Relationships Between Circulating Hormones and Brain Abnormalities in Schizophrenia and Restoration of Brain Activity by Raloxifene Treatment

Home , AP5

Relationships between circulating hormones and brain abnormalities in schizophrenia and restoration of brain activity by raloxifene treatment

Ellen F Ji

Submitted in total fulfilment of the requirements of the degree of Doctor of Philosophy

School of Psychiatry Faculty of Medicine

November 2016

‘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 (this is applicable to doctoral theses only). 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.'

Signed ……………………………………………......

13 June 2017 Date ……………………………………………......

AUTHENTICITY STATEMENT

‘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.’

Signed ……………………………………………......

Date ……………………………………………...... 13 June 2017 ...... II III

Acknowledgements

First and foremost I would like to thank my academic supervisors Professor Thomas Weickert, Professor Cyndi Shannon Weickert and Professor Rhoshel Lenroot for their support and thoughtful input into my work throughout my candidature. Thank you so much Tom and Cyndi for the opportunity to pursue my Ph.D. in the Schizophrenia Research Laboratory where you have put forth countless of hours to help me produce publications and a thesis I am proud of. The past few years have been key for my professional development and I will always look back with gratitude while cherishing the time I spent in your lab.

I would also like to extend my appreciation to all the past and current members of the Schizophrenia Research Laboratory for being such a great bunch of people to work with – your friendships and laughs have made my time here so enjoyable! And of course thanks to Wai-Kit, Mic and Andrew for all the IT help over the years.

A special thank you to Marc for being there for me unconditionally. And finally, a huge thank you to my parents - without your patience, understanding, support and most of all love, the competition of this work would not have been possible.

Abstract

Schizophrenia is a severe neuropsychiatric disorder with marked deficits in cognitive and social functioning. Mounting evidence suggests that sex steroid hormones are strongly implicated in the pathophysiological underpinnings of schizophrenia. Moreover, grey matter deficits and abnormal brain function are integral components of disease neuropathology, a hypothesis that is supported by findings from postmortem and neuroimaging studies. This thesis used a multidisciplinary approach integrating structural neuroimaging, functional neuroimaging and molecular biology with the overall aim of determining how hormones may be associated with a number of core deficits and brain abnormalities routinely described in schizophrenia. In testing the hypothesis that some people with schizophrenia exhibit abnormal circulating hormone levels, I found that male patients had decreased testosterone levels and female patients had increased testosterone levels as compared with their respective healthy control groups, while DHEA levels were increased in both sexes of patients relative to controls.

I showed for the first time that cortisol/DHEA ratios were inversely related to grey matter volume in the prefrontal cortex and hippocampus, suggesting that hypothalamic– pituitary–adrenal axis dysfunction can contribute to key neuroanatomical deficits in schizophrenia. Further, endogenous testosterone levels were associated with neural activation during facial emotion recognition in male patients and not in male controls, suggesting that higher levels of testosterone may be beneficial for cortical processing during events key for healthy social functioning in patients. Since the beneficial effects of testosterone on cortical functions could be mediated by either androgen or estrogen action in the brain, I examined the extent to which adjunctive treatment with a brain estrogen receptor agonist, raloxifene, could enhance neural activation during facial V

emotion recognition in patients. Raloxifene increased prefrontal and hippocampal activation during emotion recognition in males and females with schizophrenia, demonstrating the drug’s ability to alter neural activity associated with impairments in emotion processing of patients. The correlational and experimental studies from this thesis provide evidence that peripheral blood hormones may be informative of brain pathology in schizophrenia and further implicate the role of steroid hormones in disease neuropathology and possibly for treatment.

Publications arising from this thesis

Journal Articles

Ji, E., Weickert, C. S., Lenroot, R., Kindler, J., Skilleter, A. J., Vercammen, A., Weickert, T. W. (2016). Adjunctive selective estrogen receptor modulator increases neural activity in the hippocampus and inferior frontal gyrus during emotional face recognition in schizophrenia. Transl Psychiatry, 6, e795.

Ji, E., Weickert, C. S., Lenroot, R., Catts, S. V., Vercammen, A., White, C., Weickert, T. W. (2015). Endogenous testosterone levels are associated with neural activity in men with schizophrenia during facial emotion processing. Behavioural Brain Research, 286, 338-346.

Conference Abstracts

Ji E, Weickert CS, Weickert TW. Abnormal cortisol-dehydroepiandrosterone ratios correlate with hippocampal and dorsolateral prefrontal cortex volume changes in schizophrenia, FENS Forum of Neuroscience, Copenhagen, July 2016.

Ji E, Weickert CS, Lenroot R, Weickert TW. Cortisol/DHEA ratio as a peripheral marker of HPA axis activity: relationship to grey matter volume reduction in schizophrenia, UNSW Brain Sciences Symposium, April 2016.

Ji E, Weickert CS, Lenroot R, Vercammen A, White C, Gur R, Weickert TW. Adjunctive selective estrogen receptor modulator increases neural activity in the hippocampus and inferior frontal gyrus during emotional face recognition in schizophrenia, Schizophrenia International Research conference, Florence, April 2016.

Ji E, Weickert CS, Lenroot R, Vercammen A, White C, Gur R, Weickert TW. Adjunctive selective estrogen receptor modulator increases neural activity in the hippocampus and inferior frontal gyrus during emotional face recognition in schizophrenia, Inter-university Neuroscience & Mental Health Conference, Sydney, September 2015.

Ji E, Weickert CS, Lenroot R, Vercammen A, White C, Gur R, Weickert TW. Adjunctive selective estrogen receptor modulator increases neural activity in the hippocampus and inferior frontal gyrus during emotional face recognition in schizophrenia, Biological Psychiatry Australia meeting, Sydney, September 2015.

Ji E, Weickert CS, Lenroot R, Vercammen A, White C, Gur R, Weickert TW. Effects of adjunctive raloxifene treatment on brain activity during facial emotion processing in schizophrenia, ISN-APSN Joint Biennial Meeting, Cairns, August 2015.

Ji E, Weickert CS, Lenroot R, Vercammen A, White C, Gur R, Weickert TW. Endogenous testosterone levels are associated with neural activity in men with schizophrenia during a facial emotion processing task, Schizophrenia International Research Conference, Florence, April 2014.

VII

Abbreviations used in this thesis

ACTH: adrenocorticotropic hormone BDNF: brain-derived neurotrophic factor BOLD: blood-oxygen-level dependent CPZ: chlorpromazine CSF: cerebrospinal fluid CT: computerised tomography DHEA: Dehydroepiandrosterone DLPFC: dorsolateral prefrontal cortex DSM: Diagnostic and Statistical Manual of Mental Disorders DR2: dopamine receptor subtype 2 EEG: electroencephalography fMRI: functional magnetic resonance imaging GC-MS: gas chromatography–mass spectrometry GR: glucocorticoid receptor HPA: hypothalamic-pituitary-adrenal IFG: inferior frontal gyrus IL: interleukin MRI: magnetic resonance imaging PFC: prefrontal cortex ROI: region of interest SCID: Structured Clinical Interview for DSM Disorders SERM: Selective estrogen receptor modulator SHBG: sex hormone-binding globulin TIV: total intracranial volume VBM: voxel-based morphometry

VIII

Table of Contents

Originality Statement ...... II Acknowledgements ...... III Abstract ...... IV Publications arising from this thesis ...... VI Abbreviations used in this thesis ...... VII Table of Contents ...... VIII List of figures ...... XI List of tables ...... XII 1.0 Introduction ...... 1 1.1 Schizophrenia ...... 1 1.1.1 A brief history ...... 1 1.1.2 Onset and disease course ...... 2 1.1.3 Prognosis and symptomatology ...... 4 1.1.3.1 Cognitive deficits and social impairment ...... 5 1.2 Neuroimaging ...... 7 1.2.1 Structural MRI and voxel-based morphometry...... 7 1.2.2 Structural abnormalities in schizophrenia ...... 8 1.2.3 Functional MRI ……………………………………………………….… 10 1.3 Sex steroid hormones in schizophrenia ...... 12 1.3.1 Gender differences and role of sex steroid hormones ...... 12 1.3.2 Biosynthesis of sex hormones ...... 13 1.3.3 Estrogen hypothesis of schizophrenia ...... 14 1.3.4 Adjunctive estrogen treatment ...... 16 1.3.5 Selective estrogen receptor modulator treatment ...... 17 1.3.6 Endogenous testosterone in schizophrenia...... 19 1.3.7 Neurobiological effects of testosterone and estrogen ...... 21 1.3.8 Sex steroids and brain activity ...... 21 1.4 Dehydroepiandrosterone and response to stress ...... 22 1.4.1 Stress signalling in schizophrenia ...... 22 1.4.1.1 Early life stressors ...... 22 IX

1.4.1.2 The hypothalamic–pituitary–adrenal axis ...... 24 1.4.1.3 Impaired stress response in schizophrenia ...... 26 1.4.1.4 Cortisol levels in schizophrenia ...... 26 1.4.2 Dehydroepiandrosterone ...... 27 1.4.2.1 Secretion changes across lifespan ...... 28 1.4.2.2 Mechanisms of action ...... 29 1.4.2.3 Neurobiological effects ...... 30 1.4.2.4 Molar ratio of cortisol to DHEA ...... 31 2.0 Aims of this thesis ...... 33 3.0 Endogenous hormone levels in people with schizophrenia compared with healthy controls ...... 35 3.1 Abstract ...... 35 3.2 Introduction ...... 37 3.3 Materials and Methods ...... 41 3.4 Results ...... 44 3.5 Discussion ...... 48 4.0 Cortisol-dehydroepiandrosterone ratios correlate with hippocampal and prefrontal cortex volume reductions in schizophrenia ...... 54 4.1 Abstract ...... 54 4.2 Introduction ...... 55 4.3 Materials and Methods ...... 57 4.4 Results ...... 62 4.5 Discussion ...... 68 5.0 Endogenous testosterone levels are associated with neural activity in men with schizophrenia during facial emotion processing ...... 77 5.1 Abstract ...... 77 5.2 Introduction ...... 79 5.3 Materials and Methods ...... 82 5.4 Results ...... 87 5.5 Discussion ...... 95

6.0 Adjunctive selective estrogen receptor modulator increases neural activity in the hippocampus and inferior frontal gyrus during emotional face recognition in schizophrenia ...... 102 6.1 Abstract ...... 102 6.2 Introduction ...... 104 6.3 Materials and Methods ...... 108 6.4 Results ...... 113 6.5 Discussion ...... 120 7.0 General Discussion ...... 129 7.1 Basal hormone levels in schizophrenia – implications and mechanisms ...... 130 7.2 Can the cortisol/DHEA ratio be used as a peripheral biomarker of some aspect of schizophrenia? ...... 136 7.2.1 Brain volume, cognition and symptoms ...... 136 7.2.2 Possible mediating factors between cortisol/DHEA and brain volume ... 137 7.3 Hormone therapy for schizophrenia: where we stand now and potential for continued use ...... 139 7.3.1 Is there still therapeutic potential for DHEA augmentation in schizophrenia? ...... 140 7.3.2 Implications of circulating testosterone levels when considering the development of new treatments...... 142 7.3.3 The efficacy of raloxifene for treating schizophrenia ...... 143 7.4 Methodological considerations ...... 145 7.4.1 Quantifying hormone levels ...... 145 7.4.2 Strengths and weaknesses of MRI ...... 148 7.4.3 Clinical difficulties: recruitment, confounding variables, heterogeneity...... 150 7.5 Future directions ...... 152 7.6 Conclusions ...... 154 8.0 References ...... 156

List of figures Figure 1.1 BOLD fMRI signal 11 Figure 1.2 Details about the biochemistry of steroid synthesis 14 Figure 3.1 Comparison of hormone levels in patients versus healthy controls 46 Figure 4.1 Correlation between DHEA and cortisol in people with schizophrenia 65 Figure 4.2 Association of cortisol/DHEA ratios to grey matter volume in schizophrenia 67 Figure 5.1 Coronal slices depicting areas of significant neural activation 90 Figure 5.2 Direct comparison of neural activity between groups and correlations between beta weights and testosterone levels 93 Figure 6.1 Task-related neural activity 117 Figure 6.2 Effects of raloxifene on BOLD activity in ROIs 120

Supplementary figures SFigure 4.1 Glass brain displaying anatomical regions of interest used for analysis 75 SFigure 4.2 Comparison of grey matter volume in ROIs 76 SFigure 6.1 Trial design 127 SFigure 6.2 Emotional face recognition paradigm 127

XII

List of tables Table 3.1 Demographic variables, cognitive and clinical characteristics of the whole sample 45 Table 3.2 Comparison of symptom severity in male patients versus female patients 47 Table 3.3 Associations of testosterone and estrogen to symptom severity in patients 47 Table 3.4 Correlations between mean daily CPZ equivalents dose and peripheral hormone levels in schizophrenia 47 Table 4.1 Demographic variables, clinical characteristics and peripheral markers of the whole sample 64 Table 4.2 Correlations of PANSS scores and cognitive tests with cortisol/DHEA in patients and healthy controls 66 Table 5.1 Demographic and clinical characteristics of the whole sample 88 Table 5.2 Mean reaction times and performance accuracy for people with schizophrenia and healthy controls during the facial emotional recognition task 89 Table 5.3 Whole brain analysis showing regions of activation during viewing of angry versus non-threat facial expressions in healthy controls and people with schizophrenia 91 Table 5.4 Demographic and clinical characteristics of the healthy men and men with schizophrenia 94 Table 6.1 Demographics, clinical characteristics and blood analyses of patients 114 Table 6.2 Whole brain analysis showing regions of activation during anger processing in people with schizophrenia during raloxifene treatment and placebo condition 118 Table 6.3 Region of interest analyses showing significant treatment effects 119 Supplementary tables STable 4.1 Breakdown of antipsychotic medications 73 STable 4.2 Regions with significant group differences in grey matter volume 74 STable 6.1 Mean reaction times and performance accuracy 128 1

1.0 Introduction

1.1 Schizophrenia

1.1.1 Brief history

Two centuries ago an English tea broker, James Tilly Matthews, revealed a secret world in which he was being controlled and tormented by malicious villains who implanted a magnet into his brain and used a machine that forced thoughts into his head against his will. Details of his experience are described in John Haslam’s book Illustrations of

Madness (1810), in what came to be the first documented case of schizophrenia

(Haslam, 1810). However, there is evidence from ancient documents describing symptoms of a similar nature dating back several thousand years (Okasha & Okasha,

2000), therefore it is generally believed that the disease has prevailed throughout the history of mankind even before the separation of races (Horrobin, 1998), which may explain why the prevalence of schizophrenia is similar cross-culturally. A German psychiatrist, Dr. Emil Kraepelin, was the first to classify mental disorders into different categories and used the term "dementia praecox" (1893) for individuals who displayed symptoms, including a loss of cognitive capacity in the young, that are now associated with schizophrenia (Berrios & Hauser, 1988). However, the term dementia praecox was somewhat misleading as the condition could occur later in adulthood, as well as in adolescence or young adulthood, and did not always lead to mental deterioration.

Consequently, the disease was renamed in 1908 by a Swiss psychiatrist, Eugen Bleuler, who first coined the term “schizophrenia” - which translates roughly as "splitting of the mind" – to reflect the disconnect between thought processes, emotion, behaviour and the split from reality. Bleuler grouped general characteristics of schizophrenia into the four A’s (affectivity- a blunting of emotion, ambivalence- a lack of motivation, association- disorganised thought, and autism- withdrawal from social contact) which 2

influenced diagnosis for decades to come (McGlashan, 2011). Controversies surrounding the loosely defined diagnostic criteria led to the creation of the first edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM), which was published in 1952. Modern psychiatry went on to use research-based medical models and included a re-evaluation of schizophrenia diagnostic criteria in the DSM version III

(American Psychiatric Association, 1980) that determined that positive and/or negative symptoms of the illness must be present for a minimum of 6 months in order to meet diagnostic criteria. These significant advancements and changes have remained relatively unchanged (Wilson, 1993). Today, a diagnosis of schizophrenia is made according to guidelines laid out in the DSM-V (American Psychiatric Association, 2013) or the International Classification of Diseases (ICD-10).

1.1.2 Onset and disease course

While the aetiology of this multifactorial disease is still unclear, it is well established that schizophrenia arises from the interplay of multiple factors including genetics, environmental and psychological variables, and possibly precipitated by hormonal changes that alter the brain's chemistry (van Os & Kapur, 2009). The emergence of symptoms typically occurs during the critical years of adolescence and young adulthood, with females experiencing a second peak in first diagnosis around menopause (Häfner,

2003; van Os & Kapur, 2009). Prior to the emergence of psychotic symptoms, there is the occurrence of a prodromal phase lasting months to years (Keshavan & Cornblatt,

2010), during which individuals exhibit social withdrawal, diminished academic performance and positive symptoms at subclinical levels (Yung & McGorry, 1996;

Yung et al., 2008). Before this period, few have major intellectual difficulties whereas the majority experience only subtle or non-existent deficits in cognition and social 3

behaviour (Done et al., 1994; Jones & Cannon, 1998; Weickert et al., 2000). Only a portion of those who display prodrome symptoms will go on to develop schizophrenia; specifically, "ultra-high risk" individuals have a 30-35% risk of developing psychosis within 1-2 years (Zdanowicz et al., 2014). Although there is evidence of brain pathology including structural and functional abnormalities prior to the onset of psychosis, a diagnosis at this stage is nearly impossible, which poses a challenge for early intervention strategies. The neuroimaging studies in this thesis included participants with schizophrenia who were chronically ill; therefore, in the studies contributing to this work it is unclear when precisely the neural abnormalities appeared and how they have progressed over the course of the illness. Moreover, the clinical course of the disease is variable between patients. First-episode psychosis (emergence of psychotic symptoms) may be followed by a period of remission with the vast majority (> 80%) of diagnosed patients experiencing a relapse of further psychotic episodes (Robinson et al., 1999). The first year following diagnosis often sees a stabilisation of positive symptoms (Eaton et al., 1995) while negative symptoms become progressively worse over time (Fenton & McGlashan, 1994). Longitudinal imaging studies have shown that progressive brain tissue loss following onset may be related to relapse duration (Andreasen et al., 2013; Guo et al., 2015) but not the number of relapses, suggesting that the brain itself, in addition to behaviour, may be important to study as an indicator of disease pathology. In addition to changes in the brain during the prodrome period and throughout the disease progression, there are also changes among peripheral markers found in the blood that may fluctuate in parallel with the course of the illness. To date, there is limited research on how peripheral markers

(hormones in particular) may be related to changes in the brain in people with schizophrenia. 4

1.1.3 Prognosis and symptomatology

Schizophrenia is a severe, chronic mental illness that affects approximately 0.7% of the worldwide population (Saha et al., 2005). There is considerable variability in its clinical presentation, disease course and response to treatment. Schizophrenia is characterised by symptoms such as auditory and visual hallucinations, delusions, blunted affect, reduced motivation and impaired cognitive functioning (Peralta & Cuesta, 2001).

