuence, and the prefrontal cortex provides an integrative inﬂuence on goal-directed behaviour”.

How do we utilize these models to understand comorbidity. Ac- cording to Koziol and Budding (2009, p14-16) most behavioural processing can be categorised as stimulus-based, and higher- order control respectively. Stimulus-based control is thought to consist of habits, skills and procedures, which allows for high speed of reaction. However, when problem-solving and planning are required, higher-order processing is required. As described by Koziol and Budding (2009, p41) the dual system model is characterised by both stimulus-bound processing and higher-order control. The authors describe the cortico-striatal system as modulating cognition and behaviour through loops, projecting from cortex to basal ganglia to thalamus and back to cortex. The basal ganglia are described as gating appropriate behaviours via inhibitory projections from the globus pallidus interna to the thalamus and prefrontal cortex, following activation by the cortex. Koziol and Budding (2009) point out that analogous operations on representations in visual and auditory association areas, which project to the striatum may also be gated after activation by parietal, temporal or prefrontal cortices. This model allows the possibility of inadequate gating at striatal levels, and consequent production of inappropriate comorbid behaviours. A corollary of both the dual system and neuroanatomical models draws attention to the important role of striatal functions in comorbid syndromes. By virtue of the central location and role of the basal ganglia in gating behaviour, comorbid syndromes are frequently observed in children with behaviour problems. The dual system hypothesis would also predict that comorbid behaviours would more often re- 11.2. CIRCUIT MODELS 321

ﬂect stimulus-bound behaviours, including repetitive, impulsive, distractible behaviour, but also including anxiety, as a result of inadeqate gating at striatal and accumbens levels (Levy, 2010).

11.2.2 Interaction between cognitive and affective systems

Pavuluri and Sweeney (2008) point out that the dorsolateral prefrontal cortex (DLPFC) and the ventrolateral (VLPFC) serve as higher cortical centers of cognitive control in concert with cingulate cortex and striatum. Similarly, Pavuluri and Sweeney (2008) point out that fronto-striatal cognitive circuitry that connects the DLPFC and VLPFC to the basal ganglia (caudate and ventral striatum respectively) has been shown to be affected in paediatric bipolar disorder (PBD). The authors suggest that these observations may “lead clinicians to consider treating symptoms, such as inattention, linked to frontostriatal dysfunction in common to both ADHD and paediatric bipolar disorder (PBD), rather than target diagnosis-speciﬁc brain circuits that may overlap across disorders. Pavuluri and Sweeney (2008) predict an important direction for future research involving efforts to probe the interface between affective and cognitive processing, allowing an understanding of how thinking and feeling affect each other. Thus, studies conducted in macaque monkeys, humans without disorder, and patients with PBD suggest an interaction between cognitive and affective brain regions at three hierarchically organised tiers.

Higher cortical interactions between DLPFC and VLPFC. An intermediate level of interface between dorsal and ventral anterior cingulate cortex (ACC). Interaction at the subcortical level between dorsal and ventral striatum and amygdala. (Pavuluri and Sweeney, 2008)

Pavuluri and Sweeney (2008) have investigated the relation between affective and cognitive systems in euthymic unmedi- 322 CHAPTER 11. COMORBIDITY cated subjects with PBD, utilising a task that required subjects to match the colour of emotional words with one of two coloured dots that appeared below. In response to negative word matching, relative to neutral word matching, patients with PBD had less activation at the junction of the VLPFC and DLPFC, and greater activation of bilateral pregenual ACC and left amygdala. The dysfunction at the interface of affective (VLPFC) and cognitive (DLPFC) regions was greater in response to negative word matching when compared with positive word matching. The investigtors also propose examining the impact of implicit/automatic versus explicit emotion processing, while performing a cognitive task. “The key difference between implicit fast processing and explicit conscious processing of emotion is that the latter involves elaborate perceptual analysis and cognitive processing of stimuli and relevant context. Thus automatic and fast transfer of emotional aspects of stimuli from the visual association cortex to the amygdala via the occipitolimbic association cortex.

11.2.3 Comment

Although the above line of work is in its early phases, it raises a number of important issues. First is the importance of shared circuits and possible overlapping of symptoms in common, in a number of syndromes such as ADHD and PBD. Second is the interaction between dorsal cognitive and ventral affective circuits in a number of childhood syndromes, including ADHD, PBD (and possibly Conduct disorder), and third is the difference between fast implicit (subconscious) and explicit (conscious) processing. Anal- ogous connectionist models which utilise hidden units and differential weighting of information transfer may be useful in understanding the effects of differential arousal levels in the above circuits. Pavuluri and Sweeney (2008) suggests that in place of the current pattern in clinical practise, which considers disturbances in 11.2. CIRCUIT MODELS 323 different domains as representing “comorbid” expression of different disorders in a given patient, a more comprehensive understanding of brain function may lead to a shift in thinking from treating “multiple comorbid disorders” to treating “multiple dimensions of deﬁcit”. Thus “attentional problems” or a diagnosis of ADHD may be an earlier manifestation of frontostriatal abnormalities relative to the emergence of “fronto-limbic” abnormalities over the time course of neurodevelopment, where recruitment (or failure of recruitment) of the DLPFC by age 16 years may (or may not) resolve attentional problems. The authors suggest that speciﬁc understandings of symptom-brain function may more clearly inform treatment planning.

11.2.4 Cognitive/motor comorbidity

Comorbidity of Tourette’s Disorder with hyperkinesis and obsessive-compulsive disorder has often been described. Towbin et al. (1999) point out that leaving aside the chance that a diagnostic error has occurred, there are three possible relationships that may exist between two disorders when they are observed in one person. The first is that they are entirely separate entities, running independant courses (true comorbidity). The second is when one disorder ‘sets the stage’ for another, such as an opportunistic fungal infection, but both are aetiologically independent. The third case is when two disorders are aetiologically related, but are differentially expressed (such as peripheral neuropathy and retinal haemorrhages in diabetes). As described in Chapter 10, Leckman and Cohen (1999) points out that complex tics can be extremely difficult to distinguish from compulsions, and that 30 to 60 percent of Tourettes’ patients are also diagnosed as having Obsessive Compulsive Disorder (OCD), with some studies suggesting genetic relationships (Pauls et al., 1991). Similarly Leckman and Cohen (1999) claim that approximately 50 percent of all clinically identified Tourette’s syndrome patients also have ADHD. While ascertainment bias 324 CHAPTER 11. COMORBIDITY may account for some of these relationships, Leckman attributes the underlying pathology of these conditions to parallel cortico- striato-thalamo cortical (CSTC) circuit pathology (Alexander and Crutcher, 1990; Goldman-Rakic et al., 1990; Parent and Hazrati, 1995). According to Leckman and Cohen (1999), current consensus held that CSTC circuitry had at least three components, those initiating from and projecting back to sensory-motor cortex; orbitofrontal cortex; or association cortices. Further components were also thought to include limbic system components. Leckman and Cohen (1999) describe human studies supporting the importance of basal ganglia in Tourette’s Syndrome pathophysiology (Haber and Wolfer, 1992). The CTSC circuits are also implicated in obsessive compulsive disorder (Baxter, 1992). Leckman and Cohen (1999) describe the limbic system as consisting of the amygdala and hippocampus , the temporal lobe, the cingulum, the caudate nucleus and other basal ganglia structures in the subcortex, the hypothalamus and periaqueductal grey matter in the brain stem and connections with associative frontal cortex. The involvement of the cingulate cortex in both Tourette’s and Obsessive Compulsive Disorder is described as inhibitory. Decreases in severity of OCD symptoms have been correlated with decreases in right anterior cingulate metabolic activity during successful medication therapy (Baxter, 1992). Finally, abnormalities in striatal function and related association cortex circuitry are thought to contribute to ADHD symptomatology (Castellanos et al., 1996; Leckman and Cohen, 1999) also suggest that limbic system circuitry, including the temporal lobe may be involved in the sexual and aggressive content of many complex motor and vocal tics, while the sexual and aggressive content of many obsessions and compulsions suggests involvement of amygdala and related circuitry. Towbin et al. (1999) examined the relationships between ADHD, Obsessive Compulsive Disorder and Tourette’s syndrome. In the case of ADHD and Tourette’s syndrome, the authors believe that aetiology may be independent, but in some cases symptoms 11.2. CIRCUIT MODELS 325 of ADHD may precede Tourette’s Disorder. On the other hand, Obsessive-Compulsive disorder may in some cases be the product of the same aetiology (Pauls et al., 1986). The authors point out that a categorical classification would view the three disorders as separate, whereas a concept of overlapping mechanisms, evolving from impairment at related anatomical sites or systems (such as the cortico-striato-thalamo-cortical system) provides a better explanation of comorbidity. Leckman (2002) has described cortical-subcortical circuits, which project to the striatum and back to specific regions of the cortex. Leckman describes two structurally similar, but neuro- chemically distinct compartments in the striatum - striasomes and matrix, which differ by their cortical inputs. The striasomal medium spiny projection neurons mainly receive convergent limbic and prelimbic inputs, while neurons in the matrix receive convergent input from ipsilateral primary motor and sensory motor cortices, and contralateral motor cortices. According to Leckman (2002), the response of particlar medium spiny projection neurons in the striatum is partly dependant on perceptual cues that are judged salient, so rewarding and aversive stimuli can both serve as cues. “Tonically active neurons are very sensitive to salient perceptual cues, because they signal the networks” [...] “The fast spiking interneurons of the striatum are electronically coupled via gap junctions that connect adjacent dendrites. Once activated, these fast spiking neurons can inhibit many striatal projection neurons synchronously” (Leckman, 2002). Tics are thought to reflect a lack of balance (striasome vs. matrix). According to Leckman (2002), alterations in the structure of the basal ganglia, including loss of right-left symmetry, and cortico-thalamic inputs may be important in tic and hyperki- netic disorders. Recent neurosurgical procedures have, according to Leckman (2002), reinforced the functional importance of thalamic regions, that are part of the above cortical-subcortical loops. Additionally, ascending dopaminergic pathways are thought to have a role in the consolidation and performance of tics. Leck- 326 CHAPTER 11. COMORBIDITY man (2002) points out that early results of in-vivo neuroimaging studies have shown that voluntary tic suppression involves activation of regions of the prefrontal cortex and caudate nucleus and bilateral deactivation of the putamen and globus pallidus. Accord- ing to Leckman, the findings accord with the well-known finding that chemical or electrical stimulation of inputs into the putamen provokes motor and phonic responses resembling tics. Leckman and Cohen (1999) describe human studies supporting the importance of basal ganglia in Tourette’s Syndrome pathophysiology (Haber and Wolfer, 1992). Additionally, Leckman describes a case study, which showed that high frequency stimulation of the median and rostral intralaminar thalamic nuclei produced a reduction of tics. This effect was thought to be possibly caused by the effect of the midline thalamic nuclei on tonically active neurons, or on a broadly distributed cortical system, suggesting the possibility of circuit-based therapies. While the above syndrome descriptions appear to reflect a modular, motor syndrome, the association of premonitory sensory urges, the presence of motor and phonic tics, as well as coprolalia, and associations with hyperkinesis and obsessive-compulsive disorder suggest greater complexity. The division of tics into those with and without cognitive phenomena suggest that TD may be characterized with and without cognitive phenomena. Haber (2003), describes the mediation of cognitive and motor planning, and drive behaviours, in relation to these phenomena, by the parallel cortical-subcortical neural circuits originally described by Alexander et al. (1986). However, the Alexander model fails to address how information flows between circuits. Haber points out that recent anatomical evidence from primates demonstrates that neuro-networks within the basal ganglia are in a position to move information across functional circuits. She describes two networks: the striato-nigral, and the thalamo-cortical-thalamic network. Importantly, within each of these nets of connected structures, Haber describes both reciprocal connections, linking up regions associated with similar functions, and non-reciprocal 11.2. CIRCUIT MODELS 327 connections linking up regions that are associated with different cortico-basal ganglia connections. Thus each component of information (from limbic to motor outcome) sends feedback and feed- forward connections, allowing the transfer of information, which may help explain comorbid symptomatology in Tourette’s Disor- der. Thus Tourette’s Disorder, Obsessive Compulsive Disorder, and ADHD could be conceptualised as manifesting respectively more or less cognition, as the involvement of orbitofrontal cortex in the case of OCD increases, with more or less affective reactivity as the involvement of limbic and temporal and cingulate lobes increases. This concept also applies to “dorsal” (cognitive) and “ventral” (affective) systems and symptomatology.

11.2.5 Circuit pharmacology

Stahl (2008, p202-214) describes dopamine projections as arising predominantly, but not exclusively from brainstem neurotransmitter centers, notably the ventral tegmental area, substantia nigra and to a lesser extent cerebellum and spinal cord. They regulate movement, reward and cognition, as well as thalamic neurons, which gate the transfer of information to the neocortex, striatum, and amygdala. Norepinephrine projections arise largely from the locus coerulus in the brainstem, and are thought to regulate mood, arousal and cognition. The major serotonin projections are believed to arise from several clusters of discrete brainstem nuclei, and differentially innervate many brain areas, including cerebellum, and are thought to regulate mood, anxiety, sleep, and pain pathways. Acetylcholine projections arise from the brainstem and basal forebrain and are thought to regulate arousal and cognition. The latter projections from the nucleus basalis are thought to have a prominent role in memory. According to Heimer et al. (2008, p62) the extended amygdala is a “hotbed” for neuropeptides (e.g. opiods, neurotensin) and neuropeptide receptors (e.g. vasopressin, oxytocin) and cortiptrophin releasing factor, all 328 CHAPTER 11. COMORBIDITY of which may be important in stress and anxiety. Cortico-cortical circuits: According to Stahl (2008, p207-208), cortico-cortical circuits link neurons together into functional loops. Thus while pyramidal cortical cells are “tuned” by neurotransmitters from below, cortical circuits may allow a third intermediary neuron to have positive or negative effects. Cortico-striatal-thalamic-cortical (CTSC) circuits: Stahl (2008) describes an important cortical circuit called the “CTSC loop”, allowing information to be sent downstream, and yet get feedback on how information was processed. Thus the loop through the striatum may have a synapse through another part of the striatal complex, before it leaves to go to the thalamus, which relays back to the original area of prefrontal cortex. For example, the dorsolateral prefrontal cortex (DLPFC) projects to the rostral part of the caudate, within the striatal complex, then to the thalamus, and back to DLPFC. Loops with this structure are thought to regulate executive functions, problem solving, and cognitive tasks, such as representing and maintaining goals, and allocating attentional resources to various tasks. Similarly the same type of loop arising from the dorsal anterior cingulate gyrus (ACC) modulates selective attention and self-monitoring of performance, via the ventral area of the thalamus. A third CTSC loop arises from the ventral or subgenual part of the ACC, and projects to a part of the striatal complex, the accumbens, extending to the thalamus, and back to subgenual ACC. This loop is thought to regulate emotions, including depression and fear. A fourth CTSC loop outputs from the orbital frontal cortex (OFC) to the ventral part of the caudate nucleus in the striatal complex, to the thalamus and back to OFC, and is believed to regulate impulsivity and compulsivity. Finally, a fifth loop starts in the supplemental motor area of prefrontal motor cortex, and projects to the putamen in the lateral part of striatal complex, then to thalamus, and back to premotor cortex, and may modulate overactivity, psychomotor agitation and psychomotor retar- 11.2. CIRCUIT MODELS 329 dation. These loops might also be deemed executive, attentional, emotional and motor modules. Stahl (2008, p213-214) describes CTSC loops as a very good example of how cortical engines in different topographical brain areas, not only drive neuronal structures throughout the brain, while receiving feedback from them. Each CTSC loop starts and ends in a pyramidal cortical cell, shaped like a triangular pyra- mid, and having extensively branched spiny apical dendrites and shorter basal dendrites. Pyramidal cells are located in four of the five vertical cortical laminae. The excitatory neurotransmitter output from most pyramidal cells is glutamate, while the input from the various brain areas is inhibitory GABAergic. The inhibitory GABAergic interneurons are described as basket neurons, which input to the end of an apical dendrite in the pyramidal neuron. A second inhibitory neuron can inhibit the inhibition or disinhibit the pyramidal neuron. Thus “the presence or absence of GABA tone on the pyramidal neuron can have a profound influence on the ability of that cortical pyramidal neuron to serve as the driver of a cortical engine that delivers the behavioural program of the brain”. Stahl (2008, p219) also describes cortical pyramidal neurons as receiving many excitatory inputs, coming predominantly to synapse with the apical dendrites, and utilising the excitatory neurotransmitter glutamate. “Thus it is easy to see how the cortical pyramidal neuron can either be excited by these long-distance glutamate inputs, or be inhibited by short-distance GABA inputs”. According to Stahl (2008, p220-221), glutamate and GABA exert more of an “on-off” effect on pyramidal neurons, while monoamines and other neurotransmitters may exert more of a “fine-tuning” action on pyramidal cells. Inputs from various other neurotransmitter centers, including dopamine, norepinephrine, serotonin, and acetylcholine synapse on basilar and apical dendrites. These inputs can, according to Stahl, be either excitatory or inhibitory. “ [...] graded degrees of monoamine input can “tune” a pyramidal neuron to signal that it must prioritise, while allow- 330 CHAPTER 11. COMORBIDITY ing it to ignore competing inputs” (signal to noise ratio). Stahl points out that for optimal tuning, finding the optimal amount of receptor stimulation is necessary. Thus monoamine neurotransmitters require regulation. This is accomplished by genetic control of catalytic enzymes which degrade monoamines. He provides two examples, the dopamine metabolising enzyme, catechol-O- methyltransferase (COMT), and genes for the serotonin transporter (SERT). These enzymes are thought to be involved in the psychpathology of schizophrenia and anxiety/depression respectively. Stahl (2008, p245) points out that malfunctioning circuits can be imaged by a variety of tasks, which “provoke” circuit function, allowing visualisation of the effects of risk genes on the ef- ficency of information processing in specific neuronal circuits. He describes the strategy for utilising knowledge of circuits for categorisation of pathology:

A review by Arnsten (2007) has provided molecular understandings of the powerful ‘inverted-U’ inﬂuences of both dopamine (DA) and norepinephrine (NE) on the prefrontal cortex (PFC). Arnsten points out that optimal NE levels engage alpha- 2A adrenoreceptors, and increase “signals”, via inhibition of cAMP-hyperpolarization-activated cyclic nucleotide-gated cation channel (CAMP-HCN), signaling near preferred inputs, wheras optimal levels of DAD1 receptor stimulation decreases “noise” by increasing CAMP signaling near non-preferred inputs. Thus in ADHD, suboptimal levels of catecholamine transmission are treated with catecholamine agonists, to support delay-related signal activity. “It is assumed that the inhibition of CAMP by alpha- 2A adrenoreceptors, and the activation of CAMP by D1 receptors occurs on separate spines, receiving inputs from neurons with shared versus dissimilar spatial properties. On the other hand, Post-Traumatic Stress Disorder (PTSD) is associated with high levels of noradrenergic transmission that impairs PFC, but strengthens amygdala function, and is thus treated with agents that block alpha-1 or beta adrenoreceptors. 11.2. CIRCUIT MODELS 331

11.2.6 ADHD and Autistic spectrum disorder

The overlap of ADHD and Autistic spectrum disorder (ASD) symptoms has been investigated by Sinzig et al. (2009), who studied 83 children with ASD. They found that 53% of their sample fulﬁlled DSM-IV criteria for ADHD. A comparison of those children with and without ADHD, showed that those meeting an ADHD diagnosis were signiﬁcantly younger than 8 years, while almost 60 % of those not meeting criteria for ADHD were older than 13 years. This was thought to be consistent with the broader ADHD literature, showing a decline in hyperactivity and persistence of inattention with age. The authors concluded that the study revealed high phenotypical overlap between ASD and ADHD, with two subtypes revealed by factor analysis. These were a hyperactive-communication impaired and an inattentive stereotypic type. The association between language and overactivity symptoms is generally consistent with the importance of language development for childhood psychopathology.