Schizophrenia has an enormous and devastating impact on affected individuals, their respective families and on society. The rate of suicide among people with schizophrenia is approximately 5%, 12 times higher than that of the general population (Hor & Taylor,

2010; Saha et al., 2007). The financial burden due to schizophrenia is at least 1.45 billion Australian dollars per annum in direct health costs, lost productivity and welfare costs (Carr et al., 2003). Antipsychotics are the first line of treatment and are generally effective in reducing positive symptoms by blocking the excessive activity of dopamine at dopamine D2 receptors (Creese et al., 1976; Johnstone et al., 1978; Seeman et al.,

1976). However, many antipsychotics produce debilitating side effects including obesity and diabetes and leave many people with residual symptoms including disabling cognitive impairments (Kapur & Mamo, 2003). Comorbidity with conditions such as depression, anxiety disorders, drug and alcohol abuse is high among people with schizophrenia (Buckley et al., 2009). Substance abuse comorbidity predominates as 47% of people with schizophrenia have a lifetime diagnosis of comorbid substance abuse.

The average life expectancy for people with schizophrenia is 10-25 years less than that of healthy individuals, caused primarily by suboptimal lifestyles including unhealthy diet, excessive smoking and alcohol use, and lack of exercise (Laursen et al., 2012).

Normal lifestyles typically do not manifest in people with schizophrenia as they are 5

often unemployed (Rosenheck et al., 2006), with many suffering from poverty and homelessness.

1.1.3.1 Cognitive deficits and social impairment

Prominent deficits are observed across multiple cognitive domains in people with schizophrenia. The most described deficits include impairments in working memory, executive function, verbal memory, attention and language (Depp et al., 2007; Palmer et al., 1997; Reichenberg et al., 2009; Saykin et al., 1991). Individuals with schizophrenia typically have a lower IQ score by at least 10 points compared to unaffected individuals

(Goldberg et al., 1995; Leeson et al., 2011). Intellectual decline has been demonstrated by longitudinal studies (Lubin et al., 1962; Schwartzman & Douglas, 1962) and studies of first-episode patients (Goldberg et al., 1988). A sub-population of patients score within the normal range on neuropsychological tests (preserved intellect), although there is still indication of some cognitive deterioration in this group when compared to matched healthy controls (Weickert et al., 2000). Moreover, signs of cognitive impairment can appear prior to illness onset (Fuller et al., 2002) and among at-risk children (Dickson et al., 2014), highlighting the importance of early intervention.

Deficits in working memory, verbal memory and attention become more severe

(Corigliano et al., 2014) with multiple relapses of psychotic episodes (Eberhard et al.,

2003). Impaired cognition is long-lasting, resistant to full remission and is one of the strongest predictors of functional disability in schizophrenia as it impedes with one’s ability to function independently, resulting in impoverished quality of life (Green, 1996).

In spite of the magnitude of this problem, there is still a palpable gap in our current knowledge base concerning what mechanisms can prevent and/or reverse cognitive 6

deficits in people with schizophrenia. Therefore, there is an urgent need to identify agents that may modify cognition and the associated cortical/subcortical activity.

Social cognition is also strongly associated with functional outcome in people with schizophrenia (Couture et al., 2006; Fett et al., 2011), encompassing emotion processing, social perception, attributional bias (perceptual errors that lead to biased interpretations of one’s own actions and others) and theory of mind (intuitive understanding of one's own and other people's mental states). Deficits in social functioning are predictors of poorer quality of life, relapse and course. Efforts to improve social impairment in schizophrenia include pharmacological treatments such as adjunctive oxytocin

(Guastella et al., 2015; Woolley et al., 2014) and training approaches such as cognitive remediation therapy (Penn et al., 2005; Peyroux & Franck, 2014). Social cognition and the ability to communicate appropriately are equally as important to understand and target as normal cognition, though it has been much less investigated. The neurobiology of social cognition involves complex interactions among several brain areas beginning from basic perception of social stimuli to more advanced evaluation and response.

However, the neural bases of social deficits in schizophrenia are just beginning to be understood and have been examined using paradigms that probe social functioning including perception of social cues from faces and voices, eye tracking studies of faces and mentalising studies (inferring the mental states of others) (Asgharpour et al., 2015;

Green et al., 2015; Savla et al., 2012).

Because of the strong link between social functioning outcomes and emotional face processing (Ihnen et al., 1998), this thesis will in part examine the neurobiology of emotion processing deficits in people with schizophrenia using a face identification 7

paradigm. Impaired ability to perceive facial emotions has been correlated with severity of negative symptoms in a meta-analysis (Kohler et al., 2010). In Chapters 5 and 6, participants were administered an emotional face identification task where they were asked to identify the emotion displayed (neutral, happy, fear, anger, sad). Because people with schizophrenia have particular difficulty identifying and processing facial emotions of negative valence (Kohler et al., 2003; Mandal et al., 1998), this thesis examines the functional neural correlates of identifying angry facial expressions.

1.2 Neuroimaging

1.2.1 Structural magnetic resonance imaging and voxel-based

morphometry

Early research on brain anatomy in schizophrenia relied on postmortem methods. Now, scientists are able to study brain anatomy in vivo with the development of brain imaging technologies and computerised analytic methods. Computerized tomography (CT), a low resolution imaging technique, was the first form of computerised neuroimaging.

The earliest reports from CT studies included enlarged lateral ventricles in chronic patients with schizophrenia (Johnstone et al., 1976), which has been continuously replicated over the decades. However, CT was unable to provide adequate details of soft tissue; therefore, it was not sufficient do determine where loss of brain volume may be.

Magnetic resonance imaging (MRI), a higher resolution technique, was later developed along with the analytic method called voxel-based morphometry (VBM). Unlike manual tracing, VBM is a user-independent automated method providing unbiased evaluations.

It uses a voxel-by-voxel statistical approach to provide an evaluation of the probability of whether each voxel in the structural scan represents grey matter, white matter or cerebrospinal fluid (CSF). VBM enables users to quantify the density or concentration 8

of various tissue classes for a cohort or between groups. To overcome the large variability in human brain size and anatomy among subjects, VBM registers each brain to a template (spatial normalisation). The normalised images are then segmented into tissue classes of grey matter, white matter and CSF. Subsequently, the images are smoothed by a selected smoothing kernel so that each voxel represents the average of itself and its neighbour (Ashburner & Friston, 2000). An additional but optional step is modulation of the segmented images with the Jacobian determinants derived from spatial normalisation, in order to preserve the original tissue volume. Once the images are pre-processed, analyses can be conducted where the volumes are compared across brains on a voxel-wise basis. The studies in this thesis capitalise on MRI’s ability to localise areas of differences in brain volume with high spatial resolution.

1.2.2 Structural abnormalities in schizophrenia

VBM studies have provided valuable data on grey and white matter distribution in the brain and have highlighted the importance of conceptualising schizophrenia as a neurodegenerative disorder. Structural abnormalities throughout the brain of people with schizophrenia have been studied extensively, but there is considerable heterogeneity in the effect sizes and locations of cortical and subcortical differences across studies (Haijma et al., 2012). Longitudinal studies exploring brain morphological abnormalities alongside genetic and environmental risk factors can help identify the causes and progression of schizophrenia, and how they may contribute to clinical symptoms and impairments in social and cognitive functioning. Recent large-scale meta-analysis reports may be the most reliable approach to determining the extent of structural brain deficits in people with schizophrenia. The largest multicentre brain MRI analysis of chronically ill patients with schizophrenia to date (2028 schizophrenia 9

patients and 2540 healthy controls across 15 centres worldwide) found that the most prominent reductions in schizophrenia patients were in the hippocampus, amygdala, thalamus and accumbens (van Erp et al., 2015). Further, patients had significantly larger pallidum and lateral ventricles, which were associated with illness duration and age.

Interestingly, hippocampal volume reductions were more severe in studies with a greater proportion of unmedicated patients, suggesting that antipsychotic medication may ameliorate hippocampal volume deficits. Other factors including stress can also affect hippocampal volume, though there is currently no easily accessible biological marker that can predict hippocampal loss. Several studies have found correlations between grey matter reductions and clinical, cognitive and affective measures. A variety of structural findings have been associated with positive symptoms (especially hallucinations) including grey matter reductions in the superior temporal cortex, insula, thalamus and cerebellum (Neckelmann et al., 2006; Shapleske et al., 2002; Wright et al.,

1995), and grey matter increases in basal ganglia volume (Mamah et al., 2007). Smaller total intracranial grey matter volume has been linked with poorer cognitive outcome measures in chronically ill patients (Sigmundsson et al., 2001). Given the key role of the frontal cortex in executive processing and attention, reductions in this area may contribute to their impairments. In support of this, reductions in dorsolateral prefrontal cortex (DLPFC) volume were shown to be related to cognitive deficits in people with schizophrenia (Cannon et al., 2002). Given these findings, Chapter 4 of this thesis will replicate previous findings of prefrontal and hippocampal volumetric reductions in schizophrenia and will extend those findings by demonstrating a relationship between volumetric reductions and peripheral markers, suggesting that hypothalamic-pituitary- adrenal (HPA) activity may contribute to the illness.

1.2.3 Functional MRI

Functional magnetic resonance imaging (fMRI) is a neuroimaging method that detects changes in blood flow (Gore, 2003) and has revolutionized cognitive neuroscience by allowing researchers to examine task-dependent and task-independent neural activity, which is particularly important for diseases characterised by aberrant activity such as schizophrenia. When neurons become active they require energy and because the brain itself does not store glucose, blood flows in to transport glucose along with oxygenated hemoglobin molecules. The hemodynamic response (change in the magnetic resonance

(MR) signal from neuronal activity) lags behind the neuronal event by 1-2 seconds, the amount of time it takes for the vascular system to respond to the brain's need for glucose

(signaled via glutamate release), see Figure 1.1. Glutamate release causes a change in calcium ion concentration and affects astrocytes and supporting cells, subsequently releasing nitric oxide which causes arterioles to expand and draw in more blood. The rise in oxygen-rich blood to the active region lasts for 4-6 seconds after which oxygen levels drop slightly below the baseline level before returning to normal. Deoxygenated hemoglobin is more magnetic than oxygenated hemoglobin; therefore, the difference yields a changed magnetic resonance signal. Returning neurons to their original polarised state requires glucose-produced energy in order to actively pump ions across the neuronal cell membranes.

The underlying principle of fMRI is the coupling of cerebral blood flow and neuronal activation such that there is an increase of blood flow to regions of the brain that are in use. The blood-oxygen-level dependent (BOLD) contrast that is measured comes from oxygen-rich blood displacing oxygen-depleted blood (Roy & Sherrington, 1890). This change in blood flow occurs in close proximity to the site of neural activity (within 2-3 11

mm). fMRI has become a standard tool for research because of its high spatial resolution (ability to discriminate between nearby locations).

Figure 1.1 BOLD fMRI signal (a) The proposed relationship between synaptic activity, neurotransmitter recycling and metabolic demand and (b) the effect of deoxygenated hemoglobin on the MRI signal. Reproduced from Heeger & Ress, 2002 (Heeger & Ress, 2002). Abbreviation: Hb = hemoglobin.

fMRI is an invaluable non-invasive instrument that enables examination of the functional brain network underlying behavioral deficits manifested in schizophrenia. It 12

has been used in schizophrenia studies to determine regions of brain activity during cognitive (Stevens et al., 1998; Weinberger et al., 1996) and affective processing (Lee et al., 2014), and also at rest (resting-state fMRI) (Karbasforoushan & Woodward, 2012) - all of which show abnormalities (hypo- and/or hyper-activation in patients). These failures to appropriately activate neural systems are assumed to lead to behavioural deficits in patients. fMRI has bolstered support of a genetic component in schizophrenia because specific polymorphisms are associated with differences in brain function (Baig et al., 2010). It has also demonstrated that a variety of environmental stressors can be associated with differences in BOLD response (Mothersill & Donohoe,

2016). Successful application and analysis of fMRI may be used to help characterise schizophrenia alongside the presentation of clinical symptoms. fMRI has been particularly useful in helping me understand the pathophysiology of schizophrenia, because I can test whether quantifiable biomarkers are correlated with the underlying neural activity of cognition or social functioning that are abnormal in the disease.

1.3 Sex steroid hormones in schizophrenia

1.3.1 Gender differences and the role of sex steroid hormones

There is abundant evidence from clinical behavioural data to molecular neuropathology suggesting that sex steroid hormones are pivotal factors contributing to the development and course of schizophrenia. The onset of the illness typically occurring around the same age as puberty (Burke et al., 1990; Lieberman et al., 2001) suggests that sex steroid-triggered maturational changes may unmask vulnerability (Styne, 1994;

Weinberger, 1987). Gender differences have been widely examined in recent decades; however, the extent of the differences and relevance to improving prognosis is still a subject of debate. Differences in the incidence of the illness may differ by gender. 13

Aleman et al. performed a meta-analysis demonstrating that the incidence risk ratios for men to develop schizophrenia relative to women were 1.42 (Aleman et al., 2003).

However, there are other recent studies that did not find gender differences in the prevalence of schizophrenia (Perälä et al., 2007; Saha et al., 2005). A possible explanation for the discrepancy between the incidence and prevalence could be related to higher rates of suicide completion in men than in women. Male patients typically experience increased severity of negative symptoms compared to female patients (Gur et al., 1996; Shtasel et al., 1992). Regarding brain abnormalities, male patients generally have greater ventricular enlargement and cortical atrophy compared with healthy men than female patients compared with healthy women (Goldstein & Link, 1988; Seidman et al., 1997). Women typically respond better to antipsychotic treatment compared to men (Angermeyer et al., 1990; Salokangas, 1983). Sex differences and development during the reproductive period suggest that sex steroid signalling may be altered in the brains of individuals with schizophrenia. In order to understand how sex steroid signalling may be altered in the brain, researchers can study sex steroids in vivo in humans by examining circulating sex steroids. Indeed, peripheral androgens and estrogens have been implicated in the pathophysiology of schizophrenia, although their ability to influence/predict brain pathology is only recently being studied.

1.3.2 Biosynthesis of sex hormones

Sex hormones are synthesised from cholesterol by steroid metabolizing enzymes, see

Figure 1.2. Cholesterol is first converted into pregnenolone by the mitochondrial enzyme cholesterol side chain cleavage P450scc during the rate-limiting and hormonally regulated step of all steroid hormones. The enzyme P450c17 then assists in the conversion of pregnenolone into 17-OH pregnenolone and subsequently 14

dehydroepiandrosterone (DHEA). DHEA is quickly converted into its more stable sulfate ester DHEAS by the enzyme hydroxysteroid sulfotransferase. DHEAS can be reconverted into DHEA by steroid sulfatase. Androstenedione is converted into testosterone and estrone via 17β-hydroxysteroid dehydrogenase and aromatase, respectfully, which can both be converted into estradiol.

Figure 1.2 Details about the biochemistry of steroid synthesis Cholesterol is the precursor of all steroid hormones. DHEA converts to testosterone via androstenedione which then converts to estrogen. Reproduced from (Chin et al., 2008). Steroid abbreviations: PREG = pregnenolone; PROG = progesterone; DHEA = dehydroepiandrosterone; AE = androstenedione; T = testosterone; E1 = estrone; E2 = 17β-estradiol. Enzyme abbreviations: P450scc = Cytochrome P450 side chain cleavage; P450c17 = Cytochrome P450 17α-hydroxylase/C17,20 lyase; 3β-HSD = 3β- hydroxysteroid dehydrogenase/isomerase; 17β-HSD = 17β-hydroxysteroid dehydrogenase; Aromatase = Cytochrome P450 aromatase.

1.3.3 Estrogen hypothesis of schizophrenia

The estrogen hypothesis posits that estrogen plays a protective role against the development and course of schizophrenia, such that higher levels of estrogen may be beneficial. Among female patients, low estrogen levels are related to worse negative symptoms and greater cognitive impairment (Häfner, 2003; Ko et al., 2006). Further, 15

the rate of relapse is higher during periods when estrogen levels are low (Mendrek et al.,

2012). If estrogen indeed plays a neuroprotective role against psychosis in females because their circulating levels of estradiol are higher than males then we would expect female patients to have lower levels of circulating estradiol levels compared to healthy women. However, there is conflicting evidence as some report lower than normal estradiol levels in women with schizophrenia (Hoff et al., 2001) while others report no difference compared with healthy women, suggesting that more research evaluating the circulating levels of estrogen in women with schizophrenia may be needed to resolve these disparate reports. Importantly, it has been shown that even if hormone levels are not grossly different, they still can play a modulating role in aspects of the disease. A negative correlation between low endogenous estrogen levels and higher required dose of antipsychotic treatment in pre-menopausal women has been previously reported

(Gattaz et al., 1994). Moreover, Hoff et al. found that higher estrogen levels were positively correlated with cognitive function in females with schizophrenia (Hoff et al.,

2001). Confounding reports on circulating hormone levels and heterogeneity in estrogen response of female patients may be explained by differences in individuals' estrogen receptor genotype. Estrogen receptor-α (ESR1) polymorphisms are associated with cognitive performance in midlife and with the risk of developing cognitive impairments in non-psychotic old women (Olsen et al., 2006). Reductions in ESR1 mRNA in cortical and subcortical brain regions in patients suffering from major mental illness have been found, with frontal cortical ESR1 mRNA levels significantly correlated with age of onset of schizophrenia (Perlman et al., 2005). These studies suggest that circulating estrogen levels in schizophrenia may be partially explained by genetics and hormonal receptors in the brain. In order to characterise our sample of participants and determine whether their circulating estrogen levels were abnormal, we measured estrogen levels in 16

the blood and conducted a comparative analysis with healthy controls (Chapter 3).