11.2.7 Comment

A neuroanatomical understanding of reciprocal and non- reciprocal connections between modular circuits may provide a neuroanatomical basis for understanding previously puzzling aspects of comorbidity between internalising and externalising childhood syndromes, which have been regarded as orthogonal. In addition the switching mechanisms described for PTSD may also be involved in comorbidity, when very high arousal levels result in failure of homeostatic mechanisms at cortical levels, allowing direct transmission of fear-related stimuli from the amygdala to autonomic and motor centers. While the Stahl model has clear applications for the categorisation and treatment of fully developed adult psychopathologies, its relevance for childhood psychopathology depends on a more clear understanding of the development of cortical-subcortical circuits and the stages of development when they might be best inﬂuenced. Thus synaptic develop- 332 CHAPTER 11. COMORBIDITY ment and age-related variation, both in myelination and cortico- cortical, as well as cortical-subcortical circuit development, as well as pharmacological gating mechanisms at important synaptic sites, may all contribute to comorbidity. The paradox of comorbid internalising and externalising behavioural phenotypes is better understood in terms of excess communication between subcortical circuits, possibly as a result of insufﬁcient cortical inhibition, as well as excessive reliance on stimulus-based subcortical systems. A neuroanatomical understanding of reciprocal and non-reciprocal connections between modular circuits provides a neuroanatomical basis for understanding previously puzzling aspects of comorbidity between internalising and externalising childhood syndromes, which have been regarded as orthogonal (Levy, 2010). Chapter 12

Conclusions

The prevailing corticocentric emphasis for cognition and behaviour is shifted to a neural circuit understanding of childhood behavioural syndromes, in which dual biological systems integrate or fail to integrate automatic stimulus-based control, when this is advantageous, and higher-order control when this is needed for adaptive behaviour. Modular stimulus-controlled behaviours tend to be feedforward, while higher-order systems are serial, goal directed and subject to feedback control. This duality is reflected in the concept of stimulus-based vs higher-order control of behaviour, and provides a basis for not only the understanding of modular behavioural sequences, and possibly the frequent association of language deficits with childhood behavioural syndromes. Dennett’s metaphor of the brain functioning as a “serial virtual machine, implemented inefficiently on the parallel hardware that evolution has provided for us” provides a metaphor for the present thesis, where the relationship between modular parallel input circuits and their integration at subcortical and cortical levels is thought to be fundamental for normal development. The optimal development of cortical-sucortical circuits may be in- terrupted either by genetic, developmental or environmental processes, affecting the integration and processing of internal and environmental information and experience. It is argued that the phenomenology of a number of child-

333 334 CHAPTER 12. CONCLUSIONS hood syndromes can be understood in terms of the relationship between specialised parallel modular subsystems and higher cortical functions, in which processing of spatio-temporal and linguistic information is integrated to allow goal-orientation. While concepts of ‘global workspace’ and working memory are similar, working memory is more clearly intentional, while the global workspace concept is less defined, and appears to have a ‘google-like’ search capacity dependant on parallel re-entry processes. Working memory in humans is extended by language development. Cortico-thalamic-striato-cortical, (CTSC) loops, allowing information to be sent downstream, and providing feedback on how information was processed, are fundamental in regulating executive functions, problem solving, and cognitive tasks, and allocating attentional resources to task completion. These circuits might be deemed executive, attentional, emotional and motor modules. Learning in the cortical network appears to depend mainly on a local Hebbian learning rule guided by sequential functions from the basal ganglia and cerebellum. The ’circuit hypothesis’ implies that deficits in a particular circuit may occur at cortical or subcortical levels, but still give rise to similar symptomatology. This may account for the complexity of multiple or additive genetic effects on childhood syndromic behaviour, where multiple genes, neurotransmitters and/or receptors may be involved. A neuroanatomical understanding of reciprocal and interactive connections between modular circuits also promises to provide a basis for understanding previously puzzling aspects of comorbidity between internalising and externalising childhood syndromes, which were often regarded as orthogonal. The dual system hypothesis would predict that comorbid behaviours would more often reflect stimulus-bound behaviours, including repetitive, impulsive, distractible behaviour, as well as anxiety, as a result of inadequate gating at striatal and accumbens levels. A circuit from the dorsolateral prefrontal cortex (DLPFC) 335 projects to the rostral part of the caudate, within the striatal complex, then to the thalamus, and back to DLPFC. Loops with this structure are thought to regulate executive functions, problem solving, and cognitive tasks, such as representing and maintaining goals, and allocating attentional resources to various tasks. A second loop arising from the dorsal anterior cingulate gyrus (ACC) modulates selective attention and self-monitoring of performance, via the ventral area of the thalamus. A third CTSC loop arises from the ventral or subgenual part of the ACC, and projects to a part of the striatal complex, the accumbens, extending to the thalamus, and back to subgenual ACC regulates emotions, including depression and fear. A fourth CTSC loop outputs from the orbital frontal cortex (OFC) to the ventral part of the caudate nucleus in the striatal complex, to the thalamus and back to OFC, and is believed to regulate impulsivity and compulsivity. Finally, a fifth loop starts in the supplemental motor area of prefrontal motor cortex, and projects to the putamen in the lateral part of striatal complex, then to thalamus, and back to premotor cortex, and may modulate overactivity, psychomotor agitation and psychomotor retardation. The concept of parallel and integrative spiral circuits expands the notion of a global workspace from a primarily cortical prefrontal locus, to a circuit concept, which includes striato- thalamo-cortical and cortico-thalamo-striatal connections, which ultimately maintain a “golden mean” level of arousal at cortical effector loci. The concept of local overconnectivity and long- range underconnectivity in modular cortical-subcortical circuits, responsible for integration and control of behaviour provides a useful basis for the understanding a number of childhood syndromes, where domain-specific modular routines may become repetitive in the absence of domain-general executive integration at higher-order levels. The suggested distinction by MacLean between the limbic system and motor systems is replaced by a number of cortical- subcortical circuits which include dorsal and ventral cortico- 336 CHAPTER 12. CONCLUSIONS striatal-thalamic and cortico-amygdala-thalamic circuits. The ventral systems are thought to choose and mobilise ‘reward- guided’ behaviour, while the dorsal executive circuit (and its subcircuits) is implicated in planning and working memory. As indicated by Zahm the hypothalamus is only one of a number of autonomic and somatic motor effectors in the hypothalamus and brainstem. The hypothesis by Aronov and colleagues that distinct subcortical circuits for the production of infant behaviour may be a general feature of developmental learning in the vertebrate brain is important for the present thesis. The childhood and later adolescent development and integration of cortical/subcortical behavioural circuits, is fundamental to the present concept of cognitive development. Like song maturation, the mechanisms by which this integration is established are not well understood, but deficits in this process are believed fundamental to optimal maturation, and appear to involve integration between reflexive and sequential systems. From a developmental point of view, the achievement of representational capacity is a gradual process, involving language and cognitive development. Also, by definition, consciousness is not fully developed in early childhood, and deficits in early development will have implications for the full development of higher order cognition and consciousness of self. It is likely that the language deficits described in a number of childhood syndromes such as ADHD and Conduct Disor- der, reflect deficits in higher-order language capacity, with undue reliance on more automatic repetetive language forms. It is suggested that language disorders may be more fundamental in childhood symptomatology than recognised by current classification systems. These deficits are even more evident in autistic syndromes, where language, when present, is stilted and repetitive, as would be expected for a modular lexical language system. A basic deficit (possibly genetic) in categorical phoneme perception, or a later failure of transition to the native language second-order representational system may predispose the child to 337 further deficits of monitoring systems related to language and attention. The inability to “sound-out” may be a marker for a more general delay in the early development of feedback monitored language processes. Working memory is described as the ability to represent information, no longer in the environment, through recurrent excitation of pyramidal cells with shared stimulus properties. Adequate PFC functioning appears critical for not only mature reasoning, but also involves behavioural functions, including inhibition of task-irrelevant behaviours, processing of affect, motivation, and reward attainment by virtue of connections with wide-ranging cortical centers. It can be argued that the process of development is closely dependant on adequate PFC development, but cortical function also depends on optimal gating functions at subcortical levels. Where working memory is impaired or fails to adequately develop or be maintained, the control of behaviour is captured or vulnerable to non-goal-oriented environmental stimuli, or to repetition of localised modular behaviours, with resultant distractibility or rigidity. Both the LeDoux and Zahm models of anxiety describe reciprocal cortical/amygdala connections, providing an underlying circuit basis for both cognitive behavioural or “top-down” interventions, as well as pharmacological or “bottom-up” interventions. An immediate fast visceral response to fear stimuli is necessary for a survival response, but slower, more detailed processing is necessary, in order to consciously perceive and process the stimulus. The basal ganglia, limbic striatal, thalamocortical circuits, with a direct pathway from the cortex to striatum to internal segment of the globus pallidus, to thalamus and back to the cortex, and an indirect inhibitory pathway, via the external segment of the globus pallidus, which rejoins the common pathway to the thalamus and cortex, provides a fundamental structure. Disequilibrium in this circuit gives rise to repetitive behaviour. Imbalance between inhibitory and disinhibitory influences at thalamic level will result in a failure of appropriate decay of activity at cortical levels, ex- 338 CHAPTER 12. CONCLUSIONS perienced as repetitive thoughts and concerns. Obsessive Compulsive Disorder can be characterised as a primary circuit disorder, in which anxiety inhibits feedback systems, so that a ‘habit circuit’ is maintained. The phenomena of repetition of pathological behaviours, which has been shown to occur in autism, OCD, Tourette’s Disorder and Post-traumatic stress disorder appears to derive from cycles of modular behaviours, where higher-order cortical control fails to be established. The developing child’s behaviour is then subject, either to the control of random environmental stimuli, as in ADHD, or repetitive local modular circuits as in autism and Tourette’s Disorder, over-rigidity and compulsions as in OCD, or aberrant arousal circuits as in PTSD. ADHD symptomatology may represent the opposite of a recurrent syndrome. That is, distractible behaviour, which is controlled by environmental stimuli, suggests an inability to maintain a sustained or recurrent behaviour pattern, leaving the child (or adult) vulnerable to random sensory stimuli, which will transiently control behaviour, but control will be rapidly replaced by other salient stimuli. The cortical/subcortical models postulated by Koziol and Budding (2009); Balleine and O’Doherty (2010) suggest mechanisms, whereby appropriate modulation of task- related (higher-order) processing and stimulus-bound (default) processing may be achieved. This modulation may require both striatal and cerebellar inputs to achieve appropriate timing of prefrontal activity during task-related activity. The incorporation of stimulus-bound reactivity into the ADHD model helps to explain the overreactivity of young and of ADHD children to external stimuli, and distracting internal concerns, in terms of circuit deficits between task-related and default networks. In PTSD, the stress response system demonstrates the principle that when homeostasis of an iterative system is disrupted (in this case by an exaggerated amygdala response and diminished PFC control), a repetitive behavioural syndrome (which can go on for many years in terms of flashbacks and anxiety) results, possibly representing a continuing attempt to restore homeostasis. 339

While mechanisms of mirror processing are not fully understood, the process appears to be a rapid analogue rather than sequential process. The capacity of primates to reproduce and integrate observed emotional and behavioral states appears central to successful social functioning. The possibility of a “motor theory of empathy”, involving mirror circuits, similar to the motor theory of speech, may also be useful in understanding language and affective deficits observed in psychopathy. Mirror systems may reflect stimulus-based social skills that are necessary for successful routine interactions. These interact with higher-order social processing, allowing us to think about what to say or do, to consider the thoughts, ideas and feelings of others, to manipulate or comply with situations, and to reflect upon ourselves. Language is thought to be important in this process. The relevance of the ‘problem of consciousness’ to the present discussion is the question of whether consciousness, and in particular, self-consciousness is phenomenogically different in childhood. Certainly, at a neuroanatomical level, it is clear that the circuits involved in the development and maintenance of consciousness undergo extensive changes during childhood and adolescence, and into adulthood. While the significance of these developments is not well understood, phenomena such as childhood amnesia, as well as autistic and other childhood syndromes, point to the importance of neural circuit development and consciousness. For example, the period of early childhood amnesia suggests that the establishment of hippocampal and other cortical connections are necessary for a lasting sense of self, while adolescent development requires prefrontal circuit maturation. The integration of cognitive and emotional functions is of central importance in determining behavioural outcome. Here, communication between basal forebrain systems both subcortically and cortically is important. At the cortical level, language is thought to play a central role in allowing verbal labelling and integration of emotion, as demonstrated by the association of lan- 340 CHAPTER 12. CONCLUSIONS guage deficits with a number of behavioural syndromes (including autism, ADHD, and oppositional/conduct problems), where emotional and cognitive processes fail to be integrated. Reverse replay in the hippocampus might have a critical role in support of learning in hippocampus-dependant tasks.When awake, reverse replay occurs in situ, allowing immediately preceding events in precise temporal relation to a current anchoring event, and so may be an integral mechanism for learning about recent events. The conversion of single experiences into multiple reverse events in awake replay, may provide a model of how animals learn from experience. The paradox of comorbid internalising and externalising behavioural phenotypes is better understood in terms of excess communication between subcortical circuits, possibly as a result of insufficient cortical inhibition, as well as excessive reliance on stimulus-based subcortical systems. A neuroanatomical understanding of reciprocal and non-reciprocal connections between modular circuits provides a neuroanatomical basis for understanding previously puzzling aspects of comorbidity between internalising and externalising childhood syndromes, which have been regarded as orthogonal. The thesis has implications for future nosology in child psychiatry. Suggestions that childhood syndromes should be included with adult diagnoses may fail to recognise the special developmental aspects of childhood. Should pre-linguistic, linguistic, and prefrontal connectivity stages be acknowledged as having characteristic features and deficits? Cortical/subcortical models may provide a direction for candidate gene investigations where the same transmitter may have differing, but complementary functions at cortical vs subcortical receptor sites, such as the cortical D1 receptor, and the basal ganglia DAT receptor. While structure does not necessarily determine function, particularly in complex biological systems, a better understanding of structure, provides insights into relationships between early 341 development, cognition and behaviour, and in reverse provides insights into questions of consciousnes of self, and behavioural diagnoses, including comorbidity. 342 CHAPTER 12. CONCLUSIONS Bibliography

Achenbach, T. M., Howell, C., McConaughy, S., and Stanger, C. (1998). Six-year predictors of problems in a national sample, IV: young adult signs of disturbance. Journal of the American Academy of Child and Adolescent Psychiatry, 37:718–727. Adleman, N., Menon, V., Blasey, C. M., White, C., Warofsky, I., Glover, G., and Reiss, A. (2002). A developmental fMRI study of the Stroop color-word task. Neuroimage, 16:61–75. Adolphs, R. (2001). The neurobiology of social cognition. Current Opinion in Neurobiology, 11:231–239. Adolphs, R. (2002). Neural systems for recognizing emotion. Cur- rent Opinion in Neurobiology, 12:169–177. Adolphs, R. (2003). Cognitive neuroscience of human social behaviour. Nature Reviews, Neuroscience, 4:165–178. Adolphs, R., Damasio, H., Tranel, D., and Damasio, A. (1996). Cortical systems for the recognition of emotion in facial expressions. Journal of Neuroscience, 16:7678–7687. Adret, P. (2004). In search of the song template, pages 303–324. The New York Academy of Sciences, New York, New York. Alexander, G. E. and Crutcher, M. D. (1990). Functional architecture of basal ganglia disorders. Trends in Neurosciences, 13:266–271. Alexander, G. E., Delong, M. R., and Strick, P. L. (1986). Paral- lel organization of functionally segregated circuits linking basal ganglia and cortex. Annual Review of Neuroscience, 9:357–381. Alheid, G. F. and Heimer, L. (1988). New perspectives in basal forebrain organization of special relevance for neuropsychiatric disorders: the striatopallidal, amygdaloid, and corticopetal components of substantia innominata. Neuroscience, 27:1–39.

343 344 BIBLIOGRAPHY

Allen, A., King, A., and Hollender, E. (2003). Obsessive- compulsive spectrum disorders. Dialogues in Clinical Neuro- science: Anxiety 11, 5(3):259–271. Allen, G. and Courchesne, E. (2003). Differential effects of developmental cerebellar abnormality on cognitive and motor functions in the cerebellum: an fmri study of autism. American Journal of Psychiatry, 160:262–273. Alter, M. D. and Hen, R. (2009). Stress-induced and fear circuitry disorders: Reﬁning the research agenda for DSM-V, chapter Serotonin, sensitive periods, and anxiety, pages 159–173. American Psychiatric Association, Arlington, Virginia. Angold, A., Costello, E. J., and Erkanli, A. (1999). Comorbidity. Journal of Child Psychology and Psychiatry, 40:57–87. APA (1994). Diagnostic and Statistical Manual of Mental Disor- ders. Washington, DC, 4th edition. Aragona, B. J., Liu, Y., Cameron, A., Perlman, G., and Wang, Z. (2003a). Opposite modulation of of social attachment by d1- and d2-type dopamine receptor activation in nucleus accumbens shell. Hormones and Behavior, 44:37. Aragona, B. J., Liu, Y., Curtis, T. J., Stephan, F. K., and Wang, Z. X. (2003b). A critical role for nucleus accumbens dopamine in partner preference formation of male prairie voles. Journal of Neuroscience, 23:3483–3490. Arbib, M. (1989). Computers, Brains and Minds: Essays in Cog- nitive Science, volume 7 of Australasian Studies in History and Philosophy of Science, chapter Modularity, schemas and neurons: A critique of Fodor., pages 193–219. Kluwer Academic Publishers, Dordecht, Boston and London. Arbib, M. A. (2005). From monkey-like action recognition to human language: An evolutionary framework for neurolinguistics. Behavioral and Brain Sciences, 28:105–167. Ardila, A. (2009). Foundations in EvolutionaryCognitive Neuro- science, chapter Origins of the language: Correlation between brain evolution and language development, pages 153–174. Cambrige University Press, Cambridge and New York. Arnett, J. (1992). Reckless behavior in adolescence; a developmental perspective. Developmental Review, 12:339–373. BIBLIOGRAPHY 345

Arnsten, A. F. (2006). Fundamentals of attention- deficit/hyperactivity disorder: circuits and pathways. Journal of Clinical Psychiatry, 67(Suppl 8):7–12. Arnsten, A. F., Matthew, R., Urbriani, R., Taylor, J. R., and Bao- Ming, L. (1999). The alpha 1 noradrenergic receptor stimulation impairs prefrontal cortical function. Biological Psychiatry, 45:26–31. Arnsten, A. F. T. (1997). Catecholamine regulation of the prefrontal cortex. Journal of Psychopharmacology, 11:151–162. Arnsten, A. F. T. (2000). Stress impairs prefrontal cortical function in rats and monkeys: role of dopamine d1 and norepinephrine alpha-1 receptor mechanisms. Progress in Brain Research, 126:183–192. Arnsten, A. F. T. (2007). Catecholamine and second messenger influences on prefrontal cortical networks of ”representational knowledge”: A rational bridge between genetics and the symptoms of mental illness. Cerebral Cortex, 17:Suppl 1. Arnsten, A. F. T., Cai, J. X., and Goldman-Rakic, P. S. (1998). The alpha-2 adrenergic agonist guanfacine improves memory in aged monkeys without sedative or hypotensive side effects. Journal of Neuroscience, 8:4287– 4298. Arnsten, A. F. T. and Dudley, A. G. (2005). Methylphenidate improves prefrontal cortical cognitive function through al- pha2 adrenoceptor and dopamined1 receptor actions: Relevance to therapeutic effects in attention deficit hyperactivity disorder. Behavioral and Brain Functions, doi:10.1186/1744-9081- 1-2:2005, 1:2. Arnsten, A. F. T. and Li, B.-M. (2005). Neurobiology of executive functions: Catecholamine influences on prefrontal cortical functions. Biol Psychiatry, 57:1377–1384. Arnsten, A. F. T., Steere, J., and Hunt, R. (1996). The contribution of alpha-2 noradrenergic mechanisms to prefrontal cortical cognitive function: potential significance to attention deficit hyperactivity disorder. Archives of General Psychiatry, 53:448–455. Aronov, D., Andalman, A. S., and Fee, M. S. (2008). A specialized forebrain circuit for vocal babbling in the juvenile songbird. Sci- ence, 320(5876):630–634. 346 BIBLIOGRAPHY

Astington, J., Harris, P., and Olson, D., editors (1988). Developing Theories of Mind. Cambridge University Press, Cambridge and New York. Astington, J. W. and Jenkins, J. M. (1999). A longitudinal study of the relation between language and theory-of-mind development. Developmental Psychology, 35:1311–1320. Aston-Jones, G., Rajkowski, J., and Cohen, J. (1999). Role of locus coeruleus in attention and behavioural ﬂexibility. Biological Psychiatry, 46:1309–1320. Au-Young, S. M. W., Shen, H., and Yang, C. R. (1999). Medial prefrontal cortical output neurons to the ventral tegmental area (vta) and their responses to burst-patterned stimulation of the vta: Neuroanatomical and in vivo electrophysiological analyses. Synapse, 34:245–255. Augustine, J. R. (1996). Circuitry and functional aspects of the insular lobe in primates including humans. Brain Research Re- views of Brain Research, 22:229–244. Avery, R., Franowicz, J. S., Studholme, C., van Dyck, C. H., and Arnsten, A. F. T. (2000). The alpha-2a-adenoceptor agonist, guanfacine, increases regional cerebral blood ﬂow in dorsolateral prefrontal cortex of monkeys performing a spatial working memory task. Neuropsychopharmacology, 23:240–249. Baars, B. J. (1989). A cognitive theory of consciousness. Cambridge University Press, Cambridge. Bachevalier, J. (1991). Advances in Neuropsychiatry and Psy- chopharmacology: Volume 1, Schizophrenia Research., chapter An animal model for childhood autism: memory loss and socioemotional disturbances following neonatal damage to the limbic system in monkeys. Raven Press, New York. Baddeley, A. (2000). The episodic buffer: a new component of working memory? Trends in Cognitive Science, 4:417–423. Baddeley, A. (2001). Is working memory still working? American Psychologist, 56:851–864. Baddeley, A. and Hitch, G. (1974). The Psychology of Learning and Motivation, volume 8, chapter Working memory, pages 47– 89. Academic Press, New York. BIBLIOGRAPHY 347

Baddeley, A. D. (1986). Working Memory. Oxford University Press, Oxford. Ballaban-Gil, K. and Tuchman, R. (2000). Epilepsy and epilepti- form eeg: association with autism and language disorders. Men- tal Retardation and Developmental Disability Research Review, 6:300–308. Balleine, B. (2001). Handbook of contemporary learning theories, chapter Incentive processes in instrumental conditionng, pages 307–366. LEA, Hillsdale, NJ. Balleine, B., Delgado, M., and Hikosaka, O. (2007). The role of the dorsal striatum in reward and decision-making. The Journal of Neuroscience, 27:8161–8165. Balleine, B. and Kilcross, S. (2006). Parallel incentive processing processing: an integrated view of amygdala function. Trends in Neuroscience, 29:272–279. Balleine, B. and O’Doherty, J. (2010). Human and rodent ho- mologies in action control: Corticostriatal determinants of goal- directed and habitual action. Neuropsychopharmacology Re- views, 35:48–69. Bar-Gad, I., Morris, G., and Bergman, H. (2003). Information processing, dimensionality reduction and reinforcement learning in the basal ganglia. Progress in Neurobiology, 71:439–473. Barbas, H. (1992). Advances in Neurology, chapter Architecture and cortical connections of the prefrontal cortex in the rhesus monkey., pages 91–115. Raven Press, New York. Barkley, R. (1990). Attention Deﬁcit Hyperactivity Disorder: a Handbook for Diagnosis and Treatment. Guilford Press, New York. Barkley, R. A. (1997). ADHD and the Nature of Self-Control. The Guilford Press, New York and London. Barkow, J. Comides, L. and Tooby, J. (1992). The Adapted Mind: Evolutionary Psychology and the Generation of Culture. Oxford University Press, New York. Baron-Cohen, S. (1987). Autism and symbolic play. British Jour- nal of Developmental Pschology, 5:139–148. 348 BIBLIOGRAPHY

Baron-Cohen, S. (1988). Social and pragmatic deﬁcits in autism: Cognitive or affective? Journal of Autism and Developmental Disorders, 18:379–402.

Baron-Cohen, S. (1995). Mindblindness: An Essay on Autism and Theory of Mind. MIT Press, Cambridge, Massachusetts and London, England.

Baron-Cohen, S., Leslie, A. M., and Frith, U. (1985). Does the autistic child have a theory of mind? Cognition, 21:37–46.

Baron-Cohen, S., Leslie, A. M., and Frith, U. (1986). Mechanical, behavioural and intentional understanding of picture stories in autistic children. British Journal of Developmental Psychology, 21:113–125.

Baron-Cohen, S., Ring, H., Wheelwright, S., Bullmore, E., Bram- mer, M., Simons, A., and Williams, S. (1999). Social intelligence in the normal and autistic brain: An fmri study. European Jour- nal of Neuroscience, 11:1891–1898.

Baron-Cohen, S., Ring, H. A., Bullmore, E. T., Wheelwright, S., Ashwin, C., and Williams, C. R. (2000a). The amygdala theory of autism. Neuroscience and Biobehavioural Reviews, 24:355– 364.

Baron-Cohen, S., Tager-Flusberg, H., and Cohen, D. (2000b). Understanding Other Minds: Perspectives from Developmental Cognitive Neuroscience. Oxford University Press, 2nd edition.

Baron-Cohen, S., Tager-Flusman, H., and D., C., editors (1993). Minds: Perspectives from Autism. Oxford University Press, Ox- ford and New York.

Baumrind, D. (1971). Current patterns of parental authority. De- velopmental Psychology Monographs, 94:132–142.

Baxter, L. R. (1992). Neuroimaging studies of obsessive compulsive disorder. Psychiatric Clinics of North America, 15:871–884.

Baxter, L. R., Schwartz, J. M., Guze, B. H., Bergman, K., and Szuba, M. P. (1990). Neuroimaging in obsessive-compulsive disorder: seeking the mediating neuroanatomy. In Jenike et al, eds, pages 167–188. Year Book Medical Publishers, 2nd edition. BIBLIOGRAPHY 349

Bechtel, W. (2008). Mental Mechanisms: Philosophical Perspec- tives on Cognitive Neuroscience. Taylor and Francis Group, New York, NY. Belmonte, M. K. (2000). Abnormal attention in autism shown by steady-state visual evoked potentials. Autism, 4:269–285. Belmonte, M. K., Allen, G., Beckel-Mitchener, A., Boulanger, L. M., Carper, R. A., and Webb, S. J. (2004a). Autism and abnormal development of brain connectivity. The Journal of Neu- roscience, 24(42):9228–9231. Belmonte, M. K. and Baron-Cohen, S. (2004). Normal sibs of children with autism share negative frontal but not positive sensory abnormalities: preliminary evidence from fmri duringpro- cessing of visual distractors. Society of Neuroscience Abstracts, 30:582.10. Belmonte, M. K., Cook, E. H., Anderson, G. M., Rubenstein, J. L. R., Greenough, W. T., Beckel-Mitchener, A., Courchesne, E., Boulanger, L. M., Powel, S. B., Levitt, P. R., Perry, E. K., Jiang, Y. H., DeLorey, T. M., and Tierney, E. (2004b). Autism as a disorder of neural information processing: directions for research and targets for therapy. Molecular Psychiatry, 9:646–663. Belmonte, M. K. and Yurgelun-Todd, D. A. (2003). Functional anatomy of impaired selective attention and compensatory processing in autism. Cognition and Brain Research, 17:651–664. Bennet, L. J. (1990). Modularity of mind revisited. British Jour- nal for the Philosophy of Science, 41:429–436. Bickerton, D. (2002). Language in the modular mind? it’s a no-brainer. Behavioral and Brain Sciences, Commen- tary/Carruthers, 25:677–678. Biederman, J., Wozniak, J., Kiely, K.and Ablon, S., Faraone, S., Mick, E., Mundy, E., and Kraus, I. (1995). Cbcl clinical scales discriminate prepubertal children with structured interview- derived diagnosis of mania from those with adhd. Journal of the American Academy of Child and Adolescent Psychiatry, 34:464– 471. Bilder, R. M., Volavka, J., Lachman, H. M., and Grace, A. A. (2004). The catechol-o-methyltransferase polymorphism: Rela- tions to the tonic-phasic dopamine hypothesis and neuropsychiatric phenotypes. Neuropsychopharmacology, 29:1943–1961. 350 BIBLIOGRAPHY

Blackburn, R. (1988). Adult Abnormal Psychology, chapter Psy- chopathy and personality disorder, pages 218–244. Churchill Livingstone, Edinburgh. Blair, J. and Frith, U. (2000). Neurocognitive explanations of the antisocial personality disorders. Criminal Behaviour and Men- tal Health, 10:S66–S81. Blair, R. J. R. (1995). A cognitive developmental approach to morality: Investigating the psychopath. Cognition, 57:1–29. Blair, R. J. R. (2001). Neurocognitive models of aggression, the antisocial personality disorders and psychopathy. Journal of Neurology, Neurosurgery and Psychiatry, 71:727–731. Blair, R. J. R. (2003). Neurobiological basis of psychopathy. The British Journal of Psychiatry, 182:5–7. Blair, R. J. R. (2006). The emergence of psychopathy: Implica- tions for the neuropsychological approach to developmental disorders. Cognition, 101:414–442. Blair, R. J. R., Colledge, E., Murray, L., and Mitchell, D. G. V. (2001). A selective impairment in the processing of sad and fearful expressions in children with psychopathic tendencies. Journal of Abnormal Child Psychology, 29:491–498. Blair, R. J. R., Jones, L., and Clark, F. e. a. (1997). The psychopathic individual: A lack of responsiveness to distress cues? Psychophysiology, 34:192–198. Blair, R. J. R., Morris, J. S., Frith, C. D., Perrett, D. I., and Dolan, R. J. (1999). Dissociable neural responses to facial expressions of sadness and anger. Brain, 122:883–893. Bland, B. H. (2009). Information Processing by Neuronal Pop- ulations, chapter Anatomical, physiological, and pharmacological properties underlying hippocampal sensorimotor integration, pages 283–325. Cambridge University Press, Cambridge and New York. Boder, E. and Jarrico, S. (1982). The Boder test of Reading- Spelling Patterns. Grune & Stratton, New York. Bonda, E., Petrides, M., Ostry, D., and Evans, A. (1996). Speciﬁc involvement of human parietal systems and the amygdala in BIBLIOGRAPHY 351

the perception of biological motion. Journal of Neuroscience, 16:3737–3744. Booth, J. R., Burman, D. D., Meyer, J. R., Gitelman, D. R., Par- rish, T. B., and Mesulam, M. M. (2003). Relation between brain activation and lexical performance. Human Brain Mapping, 19:155–169. Botterill, G. and Carruthers, P. (1999). The Philosophy of Psychol- ogy. Cambridge University Press, Cambridge. Botvinick, M. M., Braver, T. S., Barch, D. M., Carter, C. S., and Cohen, J. D. (2001). Conﬂict monitoring and cognitive control. Psychological Review, 108:624–652. Boucher, J. (1981). Memory for recent events in autistic children. Journal of Autism and Developmental Disorders, 11:239–302. Boucher, J. (1989). The theory of mind hypothesis of autism: Ex- planation, evidence and assessment. Briitish Journal of Disor- ders of Communication, 24:181–198. Boucher, J. (2001). Time and Memory: Issues in Philosophy and Psychology, chapter ”Lost in a sea of time”: timing-parsing and autism, pages 111–135. Clarendon Press, Oxford. Boyer, P. and Lienard, P. (2006). Why ritualized behaviour? pre- caution systems and action parsing in developmental, pathological and cultural rituals. Behavioral and Brain Sciences, 6:595– 613. Bremer, J. F. (1992). Serious juvenile sex offenders. Psychiatric Annals, 22:320–325. Bremner, J. D. (2006). Traumatic stress: effects on the brain. Di- alogues in Clinical Neuroscience, 8(4):445–461. Bremner, J. D., Randall, P., Vermetten, E., Staib, L., Bronen, R. A., Mazure, C., Bronen, L., Mazure, C., Capelli, S., McCarthy, G., Innis, R. B., and Charney, D. S. (1997). Magnetic resonance imaging-based measurement of hippocampal volume in PTSD related to childhood physical and sexual abuse: a preliminary report. Biological Psychiatry, 41:25–32. Broca, P. (1878). Anatomie comparee des circonvolutions cere- brales. Revue deAnthopologie, 1:385–498. 352 BIBLIOGRAPHY

Brock, J., Brown, C. C., Boucher, J., and Rippon, G. (2002). The temporal binding deficit hypothesis of autism. Development and Psychopathology, 14:209–224. Brothers, L. (1990). The social brain: A project for integrating primate behaviour and neurophysiology in a new domain. Cog- nitive Science, 1:27–51. Brown, J. (1977). Mind, Brain and Consciousness: The Neuropsy- chology of Cognition. Academic Press, New York, San Francisco, London. Brown, T. A. and Barlow, D. H. (2005). Dimensional versus categorical classification of mental disorders in the fifth edition of the diagnostic and statistical manual of mental disorders and beyond: Comment on the special section. Journal of Abnormal Psychology, 4:551–556. Buccino, G., Binkofski, F., Heine, H., Fink, G. R., Fadiga, L., Fo- gassi, L., Gallese, V., and Seitz, R. J. (2001). Action observation activates premotor and parietal areas in a somatotopic manner. European Journal of Neuroscience, 13:400–404. Buccino, G., Lui, F., Canessa, N., Patteri, I., Lagravinese, G., Be- nuzzi, F., Porro, C. A., and Rizzolatti, G. (2004). Neural cir- cuts involved in the recognition of actions performed by non- conspecifics: an fMR istudy. Journal of Cognitive Neuroscience, 16:114–126. Bunge, S., Dudukovic, N., Thomason, M., Vaidya, C., and Gabrieli, J. (2002). Immature frontal lobe contributions to cognitive control in children: evidence from fMRI. Neuron, 33:301–311. Bush, G. (2010). Attention-deficit/hyperactivity disorder and attention networks. Neuropsychopharmacology Reviews, 35:278– 300. Bush, G., Luu, P., and Posner, M. I. (2000). Cognitive and emotional influences in anterior cingulate cortex. Trends in Cogni- tive Sciences, 4:215–222. Buzsaki, G. (2006). Rhythms of the Brain. Oxford University Press, New York. Byrne, R. W. and Whiten, A. (1990). Tactical deception in primates. Primate Rep, 27:1. BIBLIOGRAPHY 353

Cacioppo, J., Visser, P., and Pickett, J. e., editors (2006). Social Neuroscience: People Thinking about People. University of Cal- ifornia, Los Angeles. Calvin, W. H. (2004). A Brief History of The Mind. Oxford Uni- versity Press, Oxford and New York. Cao, F., Bitan, T., Chou, T., Burman, D. D., and Booth, J. R. (2006). Deﬁcient orthographic and phonological representations in children with dyslexia revealed by brain activation patterns. Jour- nal of Child Psychology and Psychiatry and Allied Disciplines, 47:1041–1050. Carlsson, A. (2001). A paradigm shift in brain research. Science, 294:1021–1024. Carr, L., Iacoboni, M., Dubeau, M. C., Mazziotti, J. C., and Lenzi, G. L. (2003). Neural mechanisms of empathy in humans: A relay from neural systems for imitation to limbic areas. Proceed- ings of the National Academy of Sciences of the United States of America, 100:5497–5502. Carruthers, P. (2002a). The cognitive functions of language. Be- havioral and Brain Sciences, 25:657–726. Carruthers, P. (2002b). Mental modularity and distinctively human thought, response/carruthers. Behavioral and Brain Sci- ences, 25:708. Carruthers, P. (2005). The Innate Mind: Structure and Contents, chapter Distinctively human thinking: Modular precursors and components. In, (eds) Carruthers, P. and Laurence, S. and Stich, S. (2005), pages 69–88. Oxford University Press, New York. Carruthers, P. and Chamberlain, A. (2000). In, Evolution and the Human Mind: Modularity, Language and Meta-Cognition, chapter Introduction., pages 1–12. Cambridge University Press, Cambridge, UK. Carruthers, P. and Smith, P., editors (1996). Theories of Theo- ries of Mind. Cambridge University Press, Cambridge and New York. Caspi, A., Langley, K., Milne, B., Mofﬁtt, T., O’Donovan, M., Owen, M., Tomas, M., Poulton, R., Rutter, M., Taylor, A., Williams, B., and Thapar, A. (2008). A replicated molecular genetic basis for subtyping antisocial behavior in children 354 BIBLIOGRAPHY

with attention-deﬁcit/hyperactivity disorder. Archives of Gen- eral Psychiatry, 65:203–210.

Caspi, A., Sugden, K., Mofﬁtt, T. E., Taylor, A., Craig, I. W., Har- rington, H., McClay, J., Mill, J., Martin, J., Braithwaite, A., and Poulton, R. (2003). Inﬂuence of life stress on depression: mod- eration by a polymorphism in the 5-htt gene. Science, 301:386– 389.

Castellanos, F., Fine, E., Kaysen, D., Marsh, W., Rapoport, J., and Hallett, M. (1996). Sensorimotor gating in boys with Tourette’s syndrome and adhd: Preliminary results. Biological Psychiatry, 39:33–41.

Castellanos, F., Sonuga-Barke, E., Milham, M., and Tannock, R. (2006). Characterizing cognition in ADHD : beyond executive dysfunction. Trends in Cognitive Neuroscience, 10:117–123.

Castellanos, F. and Tannock, R. (2002). Neuroscience of attention- deﬁcit/hyperactivity disorder: The search for endophenotypes. Nature Reviews Neuroscience, 3:617–628.

Castellanos, F. X. (1997). Toward a pathophysiology of attention- deﬁcit/hyperactivity disorder. Clinical Paediatrics, 36:381–393.

Castellanos, F. X., Lee, P. P., Sharp, W., Jeffries, N. O., Greenstein, D. K., Clasen, L. S., Blumenthal, J. D., James, R. S., Ebens, B. A., Walter, J. M., Zijdenbos, A., Evans, A. C., Gied, J. N., and Rapoport, J. L. (2002). Developmental trajectories of brain volume abnormalities in children and adolescents with attention- deﬁcit/hyperactivity disorder. JAMA, 288:1740–1748.

Castellanos, F. X., Margulies, D. S., Kelly, C., Uddin, L. Q., Ghaf- fari, M., Kisrsch, A., Shaw, D., Shehzad, Z., DiMartino, A., Biswal, B., Sonuga-Barke, E. J. S., Rotrosen, J., and Milham, M. P. (2008). Cingulate-precuneus interactions: a new locus of dysfunction in adult attention-deﬁcit/hyperactivity disorder. Biological Psychiatry, 63:332–337.

Cavanna, A. E. (2007). The precuneus and consciousness. CNS Spectrums, 12:545–556.

Changeux, J. P. (2005). From Monkey Brain to Human Brain, chapter Genes, brains and culture: From human to monkey. MIT Press, Cambridge. BIBLIOGRAPHY 355

Charney, D. (2004). Psychobiological mechanisms of resilience and vulnerability: implications for the successful adaption to extreme stress. American Journal of Psychiatry, 161:195–216. Cheng, K. (1986). A purely geometric model in the rat’s spatial representation. Cognition, 23:149–178. Cheng, M.-F. and Durand, S. E. (2004). Birdsong, chapter Song and the Limbic Brain: A new function for the bird’s own song., pages 611–627. The New York Academy of Sciences, New York, New York. Chomsky, N. (1988). Language and Problems of Knowledge. MIT Press., Cambridge, MA. Chomsky, N. (1995). The Minimalist Program. MIT Press. Churchland, P. (2002a). Daniel Dennett, chapter Catching consciousness in a recurrent net, pages 64–80. Cambridge Univer- sity Press. Churchland, P. (2002b). Daniel Dennett., chapter Catching consciousness in a recurrent net. Cambridge University Press, Cambridge and New York. Churchland, P. S. and Sejnowski, T. J. (1999). The Computational Brain. The MIT Press, Cambridge, Massachusetts. Clark, P. and Rutter, M. (1981). Autistic children’s response to structure and to interpersonal demands. Journal of Autism and Developmental Disorders, 11:201–217. Cohen, J., Braver, T., and O’Reilly, C. (1998). The Prefrontal Cortex: Executive and Cognitive Functions, chapter A computational approach to prefrontal cortex, cognitive control, and schizophrenia: recent developments and current challenges, pages 195–220. Oxford University Press, Oxford and New York. Cohen, J. D., Braver, T. S., and Brown, J. W. (2002). Computa- tional perspectives on dopamine function in prefrontal cortex. Current Opinion in Neurobiology, 12:223–229. Cohen, J. D., Bravers, T. S., and O’Reilly, C. (1996). A computational approach to prefrontal cortex, and schizophrenia: recent developments and current challenges. Philosophical Transac- tions of of Royal Society of London, B, 351:1515–1527. 356 BIBLIOGRAPHY

Coltheart, M. (1999). Modularity and cognition. Trends in Cogni- tive Science, 3:115–120. Coltheart, M., Posner, M., and Marin, O., editors (1985). Attention and Performance XI. Erlbaum, New York. Coltheart, M. and Rastle, K. (1994). Serial processing in reading aloud: Evidence for dual-route models of reading. Journal of Experimental Psychology: Human Perception and Performance, 20:1197–1211. Conboy, B., Sommerville, J., and Kuhl, P. (2008). Cognitive control factors in speech perception at 11 months. Developmental Psychology, 44:1505–1512. Conway, C. M. and Christiansen, M. (2001). Sequential learning in non-human primates. Trends in Cognitive Sciences, 5:539– 546. Cornil, C. A., Balthazart, J., Motte, P., Massote, L., and Seutin, V. (2002). Dopamine activates noradrenergic receptors in the preoptic area. Journal of Neuroscience, 22:9320–9330. Courchesne, E., Carper, R. A., and Akshoomoff, N. A. (2003). Ev- idence of brain over-growth in the ﬁrst year of life in autism. JAMA, 290:337–344. Creamer, M., Bell, R., and Failla, S. (2003). Psychometric properties of the impact of event scale-revised. Behavior Research and Therapy, 41:1489–1496. Crick, F. C. and Koch, C. (2005). What is the function of the claustrum? Philosophical Transctions of The Royal Society B, 360:1271–1279. Critchley, H. D., Daly, E. M., Bullmore, E. T., Williams, S. C. R., Van Amelsvoort, T., Robertson, D. M., Rowe, A., Phillips, M., Mc Alonan, G., Howlin, P., and Murphy, D. G. M. (2000). The functional neuroanatomy of social behaviour: changes in cerebral blood ﬂow when people with autistic disorder process facial expressions. Brain, 123:2203–2212. Cummings, J. L. (1993). Frontal subcortical circuits and human behavior. Archives of Neurology, 50:873–880. Dadds, M., Schwartz, S., Adams, T., and Rose, S. (1988). The effects of social context and verbal skill on the stereotypic and BIBLIOGRAPHY 357

task-involved behaviour of autistic children. Journal of Child Psychology and Psychiatry, 29:669–76.

Dadds, M. R., El Masry, Y., Wimalaweera, S., and Guastella, A. (2008). Reduced eye gaze explains "fear blindness" in childhood psychopathic traits. Journal of the American Academy of Child and Adolescent Psychiatry, 47:455–463.

Damasio, A. R. (1994). Descartes error: emotion, reason and the human brain. Grosset/Putnam, New York.

Damasio, A. R. (2000). A neurobiology for consciousness., pages 111–120. In Metzinger (2000).

Dapretto, L., Davies, M. S., Pfeifer, J. H., Scott, A., Sigman, M., Bookheime, S. Y., and Iacoboni, M. (2006). Understanding emotions in others: mirror neuron dysfunction in children with autism spectrum disorders. Nature Neuroscience, 9:28–30.

Davidson, R. J., Putnam, K. M., and Larson, C. L. (2000). Dys- function in the neural circuitry of emotion regulation - a possible prelude to violence. Science, 289:591–594.

Davis, M. and Whalen, P. J. (2001). The amygdala: vigilance and emotion. Molecular Psychiatry, 286:13–34.

Dawson, G. and Fischer, K., editors (1994). Human Behavior and the Developing Brain. Guilford Press, New York. de Villiers, J. and Pyers, J. (2002). Complements to cognition: A longitudinal study study of the relationship between complex syntax and false-belief understanding. Cognitive Development, 17:1037U1060.˝

DeGelder, B. (1987). On not having a theory of mind. Cognition, 27:285–290.

Dehaene, S. (2007). A few steps toward a science of mental life. Mind, Brain, and Education, 1(1):28–47.

Dehaene, S. (2009). Reading in the Brain. Viking, New York, NY.

Dehaene, S., Kerszberg, M., and Changeux, J. P. (1998a). A neuronal model of a global workspace in effortful cognitive tasks. Proceedings of the National Academy of Sciences USA, 95:14529–14534. 358 BIBLIOGRAPHY

Dehaene, S. and Naccache, L. (2001). Towards a cognitive neuroscience of consciousness: basic evidence and a workspace framework. Cognition, 79:1–37.

Dehaene, S., Naccache, L., Cohen, L., Sigman, M., and Vinckier, F. (2005). The neural code for written words: A proposal. Trends in Cognitive Sciences, 9:335–341.

Dehaene, S., Naccache, L., Le Clech, G., Koechlin, E., Mueller, M., Dehaene-Lambertz, G., van de Moortele, P. F., and Le Bihan, D. (1998b). Imaging unconscious semantic priming. Nature, 395:597–600.

DeLong, M. R. (1990). Primate models of movement disorders of basal ganglia origin. Trends in Neuroscience, 13:281–285.

Dennett, D. (1978). Brainstorms: philosophical essays on mind and psychology. Harvester Press., Sussex.

Dennett, D. C. (1991). Consciousness Explained. Penguin Books, London and New York.

Denny, C. and Rapport, M. (2001). Stimulant Drugs and ADHD: Basic and Clinical Neuroscience, chapter Cognitive pharmacology of stimulants in children with ADHD, pages 283–302. Ox- ford University Press, Oxford and New York.

DeWit, S., Kosaki, Y., Balleine, B., and Dickenson, A. (2006). Dor- somedial prefrontal cortex resolves response conﬂict in rats. Journal of Neuroscience, 26:5224–5229.

Diamond, A., Briand, L., Fossella, J., and Gehlbach, L. (2004). Genetic and neurochemical modulation of prefrontal cognitive functions in children. American Journal of Psychiatry, 161:125– 132.

Diamond, A., Werker, J., and Lalonde, C. (1994). Human Be- haviour and the Developing Brain, chapter Toward understanding commonalities in the development of object search, detour navigation, categorization, and speech perception, pages 380– 426. Guilford Press, New York.