Because estrogen-like compounds can improve cognitive functioning in both male and female patients (Weickert et al., 2015) and because estrogen receptors are present in the brains of both males and females, estrogen signalling is relevant for both sexes and, moreover, it will be meaningful to discern whether there is therapeutic benefit in people who do not display hypoestrogenism.

1.3.4 Adjunctive estrogen treatment

Estrogen may protect against age-associated decline in cognitive functioning. Changes in brain activity during language and memory tasks have been found in healthy middle- aged and postmenopausal women using hormone therapy (Resnick & Maki, 2001;

Senanarong et al., 2002). In a cross-sectional study comparing postmenopausal women with schizophrenia who used hormone replacement therapy versus those who were non- users, women who were taking hormones required a lower dose of antipsychotics and reported less severe negative symptoms (Lindamer et al., 2001).

These findings, along with evidence that estrogen levels influence symptom severity, led to the first clinical trial of adjunctive estrogen in women with schizophrenia over two decades ago (Kulkarni et al., 1996). In this initial trial, Kulkarni et al. discovered that 8 weeks of the synthetic 17β-estradiol derivative, ethinylestradiol, taken orally along with antipsychotics, significantly reduced positive symptoms in post-menopausal women with schizophrenia. In a later trial, the same group administered transdermal estradiol, which again resulted in significant improvements in positive symptoms as assessed by the PANSS (Kulkarni et al., 2001). In their recent and largest study to date,

183 premenopausal female patients received either 100 μg/day or 200 μg/day of 17

adjunctive transdermal estradiol for 8 weeks, with the highest dose yielding the largest effect of reduced positive symptoms (Kulkarni et al., 2015). The initial studies tested estrogen in older postmenopausal women, though it was not yet known whether it would yield the same efficacy in younger patients and/or in men. There have been reports of similar beneficial effects by other research groups in premenopausal women of childbearing age (Akhondzadeh et al., 2003); however, a placebo-controlled, double- blind study in 46 hypoestrogenic women (pre and post-menopausal) failed to replicate the beneficial effect of estradiol on symptom scores or relapse rates (Bergemann et al.,

2005). Estradiol has also been used in men with schizophrenia, with reports indicating improved general psychopathology compared to a placebo group although the study sample was relatively small (n = 53) and there may still be a need to refine the type, dose and administration route for estrogen therapy in men (Kulkarni et al., 2011).

Limitations of estrogen therapy include the associated increase risk of breast and uterine cancer (Barrett-Connor & Stuenkel, 2001); therefore, it may be suboptimal for long- term use.

1.3.5 Selective estrogen receptor modulator treatment

Due to the risks associated with estrogen (Barrett-Connor & Stuenkel, 2001), including undesirable adverse events of long-term adjuvant estrogen in men, researchers began using selective estrogen receptor modulators (SERMs) that stimulate estrogen receptors in the brain (and bone) while acting as an antagonist in other tissues (breast and uterus)

(Maximov et al., 2013). Raloxifene is a second generation SERM that works as an estrogen agonist in brain and has positive effects on brain function (by increasing neural activity) during delayed verbal memory and recognition memory in non-psychotic individuals (Goekoop et al., 2005). Favourable results for positive, negative and 18

cognitive symptoms have been demonstrated in women with schizophrenia using raloxifene (Huerta-Ramos et al., 2014; Kulkarni et al., 2010; Usall et al., 2011).

Raloxifene improved measures of attention and memory in both men and women with schizophrenia in a double-blind placebo-controlled crossover study; however, in that study there was no significant improvement in clinical symptoms compared to a placebo

(Weickert et al., 2015). Varied reports from estrogen and estrogen-like compound trials may be due to differences in severity of symptoms, treatment duration and/or dose, antipsychotic type, method of administration (oral vs. transdermal) and the overall heterogeneity of the disease. Nonetheless, data from multiple trials shows strong support for the use of estrogen-like agents as an adjunctive treatment for schizophrenia.

While treatment with raloxifene is promising, the neural mechanisms underlying its benefit are only beginning to be examined. The brain systems involved in therapeutic response can best be identified through brain imaging, which is dependent on the specific cognitive tasks employed. To date, adjunctive raloxifene appears to increase brain activity in the hippocampus during implicit learning in schizophrenia (Kindler et al., 2015). Raloxifene appears to be able to facilitate activity in schizophrenia in brain areas that are not normally employed (increasing medial temporal lobe activity during latter trial of probabilistic association learning), suggesting neural compensation in these patients. In that study, one may have expected beneficial raloxifene response in the frontal striatal brain regions, as would be the case in healthy controls, but raloxifene did not restore the fronto-striatal hypofunction. This raises the question as to if the frontal cortex of people with schizophrenia is able to respond to raloxifene.

In addition to probabilistic association learning and other cognitive tasks, emotion recognition relies on prefrontal cortex regions and may be a therapeutic target for raloxifene to treat emotion processing deficits. The facial emotion identification task used in Chapters 5 and 6 is an ideal task to test the neural response of raloxifene because emotional processing is sensitive to hormone signaling and it recruits activation in brain regions with densely packed estrogen receptors.

1.3.6 Endogenous testosterone in schizophrenia

Neuroimaging studies have also shown that testosterone may modulate neural activity underlying cognition in men with schizophrenia (Vercammen et al., 2013), but further work is needed in order to determine whether it also has a modulatory role in emotion processing. Testosterone can be converted into estrogen when received in the brain and has been implicated in schizophrenia, though less often than estrogen. Regarding endogenous testosterone levels, some studies have reported lower levels in male patients (Akhondzadeh et al., 2006; Taherianfard & Shariaty, 2004) while others have reported no difference compared to male controls (Ferrier et al., 1982). Testosterone levels have been associated with negative symptoms (Akhondzadeh et al., 2006) and cognition (Moore et al., 2013), while positive symptoms appear to be unrelated.

Beneficial effects of testosterone have been demonstrated with adjunctive testosterone treatment, yielding reduced negative symptoms in men with schizophrenia (Ko et al.,

2008). Conversely, testosterone has been found to be elevated in the CSF of young male patients and young men who are at high risk of developing psychosis (Hayes et al.,

2014). While the increased incidence of schizophrenia in men compared to women may suggest a protective effect of estrogen, it has also been suggested that increased levels of testosterone may be a risk factor of developing the illness (Purves-Tyson et al., 2015). 20

Testosterone may increase transcription of dopamine synthetic enzymes in the midbrain

(Morris et al., 2015), which may lead to and/or potentiate psychotic symptoms. A higher incidence of schizophrenia in men may indicate aberrant signalling of rising testosterone at puberty, leading to a cascade of other disease-related processes. The first study in this thesis will determine whether circulating testosterone is altered in people with schizophrenia and, subsequently, in a following chapter I will determine for the first time whether testosterone levels are associated with the underlying neural activity of emotion processing in men with schizophrenia.

1.3.7 Neurobiological effects of testosterone and estrogen

Steroid hormones bind to specific cytoplasmic receptors and translocate into the nucleus or they bind to receptors residing in the nucleus and bind to steroid responsive elements on DNA. Consequences of such actions result in changes in gene transcription.

Estrogen's effects are mediated by two estrogen receptor isotypes (ERα and ERβ).

Estrogen is involved in regulating synaptic plasticity including synaptogenesis and neurogenesis, in addition to regulating neuroendocrine and inflammatory processes.

Estrogen facilitates numerous neurotransmitter systems including serotonin (Matsuda et al., 2002; Osterlund et al., 2000), N-methyl-D-aspartate (NMDA) and catecholamines

(dopamine, epinephrine and norepinephrine) (Jacobs & D'Esposito, 2011; Menozzi et al.,

2000) - all of which have been implicated in major psychiatric illnesses. Regarding its effects on the serotonergic system, estrogen increases postsynaptic responsivity

(Halbreich et al., 1995) and the number of receptors (McEwen et al., 1997), and enhances transport and uptake (Matsumoto et al., 1985). Therefore, estrogen is a serotonergic agonist, which supports its role in mood regulation (Halbreich & Kahn,

2001). Estrogen can enhance dopamine activity in the prefrontal cortex (Jacobs & 21

D'Esposito, 2011), which may be particularly relevant to working memory impairments and hypodopaminergia in this region in people with schizophrenia. Estrogen is also known to upregulate NMDA receptors and increase NMDA agonist binding (Adams et al., 2004), which may facilitate reversal of hypoglutamatergic functioning in schizophrenia.

Testosterone exerts its direct effects by binding to androgen receptors. Testosterone regulates dopamine neurotransmission by increasing dopamine synthesis during adolescence and stimulating midbrain expression of dopamine receptor subtype 2 (DR2) mRNA. These effects would be deleterious in schizophrenia and may explain why males are more severely impacted than females (McGrath et al., 2004). Testosterone can also act indirectly at the estrogen receptor via aromatisation to estrogen. One study found that testosterone, in addition to estrogen, significantly increased serotonin mRNA and the density of serotonin binding sites in the frontal cortex and nucleus accumbus in rats (Sumner & Fink, 1998). However, this effect was not demonstrated by 5α- dihydrotestosterone (an androgen that cannot be converted to estrogen), suggesting that in this instance testosterone exerted its effects upon conversion to estrogen.

1.3.8 Sex steroids and brain activity

Functional neuroimaging may help us to understand the pathophysiology underlying abnormal brain function in schizophrenia in part by determining whether differences in neural activity are associated with other pathological markers such as peripheral hormones. If such a link is found, it would suggest that hormones measured in the blood can be used to detect whether brain function is normal or irregular. fMRI studies have reported both positive and negative associations between testosterone and neural 22

activation in people with schizophrenia; however, evidence is limited and focuses on cognition. Previous work by our group demonstrated a strong inverse association between serum testosterone levels and activation of the bilateral middle frontal gyrus and left insula during a cognitive-emotional inhibition task in men with schizophrenia

(Vercammen et al., 2013). There is also limited data on the effects of estrogen on neural activity in schizophrenia. Our group has shown that estrogen receptor modulation by raloxifene improves probabilistic association learning in people with schizophrenia by increasing fMRI BOLD signal in the parahippocampal gyrus, supporting the use of raloxifene to treat cognitive deficits in schizophrenia (Kindler et al., 2015). This thesis will, for the first time, examine how raloxifene may modulate neural activity associated with emotion processing. Findings from this study will suggest whether, in addition to facilitating neural activity related to cognition and improving cognitive functioning, raloxifene can be used to aid emotion processing deficits.

1.4 Dehydroepiandrosterone and response to stress

1.4.1 Stress signalling in schizophrenia

1.4.1.1 Early life stressors

The risk of developing schizophrenia increases with the occurrence of stressful events during critical developmental stages including gestation, birth and early post-natal life.

In a large cohort of 1.38 million cases, Khashan et al. (Khashan et al., 2008) reported a significant increase in schizophrenia risk for offspring of mothers who experienced maternal emotional trauma during the first trimester. People with schizophrenia have a higher incidence of early parental loss (prior to age 17) compared to controls (Agid et al., 1999). Children subjected to bullying by peers or physical harm by an adult have an increased risk for early development of hallucinations or delusions (by age 12) 23

(Arseneault et al., 2011; Schreier et al., 2009). Childhood trauma predicts worse prognosis of physical and psychiatric health, impaired social functioning and is associated with increased risk of substance use in people who go on to develop schizophrenia (Stanley D. Rosenberg et al., 2007). Observations that stress-related factors are strongly implicated in schizophrenia emphasizes the importance of understanding stress signaling in order to better understand the underlying causes, neurobiology and the potential for development of early intervention strategies. While there are multiple methods of measuring stress, one of the most common ways is determining levels of cortisol in the periphery, which will be included in Chapter 4.

Stress-related events are associated with timing of illness onset and commonly precipitate relapse of psychotic symptoms (Chabungbam et al., 2007; Ventura et al.,

1992). In ultra-high risk individuals who meet criteria for prodromal symptoms, stressful events have been linked to onset of psychosis. Adolescents who rate themselves as having a relatively low tolerance for normal stress have increased susceptibility to become psychotic. Stressful events (ie. loss of a loved one, relationship break-up, job termination) commonly occur 2-3 weeks prior to the onset of schizophrenia (Day et al., 1987). An individual's resilience or ability to respond appropriately to stress, rather than the nature of the stressor may determine their risk of transitioning into schizophrenia (Myin-Germeys & van Os, 2007). Self-reports have demonstrated that the number of daily 'hassles' are correlated with the risk of developing schizophrenia, while the number of severe life events shows no such relationship

(Thompson et al., 2007). These observations suggest that people who perceive themselves as being afflicted by many minor events are predisposed to develop schizophrenia. The diathesis-stress model, which proposes that the onset of 24

schizophrenia occurs during a critical period in adolescence when an impaired stress response interacts with genetic and environmental risk factors (Walker & Diforio, 1997), has been a major focus of schizophrenia research. It is currently unclear how a possible peripheral marker of stress may affect aberrant functional and structural abnormalities in people with schizophrenia.

1.4.1.2 The hypothalamic–pituitary–adrenal axis

Appropriate physiological and psychological responses to stress are important for maintenance of health and homeostasis in one's evolving environment (de Kloet et al.,

2005). The HPA axis consists of a combination of activity of the hypothalamus (H), pituitary (P) and adrenal (A) glands, releasing glucocorticoid hormones.

The following hormonal activity occurs in response to stress:

1. Activation of the HPA axis at the level of the hypothalamus

2. Parvocellular neurons of the paraventricular nucleus (PVN) of the hypothalamus

secrete corticotropin releasing hormone (CRH) into the

hypothalamohypophyseal portal vessels

3. CRH is delivered to the anterior pituitary

4. CRH receptors bind CRH and mediate the secretion of adrenocorticotropic

hormone (ACTH) and vasopressin (AVP) by the pituitary

5. Glucocorticoids are secreted by the adrenal glands via ACTH stimulation

Secretion of CRH, AVP and catecholamines results in the rapid stress response, increasing vigilance, arousal and attention, enabling one to react quickly to stressors. On the other hand, glucocorticoids act through the glucocorticoid receptor (GR) and 25

mineralocorticoid receptor (MR) on tissues throughout the body and brain to mobilise energy stores, mediate adaptation to stressors and restore homeostasis (de Kloet et al.,

2005).

Cortisol in humans and corticosterones in rodents are the primary glucocorticoids involved in stress signalling. Both are synthesised from cholesterol in the zona fasciculata of the adrenals. Circulating levels of glucocorticoids follow a circadian rhythm with levels peaking upon waking, followed by a decrease toward their lowest levels 3-4 hours after sleep onset (Kalsbeek et al., 2012).

HPA axis activity is controlled by a complex network involving appropriate signalling from the amygdala and limbic cortex toward brain-stem pathways reaching the hypothalamus following psychological stressors. A negative feedback loop is required for deactivation of the HPA axis in order to control the stress response and maintain cortisol homeostasis and prevent exceedingly high levels of cortisol. This feedback loop arises when cortisol, the end product of HPA axis activation, binds to the GR in the

PVN of the hypothalamus, pituitary, hippocampus and prefrontal cortex (Diorio et al.,

1993; Mizoguchi et al., 2003; Weiser et al., 2011). Decreased GR mRNA expression has been founded in the prefrontal cortex and amygdala (Perlman et al., 2004) in people with schizophrenia and may be the consequence of increased cortisol levels caused by stress (Patel et al., 2008; Sapolsky et al., 1984; Sinclair et al., 2012). Decreased glucocorticoid mRNA expression may also be the result of early life stress (Arabadzisz et al., 2010; Meaney et al., 1996), which may influence the responsivity of neurons to cortisol secretion and will be considered in Chapter 5 where I examine possible effects of HPA activity on brain volume. 26

1.4.1.3 Impaired stress response in schizophrenia

Both clinical and biological data indicate impaired biological response to stress in people with schizophrenia (Nuechterlein et al., 1994; Walker & Diforio, 1997), which is associated with dysregulated HPA axis (Ciufolini et al., 2014). These studies linking stress to schizophrenia have investigated negative feedback of the HPA axis using dexamethasone suppression test and cortisol levels in response to psychological or metabolic challenge.

1.4.1.4 Cortisol levels in schizophrenia

Reports of basal cortisol levels in schizophrenia are many and varied. A meta-analysis of 77 studies identified that nearly half of these studies (44.2%) found increased cortisol levels in patients relative to controls (Bradley & Dinan, 2010). This increase has been demonstrated in medicated and unmedicated patients (Meltzer et al., 2001; Muck-Seler et al., 1999; Venkatasubramanian et al., 2010), suggesting that the increase may not be due to antipsychotics. Further, cortisol levels are increased in high risk adolescents who eventually develop schizophrenia relative to those who are at risk but do not develop the illness (Walker et al., 2010; Walker et al., 2001). Increased cortisol levels have been positively correlated with positive symptoms and overall symptom severity (Walder et al., 2000). While there is a general agreement that cortisol levels in schizophrenia are abnormal with abundant evidence reflecting increased levels in patients relative to healthy controls, many studies have reported no difference in cortisol levels between groups (Bradley & Dinan, 2010).