Dias, R., Robbins, T. W., and Roberts, A. C. (1996). Dissociation in prefrontal cortex of affective and attentional shifts. Nature, 380:69–72. BIBLIOGRAPHY 359

DiChiara, G. (2002). Nucleus accumbens shell and core dopamine: differential role in behavior and addiction. Behavioral and Brain Research, 137:75–114. Doupe, A. J. and Kuhl, P. K. (1999). Birdsong and human speech: common themes and mechanisms. Annual Review of Neuro- science, 22:567–631. Durston, S., Davidson, M., Tottenham, N., Galvan, A., Spicer, J., Fossaella, N., and Casey, B. (2006). A shift from diffuse to focal cortical activity with development. Developmental Science, 9:1– 8. Dyck, M. J., Ferguson, K., and Schochet, I. M. (2001). Do autism spectrum disorders differ from each other and from non- spectrum disorders on emotion recognition tests? European Journal of Child and Adolescent Psychiatry, 10:105–116. Edelman, G. M. (2004). Wider Than The Sky. Yale University Press, New Haven and London. Edelman, G. M. and Tononi, G. (2000). Reentry and the dynamic core: Neural correlates of conscious experience, pages 139–151. In Metzinger (2000). Egan, M. F., Goldberg, T. E., Kolchana, B. S., Callicott, J. H., Maz- zanti, C. M., Straub, R. E., Goldman, D., and Weinberger, D. R. (2001). Effect of comt val 108/158 met genotype on frontal lobe function and risk for schizophrenia. Proceedings of the National Academy of Science, USA, 98:6917–6922. Eibl-Eibesfeldt, I. (1970). Ethology: the biology of behaviour. Holt, Rinehart, and Winston, New York. Eimas, P. (1975). Auditory and phonetic coding of the cues for speech: discrimination of the /r-l/ distinction by young infants. Perceptual Psychophysiology, 18:341–347. Eimas, P. D., Einar, R., Jusczyk, L., and Vigorito, L. (1971). Speech perception in infants. Science, 171:303–306. Elman, J. (1990). Finding structure in time. Cognitive Science, 14:179–211. Elman, J. (1993). Learning and development in neural networks: The importance of starting small. Cognition, 48(1):71–99. 360 BIBLIOGRAPHY

Elman, J. L. (1995). Mind as Motion: Explorations in the Dy- namics of Cognition., chapter Language as a dynamical system, pages 195–223. MIT Press, Cambridge, MA. Emery, N. J., Lorincz, E. N., Perrett, D. J., Oram, M. W., and Baker, C. I. (1997). Gaze following and joint attention in rhesus monkeys (macaca mulatta). Journal of Comparative Psychol- ogy, 3:286. Etkin, A. and Wager, T. (2007). Functional neuroimaging of anxiety: a meta-analysis of emotional processing in ptsd, social anxiety disorder, and speciﬁc phobia. American Journal of Psychi- atry, 164:1476–1488. Everitt, B. J. and Robbins, T. W. (2005). Neural systems of reinforcement for drug addiction: from actions to habits to compul- sion. Nature Neuroscience, 8:1481–1489. Eysenck, H. J. (1964). Crime and Personality. Routledge and Kegan Paul, London. Faraone, S. V., Perlis, R. H., Doyle, A. E., Smoller, J. W., Goralnick, J. J., Homgren, M. A., and Sklar, P. (2005). Molecular genetics of attention deﬁcit hyperactivity disorder. Biological Psychiatry, 57:1313–1323. Farran, D. C. (2001). Critical Thinking About Critical Periods,, chapter Critical periods and early intervention. Paul Brookes Publishing Co, Baltimore, London, Toronto, Sydney. Fassbender, C., Zhang, H., Buzy, W. M., Cortes, C. R., Mizuiri, D., Beckett, L., and Schweitzer, J. B. (2009). A lack of default network suppression is linked to increased distractibility in adhd. Brain Research, 1273:114–128. Ferrari, P. F., Maiolini, C., Addessi, E., Fogassi, L., , and Visal- berghi, E. (2005). The observation and hearing of eating actions activates motor programs related to eating in macaque monkeys. Behavior and Brain Research, 161:95–101. Fiez, J. A. and Peterson, S. E. (1998). Neuroimaging studies of word reading. Proceedings of the National Academy of Science of the United States of America, 95:914–921. Fliessbach, K., Weber, B., Trautner, P., Dohmen, T., Sunde, U., El- ger, C. E., and Falk, A. (2007). Social comparison affects reward- BIBLIOGRAPHY 361

related brain activity in the human ventral striatum. Science, 318:1305–1308. Floresco, S., West, A. R., Ash, B., Moore, H., and Grace, A. A. (2004). Extrasynaptic dopamine and phasic neuronal activity. Nature Neuroscience, 7:199. Floresco, S. B., West, A. R., Ash, B., Moore, H., and Grace, A. (2003). Afferent modulation of dopamine neuron ﬁring differentially regulates tonic and phasic dopamine transmission. Na- ture Neuroscience, 6:968–973. Foa, E. B. and Kozak, M. J. (1986). Emotional processing of fear: exposure to corrective information. Psychological Bulletin, 99:20–35. Fodor, J. (1983). The Modularity of Mind. MIT Press, Cambridge, MA. Fodor, J. (1985). Precis of the modularity of mind. Behavioral and Brain Sciences, 8(1):1–46. Fodor, J. (1998). The trouble with psychological darwinism. Lon- don Review of Books, 20(2). Fodor, J. (2000). The Mind Doesnt Work That Way: The Scope and Limits of Computational Psychology. MIT Press, Bioston and NY. Fodor, J. and Pylyshyn, Z. W. (1988). Connectionism and cognitive architecture: A critical analysis. Cognition, 28:3–71. Fogassi, L., Ferrari, P. F., Gesierich, B., Rozzi, S., Chersi, F., and Rizzolatti, G. (2005). Parietal lobe: From action organization to intention understanding. Science, 308:662–667. Forbes, E. E., Brown, S. M., Kimak, M., Frerrell, R. E., Manuck, S. B., and Hariri, A. R. (2009). Genetic variation in components of dopamine neurotransmission impacts ventral striatal reactivity associated with impulsivity. Molecular Psychiatry, 14:60– 70. Forssberg, H., Fernell, E., Waters, S., Waters, N., and Tedroff, J. (2006). Altered pattern of brain dopamine synthesis in male adolescents with attention deﬁcit hyperactivity disorder. Be- havioral and Brain Functions, 2:40. 362 BIBLIOGRAPHY

Fowles, D. C. (1988). Psychophysiology and psychopathy: a motivational approach. Psychophysiology, 25:373–391. Freud, S. (1895). The Standard of the Complete Psychological Works of Sigmund Freud, chapter On the grounds for detaching a particular syndrome from neurasthenia under the description of ¸Sanxietyneurosis”. Hogarth, London. Friedman, J. I., Adler, D. N., and Davis, K. L. (1999). The role of norepinephrine in the pathophysiology of schizophrenia and alzheimer’s disease. Biological Psychiatry, 46:1243–1252. Friedman, M. J. and Karam, E. G. (2009). Stress-Induced and Fear Circuitry Disorders: Deﬁning the Research Agenda for DSM-IV, chapter Posttraumatic Stress Disorder, pages 3–29. American Psychiatric Association, Arlington, Virginia. Frith, C. D. and Frith, U. (1999). Interacting minds-a biological basis. Science, 286:1692–1695. Frith, U. (1989a). Autism: explaining the enigma. Basil Blackwell. Frith, U. (1989b). A new look at language and communication in autism. British Journal of Disorders of Commumication, 24:123–150. Frith, U. and Happe, F. (1994). Autism: beyond ’theory of mind’. Cognition, 50:115–132. Fullana, M., Maitaix-Cols, D., Caspi, A., Harrington, H., Gr- isham, J., Mofﬁtt, T., and Poulton, R. (2009). Obsessions and compulsions in the community: prevalence, interference, help- seeking, developmental stability, and co-occurring psychiatric conditions. American Journal of Psychiatry, 166:329–336. Funahashi, S. (2007). The Cognitive Neuroscience of Working Memory, chapter The general-purpose working memory system and functions of the dorsolateral prefrontal cortex, pages 214– 229. Oxford University Press, Oxford and New York. Funahashi, S., Bruce, C. J., and Goldman-Rakic, P. S. (1989). Mnemonic coding of visual space in the monkey’s dorsolateral prefrontal cortex. Journal of Neurophysiology, 61:331–349. Fuster, J. M. (1989). Finding structure in time. Raven Press, New York. BIBLIOGRAPHY 363

Fuster, J. M. (1997). The Prefrontal Cortex: Anatomy, Physiology, and Neurophysiology of the Frontal Lobe. Lippincott-Raven, Philadelphia, PA, 3rd edition. Gallese, V. (2005). Embodied simulation: from neurons to phe- nomenal experience, phenom. cogn. sci. 4 (2005), pp. 23U48.˝ Phenomenology and the Cognitive Sciences, 4:23–48. Gallese, V. (2006). Intentional attunement: A neurophysiological perspective on social cognition and its disruption in autism. Brain Research, 1079:15–24. Gallese, V., Fadiga, L., Fogassi, L., and Rizzolatti, G. (119). Action recognition in the premotor cortex. Brain, 119:593–609. Gallese, V., Keysers, C., and Rizzolatti, G. (2004). A unifying view of the basis of social cognition. Trends in Cognitive Sciences, 8:396–403. Gat, E. (1998). Artiﬁcial Intelligence and Mobile Robots, chapter On three-layer architectures, pages 195–210. MIT Press, Cam- bridge, MA. Geier, C. and Luna, B. (2009). The maturation of incentive processing and cognitive control. Pharmacology, Biochemistry and Behavior, 89:1–10. Giedd, J. N., Blumenthal, J., Jeffries, N. O., Castellanos, F. X., Liu, H., Zijdenbos, A., Paus, T., Evans, A. C., and Rapoport, J. L. (1999). Brain development during childhood and adolescence: a longitudinal MRI study. Nature Neuroscience, 2:861–863. Giedd, J. N., Vaituzis, A. C., Hamburger, S. D., Lange, N., Ra- japakse, J. C., Kaysen, D., Vauss, Y. C., and Rapoport, J. L. (1996). Quantitative MRI of the temporal lobe, amygdala, and hippocampus in normal human development: Ages 4-18 years. The Journal of Comparative Neurology, 366:223–230. Glowinski, J., Tassin, J. P., and Thierry, A. M. (1984). The mesocortico-prefrontal dopaminergic neurons. TINS, 7:415– 418. Goldman-Rakic, P. (1987a). Development of cortical circuitry and cognitive function. Child Development, 58:601–622. Goldman-Rakic, P. (1987b). Handbook of Physiology. The Nervous System V., chapter Circuitry of primate prefrontal cortex and 364 BIBLIOGRAPHY

regulation of behavior by representational memory., pages 373– 417. American Physiological Society, Bethesda, MD.

Goldman-Rakic, P. (1988). Topography of cognition: parallel distributed networks in primate association cortex. Annual Review of Neuroscience, 11:137–156.

Goldman-Rakic, P. (1991). Psychopathology and the brain, chapter Prefrontal cortical dysfunction in schizophrenia: the relevance of working memory., pages 1–23. Raven Press.

Goldman-Rakic, P. (1994a). Motor and Cognitive Functions of the Prefrontal Cortex, chapter The issue of memory in the study of prefrontal functions, pages 112–122. Springer-Verlag.

Goldman-Rakic, P. (1994b). Working memory dysfunction in schizophrenia. Journal of Neuropsychiatry and Clinical Neu- roscience, 6:348–357.

Goldman-Rakic, P. (1995). Cellular basis of working memory. Neuron, 14:477–485.

Goldman-Rakic, P. (1996). The prefrontal landscape implications of functional architecture for understanding human mentation and the central executive. Philosophical Transactions of the Royal Society, London. B. Biological Sciience, 35:1445–1453.

Goldman-Rakic, P. (1998a). The cortical dopamine system: role in memory and cognition. Advances in Pharmacology, 42:707–711.

Goldman-Rakic, P. (1998b). The Prefrontal Cortex: Executive and Cognitive Functions, chapter The prefrontal landscape: implications of functional architecture for understanding human mentation and the central executive, pages 87–102. Oxford Univer- sity Press.

Goldman-Rakic, P. S. (1987c). Handbook of Physiology. Vol 1,sec1, chapter Circuitry of primate prefrontal cortex and regulation of behavior by representational memory., pages 373–417. Oxford University Press, New York.

Goldman-Rakic, P. S., Lidow, M. S., and Gallager, D. W. (1990). Overlap of dopaminergic, adrenergic and serotonergic receptors and complimentarity of their subtypes in primate prefrontal cortex. Journal of Neuroscience, 10:2125–2138. BIBLIOGRAPHY 365

Goldman-Rakic, P. S., Muly, E. C., and Williams, G. V. (2000). D1 receptors in prefrontal cells and circuits. Brain Research Re- views, 1:295–301.

Goodale, M. and Milner, A. (1992). Separate visual pathways for perception and action. Trends in Neuroscience, 15:20–25.

Goodyear, P. and Hynd, C. (1992). Attention-deﬁcit disorder with (add/h) and without (add/wo) hyperactivity: behavioural and neuropsychological differentiation. Journal of Clinical Child Psychology, 21:243–253.

Goto, Y. and O’ Donnell, P. (2002). Synchronous activity in the hippocampus and nucleus accumbens in vivo. Journal of Neu- roscience, 21:RC131.

Grace, A. (1995). The tonic/phasic model of dopamine system regulation; its relevance for understanding how stimulant abuse can alter basal ganglia function. Drug and Alcohol Dependence, 37:111–129.

Grace, A. (2000). The tonic/phasic model of dopamine system regulation and its implications for understanding psychostimulant craving. Addiction, 95: Suppl 2::119–28.

Grace, A. (2001). Stimulant Drugs and ADHD., chapter Psychos- timulant actions on dopamine and limbic system function: Rel- evance to the pathophysiology and treatment of ADHD: Basic and Clinical Neuroscience., pages 134–157. Oxford University Press, New York.

Graham, P. and Rutter, M. (1973). Psychiatric disorders in the young adolescent: A follow-up study. Proceedings of the Royal Society of Medicine, 6:1226–1229.

Gray, J. (1971). The Psychology of Fear and Stress. Weidenfeld and Nicolson, London.

Gray, J. (1987). The Psychology of Fear and Stress. Cambridge University Press, Cambridge, 2nd edition.

Gray, J. R., Braver, T. S., and Raichle, M. E. (2002). Integration of emotion and cognition in the lateral prefrontal cortex. Proceed- ings of the National Academy of Sciences USA, 99:4115–4120. 366 BIBLIOGRAPHY

Graybiel, A. (1995). Building action repertoires: memory and learning functions of the basal ganglia. Current Biology, 5:733– 741.

Graybiel, A. (1998). The basal ganglia and chunking of action repertoires. Neurobiology of Learning and Memory, 70:119– 136.

Graybiel, A. and Rauch, S. (2000). Toward a neurobiology of obsessive-compulsive disorder. Neuron, 28:343–347.

Grossberg, S. (1999). Neural models of normal and abnormal behavior: what do schizophrenia, parkinsonism, and attention deﬁcit disorder, and depression have in common? Progress in Brain Research, 121:375–406.

Grèzes, J., Armonya, J. L., J., R., and Passinghama, R. E. (2003). Action observation activates premotor and parietal areas in a somatotopic manner: an fmri study. European Journal of Neu- roscience, 18:928–937.

Gusnard, D., Akbudak, L., Shulman, G., and Raichle, M. (2001). Medial prefrontal cortex and self-referential mental activity: Relation to a default mode of brain function. PNASS, 98:4259– 4264.

Gusnard, D. and Raichle, M. (2001). Searching for a baseline: functional imaging and the resting brain. Nature Reviews Neu- roscience, 2:685–694.

Haber, S. (2003). The primate basal ganglia: parallel and integrative networks. Journal of Chemical Neuroanatomy, 26:317–330.

Haber, S., Kunishio, K., Mizbuchi, M., and Lynd-Balta, E. (1995). The orbital and medial prefrontal circuit through the primate basal ganglia. The Journal of Neuroscience, 15:4851–4867.

Haber, S. and McFarland, N. (2001). The place of the thalamus in frontal cortical-basal ganglia circuits. The Neuroscientist, 7(4):315–324.

Haber, S. N., Fudge, J. L., and McFarland, N. R. (2000). Stria- tonigrostriatal pathways in primates form an ascending spiral from the shell to the dorsolateral striatum. The Journal of Neu- roscience, 20:2369–2382. BIBLIOGRAPHY 367

Haber, S. N. and Wolfer, D. B. (1992). Basal ganglia peptidergic staining in tourette syndrome. a follow-up study. Advances in Neurology, 58:145–150. Hale, C. and Tager-Flusberg, H. (2003). The inﬂuence of language on theory of mind: A training study. Developmental Science, 6:346U359.˝ Hampson, M.and Driesen, N. R., Skudlarski, P., Gore, J. C., and Constable, R. T. (2006). Brain connectivity related to working memory performance. Journal of Neuroscience, 26:13338– 13343. Hare, R. (1970). Psychopathy: Theory and Research. Wiley, New York. Hare, R. (1978). Psychopathy and electrodermal response to non- signal stimulation. Biological Psychology, 6:237–246. Hare, T., Tottenham, N., Davidson, M., Glover, G., and Casey, B. (2005). Contributions of amygdala and striatal activity in emotion regulation. Biological Psychiatry, 57:624–632. Harmer, C. J., Perrett, D. I., and Cowen, P. J. e. a. (2001). Adminis- tration of the beta-adrenoreceptor blocker propranalol impairs the processing of facial expressions of sadness. Psychopharma- cology (Berlin), 154:383–389. Harnad, S. (1990). The symbol grounding problem. Physica, D42:335–346. Hauser, M. D., Chomsky, N., and Fitch, W. T. (2002). The faculty of language: what is it, who has it, and how did it evolve? Science, 298:1569–1579. Hauser, M. D. and Fitch, W. T. (2004). Computational constraints on syntactic processing in a non human primate. Science, 303:377–380. Hebb, D. (1949). The Organisation of Behavior: A Neuropsycho- logical Theory. Wiley, New York. Heimer, L. (1972). The olfactory cortex and the ventral striatum. an experimental light- and electron-microscopic study with special emphasis on the problem of terminal degeneration. Brain, Behaviour and evolution, 6:484–523. 368 BIBLIOGRAPHY

Heimer, L. (1991). The Basal Forebrain: Anatomy to Function, chapter Piecing together the puzzle of basal forebrain anatomy., pages 1–42. Plenum Press, New York. Heimer, L. (2003). A new anatomical framework for neuropsychiatric disorders and drug abuse. American Journal of Psychiatry, 160:1726–1739. Heimer, L., Switzer, R. D., and Van Hoesen, G. W. (1982). Ven- tral striatum and ventral pallidum: components of the motor system? TINS, 5:83–87. Heimer, L. and Van Hoesen, G. W. (2006). The limbic lobe and its output channels: implications for emotional functions and adaptive behavior. Neuroscience and Biobehavioral Reviews, 30:126–147. Heimer, L., Van Hoesen, G. W., Trimble, M., and Zahm, D. S. (2008). Anatomy of Neuropsychiatry: The New Anatomy of the Basal Forebrain and its Implications for Neuropsychiatric Ill- ness. Elsevier, Burlinton MA, San Diego CA and London. Heimer, L., Wilson, R., and Santini, M., editors (1975). Golgi Cen- tennial Symposium Proceedings. Raven Press, New York. Heimer, L., Zahm, D. S., Churchill, L., Kalivas, P. W., and Wohlt- man, C. (1991). Speciﬁcity in the projection patterns of accum- bal core and shell in the rat. Neuroscience, 41:89–125. Herba, C. and Phillips, M. (2004). Annotation: Development of facial expression recognition from childhood to adolescence: behavioural and neurological perspectives. Journal of Child Psy- chology and Psychiatry, 45(7):1185–1198. Herbert, M. R., Ziegler, D A. Makris, N., Filipek, P. A., Kemper, T. L., Normandin, J. J., Sanders, H. A., Kennedy, D. N., and Caviness, J. V. S. (2004). Dissociations of cerebral cortex, subcortical and cerebral white matter volumes in autistic boys. An- nals of Neurology, 55:530–540. Hermann, B. P. and Chambria, S. (1980). Interictal psychopathology in patients with ictal fear. Archives of Neurology, 37:667– 668. Hermer, L. and Spelke, E. (1994). A geometric process for spatial reorientation in young children. Nature, 370:57–59. BIBLIOGRAPHY 369

Herschkowitz, N. (2000). Neurological bases of behavioural development in infancy. Brain and Development, 22:411–416. Hezler, B. and Griffin, J. (1981). Infantile autism and the temporal lobe of the brain. Journal of Autism and Developmental Disorders, 9:153–157. Hobson, R. (1986). The autistic child’s appraisal of expressions of emotion: a further study. Journal of Child Psychology and Psychiatry, 27:671–680. Hoffman, M. L. and Saltzstein, H. D. (1967). Parent discipline and the child’s moral development. Journal of Personality and Social Psychology, 5:45–57. Hollerman, J. R., Tremblay, L., and Schultz, W. (1998). Influence of reward expectation on behavior-related neuronal activity in primate striatum. Journal of Neurophysiology, 80:947–963. Hollerman, J. R., Tremblay, L., and Schultz, W. (2000). Involve- ment of basal ganglia and orbitofrontal cortex in goal-directed behavior. Progress in Brain Research, 126:193–215. Holscher, C. and Munk, M., editors (2009a). Information Pro- cessing by Neuronal Populations. Cambridge University Press, Cambridge and New York. Holscher, C. and Munk, M., editors (2009b). Information Pro- cessing by Neuronal Populations, chapter Summary of chapters, conclusion, and future targets, pages 433–469. Cambridge Uni- versity Press, Cambridge University Press. Holscher, C. and Munk, M., editors (2009c). Information Process- ing by Neuronal Populations, chapter Functional roles of theta and gamma oscillations in the association and dissociation of neuronal networks in primates and rodents., pages 151–173. Cambridge University Press, Cambridge and New York. Hopfield, J. (1982). Neural networks and physical systems with emergent collective computational abilities. Proceedings of the National Academy of Science, USA, 79:2554–2558. Houk, J. C. and Wise, S. P. (1995). Distributed modular architectures linking basal ganglia, cerebellum, and cerebral cortex: Their role in planning and controlling action. Cerebral Cortex, 2:95–110. 370 BIBLIOGRAPHY