A multitude of factors may explain differential results that are constantly reported.

Cortisol is secreted according to a diurnal rhythm, generating considerable fluctuations in circulating levels throughout the day. Cortisol is also secreted for roughly 20 minutes each hour, changing substantially between each pulse (Lightman & Conway-Campbell,

2010; Veldhuis et al., 1989; Young et al., 2004). Therefore, the timing of obtaining a blood or salivary sample in relation to the occurrence of a pulse can affect the amount measured. Moreover, people with psychiatric illness often display shifts in their diurnal pattern (Lamont et al., 2010; Salvatore et al., 2008). Together, these factors make it extremely difficult to discern whether differences in plasma and serum levels of cortisol between study participants are disease-related or attributed to diurnal influences.

Because of these issues, it is important to develop new and improved methods of examining cortisol; therefore, this thesis uses the measurement of cortisol in combination with another regulatory hormone in the HPA axis, discussed below. The interaction of sex and stress hormones during adolescence modulates dopamine neurotransmission (Sinclair et al., 2014), a signalling pathway that has been implicated in schizophrenia. While testosterone and estrogen interact with glucocorticoids to drive changes in brain maturation and cognitive function, other sex hormones such as DHEA can also influence and modify key neurotransmitter systems by interacting with glucocorticoids.

1.4.2 Dehydroepiandrosterone

Given the implication of altered glucocorticoid signalling and stressors in schizophrenia, examination of possible anti-stress mechanisms is warranted. DHEA in particular may be an important hormone to examine owing to its potent antiglucocorticoid properties.

The first adequately controlled clinical trial of DHEA was published in 1994, 28

demonstrating the ability of DHEA to increase insulin-like growth factor-1 and improve physical and psychological well-being, with an absence of side effects, in age-advanced men and women (Morales et al., 1994). Thus, DHEA is being considered as a target for pharmacotherapy and may protect against cognitive decline in the elderly (Hirshman et al., 2003; Stangl et al., 2011; Yamada et al., 2010). Accumulating data suggest abnormal DHEA and/or DHEAS concentrations in several neuropsychiatric conditions including schizophrenia. Taken together, these findings suggest potential for therapeutic use in schizophrenia, which has recently been explored.

DHEA is found in greater concentrations in the brain compared to plasma (6.5:1 brain- to-plasma ratio). This observation, along with the synthesis of DHEA in the brain, resulted in the coining of the term "neurosteroids", leading scientists to believe that

DHEA has important actions in the central nervous system (Baulieu, 1998; Baulieu &

Robel, 1998).

1.4.2.1 Secretion changes across the lifespan

DHEA and its sulfate ester, DHEAS, serve as the most abundant steroids in the human body. High concentrations of DHEAS are secreted by the fetal zone of the adrenal gland during human gestation (Mesiano & Jaffe, 1997). DHEA(S) concentrations decline for approximately 6 months following birth and remain low until adrenarche begins at 6-8 years of age in both sexes. At this point DHEA(S) is synthesised and secreted from the zona reticularis layer of the adrenal cortex and levels start to increase (Havelock et al.,

2004; Parker, 1991). Circulating concentrations of DHEA(S) in plasma and CSF peak in the mid-20s to early 30s after which they progressively decline with age in both men and women. Levels of DHEA(S) are lowest (approximately 20% of peak concentrations) 29

at 65-75 years of age and are associated with age-related cognitive decline (Valenti et al., 2009). Such observations suggest that restoring DHEA to youthful levels may be beneficial to overall well-being and may protect the brain against damages occurring with age. Interestingly, although plasma concentrations of DHEAS continue to decrease by 1-4% per year between 40-80 years (Muller et al., 2003; Tannenbaum et al., 2004), one study demonstrated that a subgroup (15% of women and 5% of men) experience increases in DHEAS over a 10-14 year follow-up period in late adulthood (Tannenbaum et al., 2004).

1.4.2.2 Mechanisms of action

To date, no nuclear steroid receptor with high affinity for DHEA(S) has been found

(Webb et al., 2006; Widstrom & Dillon, 2004). Therefore, the exact mechanism of

DHEA(S) action is currently unknown. It is possible that DHEA(S) exerts its actions through the conversion into other sex steroids such as testosterone and estrogen, acting on their receptors in brain tissue (Labrie, 2004). DHEA(S) modulates activity of several receptors in the brain (Baulieu, 1997, 1998; Bergeron et al., 1996; Compagnone &

Mellon, 2000; Pérez-Neri et al., 2008). DHEA(S) acts as antagonists at the GABAA receptor (Majewska, 1992; Sousa & Ticku, 1997) and is a positive allosteric modulator of the NMDA receptor (Baulieu, 1998). DHEA(S) can act as a sigma subtype 1 receptor agonist, potentiating NMDA receptor function. This may be particularly relevant to

NMDA receptor hypofunction, which is thought to underlie the symptoms and course of schizophrenia (Olney et al., 1999). Use of DHEA administration and the addition of a sigma 1 receptor antagonist in the hippocampus demonstrated that a sigma 1 receptor antagonist inhibits the potentiating effect of DHEA, indicating that DHEA modulates the NMDA response through sigma 1 receptors in the hippocampus (Bergeron et al., 30

1996). One study showed that sigma receptor agonists reduce NMDA-induced dopamine release in the striatum (Gonzalez-Alvear & Werling, 1994), suggesting that

DHEA(S) inhibits glutamate neurotransmission through sigma receptors outside of the hippocampus.

1.4.2.3 Neurobiological effects

The major biological actions of DHEA(S) in the central nervous system include neuroprotection, neurite growth, neurogenesis and neuronal survival, apoptosis, catecholamine synthesis and secretion, as well as antioxidant, anti-inflammatory and anti-glucocorticoid effects (Maninger et al., 2009) – therefore, it has potential to combat oxidative stress and inflammation, among several other pathological mechanisms, in schizophrenia. Female mice treated with DHEA had better locomotor recovery, left- right coordination and fine motor control following contusive spinal cord injury compared with animals treated with a placebo (Fiore et al., 2004). The DHEA-treated mice also had reduced area of reactive gliosis surrounding the lesion and an increase of spared white matter at the epicentre of the injury. Implantation of DHEA pellets reduced neuronal injury following forebrain ischemia compared to placebo pellets in male

Wistar rats (Li et al., 2001). In vitro studies have also demonstrated neuroprotective actions of DHEA(S). Rat cerebral cortical cultures pre-treated with DHEA(S) prior to 2 hours of anoxia in an anaerobic chamber had increased neuronal survival (Marx et al.,

2000). Further, estradiol concentrations were not detectable in the culture; therefore, the increase in neuronal survival was not a result of metabolism of DHEA(S) into estradiol.

DHEA protected neurons against glutamate and beta amyloid-protein toxicity in HT-22 cells in a dose-dependent manner (Cardounel et al., 1999) . DHEA(S) protected against

NMDA toxicity in fetal rat hippocampal cultures (Kimonides et al., 1998). Both DHEA 31

and DHEAS have effects on neurite growth. DHEA increased the length of Tau- immunopositive axons but had minimal effects on microtuble-associated protein-2 immunopositive dendrites (Compagnone & Mellon, 2000). Conversely, DHEAS did not affect axonal growth, but enhanced growth of dendrites. DHEA was also reported to increase synaptic density in adult rat hippocampal neurons. Two days of DHEA administration increased CA1 spine synapse density compared to the control treatment in ovariectomized rats (Hajszan et al., 2004).

DHEA’s anti-glucocorticoid properties have been demonstrated in vivo and in vitro.

DHEA suppressed the effects of corticosterone (stimulating neurogenesis) in the dentate gyrus of male Lister Hooded rats (Karishma & Herbert, 2002). In vitro, DHEA prevented corticosterone-induced hippocampal neurotoxicity in primary rat tissue cultures (Kimonides et al., 1999). In humans, DHEA levels fluctuate in parallel to those of cortisol as a response to acute stress, protecting against the effects of hypercortisolemia. However, during periods of chronic stress, DHEA levels decline while cortisol levels increase or remain the same, resulting in elevated cortisol/DHEA ratios (Wolkowitz et al., 2001). The fourth chapter of this thesis will examine whether

DHEA levels in people with schizophrenia are abnormal and whether they are related to glucocorticoid activity.

1.4.2.4 Molar ratio of cortisol to DHEA

The cortisol to DHEA ratio is believed to be an important indicator of overall physiological and psychological wellbeing (Wolkowitz et al., 2001), and has been implicated in depression (Mocking et al., 2015), post-traumatic stress disorder (Van

Voorhees et al., 2014) and to some extent in schizophrenia (Gallagher et al., 2007). 32

Cortisol and DHEA are both powerful adrenal hormones that have opposing effects and work via a dualistic balance. Their ratio represents a balance of catabolic and anabolic activity, with higher ratios of cortisol/DHEA indicating high stress levels that are not adequately counterbalanced. Investigation of cortisol/DHEA ratios is particularly warranted in studies of schizophrenia given the impaired biological response to stress in patients. Cortisol/DHEA ratios have been associated with positive symptoms (Garner et al., 2011), negative symptoms (Garner et al., 2011) and cognition (Harris et al., 2001), supporting the argument that cortisol/DHEA ratios are clinically important and should be reported. To date, no study has investigated the relationship of cortisol/DHEA ratios to brain volume in people with schizophrenia. Exploring this association may shed new light as to whether the ratio can be effectively used to predict stress-induced brain volume changes, which may underlie a variety of symptoms in schizophrenia. 33

2.0 Aims of this thesis

Understanding factors that contribute to the onset and progression of schizophrenia is necessary for early intervention and identifying novel therapeutic targets. However, the clinical and molecular heterogeneity of the disease makes this goal an exceptionally difficult task. The earliest studies relied on clinical observations and postmortem techniques; however, the past few decades have seen a rise in advanced neuroimaging approaches, enabling us to take a closer look at structural and functional abnormalities in vivo. The onset of schizophrenia typically occurs during critical periods of hormonal changes, warranting an understanding of the involvement of steroid hormones.

Based on prior clinical and endocrinology studies, I hypothesised that patients with schizophrenia have abnormal circulating levels of steroid hormones. There have been inconsistent reports regarding circulating hormonal levels in schizophrenia; therefore, my goal was not merely to replicate previous findings, but to clarify the roles of sex hormones in relation to pathology. Thus, I further predicted that levels of peripheral blood hormones would be associated with structural deficits and altered brain activity in people with schizophrenia. Moreover, based on evidence that novel estrogen treatments can improve clinical and cognitive outcomes in patients and may modulate receptors in the brain, I hypothesised that the SERM raloxifene would modulate neural activity underlying emotion processing in schizophrenia. Therefore, the purpose of this thesis was to investigate peripheral markers (estrogen, testosterone, DHEA, cortisol/DHEA ratios) in the context of their possible relationship to brain morphometry and neural activity in a sample of chronically ill patients with schizophrenia. In order to achieve this, the following specific aims were addressed:

- To compare serum levels of hormones in people with schizophrenia versus

healthy controls, separating by gender when appropriate;

- To investigate whether the HPA axis marker (cortisol/DHEA ratio) correlates

with reductions in hippocampal and DLPFC volume in people with

schizophrenia;

- To examine the neural correlates of emotion processing in people with

schizophrenia versus healthy controls;

- To determine the extent to which serum testosterone levels are related to neural

activity in processing angry faces in men with schizophrenia;

- To determine the extent to which adjunctive raloxifene treatment alters neural

activity associated with emotion processing in men and women with

schizophrenia.

Results from this thesis may be clinically important in helping evaluate the efficacy of novel pharmacological treatments targeting sex steroid receptors. In particular, it will be important to determine whether such treatments are suitable for patients of both sexes and, moreover, whether there is therapeutic potential for participants who do not display abnormal circulating hormone levels.

3.0 Endogenous hormone levels in people with schizophrenia compared with healthy controls

3.1 Abstract

Sex differences relating to disease risk, course, and outcome have implicated sex steroids in the pathophysiology of schizophrenia. Previous studies examining circulating levels of sex hormones have reported inconsistent results in people with schizophrenia relative to healthy controls. Such varying reports warrant further investigation in a sample of males and females with schizophrenia as it is unclear whether circulating hormone levels differ by both diagnosis and gender and how hormones may play a role in pathology. Whole blood samples were collected from 87 healthy controls and 97 people with schizophrenia. We compared serum levels of estrogen, testosterone and dehydroepiandrosterone (DHEA) based on diagnosis and sex.

Circulating testosterone was significantly decreased in male patients relative to male controls and significantly increased in female patients relative to female controls. Both male and female patients had significantly increased serum DHEA levels relative to healthy controls. There were no diagnostic differences in circulating estrogen levels.

Our finding of decreased testosterone in male patients supports previous studies showing hypoandrogenism in men with schizophrenia. Increased testosterone in female patients with schizophrenia suggests a gender-relative hyperandrogenism, which may contribute to increased dopamine activity in the brain or, conversely, it may suggest more conversion of testosterone to estrogen and protect females against the more severe course of illness seen in male patients. Our finding of increased serum DHEA in patients with schizophrenia may reflect a compensatory upregulation against biological stressors present in schizophrenia. Investigation of sex hormones is particularly 36

important because they interact with major neurotransmitter systems involved in schizophrenia pathology. Moreover, there has been a recent increase in pharmacological therapies using hormone augmentation; therefore, a better basic understanding of how sex may interact with other biological mechanisms is critical for ensuring efficacy and safety.

3.2 Introduction

There is increasing evidence that sex hormones may influence the risk, development and course of schizophrenia. These hormones contribute to brain changes during adolescence (Blakemore et al., 2010), a developmental period of heightened vulnerability to the illness (Häfner, 2003). Sex steroid hormones, most notably estrogen and to a lesser extent testosterone, have been implicated in the development of schizophrenia due to gender differences observed in the disease (Abel et al., 2010;

Markham, 2012; Ochoa et al., 2012). These differences include differential age of onset in which males reach a peak in the risk of onset between 18–24 years of age and females reach a peak in the risk of onset up to 4 years later (Eranti et al., 2013; Häfner,

2003) with a second peak age of onset at 45–50 years of age following menopause

(Hafner et al., 1993). Further, the course of illness differs where women typically experience a less severe course of the illness (fewer hospitalizations and improved quality of life) (Canuso & Pandina, 2007; Hambrecht et al., 1992; Ochoa et al., 2012) and show a more favourable response to antipsychotics (Begemann et al., 2012) as compared with men. Androgens, testosterone and dehydroepiandrosterone (DHEA), and estrogens are produced primarily by the gonads and in smaller amounts by other tissues including the adrenal glands, liver and brain. Sex steroids readily cross the blood brain barrier (Banks, 2012); therefore, changes in circulating levels of hormones may be particularly relevant to brain function.

Estrogen exerts its effects by binding to estrogen receptors (Heldring et al., 2007) and plays a central role in modulating dopamine neurotransmission (Chavez et al., 2010), a major neurotransmitter implicated in schizophrenia (Grace, 2016; Howes & Kapur, 38

2009). Ovariectomy of adult female rats removes circulating estrogen, which increases dopamine D2 receptor (DRD2) and is in turn reversed by the addition of 17β-estradiol replacement (Chavez et al., 2010). Likewise, ovarectomised nonhuman primates showed increased expression of dopamine β hydroxylase, an enzyme involved in dopamine synthesis. Subsequent treatment with 17β-estradiol reversed this increase to levels comparable to intact primates (Kritzer & Kohama, 1999). Estrogen modulates neural activity underlying cognitive tasks in humans, as demonstrated by functional neuroimaging studies (Berman et al., 1997; Shaywitz et al., 1999). Several clinical trials have investigated the cognitive-enhancing efficacy of estrogen replacement therapy for women with Alzheimer's disease (AD) with findings of enhanced attention and verbal memory in postmenopausal women with AD (Asthana et al., 2001; Asthana et al., 1999).

In schizophrenia, adjunctive treatment with estrogen and selective estrogen receptor modulators has proven to be beneficial in ameliorating symptomatology in females

(Kulkarni et al., 1996; Kulkarni et al., 2010) and cognitive deficits in men and women

(Weickert et al., 2015). In light of these findings, it is hypothesised that estrogen may buffer females against the development and course of schizophrenia.

Since slightly more men are affected with schizophrenia and men are more severely affected, the predominately male hormone, testosterone, is implicated. Testosterone modulates neurotransmitter systems and exerts its effects by binding to androgen receptors (Fang et al., 2003; Vermes et al., 1979). Several lines of evidence support a neuroprotective role for testosterone as it acts protectively in neurodegenerative disorders such as Alzheimer's disease and mild cognitive impairment (Lim et al., 2003;

Papasozomenos, 1997). Testosterone supplementation improves spatial and verbal memory in healthy elderly men (Cherrier et al., 2001). Such evidence led to one clinical 39

trial using exogenous testosterone in men with schizophrenia, which reported a significant reduction of negative symptoms (Ko et al., 2008). Endogenous testosterone levels were positively associated with cognition in men with schizophrenia (Moore et al.,

2013), suggesting that higher testosterone within normal ranges is beneficial. Counter to this, there is evidence from animal studies that argue against the neuroprotective nature of testosterone. Unlike estrogen, testosterone does not protect against methamphetamine-induced neurotoxicity of the dopaminergic system in male or female mice (Myers et al., 2003). Further, unlike estrogen, testosterone does not protect against glutamate-induced neurotoxicity (a model of oxidative stress) (Sawada et al., 1998) and it enhances MK-801-induced prepulse inhibition disruption in rodents (Gogos et al.,

2012). Testosterone replacement given to gonadectomised male rats increases dopamine synthesis and stimulates expression of DR2 mRNA, dopamine transporter mRNA, dopamine transporter protein and vesicular monoamine transporter mRNA in the midbrain (Purves-Tyson et al., 2014). These findings therefore suggest that modulation of dopamine by testosterone may have deleterious effects which could potentiate symptoms of schizophrenia, though this has not been demonstrated in people with schizophrenia. It appears that clinical versus animal research illustrate opposing arguments for the role of testosterone in schizophrenia where the former suggests advantageous effects and the latter suggests unfavourable effects are possible.

Dehydroepiandrosterone (DHEA) is a precursor of testosterone and estrogen that binds to both androgen (Lu et al., 2003) and estrogen receptors (Chen et al., 2005) and exerts multiple effects on the central nervous system. DHEA protects neurons against glutamate and beta amyloid-protein toxicity (Cardounel et al., 1999) and oxidative stress (Bastianetto et al., 1999; Jacob et al., 2010). DHEA also enhances myelination 40

and synaptogenesis, regulating the growth/health of neurons and is thus neuroprotective

(Friess et al., 2000; Mao & Barger, 1998). In healthy people, fluctuations of DHEA levels rise in parallel to those of cortisol, which is increased with periods of acute stress

(McEwen, 2006). Some clinical trials have demonstrated the efficacy of DHEA augmentation to improve symptoms and cognitive functioning in people with schizophrenia (Ritsner et al., 2006; Strous et al., 2003); however, some trials have failed to replicate beneficial DHEA effects; therefore, the precise influence of DHEA on schizophrenia is uncertain.

Because androgens and estrogens have neuroprotective properties and can modulate neurotransmitter systems linked to schizophrenia pathology, an important question has been raised as to whether people with schizophrenia exhibit abnormal levels of circulating sex hormones. It has been proposed that higher or lower levels of circulating sex hormones may contribute to the onset and progression of the illness. Indeed, women with schizophrenia are often hypoestrogenic with lower circulating levels of estrogen compared with healthy women (Bergemann et al., 2005; Huber et al., 2001; Riecher-

Rossler et al., 1994), though there are also reports of no diagnostic difference. Similarly, testosterone was found to be decreased or not different in males with schizophrenia relative to healthy male controls (Akhondzadeh et al., 2006; Moore et al., 2013;

Taherianfard & Shariaty, 2004). Interestingly, there are limited reports of decreased

(Tourney & Hatfield, 1972) or no difference in (Ritsner et al., 2004) DHEA levels in individuals with schizophrenia compared to healthy controls, but the majority of studies report increased levels of DHEA (di Michele et al., 2005; Ritsner et al., 2006; Strous et al., 2004). It has therefore been speculated that increased DHEA serves as a protective or compensatory factor in the disease (Strous et al., 2004). 41

The aim of this study was to determine whether there are sex-specific hormonal differences in circulating levels of testosterone, estrogen and DHEA in men and women with schizophrenia compared with healthy controls. In considering the hormonal properties and previous reports of circulating hormonal levels in schizophrenia described above, we hypothesised that our cohort of participants with schizophrenia would exhibit abnormal levels of some endogenous circulating hormones. We predicted no diagnostic difference in testosterone levels among men because previous work by our group (Moore et al., 2013) using a subsample of the present study’s cohort did not report any difference, suggesting that hypoandrogenism which has been reported in some studies is not characteristic of the present sample. Based on several reports of hypoestrogenism in females with schizophrenia (Bergemann et al., 2005; Huber et al.,

2001; Riecher-Rossler et al., 1994), we predicted that estrogen would be decreased in female patients relative to female controls. Additionally, we predicted that DHEA levels would be increased in the entire cohort to reflect an upregulated compensatory response.

3.3 Materials and Methods

Participants

Ninety-seven adult outpatients with schizophrenia or schizoaffective disorder and 87 healthy controls 18-51 years of age were recruited to one of two sites, either

Neuroscience Research Australia in Randwick, NSW, Australia or the Lyell McEwen

Hospital in Adelaide, South Australia. Patients were recruited from local clinics and via a national television documentary on schizophrenia research. A psychiatrist or psychologist trained in using the Structured Clinical Interview for DSM-IV (First et al., 42

2007) confirmed diagnoses. All patients met DSM-IV criteria for schizophrenia or schizoaffective disorder and had been receiving antipsychotics for at least 1 year before entry into the study. Healthy control participants were recruited through advertising.

Exclusion criteria for all participants included history of a head injury with loss of consciousness, seizures, recent history of alcohol and/or substance abuse/dependence

(within the past 5 years), a central nervous system infection, uncontrolled diabetes or hypertension, structural brain abnormalities, mental retardation or a learning disability.

Further, healthy participants with a personal history or first-degree relative with DSM-

IV Axis I psychiatric diagnosis were excluded. Patients with a concurrent Axis 1 diagnoses were excluded. The protocol was approved by the Human Research Ethics

Committees from the University of New South Wales, the South Eastern Sydney and

Illawarra Area Health Service and the Queen Elizabeth Hospital, Adelaide, South

Australia. All participants provided written informed consent prior to participation in this study.

Symptom severity in participants with schizophrenia was assessed using the Positive and Negative Syndrome Scale (PANSS) (Kay et al., 1987). Positive, negative, general and total symptom scores were calculated. A four subtest version of the Wechsler Adult

Intelligence Scale, 3rd edition (WAIS-III) (Wechsler, 1997) comprised of the

Arithmetic, Digit Symbol, Similarities and Picture Completion subtests was administered to assess current intellectual functioning. The Wechsler Test of Adult

Reading (WTAR) (Wechsler, 2001) was administered to obtain an estimate of premorbid intellectual functioning in patients.

Serum collection and laboratory analyses

Fasting peripheral blood was collected between 0900 and 1100h to control for diurnal variation. Blood samples were put on ice and stored at -80°C until assayed. Serum estrogen and testosterone were assayed by South Eastern Area Laboratory Services in

Randwick, NSW, Australia using a chemiluminescent immunometric assay (Siemens

Immulite 2000). Serum DHEA was assayed by the ANZAC Research Institute in

Concord, NSW, Australia using gas chromatography–mass spectrometry (GC-MS) analysis. Inter-assay coefficient of variation (CV) was 13.5% for estradiol, 9.7% for testosterone and 13% for DHEA. The lower sensitivity limits of assays were 73 pmol/L for oestrogen, 0.7 nmol/L for testosterone and 0.05 ng/ml for DHEA. Hormone measurements were determined while blind to diagnosis.

Data analyses

Statistical analyses were performed with IBM SPSS 22 statistical software (Armonk,

NY, USA). Demographics of healthy controls and patients were analysed with independent t-tests or χ2-tests, as appropriate. Outliers were determined by Grubbs' test for outliers (GraphPad Software, online calculators) which resulted in the removal of one male patient with exceedingly elevated estradiol levels and one female patient with exceedingly elevated testosterone levels. DHEA levels were normalised through logarithmic transformation. Age was used as a covariate for hormone levels that demonstrated a significant association with age, which was the case for DHEA and not testosterone or estrogen. Raw serum testosterone and estrogen levels were compared in male and female patients and controls using Mann-Whitney U tests since the data were not normally distributed even after logarithmic transformation. ANCOVAs were performed to compare mean log-transformed DHEA levels in schizophrenia compared 44

to control participants with age as a covariate and diagnosis as a grouping factor, in males and females separately. Independent samples t-tests were performed for PANSS scores in male patients versus female patients in order to determine whether gender was associated with severity of illness. Pearson’s correlations were performed between symptom severity and testosterone and estrogen in male and female patients in order to understand if hormones, whether they display normal or abnormal levels, are associated with severity of illness. Pearson’s correlations were performed between hormone levels and mean daily chlorpromazine (CPZ) equivalents (CPZ) dose (Leucht et al., 2003;

Woods, 2003) in patients in order to determine any relationship of antipsychotic medications to hormone levels.

3.4 Results

Demographics for healthy controls and participants with schizophrenia are shown in

Table 3.1. Age at sampling, years of education, current IQ and premorbid IQ estimates differed significantly between diagnostic groups, where the schizophrenia sample was older, received less years of education and had poorer scores on premorbid and current

IQ estimates compared to controls. There were no significant differences regarding sex or ethnicity between the groups. PANSS scores reflect mild to moderate symptom severity in the patient sample.

Table 3.1 Demographic variables, cognitive and clinical characteristics of the whole sample.

Controls Schizophrenia N Mean (SD) N Mean (SD) T/X2(df) p-value Age (years) 87 31.9 (8.3) 97 35.7 (8.4) -3.1 (182) p = 0.002 Education(years) 87 14.6 (2.3) 97 12.5 (2.4) 6.0 (182) p < .001 Sex (M/F) 87 46/41 97 59/38 1.2 (1) p = 0.28 Ethnicity 87 97 4.9 (3) p = 0.18 Caucasian 69 83 Asian 12 5 Caucasian-Asian 2 5 Other 4 4 WTAR 87 108.1 (8.8) 97 102.3 (9.1) 4.4 (182) p < 0.001 WAIS-III IQ 87 107.0 (14.5) 97 90.9 (13.1) 7.9 (182) p < 0.001 Diagnosis (SZ/SA) 97 64/33 Age of onset (years) 97 22. 8 (5.6) Illness duration (years) 97 12.9 (7.5) CPZ equivalents (mg) 97 551.5 (465.7) PANSS positive 97 15.0 (4.7) PANSS negative 97 14.4 (6.2) PANSS general 97 30.6 (8.7) PANSS total 97 60.0 (16.6)

Abbreviations: CPZ, chlorpromazine; F, Female; M, male; PANSS, Positive and Negative Syndrome Scale; SZ, schizophrenia; SA, schizoaffective disorder; WAIS-III, Wechsler Adult Intelligence Scale; WTAR, Wechsler Test of Adult Reading. LNS, WAIS III Letter Number Sequencing; TMT-A, Form A of the Trail Making Test. Bold P-values indicate significant results.

Figure 3.1: Comparison of hormone levels in patients versus healthy controls Serum levels of estrogen, testosterone and DHEA were compared between patient and control groups, separated by sex. Error bars represent standard error of the mean. (*p < 0.05, ** p <0.01).

Serum levels of hormones are presented in Figure 3.1. Male and female patients with schizophrenia displayed significantly greater serum DHEA levels relative to healthy controls. Testosterone levels were significantly decreased in male patients relative to healthy males and significantly increased in female patients relative to healthy females.

There was no diagnostic difference in estrogen levels for either males or females.

Female patients displayed lower symptom severity on all PANSS ratings (positive, negative, general and total) as compared with male patients, with negative scores reaching significance (see Table 3.2). There were no strong, significant correlations between serum testosterone or estrogen levels and symptom severity in male or female patients after correcting for multiple comparisons (see Table 3.3). However, prior to correcting for multiple comparisons, estrogen was significantly positively correlated with total PANSS symptoms and there was a trend for positive correlations between estrogen and positive and general symptoms in men with schizophrenia. There were no 47

significant correlations between circulating hormone levels and mean daily CPZ dose in patients with schizophrenia (see Table 3.4).

Table 3.2 Comparison of symptom severity in male patients versus female patients Male patients Female patients N Mean (SD) N Mean (SD) T/(df) p -value Positive symptoms 59 15.6 (4.3) 38 14.1 (5.2) 1.6 (95) p = 0.13 Negative symptoms 59 15.8 (6.7) 38 12.3 (4.5) 2.9 (95) p = 0.005 General symptoms 59 30.9 (9.5) 38 30.0 (7.6) 0.5 (95) p = 0.60 Total symptoms 59 62.4 (17.4) 38 56.4 (14.8) 1.7 (95) p = 0.08 Bold P-values indicate significant results.

Table 3.3 Associations of testosterone and estrogen to symptom severity in patients Male patients Female patients df r P-value df r P-value Testosterone Positive symptoms 57 -0.02 0.90 35 0.14 0.42 Negative symptoms 57 -0.15 0.26 35 0.05 0.78 General symptoms 57 -0.06 0.66 35 -0.04 0.80 Total symptoms 57 -0.09 0.48 35 0.04 0.80

Estradiol Positive symptoms 56 0.24 0.08 36 0.10 0.55 Negative symptoms 56 0.12 0.23 36 -0.01 0.96 General symptoms 56 0.25 0.06 36 -0.21 0.21 Total symptoms 56 0.23 0.05 36 -0.07 0.66 Bonferroni correction of alpha .05/16 so that p < 0.003 to be significant.

Table 3.4 Correlations between mean daily CPZ equivalents dose and peripheral hormone levels in schizophrenia df r P-value Estradiol 95 -0.15 0.14 Testosterone 95 0.05 0.70 Log DHEA 93 0.03 0.76

3.5 Discussion

The key findings from the present study are that: (1) male patients had significantly decreased serum testosterone levels relative to male controls; (2) female patients had significantly increased serum testosterone levels relative to female controls; and (3) men and women with schizophrenia showed a significant increase in serum DHEA levels.

Our finding of decreased serum testosterone levels in males supports the hypothesis of hypoandrogenism in men with schizophrenia (Akhondzadeh et al., 2006). Previous work has demonstrated that testosterone levels are negatively associated with symptom severity (Akhondzadeh et al., 2006; Ramsey et al., 2013) in male patients. Therefore, it has been believed that lower levels of circulating testosterone in male patients may represent a risk factor of the illness such that higher levels may be associated with improved symptoms and outcomes. This is supported by findings from a clinical trial using testosterone augmentation to treat male patients; results indicated a significant improvement of negative symptoms following treatment (Ko et al., 2008). Counter to this, we did not find any significant correlations between testosterone levels in male patients and PANSS scores, which demonstrate that testosterone is not a predictor of illness severity in at least some samples of men with schizophrenia. Sample size and heterogeneity of illness (male patients in our study displayed milder symptoms as compared with most studies) may explain our negative findings.

Our finding of increased serum testosterone levels in females was unexpected and counter to our hypothesis that there would be a diagnostic difference in males but not in females. The finding is not a novel one as at least two groups have reported increased 49

circulating testosterone in females with schizophrenia (Mendrek et al., 2011; Ramsey et al., 2013). However, this is the first study to investigate whether upregulated testosterone in female patients is associated with symptom severity. While no such relationships were found, female patients in this study displayed mild to moderate symptom severity, therefore further studies are required before reaching a firm conclusion that testosterone in females is not associated with any disease parameter

Based on studies in men with schizophrenia that suggest testosterone may be beneficial

(Ko et al., 2008; Moore et al., 2013), it is possible that that upregulated testosterone in female patients may be neuroprotective by buffering them against the more severe course of the illness often seen in male patients (Häfner, 2003). In support of this, we found that female patients displayed significantly less negative symptoms compared to male patients. Further, in line with this possibility that testosterone may be neuroprotective in female patients, at a molecular level testosterone is involved in determining neuronal size, neurite growth and synaptogenesis (Beyer et al., 1994; Beyer

& Hutchison, 1997; Lustig, 1994).

In contrast to the potential favourable effects of increased testosterone in female patients, our finding may indicate a pathological risk marker in females that is detectable in the blood, owing to testosterone’s ability to increase dopamine activity. This is supported by reports of testosterone-induced psychotic episodes in steroid users, consisting of hallucinations and paranoid delusions (Pope & Katz, 1988). In addition to being influenced by levels of circulating sex hormones, symptoms are also influenced by sex hormone receptors. Previous work by our group demonstrated that increased androgen receptor mRNA is associated with increased tyrosine hydroxylase mRNA (Morris et al.,

2015), which is the rate-limiting step in dopamine biosynthesis (Nagatsu et al., 1964), 50

supporting a direct modulation and increase of dopamine neurotransmission by testosterone. Thus, increased testosterone levels in females with schizophrenia may have deleterious effects by exacerbating risk and pathology. Increased testosterone may also reflect disrupted conversion into estrogen by the aromatase enzyme. One of testosterone's notable roles in females is to provide the basic structure for estrogen production where it may exert positive effects indirectly through this conversion.

However, disruptions in the conversion process may limit the direct neuroprotective effects of estrogen. The mechanism by which testosterone is increased in females with schizophrenia may be partially explained by sex hormone-binding globulin (SHBG), which binds to sex hormones including testosterone and in turn influences their bioavailability. Previous studies have reported decreased SHBG in females with schizophrenia in parallel with increased testosterone (Ramsey et al., 2013). Despite the possibility that increased testosterone may have positive or negative effects for female patients, we did not find an association between testosterone levels and symptom severity and, moreover, the levels of testosterone do not approach the levels of males.

Therefore, the significance of such relatively low concentrations is unclear.

We also found that DHEA was upregulated in both male and female patients compared with healthy controls. Based on the many and varied neuroprotective properties of

DHEA, our findings support the hypothesis that increased DHEA serves as a protective or compensatory factor in schizophrenia. One study reported higher levels of DHEA at first onset of psychosis but not in subsequent episodes (Beyazyüz et al., 2014; Strous et al., 2004), so it has been speculated that DHEA levels diminish with chronic illness.

However, we and others (di Michele et al., 2005) found significantly elevated DHEA levels in chronically ill patients, indicating longstanding upregulation of DHEA. DHEA 51

has positive modulatory actions at NMDA receptors (Maninger et al., 2009), and may be relevant to NMDA receptor hypofunction in schizophrenia. DHEA’s precursor

(pregnenolone) prevents cognitive deficits following administration of NMDA receptor antagonists MK801 (Romeo et al., 1994) and D-AP5 (Mathis et al., 1994) in rats. There is also evidence of an association between DHEA and working memory and attention in schizophrenia from a small sample of inpatients (Harris et al., 2001). Therefore, while

DHEA may stimulate NMDA receptors (Maninger et al., 2009), it is unclear at this stage whether an increase in circulating DHEA has any beneficial impact on brain

NMDAR mediated function in schizophrenia. Longitudinal studies will be helpful in characterising how endogenous DHEA levels may fluctuate from the prodromal phase over the course of illness and may help us understand how DHEA levels may be related to changes in disease-state. Chapter 4 will examine the relationship of DHEA to schizophrenia pathology in more detail by including the context of DHEA’s anti- glucocorticoid activity.