Huttenlocher, P. (1994). Human Behavior and the Developing Brain. In Dawson, G. and Fischer, K. (eds), chapter Synapto- genesis in human cerebral cortex, pages 233–234. The Guilford Press., New York. Hyman, S. (2007). Can neuroscience be integrated into the dsm? Nature Reviews Neuroscience, 8:725–732. Iacoboni, M., Woods, R. P., Brass, M., Bekkering, H., and Mazz- iotta, J. C. (1999). Cortical mechanisms of human imitation. Science, 2:2526–2528. Insel, T. R. (2010). Faulty circuits. Scientiﬁc American, pages 44–51. Jackendoff, R. (2002). Foundations of Language: Brain,Meaning, Grammar, Evolution. Oxford University Press, Oxford and New York. Jackson, D. C., Malmstadt, J. R., Larson, C. L., and Davidson, R. J. (2000). Suppression and enhancement of emotional responses to unpleasant pictures. Psychophysiology, 37:515–522. Jenike, M. A., Baer, L., and Minichiello, W. (1990). Obsessive Com- pulsive Disorders: Practical Management. Year Book Medical Publishers, Chicago, Ill, 2nd edition. Joel, D. (1994). The organization of the basal ganglia- thalamocortical circuits: Open interconnected rather than closed segregated. Neuroscience, 63:363–379. Jorm, A. E., Prior, M., and Sanson, A. (2000). Association of a functional polymorphism of the serotonin transporter gene with anxiety-related temperament and behaviour problems in children: a longitudinal study from infancy to the mid-teens. Molec- ular Psychiatry, 5:542–547. Joseph, R. (2000). Neuropsychiatry, Neuropsychology, Clin- ical Neuroscience, chapter Limbic languge, pages 235–266. Williams and Wlkins, New York. Kandel, E. and Freed, D. (1989). Frontal lobe dysfunction and antisocial behavior: a review. Journal of Clinical Psychology, 45:404–413. Kanner, L. (1943). Autistic disturbances of affective contact. Ner- vous Child, 2:217–250. BIBLIOGRAPHY 371

Kanner, L. and Eisenberg, L. (1956). Early infantile autism. American Journal of Orthopsychiatry, 26:55–65. Kapur, S. and Lecrubier, Y. (2003). Dopamine in the Pathophysi- ology and Treatment of Schizophrenia. Martin Dunitz, London and New York. Karmilloff-Smith, A., Brown, J., Grice, S., and Paterson, S. (2003). Dethroning the myth: cognitive: cognitive dissociations and innate modularity in williams syndrome. Developmental Neu- ropsychology, 23:227–242. Karmiloff-Smith, A. (1992). Beyond Modularity: A developmental perspective on human consciousness. MIT Press, Cambridge, MA. Karmiloff-Smith, A. (1994). Precis of beyond modularity: A developmental perspective on cognitive science. Behavioral and Brain Sciences, 17(4):693–745. Keil, A., Mueller, M. M., Gruber, T., Weinbruch, C., and Elbert, T. (1999). Human large-scale oscillatory brain activity during an operant shaping procedure. Cognition and Brain Research, 12:397–407. Kellendonk, C., Simpson, E., Polan, H., Malleret, G., Vronskaya, S., Winiger, V., Moore, H., and Kandel, E. (2006). Transient and selective overexpression of dopamine d2 receptors in the striatum causes persistent abnormalities in prefrontal cortex functioning. Neuron, 49:603–615. Kenway, L. and Wilson, M. (2001). Temporally structured replay of awake hippocampal ensemble activity during rapid eye movement sleep. Neuron, 29:145–156. Kiehl, K. A., Smith, A. M., and Hare, R. D. e. a. (2001). Limbic abnormalities in affective processing by criminal psychopaths as revealed by functional magnetic resonance imaging. Biological Psychiatry, 50:677–684. Kling, A. and Brothers, L. (1992). Neurobiological aspects of emotion, memory, and mental dysfunction, chapter The amygdala and social behavior. Wiley-Liss, Inc., New York. Klorman, R., Brumaghim, J., and Borgstedts, A.D.and Salzman, L. (1991). How event-related potentials help understand the 372 BIBLIOGRAPHY

effects of stimulants on attention deﬁcit hyperactivity disorder. Advances in Psychophysiology, 4:107–153.

Kluver, H. and Bucy, P. (1939). Preliminary analysis of function of the temporal lobe in monkeys. Achives of Neurology, 42:979– 1000.

Kohlberg, L. and Gilligan, C. (1971). The adolescent as a philoso- pher: The discovery of the self in a postconventional world. Daedalus, Journal of the American Academy of Arts and Sci- ences, 100(4):1051–1086.

Kohler, E., Keysers, C., Umilta, L., and Fogassi, V. (2002). Hear- ing sounds, understanding actions: action representation in mirror neurons. Science, 297:846–848.

Kopell, B. and Greenberg, B. (2008). Anatomy and physiology of the basal ganglia: Implications for dbs in psychiatry. Neuro- science and Biobehavioral Reviews, 32:408–422.

Koziol, L. and Budding, D. (2009). Subcortical Structures and Cognition: Implications for Neuropsychological Assessment. Springer, New York.

Krebs, P. (1995). Models of cognition: Neurological possibility does not indicate neurological plausibility. In Cognitive Science Journal Archive, Mahwah, New Jersey. Lawrence Erlbaum As- sociates.

Krueger, R., Markon, K., Patrick, C., and Iacono, W. (2005). Externalising psychopathology in adulthood: A dimensional- spectrum conceptualization and its implications for DSM-V. Journal of Abnormal Psychology, 114:537–550.

Krystal, J., Southwick, S., and Charney, D. (1995). Memory Dis- tortions: How Minds, Brains and Societies Reconstruct the Past, chapter Posttraumatic stress disorder: Psychological mechanisms of traumatic remembrance. In, Schacter, D. (ed), pages 150–171. Harvard University Press, Cambridge MA.

Kuhl, P., Conboy, B., Coffey-Corina, S., Padden, D., Rivera- Gaxiola, M., and Nelson, T. (2008). Phonetic learning as a pathway to language: new data and native language magnet theory expanded (nlm-e). Philosophical Transactions of the Royal So- ciety, B, 363:979–1000. BIBLIOGRAPHY 373

Lahey, B., Urbano, R., Krueger, R., Applegate, B., Garriock, H., Chapman, D., and Waldman, I. (2008). Testing structural models of dsm-iv symptoms of common forms of child and adolescent psychopathology. Journal of Abnormal Child Psychology, 36:187–206. Lalonde, C. and Werker, J. (1995). Cognitive inﬂuences on cross- language speech perception in infancy. Infant Behaviour and Development, 18:459–475. Lamme, V. (2006). Towards a true neural stance on consciousness. Trends in Cognitive Sciences, 10(11):494–501. Lawler, A. (2001). Archeology. writing gets a rewrite. Science, 292:2418–2420. Leary, M. and Hill, D. (1996). Moving on: autism and movement disturbance. Mental Retardation, 34(1):39–53. Leckman, J. (2002). Tourette’s syndrome. The Lancet, 360:1577– 1586. Leckman, J., Bloch, M., and King, R. (2009). Symptom dimensions and subtypes of obsessive-compulsive disorder: a developmental perspective. Dialogues in Clinical Neuroscience, 11:21–33. Leckman, J. and Cohen, D. (1999). Tourette’s Syndrome-Tics, Obsessions, Compulsions: Developmental Psychopathology and Clinical Care. John Wiley and Sons, New York. Leckman, J., Pauls, D., Zhang, H., Rosario-Campos, M., Katso- vich, L., Kidd, K., Pakstis, A., Alsobrook, J., Robertson, M., McMahon, W., Walkup, J., van de Wettering, B., King, R., and Cohen, D. (2003). Obsessive compulsive symptom dimensions in affected sibling pairs diagnosed with gilles de la tourette syndrome. American Journal of Medical Genetics B Neuropsychi- atric Genetics, 116:60–68. LeDoux, J. (1995). Emotion:clues from the brain. Annual Review of Psychology, 46:209–235. LeDoux, J. (2002). Synaptic Self: How our Brains Become Who We Are. Viking Penguin, New York. LeDoux, J. and Gorman, J. (2001). A call to action; overcoming anxiety through active coping. American Journal of Psychiatry, 158:1953–1955. 374 BIBLIOGRAPHY

LeDoux, J. E. ., Romanski, L., and Xagoraris, A. (1989). Indeli- bility of subcortical emotional memories. Journal of Cognitive Neuroscience, 1:238–243. Lee, A. and Wilson, M. (2002). Memory of sequential experience in the hippocampus during slow wave sleep. Neuron, 36:1183– 1194. Leiner, H., Leiner, A., and Dow, R. (1991). The human cerebro- cerebellar system: its computing, cognitive, and language skills. Behavioural Brain Research, 44:113–128. Leiner, H., Leiner, A., and Dow, R. (1993). Cognitive and language functions of the human cerebellum. TINS, 16:444–447. Leslie, A. (1987). Pretence and representation: the origins of “theory of mind”. Psychological Review, 4:412–426. Leslie, A. (1994). Mapping the Mind: Domain-Specificity in Cog- nition and Culture, chapter ToMM, ToBY and agency: Core architecture and domain specificity., pages 119–148. Cambridge University Press, Cambridge. Leslie, A. and Frith, U. (1988). Autistic children’s understanding of seeing, knowing, and believing. British Journal of Develop- mental Psychology, 6:315–324. Leslie, A. and Roth, D. (1993). Minds: Perspectives from Autism., chapter What autism teaches us about metarepresentation. Ox- ford University Press, Oxford and New York. Leslie, K., Johnson-Frey, S., and Grafton, S. (2004). Functional imaging of face and hand imitation: towards a motor theory of empathy. Neuroimage, 21:601–607. Leung, H., Skudlarski, P., Gatenby, J., Peterson, B., and Gore, J. C. (2000). An event-related functional MRI study of the Stroop color word interference task. Cerebral Cortex, 10:552–560. Levy, F. (1980). The development of sustained attention (vigilance) and inhibition in children: some normative data. Journal of Child Psychology and Psychiatry, 21(21):77–84. Levy, F. (1989). Reading, spelling, and vigilance in attention deficit and conduct disorder. Journal of Abnormal Child Psy- chology, 17:291–298. BIBLIOGRAPHY 375

Levy, F. (2004). Synaptic gating and ADHD : A biological theory of comorbidity of ADHD and anxiety. Neuropsychopharmacology, 29:1589–1596.

Levy, F. (2007a). Theories of autism: A review. Australian and New Zealand Journal of Psychiatry, 41:859–868.

Levy, F. (2007b). What do dopamine transporter and catechol-o- methyltransferase tell us about attention deﬁcit-hyperactivity disorder? Pharmacogenomic implications. Australian and New Zealand Journal of Psychiatry, 41:10–14.

Levy, F. (2007c). Working memory, catecholamines and psychosis: Illustrative case. Australian and New Zealand Journal of Psy- chiatry, 41:74–77.

Levy, F. (2008). Pharmacological and therapeutic directions in adhd: Speciﬁcity in the PFC. Behavioral and Brain Functions, 4:12.

Levy, F. (2010). Internalising versus externalising comorbidity: neural circuit hypothesis. Australian and New Zealand Journal of Psychiatry, 44:399–409.

Levy, F. and Hobbes, G. (1981). The diagnosis of attention deﬁcit disorder (hyperkinesis) in children. Journal of the American Academy of Child and Adolescent Psychiatry, 20:376–384.

Levy, F., Horn, K., and Dalglish, R. (1987). Relation of attention deﬁcit and conduct disorder to vigilance and reading lag. Aus- tralian and New Zealand Journal of Psychiatry, 21:242–245.

Levy, F. and Krebs, P. (2006). Cortical-subcortical reentrant circuits and recurrent behaviour. Australian and New Zealand Journal of Psychiatry, 40(9):752–758.

Levy, F., Young, D., McDougal, M. R., Martin, N. C., Piek, J. P., and Hay, D. A. (2009). Comorbidity of internalising and externalising disorders with ADHD and reading disorder. International Society for Research in Child and Adolescent Psychopathology.

Levy, R. and Goldman-Rakic, P. (1999). Association of of storage and processing functions in the dorsolateral prefrontal cortex of the non-human primate. Journal of Neuroscience, 19:5149– 5158. 376 BIBLIOGRAPHY

Lewis, D. (1997). Development of the prefrontal cortex during adolescence: Insights into vulnerable neural circuits in schizophrenia. Neuropsychopharmacology, 16:385–398.

Lewis, V. and Boucher, J. (1988). Spontaneous, instructed and elicited play in relatively able autistic children. British Journal of Developmental Psychology, 6:325–339.

Liberman, A. and Mattingly, I. (1985). The motor theory of speech perception revised. Cognition, 21:1–36.

Liberman, A. and Wahlen, D. (2000). On the relation of speech to language. Trends in Cognitive Neuroscience, 4:187–196.

Liberman, A. M., Cooper, F., Shankweiler, D., and Stuart- Kennedy, M. (1967). Perception of the speech code. Psychology Review, 74:431–461.

Liberman, I. (1974). Explicit syllable and phoneme segmentation in the young child. Journal of Experimental Child Psychology, 18:201–212.

Liberzon, I. and Sripada, C. (2008). The functional neuroanatomy of ptsd: a critical review. Progress in Brain Research, 167:151– 169.

Lichter, D. G. and Cummings, J. L. (2001). Frontal-Subcortical Circuits in Psychiatric and Neurological Disorders. The Guil- ford Press, New York.

Lieberman, M. and Eisenberger, N. (2006). Social Neuroscience, chapter A pain by any other name (rejection, exclusion, os- tracism) still hurts the same: The role of dorsal anterior cingulate cortex in social and physical pain, pages 167–187. MIT Press, Cambridge, Massachusetts and London, England.

Lieberman, M., Eisenberger, N., Crockett, M., Tom, S., Pfeifer, J., and Way, B. (2007). Putting feelings into words: Affect labeling disrupts amygdala activity in response to affective stimuli. Psychological Science, 18:421–427.

Lieberman, P. (2008). Cortical-striatal-cortical neural circuits, reiteration, and the “narrow faculty of language”. Behavioural and Brain Sciences, 31:527–528. BIBLIOGRAPHY 377

Livingstone, M., Rosen, G., Drislane, F., and Galaburda, A. (1991). Physiological and anatomical evidence for a magnocellular de- fect in developmental dyslexia. Proceedings of the National Academy of Sciences of the United States of America, 88:7943– 7947.

Lock, A. and Peters, C. (1996). Cognitive abilities in a comparative perspective, chapter In, Lock, A. and Colombo, M. (eds) Hand- book of Human Symbolic Evolution., pages 596–643. Clarenden Press, Gloucestershire, UK.

Lorenz, K. (1981). The Foundation of Ethology. Springer-Verlag, New York.

Luna, B., Thulborn, K., Munoz, D., Merriame, E., Garvera, K., Minshew, N., Keshavan, M., Genovese, C., Eddy, W., and Sweeney, J. (2001). Maturation of widely distributed brain function subserves cognitive development. Neuroimage, 13:786– 793.

MacLean, P. (1952). Some psychiatric implications of physiological studies on frontotemporal portions of limbic system (visceral brain). Electroencephalography and Clinical Neurophysiology, 4:407–418.

MacLean, P. (1972). Cerebral evolution and emotional processes: New ﬁndings on the striatal complex. Annals of the New York Academy of Sciences, 193:137–149.

MacLean, P. (1990). MacLean, P.D. Plenum, New York.

MacLean, P. D. (1949). Psychsomatic disease and the ’visceral brain’, recent developments bearing on the papez theory of emotion. Psychosomatic Medicine, 11:338–353.

MacWhinney, B. (1998). Models of the emergence of language. Applied Psycholinguist, 49:199–227.

Mann, E. O. and Paulsen, O. (2009). Information Processing by Neuronal Populations, chapter Cellular mechanisms underlying network synchrony in the medial temporal lobe, pages 21– 48. Cambridge University Press, Cambridge.

Marcus, G. F. (2006). Cognitive architecture and descent with modiﬁcation. Cognition, 101:443–465. 378 BIBLIOGRAPHY

Marcus, M. (2004). The Birth of the Mind, pages 115–124. Basic Books, New York.

Marios, R. and Ivanoff, J. (2005). Capacity limits of information processing in the brain. Trends in Cognitive Science, 9:296–305.

Marsh, R., Maia, T., and Peterson, B. (2009). Functional disturbances within frontostriatal circuits across multiple childhood psychopathologies. American Journal of Psychiatry, 166:664– 674.

Marsh, R., Zhu, H., Schultz, R., Quackenbush, G., Royal, J., Skud- larski, P., and Peterson, B. (2006). A developmental fmri study of self-regulatory control. Human Brain Mapping, 27:664–674.

Mason, J., Wang, S., and Yehuda, R. e. a. (2001). Psychogenic lowering of urinary cortisol levels linked to increased emotional numbing and a shame-depressive syndrome in combat- related posttraumatic stress disorder. Psychosomatic Medicine, 63(3):387–401.

Mattay, V. S., Goldberg, T. E., Fera, F., Hariri, A. T., Egan, M. F., Kolachana, B., Callicott, J. H., and Weinberger, D. R. (2003). Catechol o-methyltransferase val 158-met genotype and individual variation in brain response to amphetamine. Proceed- ings of the National Academy of Science, USA, 100(10):6186– 6191.

Mayberg, H., Liotti, M., Brannan, S., McGinnis, S., Mahurin, R., Jerabek, P., Silva, P., Arturo, S., Tekell, J., Martin, C., Lan- caster, J., and Fox, P. (1999). Reciprocal limbic-cortical function and negative mood: Converging pet ﬁndings in depression and normal sadness. The American Journal of Psychiatry, 156:675– 682.

McCormick, D.A.and Shu, Y., Hasentaub, A., Sanchez-Vives, M., Badoual, M., and Bal, T. (2003). Persistent cortical activity: Mechanisms of generation and effects on neuronal excitability. Cerebral Cortex, 13:1219–1231.

McEvoy, R., Loveland, K., and Landry, S. H. (1988). The functions of immediate echolalia in autistic children: A developmental perspective. Journal of Autism and Developmental Disorders, 18:657–668. BIBLIOGRAPHY 379

McFarlane, A., Yehuda, R., and Clark, C. (2002). Biologic models of traumatic memories and post-traumatic stress disorder. the role of neural networks. Psychiatric Clinics of North America, 25:253–270.

McGowan, P., Sasaki, A., D‘Alessio, A., Dymov, S., Labonte, B., Szyf, M., Turecki, G., and Meaney, M. (2009). Epigenetic regulation of the glucocorticoid receptor in human brain associates with childhood abuse. Nature Neuroscience, 12:342–348.

McLoughlin, G., Ronald, A., Kunsti, J., Asherson, P., and Plomin, R. (2007). Genetic support for the dual nature of attention deﬁcit hyperactivity disorder: Substantial genetic overlap between the inattentive and hyperactive impulsive components. Journal of Abnormal Child Psychology, DOI 10.1007/s:10802– 007–9149–9.

McMaster, F., O’Neill, J., and Rosenberg, D. (2008). Brain imaging in obsessive-compulsive disorder. Journal of the American Academy of Child and Adolescent Psychiatry, 47:1262–1272.

Mehler, J. and Dupoux, E. (1994). What Infants Know: The New Cognitive Science of Early Development. Blackwell, Oxford.

Mervis, C., Morris, C., Klein-Tasman, B., Bertrand, J., Kwitny, S., Applebaum, L., and Rice, C. (2003). Attentional characteristics of infants and toddlers with Williams syndrome during triadic interactions. Developmental Psychopathology, 23:243U268.˝

Mesulam, M. and Mufson, E. (1984). Neural inputs into the nucleus basalis of the substantia innominta (ch4) in the rhesus monkey. Brain, 107:253–274.

Metzinger, T. (2000). Neural Correlates of Consciousness. MIT Press.

Middelton, F. and Strick, P. (2002). Basal-ganglia ’projections’ to the prefrontal cortex of the primate. Cerebral Cortex, 12:926– 935.

Middleton, F. A. and Strick, P. L. (2001). Revised neuroanatomy of frontal-subcortical loops. In, Lichter, D.G. and Cummings, J. L. Eds. Frontal-Subcortical Circuits in Psychiatric and Neuro- logical Disorders., pages 44–58. The Guilford Press. 380 BIBLIOGRAPHY

Miguel, E. C., Baer, L., and Coffey, B. J. (1997). Phenomeno- logical differences appearing with repetitive behaviours in obsessive-compulsive disorder and gilles de la tourette’s syndrome. British Journal of Psychiatry, 170:104–145.

Miller, E., Erikson, C., and Desimone, R. (1996). Neural mechanisms of visual working memory in prefrontal cortex of the macque. Journal of Neuroscience, 16:5154–5167.

Miller, E. K. and Cohen, J. D. (2001). An integrative theory of prefrontal cortex function. Annual Review of Neuroscience, 24:167– 202.

Mills, S., Langley, K., van den Bree, M., Street, E., Turic, D., Owen, M., O’Donovan, M., and Thapar, A. (2004). No evidence of of association between catechol-o-methyltransferase (comt) val158met genotype and performance on neuropsychological tasks in children with adhd: a case control study. BMC Psychiatry, 4:15–19.

Milne, E., Swetenham, J., Hansen, P., Campbell, R., Jeffries, H., and Plaisted, K. (2002). High motion coherence thresholds in children with autism. Journal of Child Psychology and Psychi- atry, 43:255–263.

Milner, A. and Goodale, M. (1995). The Visual Brain in Action. Oxford University Press.

Minsky, M. (2006). The Emotion Machine. Simon and Schuster, New York, London, Toronto and Sydney.

Mirenowicz, J. and Schultz, W. (1997). Importance of unpredictability for reward responses in primate dopamine neurons. Journal of Neurophysiology, 72:1024–1027.

Miyake, A. and Shah, P. (1999). Models of Working Memory: Mech- anisms of Active Maintenance and Executive Control, chapter Toward uniﬁed theories of working memory: emerging general consensus, unresolved theoretical issues, and future directions., pages 442–481. Cambridge University Press, Cambridge.

Modell, J., Mountz, J., Curtis, G., and Greden, J. (1989). Neuro- physiologic dysfunction in basal ganglia/limbic striatal and thalamocortical circuits as a pathogenic mechanism of obsessive- compulsive disorder. Journal of Neuropsychiatry, 1:27–36. BIBLIOGRAPHY 381

Mogenson, G., Jones, D., and Yim, C. (1980). From motivation to action: functional interface between the limbic system and the motor system. Progress in Neurobiology, 14:69–92.

Mogenson, G. J. and Wu, M. (1988). Disruption of food hoarding by injections of procaine into mediodorsal thalamus. gaba into subpallidal regions, and haloperidol into accumbens. Brain Re- search Bulletin, 20:247–251.

Moldin, S. and Rubenstein, J., editors (2006). Understanding Autism: From Basic Neuroscience to Treatment. Taylor and Francis, London and New York.

Mooney, R. (2004). Birdsong, chapter Synaptic mechanisms for auditory-vocal integration and the correction of vocal errors, pages 476–494. The New York Academy of Sciences, New York, New York.

Moore, J. and Guan, Y. (2001). Cytoarchitectural and axonal maturation in human auditory cortex. Journal of the Association for Research in Otolaryngology, 2:297–311.

Moron, J. A., Brockington, A., Wise, R. A., Rocha, B. A., and Hope, B. T. (2002). Dopamine uptake through norepinephrine transporter in brain regions with low levels of the dopamine transporter: evidence from knockout mouse lines. Journal of Neuro- science, 22:389–395.

Morris, J. S., Frith, C. D., Perrett, D. I., Rowland, D., Young, A. W., Calder, A. J., and Dolan, R. J. (1996). A differential neural response in the human amygdala to fearful and happy facial expressions. Nature, 383:812–815.

Morris, R., Pandya, D. N., and Petrides, M. (1999). Fiber system linking the mid-dorsal frontal cortex with the retrosplenial/presubicular region in the rhesus monkey. Journal of Com- parative Neurology, 331:68–70.

Moscovich, M. (1995). Recovered consciousness: A hypothesis con- ﬁrming modularity and episodic memory and episodic memory. Journal of Clinical and Experimental Neuropsychology, 17(2):276–290.

Munakata, Y., Morton, B. J., and Stedron, J. M. (2003). Connec- tionist Models of Development, chapter The role of prefrontal 382 BIBLIOGRAPHY

cortex in perseveration: Developmental and computational explorations, pages 83–111. Psychology Press, Hove and New York.

Munakata, Y. and Stedron, J. M. (2001). Handbook of Develop- mental Neuroscience., chapter Neural network models of cognitive development, pages 159–171. MIT Press., Camridge, MA.

Murphy, B. L., Arnsten, A. F. T., Goldman-Rakic, P. S., and Roth, R. H. (1996). Increased dopamine turnover in the prefrontal cortex impairs spatial working memory performance in rats and monkeys. Proceedings of the National Academy of Science USA, 93:1325–1329.

Nakamura, K., Hara, N., Kouider, S., Takayama, Y., Hanajima, R., Sakai, K., and Ugawa, Y. (2006). Task-guided selection of the dual neuron pathways for reading. Neuron, 52:557–564.

Newell, A. (1980). Physical symbol systems. Cognitive Science, 4:135–183.

Nigg, J. (2003). Is adhd a disinhibitory disorder? Psychological Buletin, 1008:170–182.

Nigg, J. (2006). Temperament and developmental psychopathology. Journal of Child Psychology and Psychiatry, 47:395–422.

Nigg, J. T. and Casey, B. J. (2005). An integrative theory of attention-deﬁcit hyperactivity disorder, based on the cognitive and affective neurosciences. Development and Psychopathology, 17:785–806.

Norman, D. and Shallice, T. (1980). Attention to action: willed and automatic control of behaviour. Technical report, Center for Human Information processing, Technical Report No. 99., University of California, San Diego.

Norris, D. (1990). How to build a connectionist idiot (savant). Cognition, 35:277–291.

O’Doherty, J., Kringlebach, M., Rolls, E., Hornak, J., and An- drews, C. (2001). Abstract reward and punishment representations in the human orbitofrontal cortex. Nature Neuroscience, 4:95–102. BIBLIOGRAPHY 383

Ohnishi, T., Matsuda, H., Hashimoto, T., Kunihiro, T., Nishikawa, M., Uema, T., and Sasaki, M. (2000). Abnormal regional cerebral blood ﬂow in childhood autism. Brain, 123:1838–1844.

O’Keefe, J. (1991). Brain and Space, chapter The hippocampal cognitive map and navigational strategies., pages 273–295. Ox- ford University Press.

Olsson, A. and Phelps, E. (2007). Social learning of fear. Nature Neuroscience, 10:1095–1124.

Oschner, K., Bunge, S. A., Gross, J. J., and Gabrieli, J. E. (2002). Rethinking feelings: An fMRI study of the cognitive regulation of emotions. Journal of Cognitive Neuroscience, 14:1215–1229.

Ostlund, S. and Balleine, B. (2007). Orbitofrontal cortex mediates outcome encoding in pavlovian but not instrumental conditioning. Journal of Neuroscience, 27:4819–4825.

Owen, A. M., Evans, A., and Petrides, M. (1996). Evidence for a two-stage model of spatial working memory processing with the lateral frontal cortex. Cerebral Cortex, 6:31–38.

Paillard, J., editor (1991). Brain and Space. Oxford University Press, Oxford and New York.

Panksepp, J. (1998). Affective Neuroscience: The Foundations of Human and Animal Emotions. Oxford University Press, New York and Oxford.

Papez, J. W. (1937). A proposed mechanism of emotion. Archives of Neurology and Psychiatry, 38:725–743.

Parent, A. and Hazrati, L. (1995). Functional anatomy of the basal ganglia. i. the cortico-basal ganglia-thalamo-cortical loop. Brain Research Reviews, 20:91–127.

Pascual-Leone, A. and Walsh, V. (2001). Fast backprojections from the motion to the primary visual visual area necessary for visual awareness. Science, 292:510–512.

Pasher, H. (1994). Dual-task interference in simple tasks. data and theory. Psychological Bulletin, 116:220–244.

Patrick, C. (1994). Emotion and psychopathy: startling new insights. Psychophysiology, 31:319–330. 384 BIBLIOGRAPHY

Patrick, C. J., Bradley, M. M., and Lang, P. J. (1993). Emotion and the criminal psychopath: startle reﬂex modulation. Journal of Abnormal Psychology, 102:82–92.

Patterson, C. M. and Newman, J. P. (1993). Reﬂectivity and learning from aversive events: toward a psychological mechanism for the syndromes of disinhibition. Psychological Review, 100:716– 736.

Pauls, D., Raymond, C. L., Stevenson, J. M., and Leckman, J. F. (1991). A family study of gilles de la tourette syndrome. Amer- ican Journal of Human Genetics, 48:154–163.

Pauls, D. L., Towbin, K. E., Leckman, J. F., Zahnner, G. E. P., and Cohen, D. J. (1986). Gilles de la tourette’s syndrome and obsessive compulsive disorder. evidence supporting a genetic relationship. Archives of General Psychiatry, 43:1180–1182.

Paus, T., Zijdenbos, A., Worsley, K., Collins, D. L., Blumenthal, J., Giedd, J., Rapoport, J., and Evans, A. C. (1999). Structural maturation of neural pathways in children and adolescents: In vivo study. Science, 283:1908–1911.

Pavuluri, M. N. and Sweeney, J. A. (2008). Integrating functional brain neuroimaging and developmental cognitive neuroscience in child psychiatry research. Journal of the American Academy of Child and Adolescent Psychiatry, 47:1273–1288.

Penn, D., Holyoak, K. J., and D.J., P. (2008). Darwin’s mistake: Explaining the discontinuity between human and nonhuman minds. Behavioral and Brain Sciences, 31:109–130.

Penn, D. C. and Povinelli, D. J. (2007). On the lack of evidence that non-human animals possess anything remotely resembling a theory of mind. Philosophical Transactions of the Royal Soci- ety B, Biological Sciences, 362:731–744.

Pennington, B. and Ozonoff, S. (1996). Executive functions and developmental psychopathology. Journal of Child Psychology and Psychiatry, 37:51–87.

Perner, J. (1988). Developing Theories of Mind., chapter Devel- oping semantics for theories of mind: From propositional attitudes to mental representation. Cambridge University Press, Cambridge and New York. BIBLIOGRAPHY 385

Peterson, B., Kane, M., Alexander, G., and Lacadie, C. (2002). An event-related functional mri study comparing interference effects in the simon and stroop tasks. Cognitive Brain Research, 13:427–440. Petrides, M. (1994). Handbook of Neuropsychology, volume 9, chapter Frontal lobes and working memory: evidence from investigations of the effects of cortical excisions in nonhuman primates, pages 17–58. Elsevier, Amsterdam. Petrides, M. (1998). The Prefrontal Cortex: Executive and Cogni- tive Functions, chapter Specialized systems for the processing of mnemomic information within the primate frontal cortex, pages 103–116. Oxford University Press, Oxford and New York. Petrides, M., Alivisatos, B., Meyer, E., and Evans, A. (1993). Func- tional activation of the human frontal cortex during the performance of verbal working memory tasks. Proceedings of the National Academy of Science USA, 90:874–887. Petrides, M. and Pandya, D. N. (2002). Principles of Frontal Lobe Function, chapter Association pathways of the prefrontal cortex and functional observations, pages 31–50. Oxford University Press, New York. Phelps, E. (2004). Human emotion and memory: interactions of the amygdala and hippocampal complex. Current Opinion in Neurobiology, 14:198–202. Phillips, M. L., Drevets, W., Rauch, S. L., and Lane, R. (2003a). Neurobiology of emotion perception ii: Implications for major psychiatric disorders. Biological Psychiatry, 54(5):515–528. Phillips, M. L., Drevets, W. C., Rauch, S. L., and Lane, R. (2003b). Neurobiology of emotion perception i: The neural basis of normal emotion perception. Biological Psychiatry, 54(5):504–514. Phillips, M. L. and Young, A. W. (1998). Neural responses to facial and vocal expressions of fear and disgust. Proceedings of the Royal Society B. Biological Sciences, 265:1809–1817. Phillips, M. L., Young, A. W., Senior, C., Brammer, M., Calder, A. J., Bullmore, E. T., Perrit, D. I., Rowland, D., Williams, S. C., Gray, J. A., and David, A. S. (1997). A speciﬁc neural substrate for perceiving facial expressions of disgust. Nature, 389:495– 498. 386 BIBLIOGRAPHY

Piaget, J. (1955). The Child’s Construction of Reality. Routledge and Kegan Paul.

Pierce, K., Muller, R. A., Ambrose, J., Allen, G., and Courchesne, E. (2001). Face processing occurs outside the fusiform ’face area’ in autism: evidence from functional mri. Brain, 124:2059–2073.

Pinker, S. (1997). How the Mind Works. Penguin Books, London.

Pinker, S. and Prince, A. (1988). On language and connectionism: Analysis of a parallel distributed processing model of language acquisition. Cognition, 28:73–193.

Plaisted, K., Saksida, L., Alcantara, J. I., and Weisblatt, E. J. L. (2003). Towards an understanding of the mechanisms of weak central coherence effects: experiments in visual conﬁgural learning and auditory perception. Philosophical Transactions of the Royal Society London, Biological Science, 358:375–386.

Pollack, S. D. and Kistler, D. J. (2002). Early experience is associated with the development of categorical representations for facial expressions of emotion. Proceedings of the National Academy of Sciences of the United States of America, 99:9072– 9076.

Ponzi, A. (2008). Dynamical model of salience gated working memory, action selection ans reinforcement based on basal ganglia and dopamine feedback. Neural Networks, 21:322–330.

Posner, J., Russell, J. A., Gerber, A., Gorman, D., Colibazzi, T., Yu, S., Wang, Z., Kangarlu, A., Zhu, H., and Peterson, B. S. (2009). The neurophysiologica bases of emotion: An fmri study of the affectve circumplex, using emotion-denoting words. Hu- man Brain Mapping, 30:883–895.

Posner, M. (1994). Attention: the mechanisms of consciousness. Proceedings of the National Academy of Sciences USA, 91:7398– 7403.

Premack, D. and Woodruff, G. (1978). Does the chimpanzee have a theory of mind? Behaviour and Brain Sciences, 4:515–526.

Price, C. J. and Devlin, J. T. (2004). The pro and cons of labelling a left occipitotemporal region: “the visual word form area.”. Neu- roimage, 22:477–479. BIBLIOGRAPHY 387

Puelles, L. (1993). Thoughts on the development, structure, and evolution of the mammalian and avian telencephalic palliium. Philospophical transactions of the royal society London, Series B356:1583–1589.

Pulvermuller, F. (2002). A brain perspective on language mechanisms: from discrete neuronal ensembles to serial order. Progress in Neurobiology, 67:85–111.

Pulvermuller, F., Lutzenberger, W., Preissl, H., and Birbaumer, N. (1995). Spectral responses in the gamma-band: Physiological signs of higher cognitive processes? NeuroReport, 6:2059–2064.

Pylyshyn, Z. (1978). When is attribution of beliefs justiﬁed? Be- havioural and Brain Sciences, 1:127–142.

Quay, H. (1993). The psychology of undersocialized aggressive conduct disorder: a theoretical perspective. Development and Psychpathology, 5:165–180.

Quinlan, P. (2003). Connectionist Models of Development. Psy- chology Press., Hove and New York.

Raichle, M. E., MacLeod, A. M., Snyder, A. Z., Powers, W. J., Gun- sard, D. A., and Schulman, G. L. (2001). A default mode of brain function. Proceedings of the National Academy of Science USA, 98:80–93.

Raine, E. (1997). The Psychopathology of Crime. Academic Press, New York.

Rakic, P. (1988). Speciﬁcation of cerebral cortical areas. Science, 241:170–176.

Ramos, B. P. and Arnsten, A. F. (2007). Adrenergic pharmacology and cognition. Pharmacology and Therapeutics, 113:523–536.

Rauch, S. and Savage, C. (1997). Neuroimaging and neuropsychology of the striatum. Psychiatric Clinics of North America, 10:863–883.

Rauch, S., Shin, L. M., and Phelps, E. (2006). Neurocircuitry models of posttraumatic stress disorder and extinction: human neuroimaging research-past, present and future. Biological Psychi- atry, 60:376–382. 388 BIBLIOGRAPHY

Rauch, S., van der Kolk, B., Fisler, R., Alpert, N., Orr, S., and Sav- age, C. (1996). A symptom provocation study of posttraumatic stress disorder using positron emission tomography and script- driven imagery. Archives of General Psychiatry, 53:380–387. Rauch, S. L. and Drevets, W. C. (2009). Stress-Induced and Fear Circuitry Disorders: Deﬁning the Research Agenda for DSM- IV, chapter Neuroimaging and neuroanatomy of stress-induced and fear-circuitry disorders, pages 215–254. American Psychi- atric Association, Arlington, Virginia. Reiner, A. (1993). Neurotransmitter organisation and connections of turtle cortex: implications for the evolution of mammalian isocortex. Comparative Biochemistry and Physiology, 104A:735–748. Richardson, M. P., Strange, B., and Dolan, R. J. (2004). Encoding of emotional memories depends on the amygdala and hippocampus and their interactions. Nature Neuroscience, 7:278–285. Rimland, B. and Hill, A. L. (1984). Mental retardation and developmental disabilities, volume 13, chapter Idiot savants, pages 15–169. Plenum Press. Rizzolatti, G., Fadiga, L., Gallese, V., and Fogassi, L. (1996). Pre- motor cortex and the recognition of motor actions. Cognitive Brain Research, 3:131–141. Rizzolatti, G., Fogassi, L., and Gallese, V. (2006). Mirrors of the mind. Scientiﬁc American, 295:54–61. Robbins, T. (1998). The Prefrontal Cortex: Executive and Cog- nitive Functions, chapter Dissociating executive functions of the prefrontal cortex. In, Roberts, A.C. and Robbins, T.W. and Weiskrantz, L. (eds), pages 117–130. Oxford UniversityPress. Roberts, A. C., Robbins, T. W., and Weiskrantz, L. (1998). The Prefrontal Cortex: Executive and Cognitive Functions. Oxford University Press, Oxford and New York. Rolls, E. (1997). The orbitofrontal cortex. Philosophical Transac- tions of the Royal Society B, 351:1433–1443. Rolls, E. (1998). The Prefrontal Cortex: Executive and Cognitive Functions, chapter The orbitofrontal cortex. In, Roberts, A. C., Robbins, T. W., Weiskrantz, L. (eds), pages 67–86. Oxford Uni- versity Press, Oxford and New York. BIBLIOGRAPHY 389

Rolls, E. (2007). Emotion Explained. Oxford University Press, Oxford and New York. Rolls, E. T., Critchley, H. D., Mason, R., and Wakeman, E. A. (1996). Orbitofrontal cortex neurons: role in olfactory and visual association learning. Journal of Neurophysiology, 75:1970– 1981. Rolls, E. T., Hornak, J., and McGrath, J. (1994). The orbitofrontal cortex. Philosophical Transactions of the Royal So- ciety, 351:1433–1524. Rosenberg, D. R., Keshavan, M. S., and Bennett, A. E. (1998). To- ward a neurodevelopmental model of obsessive-compulsive disorder. Biological Psychiatry, 43:623–640. Rovee-Collier, C. (1993). The capacity for long-term memory in infancy. Current Directions in Psychological Science, 2:130–135. Rubia, K., Overmyer, S., Taylor, E., Brammer, M., Williams, S. C. R., Simmons, A., Andrew, C., and Bullmore, E. T. (2000). Functional frontalisation with age: Mapping neurodevelopmental trajectories with fmri. Neuroscience and Biobehavioral Re- views, 24:13–19. Rubia, K., Overmyer, S., Taylor, E., Brammer, M., Williams, S. C. R., Simmons, A., and Bullmore, E. T. (1999). Hypofrontality in attention deﬁcit hyperactivity disorder during higher-order motor control: a study with functional mri. American Journal of Psychiatry, 156:891–896. Rumelhart, D. and McClelland, J. (1986). Parallel Distributed Processing: Explorations in the Microstructure of Cognition, volume vol1. MIT Press, Cambridge, Massachusetts. Rummelhart, D. E., Hinton, G. E., and McClelland, J. L. (1987). Parallel Distributed Processing, Vol, 1, Foundations (In, Rum- melhart, D.E. and McClelland, J.L. eds), chapter A general framework for parallel distributed processing., pages 45–76. Cambridge, MA: MIT Press. Russell, J., editor (1997). Autism as an executive disorder. Oxford University Press, Oxford and New York. Rutter, M. (1983). Cognitive deﬁcits in the pathogenesis of autism. Journal of hild Psychology and Psychiatry, 24:513–531. 390 BIBLIOGRAPHY

Sachdev, P. S.and Malhi, G. S. (2005). Obsessive-compulsive behaviour: a disorder of decision-making. Australian and New Zealand Journal of Psychiatry, 39:757–763.

Sagvolden, T., Johansen, E. B. Aase, H., and Russell, V. A. (2005). A dynamic developmental theory of attention-deﬁcit hyperactivity disorder (adhd) predominantly hyperactive/impulsive and combined subtypes. Behavioral and Brain Sciences, 28(3):397–419.

Samuels, R. (2005). The Innate Mind: Structure and Contents, chapter The complexity of cognition: Tractability arguments for massive modularity. In, Carruthers, P., Laurence, S., Stich, S. (eds), pages 107–121. Oxford University Press, Oxford and New York.

Sanides, F. (1970). chapter Functional architecture of motor and sensory cortices in primates in the light of a new concept of neocortex evolution. In, The Primate Brain ( Eds). Noback, C., and Montagna, W. Appleton, New York.

Santini, M. (1975). Golgi Centennial Symposium Proceedings, chapter The subcortical projections of allocortex: Similarities in the neuronal associations of the hippocampus, the piriform cor- texand the neocortex, pages 173–193. Raven Press, New York.

Saunders, R. C., Rosene, D. L., and Van Hoesen, G. W. (1988). Comparison of the efferents of the amygdala and hippocampal formation in the rhesus monkey: reciprocal and non-reciprocal connections. Journal of Comparative Neurology, 271:185–207.

Sawaguchi, T. (1998). Attentuation of delay-period activity of monkey prefrontal neurons by an alpha2-adrenergic agonist during an oculomotor delayed-response task. Journal of Neu- rophysiology, 80:2200–2205.

Sawaguchi, T. and Goldman-Rakic, P. S. (1994). The role of d1-dopamine receptors in working memory: Local injections of dopamine antagonists into the prefrontal cortex of rhesus monkeys performing an oculomotor delayed response task. Journal of Neurophysiology, 71:515–528.