This study has a number of limitations. All patients in our sample were receiving antipsychotic medication. There is evidence suggesting that antipsychotics may influence testosterone (Konarzewska et al., 2009) and estrogen (Wieck & Haddad, 2003) levels. There is some evidence from animal models suggesting that clozapine is associated with decreases in DHEA (di Michele et al., 2005; Nechmad et al., 2003). If this occurs in humans then such an effect would have lowered DHEA levels in our patient sample, yet we detected significantly elevated DHEA levels. Further, DHEA levels are elevated in drug-naive first-episode patients, further suggesting that increased

DHEA may be related to pathology and is not a result of antipsychotic use. We did not find an association of daily chlorpromazine equivalent dose with hormone levels, 52

though we cannot entirely dismiss the possible confounding effects of chronic antipsychotic use given that our patient sample was receiving different antipsychotics.

Moreover, although hormone levels were not significantly associated with symptom severity, future studies should examine the role of circulating estrogen levels in men with schizophrenia because of the trending and significant correlations between estrogen and positive, general and total PANSS symptoms prior to correcting for multiple corrections. Lastly, our sample had the typically characteristic fewer female patients as compared with male patients; thus, larger cohorts including more females will be useful to determine sex differences in relation to hormonal influences on schizophrenia.

Notwithstanding the limitations, the present study found significant differences in peripheral markers that may be relevant to schizophrenia pathology. Because of their ability to cross the blood brain barrier, abnormal hormone levels of peripheral origin in may be particularly relevant to brain pathology. Investigation of sex hormones is particularly important because they can drive brain changes during adolescence and interact with major neurotransmitter systems involved in schizophrenia. Moreover, future studies should take into consideration sex-specific differences in schizophrenia as certain pathologies may be specific to a given sex.

Implications for this thesis

Our finding of increased DHEA levels in both male and female patients allows further investigation in a combined-sex cohort. Because of DHEA’s anti-glucocorticoid properties and of the relationship between DHEA and cortisol (Kalimi et al., 1994; 53

Kroboth et al., 2003; McNelis et al., 2013), the following chapter investigates the molar ratio of cortisol to DHEA, which may be a more complete measure of functional hypercortisolaemia and therefore more informative than the absolute concentrations of

DHEA or cortisol alone. In order to understand the involvement of DHEA in schizophrenia pathology, we will examine the relationship of cortisol/DHEA ratios to symptom severity, cognitive functioning and grey matter deficits. Because sex was a modulating factor of testosterone levels such that the direction (decreased/increased in males/females respectively) of change differed by sex, Chapter 5 of this thesis focuses on men with schizophrenia in order to test whether circulating levels may influence the underlying neural activity of emotion processing which is associated with social functioning outcomes, another important and prominent feature of schizophrenia.

Finally in Chapter 6, we test whether raloxifene can restore abnormal neural activity in both male and female patients, because therapeutic use of raloxifene has shown benefit for both sexes (Weickert et al., 2015).

4.0 Cortisol-dehydroepiandrosterone ratios correlate with hippocampal and prefrontal cortex volume reductions in schizophrenia

4.1 Abstract

Dehydroepiandrosterone (DHEA) is a hormone that has anti-glucocorticoid properties.

The cortisol/DHEA ratio is a balance of anabolic to catabolic hormones and has been implicated in schizophrenia and other diseases in which stress may play a role. Previous studies investigating the cortisol/DHEA ratio in schizophrenia have analysed it in regards to its relationship with behavioural phenotypes such as symptomatology, though no study to date has investigated whether cortisol/DHEA ratios are associated with underlying biological changes that may contribute to these phenotypes. The present study investigated whether cortisol/DHEA ratios correlate with hippocampal and prefrontal cortex (PFC) volume, cognitive function and symptom severity. Serum

DHEA and cortisol were assayed in 94 people with schizophrenia and 81 healthy controls. T1-weighted high-resolution anatomical scans were obtained using a 3T

Achieva scanner on a subset of 59 people with schizophrenia and 61 healthy controls.

Imaging data were preprocessed and analysed using SPM12. As expected, hippocampal and PFC volumes were significantly reduced in people with schizophrenia relative to healthy controls (FWE corrected p < 0.05). Cortisol/DHEA ratios were inversely correlated with hippocampal (r = -0.33, p = 0.01) and PFC (r = -0.41, p = 0.001) volume in schizophrenia, but not in healthy controls. There were no significant associations of cortisol/DHEA ratios to symptom severity or cognitive variables in patients. Our findings suggest that the cortisol/DHEA ratio may be a molecular signal of increased sensitivity to brain insults that is detectable in the blood. These results further 55

implicate the role of DHEA and HPA axis dysfunction in the pathophysiology of schizophrenia.

4.2 Introduction

Dehydroepiandrosterone (DHEA) is a critical hormone that exerts multiple neuroprotective effects on the central nervous system. DHEA readily crosses the blood brain barrier and levels of DHEA in blood and cerebrospinal fluid are strongly correlated (Guazzo et al., 1996; Kancheva et al., 2011); therefore, changes in circulating levels of DHEA may be particularly significant to brain function. DHEA concentrations increase during puberty and peak in the mid-twenties to early thirties after which they markedly decrease with age in both men and women (Sulcova et al., 1997). In healthy people, fluctuations of DHEA levels occur in parallel to those of cortisol, which is increased with periods of acute stress (McEwen, 2006). This co-release of DHEA with cortisol protects against the damaging effects of hypercortisolemia. Both clinical and biological data indicate impaired biological response to stress in people with schizophrenia (Nuechterlein et al., 1994; Walker & Diforio, 1997), which is associated with dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis (Ciufolini et al.,

2014). Therefore, owing to the anti-glucocorticoid properties of DHEA, the cortisol/DHEA ratio may be a useful marker of HPA axis activity and more informative than the absolute concentrations of DHEA and cortisol alone (Maninger et al., 2009;

Wolkowitz et al., 2001). Previous research suggests that the cortisol/DHEA ratio may be abnormal in people with schizophrenia (Ritsner et al., 2004), however there are conflicting reports with some studies finding no difference in cortisol/DHEA ratio compared with healthy people (Gallagher et al., 2007; Garner et al., 2011). There is 56

some work reporting that the cortisol/DHEA ratio is associated with negative (Garner et al., 2011) and depressive symptoms (Garner et al., 2011; Ritsner et al., 2004) in schizophrenia. While these studies have assessed the relationship of cortisol/DHEA ratio to behavioural phenotypes, no study to date has investigated whether cortisol/DHEA ratios are associated with underlying biological changes such as neuroanatomical abnormalities that may contribute to these phenotypes.

Abnormalities in the hippocampus and prefrontal cortex (PFC) are well-documented in schizophrenia and are related to cognitive impairments, a prominent feature of the illness (Cannon et al., 2002; Lee & Park, 2005; Petrides, 1995; Weickert et al., 2000).

Neuroimaging and post-mortem studies have reported functional abnormalities and volumetric reductions in these regions of people with schizophrenia (Goldman-Rakic &

Selemon, 1997; Nelson et al., 1998; Selemon et al., 2002; Selemon et al., 2003;

Selemon et al., 1998). Importantly, these brain regions are principal targets for glucocorticoids. Prolonged exposure to glucocorticoids is associated with neurotoxicity, atrophy, inhibition of neurogenesis and death of hippocampal neurons in rodents

(Sapolsky, 1985; Sapolsky et al., 1985). Chronic corticosterone administration results in dendritic reorganization in pyramidal neurons in the prefrontal cortex of rodents

(Wellman, 2001). Increased corticosteroids are also associated with structural changes

(usually decreased volumes) in the hippocampus (Brown et al., 2004; Chapman et al.,

2006; Sapolsky et al., 1990) and prefrontal cortex (Carrion et al., 2010). Studies linking circulating glucocorticoids, such as cortisol, with structural changes in people with first episode psychosis demonstrate the deleterious effect early in the illness (Mondelli et al.,

2010). However, studies of cortisol alone do not take into account the possible role of

DHEA in mediating the adverse effects of elevated glucocorticoids. As a peripheral 57

marker of HPA axis activity, the cortisol/DHEA ratio may be useful in examining stress-induced morphological changes in the brain.

The aim of this study was to determine the extent to which cortisol/DHEA ratios correlate with hippocampal and PFC volume, cognition and symptom severity in chronically ill people with schizophrenia. Based on previous work demonstrating an inverse relationship of cortisol/DHEA to behavioural phenotypes, we predicted that cortisol/DHEA ratios would be negatively correlated with symptom severity and cognition. Further, based on previous studies showing detrimental effects of glucocorticoid activity on brain volume, we predicted an inverse relationship between cortisol/DHEA ratios and regional brain volumes in patients with schizophrenia, with no such relationship in healthy controls. The hippocampus and PFC are particularly vulnerable targets of the disease process; therefore, identifying peripheral markers that signify processes contributing to the regionally distinct volumetric reduction in schizophrenia would be important for understanding the underlying pathophysiology of schizophrenia.

4.3 Materials and Methods

Participants

Ninety-four outpatients with schizophrenia or schizoaffective disorder and 81 healthy controls 18-51 years of age met inclusion criteria for the present study and were recruited at two sites: Neuroscience Research Australia in Randwick, NSW, Australia or the Lyell McEwen Hospital in Adelaide, South Australia. Patients were recruited from 58

local clinics and via a national television documentary on schizophrenia research.

Clinical diagnostic interviews using the Structured Clinical Interview for DSM-IV (First et al., 2007) were conducted by a trained psychologist or psychiatrist. All patients met

DSM-IV criteria for schizophrenia or schizoaffective disorder and had been receiving antipsychotics for at least 1 year before entry into the study (see Supplementary Table

4.1 for numbers of patients receiving antipsychotics alone or in combination). Patients with a concurrent Axis I diagnosis, head injuries with loss of consciousness, seizures, recent history of alcohol and/ or substance abuse/dependence (within past 5 years), a central nervous system infection, uncontrolled diabetes or hypertension, structural brain abnormalities, mental retardation or a learning disability were excluded.

Healthy control participants were recruited through advertising. Healthy participants with a personal history or first-degree relative with DSM-IV Axis I psychiatric diagnosis, history of a head injury with loss of consciousness, seizures, recent history of alcohol and/or substance abuse/dependence (within the past 5 years), a central nervous system infection, uncontrolled diabetes or hypertension, structural brain abnormalities, mental retardation or a learning disability were excluded.

The protocol was approved by the Human Research Ethics Committees from the

University of New South Wales, the South Eastern Sydney and Illawarra Area Health

Service and the Queen Elizabeth Hospital, Adelaide, South Australia. All participants provided written informed consent prior to participation in this study.

Symptom measures and cognitive assessments

Symptom severity in participants with schizophrenia were assessed with the Positive 59

and Negative Syndrome Scale (PANSS) (Kay et al., 1987). Positive, negative, general and total symptom scores were calculated. Inter-rater reliability for administering the

PANSS was achieved with an average intraclass correlation coefficient = 0.90. All participants were assessed with a four subtest version of the Wechsler Adult Intelligence scale, 3rd edition (WAIS-III) (Wechsler, 1997) comprised of the Arithmetic, Digit

Symbol, Similarities and Picture Completion subtests to assess current intellectual functioning. The Wechsler Test of Adult Reading (WTAR) (Wechsler, 2001) was administered to obtain an estimate of premorbid intellectual functioning in patients.

Working memory and verbal fluency were assessed using the WAIS-III Letter-Number

Sequencing (LNS) (Wechsler, 1997) and the Controlled Oral Word Association Test

(COWAT) (Lezak, 2004), respectively.

Serum collection and laboratory analyses

Fasting peripheral blood was collected from participants between 0900 and 1100h to control for diurnal variation of hormone levels. Blood samples were put on ice and stored at -80°C until assayed. Serum DHEA levels were analysed by the ANZAC

Research Institute in Concord, NSW, Australia with gas chromatography–mass spectrometry (GC-MS) analysis, which profiles and quantifies minute amounts of steroidal compounds with high accuracy. Serum cortisol levels were assayed by South

Eastern Area Laboratory Services in Randwick, NSW, Australia using a chemiluminescent immunometric assay (Siemens Immulite 2000). Interassay coefficient of variation were 13% for DHEA and X% for cortisol. The lower sensitivity limits of assays were 0.05 ng/ml for DHEA and 27.6 nmol/L for cortisol. Serum cortisol and

DHEA levels were determined while blind to diagnosis.

Structural scan acquisition

A subset of participants at the Sydney site (59 patients and 61 healthy controls) also received magnetic resonance imaging (MRI) scans and were included in imaging analyses. Structural MRI scans were acquired using a 3-Tesla Phillips

Achieva scanner with an 8 channel bird-cage type head coil at Neuroscience Research

Australia, Randwick, NSW, Australia. Each participant received a T1-weighted high- resolution anatomical scan (TR: 5.4 ms; TE: 2.4 ms; FOV: 256 mm; matrix: 256 × 256; sagittal plane; slice thickness: 1 mm, no gap; 180 slices).

Image processing

All scans were processed and analysed using the VBM8 toolbox (http://www.neuro.uni- jena.de/vbm) implemented in Statistical Parametric Mapping software (SPM12; http://www.fil.ion.ucl.ac.uk/spm) running under MATLAB version 2012b. The T1- weighted images were segmented into tissue classes of grey matter, white matter and cerebrospinal fluid. Following this, a high-dimensional DARTEL normalisation

(Ashburner, 2007) was performed for optimal registration of individual segments to a group mean template. The voxel-based morphometry analysis was restricted to differences in grey matter, therefore the resulting grey matter volume segments were modulated by the Jacobian determinants to correct for local volume changes introduced by normalisation. Finally, the DARTEL-normalised modulated grey matter segments were smoothed using a 8 mm full-width at half-maximum Gaussian kernel.

Homogeneity using covariance was performed on the entire sample to help identify outliers, followed by visual inspection for artifacts.

A priori regions of interest (ROI) were used for subsequent analyses. An ROI mask of the PFC was created by combining Brodmann areas 9 and 46 using WFU PickAtlas

(http://fmri.wfubmc.edu/software/pickatlas), dilated x1. An ROI mask of the hippocampus was created and defined by the Automated Anatomic Labelling system in

WFU PickAtlas, dilated x1 (See Supplementary Figure 4.1 for ROIs). The masks created in WFU Pick Atlas were resliced from the default 2x2x2 voxel dimension to

1.5x1.5.1.5 voxels in order to match the dimension of DARTEL-processed images.

Total intracranial volume (TIV) was calculated by adding up the native space volumes of the grey matter, white matter and cerebrospinal maps using the Tissue Volumes tool.

Two-sample t-tests, with age and TIV as nuisance parameters, were performed to compare gray matter volume in patients and controls. Analyses were restricted to the

PFC and hippocampus using the explicit masks. Results were generated using a family- wise error rate (FWE) correction with a p < 0.05 at the cluster-level.

Lastly, using the segmented, normalised, modulated and unsmoothed images, grey matter volume within the ROIs were obtained using get_totals.m Matlab script (Ged

Ridgway; http://www0.cs.ucl.ac.uk/staff/g.ridgway/vbm/get_totals.m) with a signal threshold ≥0.1 in order to perform correlation analyses with cortisol/DHEA ratios.

Hippocampal and PFC volumes were regressed onto TIV in order to control for differences in overall brain size and the residuals were used for the correlation analyses with cortisol/DHEA ratios.

Data Analysis

Statistical analyses were performed with IBM SPSS 22 statistical software (Armonk,

NY, USA). Demographics of healthy controls and patients were analysed with t-tests or 62

X2-tests, as appropriate. Raw DHEA, cortisol and cortisol/DHEA values were substantially skewed, therefore they were normalised through log10 transformation.

ANCOVAs were performed to compare mean DHEA, cortisol and cortisol/DHEA values between healthy controls and patients, with age as a covariate. Partial correlations between DHEA and cortisol levels were performed separately in controls and patients with age as a covariate. Pearson’s correlations of cortisol/DHEA ratios to cognitive measures were performed in patients and controls, separately, to determine the relationship of cortisol/DHEA ratios with cognitive functioning. Pearson’s correlations of cortisol/DHEA ratios to PANSS scores and were performed in patients to determine whether cortisol/DHEA ratios are associated with severity of illness.

Partial correlations of cortisol/DHEA ratios to PFC and hippocampal ROI volumes with age as a covariate were performed in healthy controls and patients separately. Partial correlations of mean daily chlorpromazine equivalent dose to PFC and hippocampal volumes and cortisol/DHEA ratio were performed in patients only. A Fisher’s r-to-z transformation was performed to test for potential significant differences in correlations between diagnostic groups, using a one-tailed test of significance given our a priori directional hypothesis that the cortisol/DHEA ratio would be associated with brain volume in patients but not in controls.

4.4 Results

Demographics, clinical characteristics and age-adjusted means for log-transformed cortisol, DHEA and cortisol/DHEA ratio are presented in Table 4.1. The patients showed mild to moderate symptom severity based on the PANSS scores. There was a 63

significant difference in age and an expected significant difference in education between the groups in which healthy controls were younger and had more years of education relative to patients. There were no significant differences in relation to the sex or ethnicity ratios between the groups. Patients with schizophrenia displayed significantly greater serum DHEA levels relative to healthy controls. There was no significant difference in serum cortisol levels between diagnostic groups. Cortisol/DHEA ratios were significantly decreased in people with schizophrenia as compared with healthy controls.