Saxena, S., Brody, A. L., Maidment, K. M., Dunkin, J. J., Col- gan, M., Alborzian, S., Phelps, M. E., and Baxter, L. R. (1999). Localized orbitofrontal and subcortical metabolic changes and BIBLIOGRAPHY 391

predictors of response to paroxetine treatment in obsessive- compulsive disorder. Neuropsychpharmacology, 21:683–693. Saxena, S., Brody, A. L., Schwartz, J. M., and Baxter, L. R. (1998). Neuroimaging and frontal-subcortical circuitry in obsessive- compulsive disorder. British Journal of Psychiatry (suppl 35), 173:26–37. Saxena, S. and Rauch, S. L. (2000). Functional neuroimaging and the neuroanatomy of obsessive-compulsive disorder. The Psy- chiatric Clinics of North America, 23(3):563–586. Schacter, D., editor (1995). Memory Distortion: How Minds, Brains, and Societies Reconstruct the Past. Harvard University Press, Cambridge, Massachusetts. Scherer, K., Schorr, A., and Johnstone, T., editors (2001). Ap- praisal Processes in Emotion. Oxford University Press, New York. Scherf, K., Sweeney, J., and Luna, B. (2006). Brain basis of developmental change in in visuospatial working memory. Journal of Cognitive Neuroscience, 18:1045–1058. Schienle, A., Stark, R., Walter, B., Blecker, C., Ott, U., Kirsch, P., Sammer, G., and Vaitl, D. (2002). The insula is not speciﬁ- cally involved in disgust processing: an fmri study. Neuroreport, 13:2023–2026. Schneider, W. and Shiffrin, R. M. (1977). Controlled and automatic human information processing. 1. detection, search, and attention. Psychological Review, 84:1–76. Schopler, E. and Mesibov, G. B., editors (1988). Diagnosis and Assessment in Autism. Plenum, New York. Schopler, E., Reichlet, R. J., and DeVellis, R. F. e. a. (1980). To- ward objective classiﬁcation of childhood autism: Childhood autism rating scale (cars). Journal of Autism and Developmen- tal Disorder, 10:91–103. Schore, A. (2003). Afffect Dysregulatin and Disorders of the Self. W.W. Norton & Company, New York and London. Schore, A. N. (1994). Afffect Regulation and the Origin of the Self. Lawrence Erlbaum Associates Publishers, Hillsdale, New Jer- sey and Hove, UK. 392 BIBLIOGRAPHY

Schulkin, J., McEwan, B. S., and Gold, P. W. (1994). Allostasis, amygdyla, and anticipatory angst. Neuroscience and Biobehav- ioral Reviews, 18(3):385–396. Schultz, W., Dayan, P., and Montague, P. R. (1997). A neural substrate of prediction and reward. Science, 275:1593–1599. Schultz, W., Tremblay, L., and Hollerman, J. R. (2000). Reward processing in primate orbitofrontal cortex and basal ganglia. Cerebral Cortex, 10:272–283. Schwartz, J. (1998). Neuroanatomical aspects of cognitive- behavioural therapy response in obsessive-compulsive disorder: An evolving perspective on brain and behaviour. British Jour- nal of Psychiatry (suppl 35), 173:38–44. Segal, G. (1996). Theories of Theories of Mind., chapter The modularity of theory of mind. In, Carruthers, P., Smith, P.K. (eds) Theories of Theories of Mind, 1996, page 151. Cambridge Uni- versity Press, Cambridge and New York. Selemon, L. and Goldman-Rakic, P. S. (1985). Longitudinal topography and interdigitation of corticostriatal projections in the rhesus monkey. Journal of Neuroscience, 5:776–794. Sesack, S. and Grace, A. (2010). Cortico-basal ganglia reward network: Microcircuitry. Neuropsychopharmacology Reviews, 35:27–47. Shah, A. and Frith, U. (1983). An islet of ability in autistic children: A research note. Journal of Child Psychology and Psychi- atry, 24:613–620. Shallice, T. (1988). From Neuropsychology to Mental Structure. Cambridge University Press, Cambridge. Shanon, B. (1988). Remarks on the modularity of mind. British Journal for the Philosophy of Science, 39:331–352. Shaw, P., Lalonde, F., Lepage, C., Rabin, C., Eckstrand, K., Sharp, W., Greenstein, D., Evans, A., and Rapaport, J. (2009). Develop- ment of cortical asymmetry in typically developing children and disruption in attention-deﬁcit/hyperactivity disorder. Archives of General Psychiatry, 66(8):888–896. Shaywitz, B., Skudlarski, P., Holahan, J., Marchione, K., Consta- ble, R., Fulbright, R., Zelterman, D., Lacadie, C., and Shaywitz, BIBLIOGRAPHY 393

S. E. (2007). Age-related changes in reading systems of dyslexic children. Annals of Neurology, 61:363–370. Shaywitz, S. (2003). Overcoming dyslexia: a new and complete science-based program for reading problems at any level. Alfred A. Knopf, New York. Shaywitz, S. E. and Shaywitz, B. A. (2008). Paying attention to reading: The neurobiology of reading and dyslexia. Develop- ment and Psychopathology, 20:1329–1349. Shima, K., Aya, K.and Mushiake, H., Inase, M., Aizawa, H., and Tanji, J. (1991). Two movement-related foci in the primate cingulate cortex observed in signal-triggered and self-paced fore- limb movements. Journal of Neurophysiology, 65(2):188–202. Shin, L. and Liberzon, I. (2010). The neurocircuitry of fear, stress, and anxiety disorders. Neuropsychpharmacology Re- views, 35:169–191. Siegel, D. (2001). Memory: an overview, with emphasis on developmental, interpersonal, and neurobiological aspects. Journal of the American Academy of Child and Adolescent Psychiatry, 40(9):997–1011. Silberstein, R. B., Farrow, M., Levy, F., Pipingas, A., Hay, D. A., and Jarman, J. C. (1998). Functional brain electrical activity mapping in boys with attention-deﬁcit/hyperactivity disorder. Archives of General Psychiatry, 55:1105–1112. Simpson, T., Carruthers, P., Laurence, S., and Stitch, S. (2005). The Innate Mind: Structure and Contents, chapter Introduction, pages 1–19. Oxford University Press, Oxford and New York. Sinzig, J., Walter, D., and Doepfner, M. (2009). Attention deﬁcit/hyperactivity disorder in children and adolescents with autism spectrum disorder: Symptom or syndrome? Journal of Attention Disorders., 13:117–126. Skuse, D. (2003). Fear recognition and the neural basis of social cognition. Child and Adolescent Mental Health, 8:50–60. Smith, E. and Jonides, J. (1999). Storage and executive processes in the frontal lobes. Science, 283:657–660. Smith, I. M. and Bryson, S. E. (1994). Imitation and action in autism: A critical review. Psychological Bulletin, 116:259–273. 394 BIBLIOGRAPHY

Smolensky, P. (1988). On the proper treatment of connectionism. Behavioural and Brain Sciences, 11:1–74.

Smolka, M. N., Schumann, G., Wrase, J., Grusser, S. M., Flor, H., Mann, K., Braus, D. F., Goldman, D., Buchel, C., and Heinz, A. (2005). Catechol-o-methyltransferase val 158 met genotype affects processing of emotional stimuli in the amygdala and prefrontal cortex. The Journal of Neuroscience, 4:836–842.

Solanto, M. (2002). Dopamine dysfunction in ad/hd. integrating clinical and basic neuroscience research. Behavioral and Brain Research, 130:65–71.

Sonuga-Barke, E. (20002). Psychological heterogeneity in ad/hd-a dual pathway model of behaviour and cognition. Behavior and Brain Research, 130:29–36.

Sonuga-Barke, E. (2003). The dual pathway model of ad/hd: an elaboration of neuro-developmental characteristics. Neuro- science and Biobehavioral Reviews, 27:593–604.

Sonuga-Barke, E. (2005). Causal models of attention- deﬁcit/hyperactivity disorder: from common simpledeﬁcits to multiple developmental pathways. Biological Psychiatry, 57:1231–1238.

Sonuga-Barke, E., Bitsakou, P., and Thompson, M. (2010). Be- yond the dual pathway model: Evidence for the dissociation of timing, inhibitory, and delay-related impairments in attention- deﬁcit/hyperactivity disorder. Journal of the American Academy of Child and Adolescent Psychiatry, 49:345–355.

Sonuga-Barke, E. J. S. and Castellanos, F. X. (2007). Sponta- neous attentional ﬂuctuations in impaired states and pathological conditions. Neuroscience and Biobehavioral Reviews, 31:977–986.

Southwick, S. M., Krystal, J. H., Morgan, C. A., Johnson, D., Nagy, L. M., and Nicolaou, A. e. a. (1993). Abnormal noradrenergic function in posttraumatic stress disorder. Archives of General Psychiatry, 50:266–274.

Spencer, R. M., Zelaznik, H. N., Diedrichsen, J., and Ivry, R. B. (2003). Disrupted timing of discontinuous but not continuous movements by cerebellar lesions. Science, 300:1437–1439. BIBLIOGRAPHY 395

Sperber, D. (1994). Mapping the Mind: Domain -Speciﬁcity in Cognition and Culture, chapter The modularity of thought and the epidemiology of representations., pages 39–67. Cambridge University Press, Cambridge.

Sperber, D. (1997). Intuititive and reﬂective beliefs. Mind and Language, 12:67–83.

Sperber, D. (2000a). chapter Metarepresentations in an evolutionary perspective. In, Sperber, D. (ed) Metarepresentations: A Multidisciplinary Perspective, pages 117–138. Oxford Uni- versity Press.

Sperber, D., editor (2000b). Metarepresentations: A Multidisci- plinary Perspective. Oxford University Press.

Sprengelmeyer, R., Rausch, M., Eysel, U. T., and Przuntek, H. (1998). Neural structures associated with recognition of facial expressions of basic emotions. Proceedings of the Royal Society London. B. Biological Sciences, 265:1927–1931.

Stahl, S. (2008). Stahl’s Essental Psychpharmacology: Neurosci- entiﬁc Basis and Practical Applications. Cambridge University Press, Cambridge and New York.

Steinberg, L. (2007). Risk taking in adolescence new perspectives from brain and behavioral science. Current Directions in Psy- chological Science, 16:55–59.

Still, G. (1902). Some abnormal psychical conditions in children. Lancet, 1:1008–1168.

Striedter, G. F. (2004). Principles of Brain Evolution. Sinauer Associates.

Stuss, D. (1992). Biological and psychological development of executive functions. Brain and Cognition, 20:8–23.

Swanson, J. M., Kinsbourne, M., Nigg, J., Lanphear, B., Ste- fanatos, G. A., Volkow, N., Taylor, E., Casey, B. J., Castellanos, F. X., and Wadhwa, P. D. (2007). Etiologic subtypes of attention- deﬁcit hyperactivity disorder: Brain imaging, molecular genetic and environmental factors and the dopamine hypothesis. Neu- ropsychological Review, 17(1):39–59. 396 BIBLIOGRAPHY

Swanson, L. (2003). The Amygdala in Brain Function, chapter The amygdala and its place in the cerebral hemisphere., pages 174–184. New York Academy of Sciences, New York.

Taerk, E., Grizenko, N., Ben Amor, L., Lageix, P., Mbekou, V., and Deguzman, R. (2004). Catechol-o-methyltransferase (comt) val108/158 met polymorphism does not modulate executive function in children with adhd. 2004, 5: 30. BMC Medical Ge- netics, 5:30.

Tager-Flusberg, H. (2000). Understanding Other Minds: Perspec- tives from Developmental Cognitive Neuroscience, chapter Lan- guage and understanding Minds: Connections in autism, pages 3–46. Oxford University Press, 2nd edition.

Tager-Flusberg, H. (2005). The Innate Mind, chapter What neurodevelopmental disorders can reveal about cognitive architecture? In, Carruthers, P. and Laurence, S. and Stich, S. (eds) The Innate Mind, pages 272–288. Oxford University Press, Oxford and New York.

Tallal, P. (1980). Auditory temporal perception, phonics, and reading disabilities in children. Brain and Language, 9:182–198.

Tallal, P., Miller, S., and Fitch, R. H. (2006). Neurobiological basis of speech: A case for the preeminence of temporal processing. Annals of the New York Academy of Sciences, 682:27–47.

Tannock, R., Ickowicz, A., and Schachar, R. (1995). Differential effects of methylphenidate on working memory in adhd children with and without comorbid anxiety. Journal of the American Academy of Child and Adolescent Psychiatry, 34:886–896.

Tau, G. and Peterson, B. (2010). Normal development of brain circuits. Neuropsychopharmacology Reviews, 35:147–168.

Teitelbaum, P., Teitelbaum, O., Nye, J., Fryman, J., and Maurer, R. G. (1998). Movement analysis in infancy may be useful for early diagnosis of autism. Proceedings of the National Academy of Science, USA., 95:13982–13987.

Thapar, A., Harrington, R., and McGufﬁn, P. (2001). Examining the comorbidity of adhd-related behaviours and conduct problems using a twin study design. British Journal of Psychiatry, 179:224–229. BIBLIOGRAPHY 397

Tiihonen, J., Hodgins, S., Vaurio, O., Laakso, M., and Repo, E. (2000). Amygdaloid volume loss in psychopathy. Society for Neuroscience Abstracts, page 2017.

Tononi, G. and Edelman, G. (1998). Consciousness and complexity. Science, 282:1846–1851.

Towbin, K. E., Peterson, B., Cohen, D., and Leckman, J. (1999). Tourette’s Syndrome-Tics, Obsessions, Compulsions: Develop- mental Psychopathology and Clinical Care, chapter Differential Diagnosis. In, Leckman, J.F. and Cohen, D.J. (eds) 1999, pages 118–139. John Wiley and Sons, New York.

Turing, A. (1938). On computable numbers, with an application to the entscheidungsproblem: A correction. Proceedings of the London Mathematical Society, 43:544–546.

Turner, M. (1997). Autism as an Executive Disorder, chapter To- wards an executive dysfunction account of repetitive behaviour in autism. In, Russell,J. (ed) Autism as an Executive Disorder, pages 57–100. Oxford University Press, Oxford and New York.

Umilta, M. A., Kohler, M., Gallese, V., and Fogassi, L. (2001). ’i know what you are doing’: a neurophysiological study. Neuron, 32:91–101.

Van Ameringen, M., Oakman, J. M., Mancini, C., and Farvolden, P. (2001). Obsessive Compulsive Disorder: A Practical Guide, chapter Obsessive compulsive spectrum disorders: From serotonin to dopamine and back again, pages 23–36. Martin Dunitz, London.

Varley, R. and Siegal, M. (2002). Language, cognition, and the nature of modularity: Evidence of aphasia. Behavioral and Brain Sciences, Commentary/Carruthers, 25:702–703.

Vijayraghavan, S., Wang, M., Birnbaum, S. G., Williams, G. V., and Arnsten, A. F. T. (2007). Inverted-u dopamine d1 receptor actions on prefrontal neurons engaged in working memory. Nature Neuroscience, 10:376–384.

Waldman, I. D. and Lilienfeld, S. O. (1991). Diagnostic efficiency of symptoms for oppositional defiant disorder and attention- deficit hyperactivity disorder. Journal of Consulting and Clini- cal Psychology, 39:732–738. 398 BIBLIOGRAPHY

Walenski, M., Tager-Flusberg, H., and Ullman, M. T. (2006). Un- derstanding Autism: From Basic Neuroscience to Treatment, chapter Language in Autism. In (eds) Moldin, S.O. and Ruben- stein, J.L.R., pages 175–203. Taylor and Francis, London and New York. Wallentin, M., Roepstorff, A., Glover, R., and Burgess, N. (2006). Parallel memory systems for talking about location and age in precuneus, caudate and Broca’s area. NeuroImage, 32:1850– 1864. Wang, A., Ting, M. A., Dapretto, M., Hariri, A., Sigman, M., and Bookheimer, S. (2004a). Neural correlates of facial affect processing in children and adolescents with autism spectrum disorder. Journal of the American Academy of Child & Adolescent Psychiatry, 43:481–490. Wang, M., Ramos, B., Paspalas, C., Shu, Y., Simen, A., Duque, A., Vijayraghavan, S., Brennan, A., Dudley, A., and Nou, E. (2007). Alpha2a-adrenoceptor stimulation strengthens working memory networks by inhibiting c amp-hcn channel signaling in prefrontal cortex. Cell, 129:397–410. Wang, M., Vijayraghavan, S., and Goldman-Rakic, P. (2004b). Se- lective D2 receptor actions on the functional circuitry of working memory. Science, 303:853–856. Weinberger, D. (2003). Dopamine in the Pathophysiology and Treatment of Schizophrenia, chapter Dopamine, the prefrontal cortex, and a genetic mechanism of schizophrenia., pages 129– 154. In Kapur and Lecrubier (2003). Weingarten, S. M. and Cummings, J. L. (2001). Frontal- Subcortical Circuits in Psychiatric and Neurological Disorders, chapter Psychosurgery of frontal-subcortical circuits., pages 421–435. In Lichter and Cummings (2001). Weiskrantz, L. (1997). Consciousness lost and found: a neuropsychological exploration. Oxford University Press, New York. Werker, J. and Tees, R. (2005). Speech perception as a window for understanding plasticity and committment in language systems of the brain. Developmental Psychobiology, 46:233–251. Werker, J. F. and Lalonde, C. E. (1988). Cross-language speech perception: initial capabilities and developmental change. De- velopmental Psychology, 52:672–683. BIBLIOGRAPHY 399

Werker, J. F. and Tees, R. C. (1999). Influences on infant speech processing: Toward a new synthesis. Annual Review of Psychol- ogy, 50:509–535. White, N. M. and McDonald, R. J. (2002). Multiple parallel memory memory systems in the brain of the rat. Neurobiology, Learning, and Memory, 77:125–184. WHO (1993). ICD-10 Classification of Mental and Behavioural Disorders. Geneva. Wicker, B., Keysers, C., Plailly, J., Royet, J., Gallese, V., and Rizzo- latt, G. (2003). Both of us disgusted in my insula: The common neural basis of seeing and feeling disgust. Neuron, 40:655 – 664. Wilens, T. E., Faraone, S. V., and Biederman, J. (2004). Attention- deficit/hyperactivity didorder in adults. JAMA, 292:619–623. Williams, L. M., Brown, K. J., Das, P., Boucsein, W., Sokolov, E. N., Brammer, M. J., Olievieri, G., Peduto, A., and Gordon, E. (2004). The dynamics of cortico-amygdala and autonomic activity over the experimental time course of fear perception. Cognitive Brain Research, 21:114–123. Williams, L. M., Phillips, M. L., Brammer, M. J., Skerrett, D., Lagopoulos, J., Rennie, C., Bahramali, H., Olivieri, G., David, A. S., Peduto, A., and Gordon, E. (2001). Arousal dissociates amygdala and hippocampal fear responses: Evidence from simultaneous fMRI and skin conductance recording. Neuroimage, 14:1070–1079. Wing, L. (1988). Diagnosis and assessment in autism, chapter The continuum of autistic characteristics. In, (eds) Schopler, E., Mesibov, G,B. 1988, pages 91–110. Plenum, New York. Wingo, A. and Ghaemi, S. (2007). A systematic review of rates in diagnostic validity of comorbid adult attention- deficit/hyperactivity disorder and bipolar disorder. Journal of Clinical Psychiatry, 68:1776–1784. Woods, R., Dodrill, C. B., and Ojemann, G. A. (1988). Brain injury, handedness, and speech lateralization in a series of amybarbi- tal studies. Annals of Neurology, 23(5):1255–1260. Wortis, J., editor (1984). Mental retardation and developmental disability. Bruner/Mazel, London. 400 BIBLIOGRAPHY

Wynn, T. and Coolidge, F. (2002). The role of working memory in skilled and conceptual thought. Behavioral and Brain Sciences, Commentary/Carruthers, 25:703–704.

Xu, B., Grafman, J., Gaillard, W. D., Ishii, K., Vega-Bermudez, F., Pietrini, P., Reeves-Tyer, P., DiCamillo, P., and Theodore, W. (2001). Conjoint and extended neural networks for the computation of speech codes: The neural basis of selective imparment in reading words and pseudowords. Cerebral Cortex, 11:267– 277.

Yang, C. R. and Seamans, J. K. (1996). Dopamine d1 receptor actions in layers v-v1 rat prefrontal cortex neurons in vitro: modulation of dendritic-somatic signal integration. The Journal of Neuroscience, 16:1922–1935.

Yehuda, R. (2001). Biology of posttraumatic stress disorder. Jour- nal of Clinical Psychiatry, 62(Suppl 17):41–46.

Yehuda, R. (2002). Current concepts: Post-traumatic stress disorder. The New England Journal of Psychiatry, 346(2):108–114.

Yehuda, R., Bierer, L. M., and Scmeidler, J. (2000). Low cortisol and risk for ptsd in adult offstring of holocaust survivors. The American Journal of Psychiatry, 157(8):1252–1259.

Yelnik, J. (2002). Functional anatomy of the basal ganglia. Move- ment Disorders, 17(Suppl3):S15–21.

Yin, H. H. and Knowlton, B. (2006). The role of the basal ganglia in habit formation. Nature Reviews/Neuroscience, 7:464–476.

Young, L. J. and Wang, Z. (2004). The neurobiology of pair bonding. Nature Neuroscience, 7:1048–1054.

Young, S. E., Friedman, N. P., Miyake, A., Wilcutt, E. G., Corley, R. P., and Haberstick, B. C .and Hewitt, J. K. (2009). Behav- ioral disinhibition: Liability for externalizing spectrum disorders and its genetic and enironmental relation to response inhibition across adolescence. Journal of Abnormal Psychology, 118:117–130.

Zahm, D. (2006). The evolving theory of basal forebrain functional-anatomical “macrosystems”. Neuroscience and Biobehavioral Reviews, 30:148–172. BIBLIOGRAPHY 401

Zahm, D. S. and Brog, J. S. (1992). On the significance of sub- territories in the ”accumbens” part of the rat ventral striatum. Neuroscience, 50:751–767. Zald, D. H. and Kim, S. W. (1996). Anatomy and function of the orbital frontal cortex, 1: Anatomy, neurocircuitry, and obsessive- compulsive disorder. Journal of Neuropsychiatry and Clinical Neurosciences, 8:125–138. Zametkin, A. J., Nordahl, T. E., Gross, M., King, A., Semple, W. E., Hamburger, S., and Cohen, R. M. (1990). Cerebral glucose metabolism in adults with hyperactivity of childhood onset. New England Journal of Medicine, 323:1361–1366. Zeigler, H. (2008). Neuroscience of Birdsong. Cambridge Univer- sity Press, Cambridge and New York. Zeigler, P. H. and Marler, P. (2004). Behavioral Biology of Bird- song, volume 1016 of Annals of the New York Academy of Sci- ences. The New York Academy of Sciences, New York, New York. Zelazo, P., Burack, J., Benedetto, E., and Frye, D. (1996). Theory of mind and rule use in individuals with down’s syndrome: A test of the uniqueness and specificity claims. Journal of Child Psychology and Psychiatry, 37(4):479–84. Zepf, F. (2009). Attention deficit-hyperactivity disorder and early- onset bipolar disorder: two faces of one entity? Dialogues in Clinical Neuroscience, 11:63–72. 402 BIBLIOGRAPHY Author Index

APA (1994), 2 Arnsten et al. (1998), 138, Achenbach et al. (1998), 204, 215 213 Arnsten et al. (1999), 298 Adleman et al. (2002), 99 Arnsten (1997), 215 Adolphs et al. (1996), 229 Arnsten (2000), 156 Adolphs (2001), 183 Arnsten (2006), 196, 197, 215 Adolphs (2002), 255–257 Arnsten (2007), 127, 148, Adolphs (2003), 89 218, 330 Adret (2004), 79 Aronov et al. (2008), 79–82 Alexander and Crutcher Astington and Jenkins (1990), 33, 109, 112, (1999), 245 117, 287, 324 Aston-Jones et al. (1999), 298 Alexander et al. (1986), 33, Au-Young et al. (1999), 126 107, 109, 110, 135, Augustine (1996), 258 142, 165, 274, 290, Avery et al. (2000), 207 326 Baars (1989), 16, 17 Alheid and Heimer (1988), Bachevalier (1991), 258 168 Baddeley and Hitch (1974), Allen and Courchesne (2003), 137 263 Baddeley (1986), 17, 87, 89 Allen et al. (2003), 242, 268, Baddeley (2000), 137 277, 278 Baddeley (2001), 87 Alter and Hen (2009), 303, Ballaban-Gil and Tuchman 310 (2000), 263 Angold et al. (1999), 313, 316 Balleine and Kilcross (2006), Aragona et al. (2003a), 234 105 Aragona et al. (2003b), 234 Balleine and O’Doherty Arbib (1989), 24 (2010), 45, 49, 104– Arbib (2005), 66, 67, 235 106, 338 Ardila (2009), 54 Balleine et al. (2007), 45 Arnett (1992), 100 Balleine (2001), 105 Arnsten and Dudley (2005), Bar-Gad et al. (2003), 43 215 Barbas (1992), 138 Arnsten and Li (2005), 206 Barkley (1990), 191 Arnsten et al. (1996), 207 Barkley (1997), 148, 213

403 404 AUTHOR INDEX

Barkow and Tooby (1992), 25 Blair (2006), 227, 228 Baron-Cohen et al. (1985), Bland (2009), 118 241, 247, 259 Boder and Jarrico (1982), 92 Baron-Cohen et al. (1986), Bonda et al. (1996), 253 247 Booth et al. (2003), 95 Baron-Cohen et al. (1999), Botterill and Carruthers 253 (1999), 26 Baron-Cohen et al. (2000a), Botvinick et al. (2001), 170 255, 258 Boucher (1981), 243 Baron-Cohen et al. (2000b), Boucher (1989), 246–248 230, 241 Boucher (2001), 244 Baron-Cohen (1987), 247 Boyer and Lienard (2006), Baron-Cohen (1988), 243, 277 244, 248 Bremer (1992), 279 Baron-Cohen (1995), 23, 26, Bremner et al. (1997), 299 252 Bremner (2006), 294, 295 Baumrind (1971), 226 Broca (1878), 151, 163 Baxter et al. (1990), 278 Brock et al. (2002), 266 Baxter (1992), 287, 288, 324 Brothers (1990), 253, 255, Bechtel (2008), 30, 31, 46 258 Belmonte and Baron-Cohen Brown and Barlow (2005), 6 (2004), 263 Brown (1977), 55 Belmonte and Yurgelun-Todd Buccino et al. (2001), 67 (2003), 263, 265 Buccino et al. (2004), 68 Belmonte et al. (2004a), 249, Bunge et al. (2002), 99 262–264 Bush et al. (2000), 170 Belmonte et al. (2004b), 264– Bush (2010), 191, 202, 210, 266 211, 319 Belmonte (2000), 263 Buzsaki (2006), 100–104 Bennet (1990), 24 Byrne and Whiten (1990), Bickerton (2002), 87 250 Biederman et al. (1995), 205 Calvin (2004), 56, 57 Bilder et al. (2004), 195 Cao et al. (2006), 95 Blackburn (1988), 226 Carlsson (2001), 218 Blair and Frith (2000), 222, Carr et al. (2003), 68, 176, 224–226 177 Blair et al. (1997), 223, 226 Carruthers and Chamberlain Blair et al. (1999), 224, 227 (2000), 24, 25 Blair et al. (2001), 223 Carruthers (2002a), 26, 85, Blair (1995), 224, 226 86 Blair (2001), 221–223 Carruthers (2002b), 88 Blair (2003), 224 Carruthers (2005), 26, 27, 64 AUTHOR INDEX 405

Caspi et al. (2003), 310 Damasio (1994), 153, 222 Caspi et al. (2008), 232 Damasio (2000), 12 Castellanos and Tannock Dapretto et al. (2006), 175 (2002), 190–192 Davidson et al. (2000), 184 Castellanos et al. (1996), 287, Davis and Whalen (2001), 324 309 Castellanos et al. (2002), 203 DeGelder (1987), 248 Castellanos et al. (2006), 202 DeLong (1990), 111 Castellanos et al. (2008), 200, DeWit et al. (2006), 105 210, 214 Dehaene and Naccache Castellanos (1997), 207 (2001), 13–20 Cavanna (2007), 120 Dehaene et al. (1998a), 17, Changeux (2005), 56 18, 48 Charney (2004), 309 Dehaene et al. (1998b), 17, 31 Cheng and Durand (2004), 80 Dehaene et al. (2005), 94 Cheng (1986), 86 Dehaene (2007), 21, 47, 48 Chomsky (1988), 25 Dehaene (2009), 19, 95, 96 Chomsky (1995), 85 Dennett (1978), 244, 246 Churchland and Sejnowski Dennett (1991), 6, 8, 9, 27, 48 (1999), 37–39 Denny and Rapport (2001), Churchland (2002a), 6, 9, 35, 206 36 DiChiara (2002), 233 Clark and Rutter (1981), 261 Diamond et al. (1994), 78 Cohen et al. (1996), 139 Diamond et al. (2004), 204 Cohen et al. (1998), 207 Dias et al. (1996), 223 Cohen et al. (2002), 125, 126 Doupe and Kuhl (1999), 80 Coltheart and Rastle (1994), Durston et al. (2006), 137 93 Dyck et al. (2001), 230 Coltheart et al. (1985), 93 Edelman and Tononi (2000), Coltheart (1999), 22 13, 18 Conboy et al. (2008), 78 Edelman (2004), 10–12, 61 Conway and Christiansen Egan et al. (2001), 218, 219 (2001), 75, 76 Eibl-Eibesfeldt (1970), 223 Cornil et al. (2002), 216 Eimas et al. (1971), 76, 91 Courchesne et al. (2003), 252, Eimas (1975), 78 264 Elman (1990), 36, 37 Creamer et al. (2003), 309 Elman (1993), 31 Crick and Koch (2005), 99 Elman (1995), 39, 40 Critchley et al. (2000), 264 Emery et al. (1997), 251 Cummings (1993), 33, 34, 109 Etkin and Wager (2007), 304 Dadds et al. (1988), 261 Everitt and Robbins (2005), Dadds et al. (2008), 269 185, 186 406 AUTHOR INDEX

Eysenck (1964), 225 176 Faraone et al. (2005), 217 Gat (1998), 46, 47 Farran (2001), 130 Geier and Luna (2009), 137 Fassbender et al. (2009), 210, Giedd et al. (1996), 132 212 Giedd et al. (1999), 133 Ferrari et al. (2005), 70 Glowinski et al. (1984), 206 Fiez and Peterson (1998), 94 Goldman-Rakic et al. (1990), Fliessbach et al. (2007), 174 144, 287, 324 Floresco et al. (2003), 195, Goldman-Rakic et al. (2000), 196, 298, 308 126 Floresco et al. (2004), 195 Goldman-Rakic (1987a), 159 Foa and Kozak (1986), 310 Goldman-Rakic (1987b), 15, Fodor and Pylyshyn (1988), 41, 124, 125, 143, 148, 30, 31, 53, 65 205 Fodor (1983), 16, 22, 23, 26, Goldman-Rakic (1987c), 13, 92 159, 199 Fodor (1985), 30 Goldman-Rakic (1988), 19, Fodor (1998), 25 20, 296 Fodor (2000), 26 Goldman-Rakic (1991), 148 Fogassi et al. (2005), 176 Goldman-Rakic (1994a), 136 Forbes et al. (2009), 231, 235 Goldman-Rakic (1994b), 135 Forssberg et al. (2006), 204 Goldman-Rakic (1995), 148 Fowles (1988), 227 Goldman-Rakic (1996), 205 Freud (1895), 292 Goldman-Rakic (1998a), 123 Friedman and Karam (2009), Goldman-Rakic (1998b), 144, 309, 311 199, 200 Friedman et al. (1999), 298 Goodale and Milner (1992), Frith and Frith (1999), 244, 96 250, 251 Goodyear and Hynd (1992), Frith and Happe (1994), 259 313 Frith (1989a), 249, 266 Goto and O’ Donnell (2002), Frith (1989b), 248, 249, 261 308 Fullana et al. (2009), 283 Grace (1995), 306 Funahashi et al. (1989), 138 Grace (2000), 306 Funahashi (2007), 137–139 Grace (2001), 156, 307 Fuster (1989), 15 Graham and Rutter (1973), Fuster (1997), 135, 139 316 Gallese et al. (119), 173 Gray et al. (2002), 179 Gallese et al. (2004), 67, 68, Graybiel and Rauch (2000), 173 45, 48 Gallese (2005), 71, 175 Graybiel (1995), 47, 242, 305 Gallese (2006), 69, 70, 173– Graybiel (1998), 49, 50 AUTHOR INDEX 407

Gray (1971), 226 Heimer (2003), 107–109, 117, Gray (1987), 227 164, 167 Grossberg (1999), 298 Herba and Phillips (2004), Grèzes et al. (2003), 67 157, 229, 230, 235 Gusnard and Raichle (2001), Herbert et al. (2004), 263 211 Hermann and Chambria Gusnard et al. (2001), 211 (1980), 74 Haber and McFarland (2001), Hermer and Spelke (1994), 111, 112, 115, 116, 86 317, 318 Herschkowitz (2000), 229 Haber and Wolfer (1992), Hezler and Grifﬁn (1981), 286, 324, 326 258 Haber et al. (1995), 113 Hobson (1986), 243 Haber et al. (2000), 106, 113, Hoffman and Saltzstein 114 (1967), 226 Haber (2003), 110, 114–117, Hollerman et al. (1998), 155 166, 172, 202, 290, Hollerman et al. (2000), 154, 317, 318, 326 155, 230, 231 Hale and Tager-Flusberg Holscher and Munk (2009b), (2003), 253 118, 119 Hampson et al. (2006), 201 Holscher and Munk (2009c), Hare et al. (2005), 199 119 Hare (1970), 225 Hopﬁeld (1982), 42, 142 Hare (1978), 226 Houk and Wise (1995), 42 Harmer et al. (2001), 224 Huttenlocher (1994), 130 Harnad (1990), 35 Hyman (2007), 5 Hauser and Fitch (2004), 70, Iacoboni et al. (1999), 67 175 Insel (2010), 7 Hauser et al. (2002), 63, 64, Jackendoff (2002), 85 70, 73, 175 Jackson et al. (2000), 181 Hebb (1949), 35, 119, 228 Joel (1994), 319 Heimer and Van Hoesen Jorm et al. (2000), 310 (2006), 164–167 Joseph (2000), 73–75 Heimer et al. (1982), 110 Kandel and Freed (1989), 222 Heimer et al. (1991), 168 Kanner and Eisenberg Heimer et al. (2008), 11, 33, (1956), 259 45, 107, 151, 152, Kanner (1943), 259, 260 162–164, 168, 170, Karmilloff-Smith et al. 171, 311, 318, 319, (2003), 229 327 Karmiloff-Smith (1992), 28, Heimer (1972), 108 48, 71, 72 Heimer (1991), 168 Karmiloff-Smith (1994), 28, 408 AUTHOR INDEX

30 283 Keil et al. (1999), 263 Leckman (2002), 285, 286, Kellendonk et al. (2006), 288, 289, 325 134–136 Lee and Wilson (2002), 123 Kenway and Wilson (2001), Leiner et al. (1991), 20 122 Leiner et al. (1993), 191, 192 Kiehl et al. (2001), 224 Leslie and Frith (1988), 246, Kling and Brothers (1992), 247 257 Leslie and Roth (1993), 252 Klorman et al. (1991), 208 Leslie et al. (2004), 178 Kluver and Bucy (1939), 257 Leslie (1987), 244, 247 Kohlberg and Gilligan (1971), Leslie (1994), 23 129 Leung et al. (2000), 99 Kohler et al. (2002), 173 Levy and Goldman-Rakic Kopell and Greenberg (2008), (1999), 112 284 Levy and Hobbes (1981), 212 Koziol and Budding (2009), Levy and Krebs (2006), 37, 31, 32, 43–45, 49, 50, 50, 269 54, 55, 90, 106, 107, Levy et al. (1987), 91, 92, 97, 146, 147, 174, 180, 236 211, 213, 214, 320, Levy et al. (2009), 92 338 Levy (1980), 211 Krebs (1995), 42 Levy (1989), 91, 92, 97, 236 Krueger et al. (2005), 313 Levy (2004), 50, 196, 305 Krystal et al. (1995), 293, 294 Levy (2007a), 269 Kuhl et al. (2008), 77, 78 Levy (2007b), 204, 206, 217 Lahey et al. (2008), 315 Levy (2007c), 204, 267 Lalonde and Werker (1995), Levy (2008), 218 78 Levy (2010), 118, 321, 332 Lamme (2006), 12, 28, 29 Lewis and Boucher (1988), Lawler (2001), 93 248 LeDoux and Gorman (2001), Lewis (1997), 140 301 Liberman and Mattingly LeDoux et al. (1989), 158 (1985), 59, 60, 77, 78, LeDoux (1995), 227 92, 132, 179 LeDoux (2002), 157, 158 Liberman and Wahlen Leary and Hill (1996), 244 (2000), 179, 235 Leckman and Cohen (1999), Liberman et al. (1967), 58, 286, 287, 323, 324, 59, 179 326 Liberman (1974), 59 Leckman et al. (2003), 283 Liberzon and Sripada (2008), Leckman et al. (2009), 282, 303 AUTHOR INDEX 409

Lieberman and Eisenberger Miller and Cohen (2001), (2006), 170 140–142, 181 Lieberman et al. (2007), 184 Miller et al. (1996), 141 Lieberman (2008), 32, 33 Mills et al. (2004), 232 Livingstone et al. (1991), 92 Milne et al. (2002), 262 Lock and Peters (1996), 76 Milner and Goodale (1995), Lorenz (1981), 223 251 Luna et al. (2001), 132 Minsky (2006), 57 MacLean (1949), 151, 167 Mirenowicz and Schultz MacLean (1952), 151, 164 (1997), 155 MacLean (1972), 55, 61 Miyake and Shah (1999), 137 MacLean (1990), 162 Modell et al. (1989), 274–276 MacWhinney (1998), 77 Mogenson and Wu (1988), Mann and Paulsen (2009), 274 118 Mogenson et al. (1980), 161– Marcus (2004), 83–85 163, 305 Marcus (2006), 22 Mooney (2004), 79 Marios and Ivanoff (2005), 21 Moore and Guan (2001), 78 Marsh et al. (2006), 99 Moron et al. (2002), 216 Marsh et al. (2009), 100 Morris et al. (1996), 227, 257 Mason et al. (2001), 295, 302 Morris et al. (1999), 139 Mattay et al. (2003), 219 Moscovich (1995), 22 Mayberg et al. (1999), 155 Munakata and Stedron McCormick et al. (2003), 120, (2001), 41 121 Munakata et al. (2003), 40 McEvoy et al. (1988), 59 Murphy et al. (1996), 215 McFarlane et al. (2002), 296– Nakamura et al. (2006), 95 299, 311 Newell (1980), 65 McGowan et al. (2009), 304 Nigg and Casey (2005), 191, McLoughlin et al. (2007), 190 198, 199 McMaster et al. (2008), 276, Nigg (2003), 315 277 Nigg (2006), 315 Mehler and Dupoux (1994), Norman and Shallice (1980), 71, 253 261 Mervis et al. (2003), 252 Norris (1990), 250 Mesulam and Mufson (1984), O’Doherty et al. (2001), 105 258 O’Keefe (1991), 227 Middelton and Strick (2002), Ohnishi et al. (2000), 258 117 Olsson and Phelps (2007), Middleton and Strick (2001), 300–302 109 Oschner et al. (2002), 181– Miguel et al. (1997), 280 184 410 AUTHOR INDEX

Ostlund and Balleine (2007), Plaisted et al. (2003), 262 105 Pollack and Kistler (2002), Owen et al. (1996), 124 230 Panksepp (1998), 61 Ponzi (2008), 147 Papez (1937), 151 Posner et al. (2009), 171 Parent and Hazrati (1995), Posner (1994), 17 287, 324 Premack and Woodruff Pascual-Leone and Walsh (1978), 69, 173, 244, (2001), 29 246 Pasher (1994), 21 Price and Devlin (2004), 94 Patrick et al. (1993), 226 Puelles (1993), 58 Patrick (1994), 224, 225 Pulvermuller et al. (1995), Patterson and Newman 245, 246 (1993), 226 Pulvermuller (2002), 89, 90 Pauls et al. (1986), 288, 325 Pylyshyn (1978), 244 Pauls et al. (1991), 286, 323 Quay (1993), 225–227 Paus et al. (1999), 131, 132 Raichle et al. (2001), 120, 210 Pavuluri and Sweeney Raine (1997), 225 (2008), 6, 7, 132, Rakic (1988), 134 225, 321, 322 Ramos and Arnsten (2007), Penn and Povinelli (2007), 61, 308 64, 65 Rauch and Drevets (2009), Penn et al. (2008), 61, 63 310, 311 Pennington and Ozonoff Rauch and Savage (1997), (1996), 222 281, 282 Perner (1988), 246 Rauch et al. (1996), 302 Peterson et al. (2002), 99 Rauch et al. (2006), 303 Petrides and Pandya (2002), Reiner (1993), 58 139 Richardson et al. (2004), 300 Petrides et al. (1993), 136 Rimland and Hill (1984), 259 Petrides (1994), 139 Rizzolatti et al. (1996), 173 Petrides (1998), 145 Rizzolatti et al. (2006), 174 Phelps (2004), 299, 300 Robbins (1998), 143 Phillips and Young (1998), 68 Roberts et al. (1998), 123, 145 Phillips et al. (1997), 68 Rolls et al. (1994), 222 Phillips et al. (2003a), 229 Rolls et al. (1996), 112 Phillips et al. (2003b), 156, Rolls (1997), 223 229 Rolls (1998), 152, 153 Piaget (1955), 28 Rolls (2007), 140 Pierce et al. (2001), 254 Rosenberg et al. (1998), 276 Pinker and Prince (1988), 65 Rovee-Collier (1993), 292 Pinker (1997), 25 Rubia et al. (1999), 230 AUTHOR INDEX 411

Rubia et al. (2000), 230 (2008), 92–95 Rumelhart and McClelland Shaywitz et al. (2007), 94 (1986), 34 Shaywitz (2003), 94 Rummelhart et al. (1987), 42 Shima et al. (1991), 251 Rutter (1983), 243 Shin and Liberzon (2010), Sachdev (2005), 273 303, 304 Sagvolden et al. (2005), 192, Siegel (2001), 292 196, 206 Silberstein et al. (1998), 208 Samuels (2005), 25, 26 Simpson et al. (2005), 23 Sanides (1970), 58 Sinzig et al. (2009), 331 Santini (1975), 163 Skuse (2003), 157 Saunders et al. (1988), 166 Smith and Bryson (1994), 244 Sawaguchi and Goldman- Smith and Jonides (1999), Rakic (1994), 215 181 Sawaguchi (1998), 138 Smolensky (1988), 35, 53, 54 Saxena and Rauch (2000), Smolka et al. (2005), 233 272, 273, 284 Solanto (2002), 198, 206 Saxena et al. (1998), 274, 276 Sonuga-Barke and Castel- Saxena et al. (1999), 184 lanos (2007), 210, 214 Scherer et al. (2001), 182 Sonuga-Barke et al. (2010), Scherf et al. (2006), 137 194, 195 Schienle et al. (2002), 68 Sonuga-Barke (0002), 193 Schneider and Shiffrin Sonuga-Barke (2003), 191, (1977), 17 193 Schopler et al. (1980), 258 Sonuga-Barke (2005), 193 Schore (1994), 159, 160 Southwick et al. (1993), 299 Schore (2003), 160 Spencer et al. (2003), 199 Schulkin et al. (1994), 294 Sperber (1994), 23 Schultz et al. (1997), 153, Sperber (1997), 245 154, 199 Sperber (2000b), 245 Schultz et al. (2000), 109, Sprengelmeyer et al. (1998), 166, 172, 264 68 Schwartz (1998), 283 Stahl (2008), 29, 327–330 Segal (1996), 22 Steinberg (2007), 100 Selemon and Goldman-Rakic Still (1902), 191 (1985), 112 Striedter (2004), 58 Sesack and Grace (2010), 319 Stuss (1992), 229 Shah and Frith (1983), 249 Swanson et al. (2007), 217 Shallice (1988), 16, 17, 225 Swanson (2003), 166 Shanon (1988), 23, 24 Taerk et al. (2004), 203, 232 Shaw et al. (2009), 133, 134 Tager-Flusberg (2000), 245, Shaywitz and Shaywitz 252 412 AUTHOR INDEX

Tager-Flusberg (2005), 245, 89 252, 253 Wicker et al. (2003), 68 Tallal et al. (2006), 132 Wilens et al. (2004), 204, 205 Tallal (1980), 92 Williams et al. (2001), 159 Tannock et al. (1995), 313 Williams et al. (2004), 158 Tau and Peterson (2010), 99, Wingo and Ghaemi (2007), 136 205 Teitelbaum et al. (1998), 262 Wing (1988), 249 Thapar et al. (2001), 313 Woods et al. (1988), 133 Tiihonen et al. (2000), 224 Wynn and Coolidge (2002), 87 Tononi and Edelman (1998), Xu et al. (2001), 95 17 Yang and Seamans (1996), Towbin et al. (1999), 286, 288, 171 323, 324 Yehuda et al. (2000), 302 Turing (1938), 96 Yehuda (2001), 302 Turner (1997), 259–261 Yehuda (2002), 302 Umilta et al. (2001), 173 Yelnik (2002), 110 Van Ameringen et al. (2001), Yin and Knowlton (2006), 279 280–282 Young and Wang (2004), 233, Varley and Siegal (2002), 87 234 Vijayraghavan et al. (2007), Young et al. (2009), 314, 315 201 Zahm and Brog (1992), 319 WHO (1993), 5 Zahm (2006), 168, 170 Zald and Kim (1996), 274 Waldman and Lilienfeld Zametkin et al. (1990), 205 (1991), 313 Zeigler (2008), 80 Walenski et al. (2006), 240– Zelazo et al. (1996), 252 242, 244, 267 Zepf (2009), 205 Wallentin et al. (2006), 88, de Villiers and Pyers (2002), 89, 91 245 Wang et al. (2004a), 175 Wang et al. (2004b), 138 Wang et al. (2007), 207, 216 Weinberger (2003), 218 Weingarten and Cummings (2001), 109 Weiskrantz (1997), 17 Werker and Lalonde (1988), 76 Werker and Tees (1999), 76, 77, 91, 92 Werker and Tees (2005), 78 White and McDonald (2002), List of Figures

1.1 Structual anatomy of the brain ...... 5 1.2 Revised Alexander Model, Lichter and Cummings, 2001.Copyright Guilford Press. Reprinted with permission of The Guilford Press...... 34 1.3 Recurrent Network ...... 38

3.1 Isocortical and greater limbic lobe projections. Reprinted with permission from Heimer et al. 2008 108 5.1 Major basal forebrain macrosystems, Adapted from Zahm, 2006, Heimer et al. 2008 ...... 165 5.2 Striatopallidum and Extended amygdala. Adapted from Zahm 2006, Alheid and Heimer, 1988, Heimer et al. 2008 ...... 169

6.1 Age vs CPT Omission Errors: From Levy, F. 1980 . . 212 6.2 Age vs Mean Reaction Time: From Levy, F. 1980 . . 212 9.1 Recurrent Neural Network ...... 273

413