Table 4.1 Demographic variables, clinical characteristics and periphe ral markers of the whole sample Controls Schizophrenia

N Mean (SD) N Mean (SD) T/X2(df) p-value Age (years) 81 31.8 (8.4) 94 35.7 (8.5) 3.0 (173) p = 0.003 Education(years) 81 14.6 (2.2) 94 12.4 (2.3) 6.5 (173) p < .001 Sex (number) 81 94 2.2 (1) p = 0.14 Male 41 58 Female 40 36 Ethnicity (number) 81 94 5.0 (3) p = 0.17 Caucasian 65 82 Asian 10 4 Caucasian-Asian 2 5 Other 4 3

Premorbid and current IQ WTAR 81 107.8 (8.8) 94 102.2 (9.2) 4.1 (173) p < 0.001 WAIS-III 81 107.3 (14.9) 94 91.0 (12.7) 7.8 (173) p < 0.001

Cognitive tests LNS (Working memory) 81 10.9 (2.9) 94 8.1 (2.7) 6.7 (173) p < 0.001 COWAT (Verbal fluency) 81 42.0 (10.8) 94 37.2 (10.6) 3.0 (173) p = 0.003

Age of onset (years) 94 22. 7 (5.6) Illness duration (years) 94 13.0 (7.6) CPZ equivalents (mg) 94 555.5 (470.0) PANSS positive 94 15.0 (4.7) PANSS negative 94 14.3 (6.2) PANSS general 94 30.6 (8.9) PANSS total 94 60.0 (16.9)

Mean adj. Mean adj. (SE) (SE) Log cortisol 81 2.44 (0.02) 94 2.46 (0.02) 0.4 (172) p = 0.53 Log DHEA 81 0.54 (0.02) 94 .64 (0.02) 9.1 (172) p = 0.003 Log cortisol/DHEA 81 1.91 (0.03) 94 1.82 (0.02) 5.5 (172) p = 0.02

Abbreviations: CPZ, chlorpromazine; F, Female; M, male; LNS, letter number sequencing; PANSS, Positive and Negative Syndrome Scale; COWAT, Controlled Oral Word Association Test; WAIS-III, Wechsler Adult Intelligence Scale; WTAR, Wechsler Test of Adult Reading. Mean adj, mean adjusted for age; SE, standard error. Bold P-values indicate significant results.

Serum cortisol and DHEA levels were positively and significantly correlated in healthy controls (r = 0.23 (78), p = 0.04), and there was a stronger, significant positive correlation found in patients with schizophrenia (r = 0.43 (91), p < 0.001), see Figure

4.1.

Figure 4.1: Correlation between DHEA and cortisol in people with schizophrenia. DHEA was positively correlated with cortisol in our patient cohort (n = 81).

There were no strong, significant correlations between PANSS scores or cognitive test scores with cortisol/DHEA ratios in participants with schizophrenia (see Table 4.2).

Cortisol/DHEA ratios were not associated with cognitive test scores in healthy controls.

Table 4.2 Correlations of PANSS scores and cognitive tests with cortisol/DHEA i n patients and healthy controls Controls Schizophrenia df r p-value df r p-value cortisol/DHEA PANSS Positive - - - 92 0.06 0.58 PANSS Negative - - - 92 0.07 0.53 PANSS General - - - 92 0.02 0.89 PANSS Total - - - 92 0.05 0.65 Verbal fluency 81 0.11 0.33 92 0.06 0.61 Working memory 81 -0.04 0.72 92 0.13 0.20 Abbreviations: PANSS, Positive and Negative Syndrome Scale.

Relationship of PFC and hippocampus ROI volumes to cortisol/DHEA ratio

As expected, after covarying for age and TIV, people with schizophrenia demonstrated significant grey matter volume reductions relative to healthy controls in the hippocampus and PFC, FWE corrected at p < 0.05 (see Supplementary Figure 4.2,

Supplementary Table 4.2). There were no significant clusters in the hippocampus or

PFC where patients displayed greater volume than healthy controls. There were moderately strong, significant inverse correlations between cortisol/DHEA ratios and hippocampal, r = -0.33 (56), p = 0.012, and PFC, r = -0.41 (56), p = 0.001, volumes in patients (see Figure 4.2). There were no strong, significant correlations between mean daily CPZ dose and hippocampal, r = -0.002 (59), p = 0.99, or PFC r = -0.12 (59), p =

0.41, volumes in patients.

Figure 4.2: Association of cortisol/DHEA ratios to grey matter volume in schizophrenia. Cortisol/DHEA ratios are negatively correlated with hippocampal and PFC volume in people with schizophrenia (n=59).

Cortisol/DHEA ratios showed no strong, significant correlations with hippocampal (r =

-.01 (57), p = 0.93) or PFC (r = -.04 (57), p = 0.76) volumes in healthy controls. The correlation between cortisol/DHEA ratios and hippocampal volume in people with schizophrenia was significantly different from the same correlation observed in healthy controls (Fisher's Z = 1.74, one-tailed p = 0.04). Likewise, the correlation between cortisol/DHEA ratios and PFC volume in people with schizophrenia was significantly different from the same correlation obtained in healthy controls (Fisher's Z = 2.09, one- tailed p = 0.02).

4.5 Discussion

The key findings from the present study are that people with schizophrenia showed: (1) significantly higher serum DHEA levels; (2) significantly lower cortisol/DHEA ratios; and (3) a significant inverse relationship of cortisol/DHEA ratios with hippocampal and

PFC volumes. Although it was not a primary aim of the present study, we also replicated well-established reports of hippocampal and PFC volume reduction in people with schizophrenia (Nelson et al., 1998; Selemon et al., 2002; Selemon et al., 2003).

In Chapter 3 of this thesis we found significantly increased serum DHEA levels in both male and female patients relative to controls, suggesting a compensatory effect. In support of this, we found a moderately strong, positive relationship between DHEA and cortisol in participants with schizophrenia, reflecting a rise in DHEA levels related to antiglucocorticoid activity. However, interestingly, there was no significant diagnostic difference in cortisol levels in our sample, indicating that elevated DHEA levels in participants with schizophrenia may be a compensatory consequence of additional 69

stressors, such as inflammation and oxidative stress. Oxidative damage has been found in the hippocampus (Che et al., 2010; Nishioka & Arnold, 2004) and prefrontal cortex

(Gawryluk et al., 2011) in people with schizophrenia. In rodents and humans, administration of agents known to induce oxidative stress results in increased DHEA formation in the brain (Brown et al., 2003; Cascio et al., 2000). In people with schizophrenia, oxidative stress may precede and precipitate elevated DHEA levels; thus, elevated DHEA may also be indicative of neuroprotection against oxidative stressors.

Our group has reported that interleukin (IL)-1β, IL-6 and IL-8 mRNA are upregulated in the dorsolateral PFC in a subset of people with schizophrenia (Fillman et al., 2013).

DHEA decreases pro-inflammatory cytokine production in vivo (Kimura et al., 1998) and in vitro (Straub et al., 1998); therefore, increased DHEA levels in schizophrenia may represent anti-inflammatory processes and reflect the body’s attempt to protect against oxidative stress and inflammation. This may explain the elevated DHEA in the absence of elevated cortisol, resulting in decreased cortisol/DHEA.

Cortisol/DHEA ratios were not associated with positive or negative symptoms in our patient sample. Similarly, Silver et al. (2005) (Silver et al., 2005) and Ritsner et al.

(2004) (Ritsner et al., 2004) found no significant association between cortisol/DHEA ratios and symptomatology as assessed by the PANSS in people with schizophrenia.

Inconsistent reports regarding the relationship between cortisol/DHEA and clinical symptoms (Garner et al., 2011; Harris et al., 2001; Ritsner et al., 2004; Silver et al.,

2005; Yildirim et al., 2011) may be attributed to schizophrenia being a heterogeneous disorder with multiple symptom profiles and comorbidities. Methodological differences between studies may also partially account for differences in results: our study included

94 participants with schizophrenia while a study that reported an association between 70

cortisol/DHEA ratios and symptomatology only included 39 patients (Garner et al.,

2011). Harris et al. (2001)(Harris et al., 2001) found an association between

DHEA/cortisol ratios and symptomatology, in a sample of only 17 patients. Further, there were differences in symptom assessment tools where the PANSS was used to assess symptom severity in our sample while the Brief Psychiatric Rating Scale and the

Scale for the Assessment of Negative Symptoms were used by Harris et al.

(2001)(Harris et al., 2001) and Garner et al. (2011)(Garner et al., 2011), respectively.

Lastly, patients in the present study displayed relatively mild to moderate illness severity and were considerably less ill as compared with patients in some previous studies (Garner et al., 2011; Harris et al., 2001).

Cortisol/DHEA ratios were not associated with cognitive functioning in our patient sample. There is strong evidence that there are distinct cognitive subtypes in people with schizophrenia, where some patients show marked cognitive decline and others show preserved intellect (Weickert et al., 2000). Moreover, cognitive subtypes of schizophrenia are associated with meaningfully distinct differences in grey mater volume (Geisler et al., 2015). Therefore, appropriate phenotype refinement through classifying patients into more homogeneous subgroups of cognitive ability may be more informative regarding the relationship of cortisol/DHEA ratios and cognitive functioning, and should be performed when the sample size allows.

This study is the first to report an association between cortisol/DHEA ratios and brain volume reductions in people with schizophrenia. In addition to schizophrenia, reduced hippocampal volume has been demonstrated in other disorders in which stress may play a role including posttraumatic stress disorder (Gurvits et al., 1996), Cushing’s syndrome 71

(Starkman et al., 1992) and depression (Sheline et al., 1999), implicating glucocorticoid-induced brain atrophy. As discussed above, the cortisol/DHEA ratio may represent a response to multiple insults (stress, oxidative stress, inflammation), all of which have been associated with brain tissue damage in schizophrenia (Cannon et al.,

2015; Fraguas et al., 2012; Mondelli et al., 2010). Therefore, the molar ratio of cortisol to DHEA may be a molecular signal of increased sensitivity to brain insults that is detectable in the blood, contributing to dysfunctional outcomes observed in the illness.

This is supported by evidence that DHEA levels are associated with cortical thickness during early brain development (Nguyen et al., 2013). Therefore, our finding of a negative correlation of brain volume with cortisol/DHEA ratios in schizophrenia may be particularly relevant to the view that schizophrenia is a progressive brain disease (van

Haren et al., 2008).

Previous studies examining cortisol and DHEA in schizophrenia have assessed global outcome measures and behavioural phenotypes such as positive and negative symptoms, anxiety and depression. Our finding of a relationship between cortisol/DHEA ratios and brain volume highlights the importance of examining relationships with core biological abnormalities embedded within these conditions, warranting investigation of cortisol/DHEA ratios with other mechanisms and processes such as oxidative stress and inflammation. Further, we demonstrate that cortisol/DHEA ratios are informative of neuroanatomical volume and should be included in clinical studies when available.

This study is not without limitations. All patients in our sample were receiving antipsychotic treatment. There are reports that first and second-generation antipsychotics are associated with grey matter loss (Fusar-Poli et al., 2013; Vita et al., 72

2015). While we did not find an association of daily chlorpromazine equivalent dose with brain volume, we cannot entirely dismiss the possible confounding effects of chronic antipsychotic use. Another potential limitation of the current study is the measurement of DHEA without its sulfate ester, DHEA-S. DHEA is a short-life molecule and is quickly metabolised into its sulfated form, which is the most abundant circulating neurosteroid. Previous studies in schizophrenia have found differential results of DHEA and DHEA-S (Strous et al., 2004), suggesting their biological actions may differ. Future studies should determine how DHEA and its sulfate may differentially influence the course of illness.

Conclusions

This is the largest study to date to compare circulating levels of DHEA and cortisol/DHEA ratios in people with schizophrenia to healthy controls and the first study to link cortisol/DHEA levels with structural brain abnormalities. These results further implicate the role of steroids and HPA axis dysfunction in the pathophysiology of schizophrenia, with implications for treatment strategies that modulate these steroids.

Supplementary Table 4.1 Breakdown of antipsychotic medications Number of Antipsychotic subjects amisulpride 4 aripiprazole 3 clozapine 21 flupentixol 1 haloperidol 1 olanzapine 11 paliperidone 5 quetiapine 3 risperidone 10 ziprasidone 3 zuclopenithixol 3 amisulpride/clozapine 6 amisulpride/quetiapine 1 amisulpride/risperidone 1 aripiprazlone/asenapine 1 aripiprazole/clozapine 3 clozapine/haloperidol 1 clozapine/paliperidone 2 clozapine/chlorpromazine 1 clozapine/risperidone 2 olanzapine/paliperidone 1 olanzapine/risperidone 1 olanzapine/zuclopenithixol 1 paliperidone/quetiapine 2 quetiapine/risperidone 2 quetiapine/ziprasidone 1 quetiapine/zuclopenithixol 3

Supplementary Table 4.2 Regions with significant group differences in grey matter volume (healthy controls > schizophrenia) Brain region Cluster size Right/Left MNI stereotactic coordinates p(FWEcorr) T x y z PFC 448 R 9 39 21 0.000 7.38 124 L -22 38 39 0.006 5.46 339 L -6 45 18 0.001 5.43 179 R 46 39 14 0.003 5.30 37 R 9 26 33 0.020 4.99 36 L -51 14 32 0.020 4.84 Hippocampus 39 R 12 -32 8 0.024 5.55 1391 L -14 -24 -8 0.000 5.54 782 R 20 -15 -14 0.000 5.50 51 R 44 -14 -15 0.021 4.61 19 R 33 -36 -8 0.032 3.73 Cluster-level: p < 0.05, FWE corrected; Cluster size = number of voxels; L/R = left/right hemisphere; MNI: Montreal Neurological Institute; T=peak t-value

Supplementary Figure 4.1: Glass brain displaying anatomical ROIs used for analysis. Cyan = hippocampus. Red = prefrontal cortex (Brodmann areas 9/46).

Supplementary Figure 4.2: Comparison of grey matter volume in ROIs. Healthy controls > schizophrenia in the (a) hippocampus and (b) PFC. p <0.05, FWE corrected, controls n = 61, schizophrenia n = 59, Two-sample t-test, covariates age and TIV.

5.0 Endogenous testosterone levels are associated with neural activity in men with schizophrenia during facial emotion processing

5.1 Abstract Growing evidence suggests that testosterone may play a role in the pathophysiology of schizophrenia given that testosterone has been linked to cognition and negative symptoms in schizophrenia. Here, we determine the extent to which serum testosterone levels are related to neural activity in affective processing circuitry in men with schizophrenia. Functional magnetic resonance imaging was used to measure blood- oxygen-level-dependent signal changes as 32 healthy controls and 26 people with schizophrenia performed a facial emotion identification task. Whole brain analyses were performed to determine regions of differential activity between groups during processing of angry versus nonthreatening faces. A follow-up ROI analysis using a regression model in a subset of 16 healthy men and 16 men with schizophrenia was used to determine the extent to which serum testosterone levels were related to neural activity. Healthy controls displayed significantly greater activation than people with schizophrenia in the left inferior frontal gyrus (IFG). There was no significant difference in circulating testosterone levels between healthy men and men with schizophrenia. Regression analyses between activation in the IFG and circulating testosterone levels revealed a significant positive correlation in men with schizophrenia

(r = .63, p = .01) and no significant relationship in healthy men. This study provides the first evidence that circulating serum testosterone levels are related to IFG activation during emotion face processing in men with schizophrenia but not in healthy men, which suggests that testosterone levels modulate neural processes relevant to facial 78

emotion processing that may interfere with social functioning in men with schizophrenia.

5.2 Introduction

Impaired social functioning is a hallmark of schizophrenia that is gaining increasing attention. Poor social functioning in individuals with schizophrenia has been linked to abnormal processing of facial emotions (Ihnen et al., 1998). A meta-analysis reported a significant positive correlation between severity of negative symptoms and facial emotion perception impairment in schizophrenia (Kohler et al., 2010). People with schizophrenia also have difficulty identifying and discriminating among different facial expressions (Edwards et al., 2001), in particular those of negative valence (Kohler et al.,

2003, Mandal et al., 1998).

The underlying neural network of this impairment is not fully understood although functional neuroimaging studies show abnormalities in the limbic system and frontal cortex when people with schizophrenia are exposed to affective stimuli (Mandal et al.,

1998; Gur et al., 2002; Holt et al., 2006). Facial affect processing has been frequently studied as a method of exploring the neural substrates of social impairment in schizophrenia. In functional Magnetic Resonance Imaging (fMRI) studies, people with schizophrenia predominantly display hypoactivation in regions normally recruited by healthy individuals during the processing of facial emotions. One study showed that individuals with schizophrenia had diminished limbic response in the left amygdala and bilateral hippocampus during facial emotional discrimination relative to healthy adults

(Gur et al., 2002). In another study, people with schizophrenia showed significantly less activation than healthy controls in response to angry expressions in the inferior frontal gyrus (IFG), putamen and cerebellum (Phillips et al., 1999). The first part of the present study was designed to use fMRI to confirm the neural activation patterns underlying the 80

identification of angry compared to non-threatening facial expressions in a sample of men and women with schizophrenia relative to healthy men and women.

Steroid hormones and particularly testosterone have been implicated in emotion processing. Studies have reported both positive and negative associations between testosterone levels and emotion-related neural activation in healthy people. Derntl et al.,

(2009) found a significant positive correlation in healthy men between endogenous testosterone and amygdala response to fearful and angry facial expressions, but no correlation with nonthreatening expressions such as sadness and happiness. Similar findings have been reported in healthy women whose endogenous testosterone levels correlated positively with amygdala activity during the processing of fearful and angry faces (Van Wingen et al., 2009). Stanton et al., (2009) found that endogenous testosterone levels were negatively correlated with amygdala BOLD activity and positively correlated with ventromedial prefrontal cortex BOLD activity during the processing of angry faces; however, both findings occurred only in healthy males and were not found in healthy females. When middle-aged women were given a single dose of testosterone, Van wingen et al., (2009) reported positive correlations between exogenous testosterone and activity in the amygdala and superior frontal cortex along with a negative correlation between exogenous testosterone and neural activity in the orbitofrontal cortex and occipital gyrus in response to angry and fearful facial stimuli.

Another study reported a significant increase in neural activity in the amygdala and hypothalamus while viewing angry faces after healthy female participants received a .5 mL dose of testosterone (Hermans et al., 2008). Altogether, these findings generally suggest that endogenous testosterone modulates neural activity during processing of 81

negative facial emotion in healthy people, particularly males, although the exact mechanism is unknown and some conflicting results have been reported.

Steroid hormones may play a role in symptom onset, severity and the disease process associated with schizophrenia and have been linked to both cognition and negative symptoms (Moore et al., 2013; Ko et al., 2008; Akhondzadeh et al., 2006). However, few studies examining a possible relationship between testosterone and emotion processing in people with schizophrenia have been reported. Previous work by our lab demonstrated a strong inverse association between serum testosterone levels and activation of the bilateral middle frontal gyrus and left insula during an emotional word inhibition task in men with schizophrenia, but not in healthy men (Vercammen et al.,

2013), suggesting that testosterone may play a moderating role in the (frontal) hypoactivity observed in schizophrenia. In the second part of the present study we sought to determine the extent to which serum testosterone levels are related to neural activity in emotion processing circuitry in men with schizophrenia compared to healthy men, using a region of interest approach.

We predicted abnormal prefrontal and amygdala activity in men and women with schizophrenia during the processing of emotional faces. Given that most studies report a positive relationship between neural activity during processing of angry faces and testosterone levels in healthy men and women and abnormal neural activity during affective facial identification in schizophrenia, we predicted that the positive relationship between neural activity during affective facial processing and circulating testosterone levels in healthy men would be disrupted in men with schizophrenia.

5.3 Material and methods

Effects of emotional faces on neural activity in men and women

Participants

The study sample included 27 people with schizophrenia or schizoaffective disorder (17 male, 10 female) and 37 healthy comparison participants (20 male, 17 female). Patients were recruited from a national television documentary, the Kiloh Centre at the Prince of

Wales Hospital and other clinics from the South Eastern Sydney and Illawarra Area

Health Service. All patients were between 21 and 51 years of age and were receiving antipsychotic medication for at least one year prior to taking part in the study. Clinical diagnostic interviews using the Structured Clinical Interview for DSM-IV (SCID) (First et al., 2007) were conducted by a trained psychologist or psychiatrist. Symptom severity was assessed using the Positive and Negative Syndrome Scale (PANSS) (Kay et al.,

1987).

Healthy comparison participants between 20 and 42 years of age were recruited from the local community via advertisements. Exclusion criteria for all participants included substance abuse or dependency within the past 5 years, seizures, central nervous system infection, uncontrolled diabetes or hypertension, a history of neurological illness, head injury with loss of consciousness and structural brain abnormalities as assessed by MRI scan. Additional exclusion factors were a concurrent DSM-IV Axis I disorder in patients and any history of DSM-IV Axis I disorder in healthy controls.

All participants were assessed with the Wechsler Test of Adult Reading (WTAR)

(Wechsler, 2001) to obtain an estimate of (premorbid) intellectual functioning in patients and a four subtest version of the Wechsler Adult Intelligence scale, 3rd edition 83

(WAIS-III) (Wechsler, 1997) comprised of the Arithmetic, Digit Symbol, Similarities and Picture Completion subtests to assess current intellectual functioning. The study procedures were approved by the University of New South Wales and the South Eastern

Sydney and Illawarra Area Health Service Ethic Committees. All participants provided written informed consent prior to participation in the study and received reimbursement for their time and expenses.

Facial stimuli and procedure

Participants were presented with 60 colour pictures of human faces depicting emotional expressions (12 happy, 12 sad, 12 angry, 12 fear, 12 neutral) that alternated with a fixation cross (Gur et al., 2002). Each face was presented for 5.5s and on presentation, participants were required to choose the emotion displayed using a button box. Given that angry facial expressions more robustly elicit differential neural activity in schizophrenia relative to controls, for the purpose of this study, we focused on neural activation for angry versus non-threat (happy and neutral) faces.

Image acquisition

Echoplanar MR brain images were acquired using a 3 Tesla Phillips Achieva MRI scanner with an 8 channel bird-cage type head coil at Neuroscience Research Australia,

Randwick, NSW, Australia. First, each participant received a T1-weighted high- resolution anatomical scan in order to screen for structural abnormalities and for co- registration (TR: 5.4 ms; TE: 2.4 ms; FOV: 256 mm; matrix: 256 x 256; sagittal plane; slice thickness: 1 mm, no gap; 180 slices). During the emotional face recognition task,

210 T2*-weighted MR images providing blood-oxygen-level-dependent (BOLD) contrast were also acquired, TR/TE = 3000/30; 21 interleaved slices, slice thickness = 84

3.0 mm, gap = 1.0 mm, voxel size = 3 x 3x 3 mm; flip angle = 90°; field of view = 24 cm.

Demographic and task statistical analyses

Patient and control demographic factors were analysed using t-tests or Chi Square analyses as appropriate. Percentage correct and reaction time (RT) for the emotional face recognition task were analysed using Komogorov-Smirnov two sample tests since the data were not normally distributed.

Image processing and analyses

BOLD fMRI data were preprocessed and analysed using Statistical Parametric Mapping software (SPM8; http://www.fil.ion.ucl.ac.uk/spm) running under MATLAB version

2012b. For each participant, the 210 volume functional time series were motion corrected, transformed into stereotactic space (Montreal Neurological Institute, MNI) and smoothed with a 10 mm FWHM Gaussian filter. Each data set was then screened for excessive movement exceeding 3mm translation on x, y or z axes, artefacts, unsuccessful normalisation and for deactivation greater than three standard deviations away from the mean, which resulted in removal of 1 male patient and 3 male and 2 female healthy controls providing totals of 26 people with schizophrenia and 32 healthy controls for analysis.

At the first level of analysis, faces displayed with a happy or neutral affect were modelled together as “non-threat” and contrasted with angry faces for each participant.

At the second level, we first constructed single sample t-test models for the healthy control and schizophrenia groups separately to assess the main task effect for the 85

contrast of interest (angry vs. non-threat). Next, a two-sample t-test applying an explicit mask including the activated frontal and limbic regions (main task effect) from the healthy control group, i.e., a functional region of interest (fROI) analysis was performed to assess activation differences between groups. These regions included as part of the fROI have also been shown to be active during the identification of emotional faces in other studies (Mandal et al., 1998; Gur et al., 2002; Holt et al., 2006; Phillips et al.,

1999). Performance scores for angry and nonthreat faces were entered as covariates of no interest as they differed significantly between the diagnostic groups (see Results).

Age was entered as a covariate to account for significant differences in age between the groups. Lastly, two-sample t tests between male and female participants, in the healthy control and schizophrenia groups separately, were performed using the fROI to assess for any sex differences in activation. False discovery rate (FDR) corrections (p<.05) were applied for both the one-sample and two-sample t tests.

Relationship between testosterone and neural activity in the inferior frontal gyrus

(IFG) in men

In the second part of the study, one healthy male from the whole sample was excluded since no hormone levels were obtained from this participant. Thus, to assess the relationship between endogenous testosterone and neural activity in task-relevant areas, a subsample from the whole study, consisting of 16 males with schizophrenia (21 to 50 years old) and 16 healthy male controls (20 to 42 years old) were assessed.

Facial task, imaging acquisition and processing

The facial emotion task and stimuli, image acquisition and processing were identical to that described in the study of the whole sample above. 86

Hormonal assays and analyses

Samples of fasting peripheral blood were collected between 9 and 11 am to control for diurnal variation in hormone levels. Clotted and heparinised blood were delivered on ice to the Prince of Wales Hospital South Eastern Area Laboratory Services Pathology Unit immediately following collection. Circulating serum hormone levels were obtained as measures of “total” circulating hormone levels. Testosterone, estradiol and prolactin were assayed using a chemiluminescent immunometric assay (Siemens Immulite 2000).

Interassay coefficient of variation (CV) were 9.7% for testosterone and 13.5% for estradiol. Intra- and interassay coefficients of variation for prolactin were 3.4% and

6.8%, respectively. The lower sensitivity limits of assays were .7 nmol/l for testosterone,

73 pmol/l for oestrogen and 11 U/ml for prolactin. Hormone levels were compared between the groups using unmatched t-tests with alpha set at .05.

Relationship of testosterone to brain activity

We first performed correlation analyses using the region that yielded a significantly greater response in healthy controls relative to patients as our a priori fROI, which included the left IFG (maximum z: 4.13; voxels: 116; coordinates: -38, 24, 6). To assess the relationship between serum testosterone and BOLD response within the fROI, individual beta weights were extracted from the fROI using the MarsBar toolbox in

SPM (Bret et al., 2002). The beta weights were based on the mean signal of all voxels within the fROI. Partial correlation analyses between serum testosterone levels and activity in the fROI were performed separately in both groups with age as a covariate due to a significant difference in age between groups. Associations between testosterone 87

and neural activity in the fROI found in men with schizophrenia and healthy men were directly compared by performing a Fisher Z transformation of the Pearson’s r-values.

5.4. Results

Effects of emotional faces on neural activity in men and women

Demographic and clinical characteristics of the two groups are presented in Table 5.1.

There were significant differences in age and current and premorbid IQ estimates between groups. There was a trend toward a significant group difference in education.

There were no significant differences between groups on the basis of sex, handedness and ethnicity ratios. The people with schizophrenia in this sample were chronically ill, treated primarily with second generation antipsychotics, and displayed mild to moderate symptom severity based on PANSS scores.

Table 5.1: Demographic and clinical characteristics of the whole sample

schizophrenia controls T/χ2 df p n = 26 n = 32 Age (years) 37.3 (9.7) 29.5 (7.1) 3.56 56 < 0.001 Education (years) 13.5 (2.8) 14.7 (2.1) 1.9 56 0.06 WAIS-III FSIQ estimate 93.8 (12.2) 105.7 (11.1) 3.9 56 < 0.001 WTAR premorbid IQ estimate 103.8 (7.2) 110.3 (6.3) 3.64 56 < 0.001 Sex (M/F) 16/10 17/15 0.41 1 0.52 Ethnicity 2.95 3 0.40 Caucasian 18 23 Asian 3 7 Caucasian/Asian 3 1 Other 2 1 Handedness (R/L) 22/2 26/6 1.22 1 0.27 Diagnosis schizophrenia 13 schizoaffective disorder 13 Age of illness onset (years) 23.7 (6.6) Illness duration (years) 13.8 (7.5) Antipsychotic CPZ equivalent dose 619.2 (543.5) PANSS scores Positive 15.5 (5.2) Negative 14.8 (5.3) Total 60.7 (14.0) Second Generation Antipsychotics risperidone 2 clozapine 2 olanzapine 3 amisulpride 2 ziprasidone 2 quetiapine 1 clozapine + paliperidone 2 clozapine + amisulpride 2 clozapine + aripiprazole 1 clozapine + risperidone 1 risperidone + olanzapine 1 asenapine + aripiprazole 1 Second and First Generation Antipsychotics clozapine + chlorpromazine 1 clozapine + haloperidol 1 zuclopenthixol + olanzapine 1 zuclopenthixol + quetiapine 1 flupentixol 1 haloperidol 1 Notes: Standard deviation in parentheses. WAIS-III FSIQ = Wechsler Adult Intelligence Scale 3rd Edition Full Scale Intelligence Quotient, WTAR = Wechsler Test of Adult Reading, CPZ = chlorpromazine, PANSS = Positive and Negative Syndrome Scale.

Emotional face recognition task

The facial emotion identification task performance measures (Table 5.2) included reaction times (RTs) and response accuracy (% correct). The healthy control and schizophrenia groups did not differ significantly in reaction time. However, significant accuracy differences were found between the groups in relation to judging angry faces and non-threat faces, with the healthy controls performing better than the people with schizophrenia in both conditions.

Table 5.2: Mean reaction times and performance accuracy for people with schizophrenia and healthy controls during the facial emotional recognition task

Patients Controls U p Angry RT 2662 (491) 2636 (515) 458 0.69 Angry Accuracy 79% (15%) 90% (9%) 239.5 0.003 Nonthreat RT 2374 (429) 2169 (288) 557 0.06 Nonthreat Accuracy 84% (9%) 94% (9%) 159 <.001 SD in parentheses.

Imaging

Whole brain one-sample t-tests for the angry versus non-threat comparison revealed similar regions of activation in both groups. The healthy control group showed increased activity in the expected regions, including frontal (inferior, medial and superior frontal gyri, insula and anterior cingulate cortex), limbic (hippocampus, amygdala, parahippocampal gyrus), parietal (left superior parietal lobule), temporal

(superior temporal gyrus) and occipital (fusiform gyrus, lingual gyrus, inferior and middle occipital gyrus) regions (FDR p<.05; cluster ≥ 18; see Figure 5.1a and Table

5.3). People with schizophrenia showed increased activity in a similar but somewhat 90

more restricted network that included occipital (fusiform gyrus, lingual gyrus, inferior and middle occipital gyrus), temporal (fusiform gyrus) and frontal (middle and inferior frontal gyrus) regions (see Figure 5.1b and Table 5.3).

Figure 5.1: Coronal slices depicting areas of significant neural activation during the processing of angry versus non-threat in (A) 32 healthy control subjects and (B) 26 people with schizophrenia. 91

Table 5.3: Whole brain analysis showing regions of activation during viewing of angry versus non-threat facial expressions in healthy controls and people with schizophrenia

Brodmann Peak coordinate Brain region Side Voxels t value area X Y Z Healthy comparison group: Main effect Inferior occipital gyrus 18 L 1303 -30 -84 -10 6.97 (extends to lingual gyrus and fusiform gyrus) Middle occipital gyrus 18 R 2236 30 -90 -4 8.49 (extends to fusiform gyrus, cuneus and cerebellum) Parahippocampal gyrus L 78 -22 -10 -14 4.76 (extends to hippocampus, and amygdala) Inferior frontal gyrus 45 L 2019 -50 22 4 7.89 (extends to middle frontal gyrus and insula) Superior temporal gyrus 22 R 41 48 -36 4 4.56 Inferior frontal gyrus (extends to insula) 13 R 99 40 26 6 4.28 Precuneus 19 L 35 -26 -70 30 3.97

Superior parietal lobule 7 L 100 -26 -56 44 4.56 Medial frontal gyrus (extends to superior frontal gyrus, limbic lobe and 8 L 248 -6 16 52 5.80 cingulate gyrus)

Schizophrenia group: Main effect Middle occipital gyrus 18 R 1034 28 -82 -10 6.47 (extends to lingual gyrus and fusiform gyrus) Fusiform gyrus 37 L 68 -44 -58 -16 4.73 Middle occipital gyrus 18 L 317 -28 -82 2 4.93 (extends to inferior occipital gyrus) Inferior frontal gyrus 46 L 66 -46 26 12 4.89

Middle frontal gyrus 9 L 216 -40 18 28 5.96 Data are MNI coordinates for activation significant at FDR p<.05.

In the fROI between-groups analysis of the angry versus non-threat contrast, healthy control participants demonstrated significantly greater activation than people with schizophrenia in the left inferior frontal gyrus (IFG), FDR p<.05 (see Figure 5.2a).

There were no areas in which people with schizophrenia showed greater activation than controls. Using age as a covariate yielded the same region of activation but with fewer voxels (results not shown). There were no significant sex differences in activity during processing of angry versus non-threat faces in the healthy controls or the people with schizophrenia (results not shown).

Relationship between testosterone and IFG neural activity in men

Table 5.4 presents demographic comparisons in healthy men relative to men with schizophrenia. Men with schizophrenia differed significantly from healthy men in age and current and premorbid IQ estimates (see Table 5.4). There were no significant differences between these groups in years of education, handedness and ethnicity ratios.

Figure 5.2: Direct comparison of neural activity between groups and correlations between beta weights and testosterone levels. (A) Greater response in the left inferior frontal gyrus (maximum z: 4.13; voxels: 116; coordinates: -38, 24, 6) for anger versus non-threat response seen in healthy controls relative to patients and (B) Correlation between residual beta weights within the fROI and residual testosterone levels in men with schizophrenia and healthy men. 94

Table 5.4: Demographic and clinical characteristics of the healthy men and men with schizophrenia

schizophrenia controls T/χ2 df p n = 16 n = 16 Age (years) 36.0 (10.1) 27.8 (7.1) 2.7 30 0.01 Education (years) 12.9 (2.6) 14.0 (2.2) 1.25 30 0.22 WAIS-III FSIQ estimate 89.9 (10.4) 107.4 (12.0) 4.4 30 < 0.001 WTAR premorbid IQ estimate 101.9 (8.3) 112.1 (6.3) 3.93 30 < 0.001 Ethnicity 4.06 3 0.25 Caucasian 13 10 Asian 1 5 Caucasian/Asian 1 1 Other 1 0 Handedness (R/L) 12/2 12/4 0.54 1 0.46 Hormone assays testosterone (nmol/L) 12.4 (4.3) 15.2 (5.5) 1.6 30 0.13 estrodiol (pmol/L) 142.4 (45.6) 121.8 (37.8) 1.4 30 0.17 prolactin (mlU/ml) 323.2 (382.7) 156.4 (59.0) 1.7 30 0.10 Diagnosis schizophrenia 9 schizoaffective disorder 7 Age of Illness onset (years) 21.2 (4.3) Illness duration (years) 14.8 (8.2) Antipsychotic CPZ equivalent dose 722.0 (594.0) PANSS scores Positive 17.1 (4.8) Negative 16.0 (5.0) Total 64.6 (14.1) Second Generation Antipsychotics clozapine 2 olanzapine 2 amisulpride 2 clozapine + paliperidone 1 clozapine + amisulpride 2 clozapine + risperidone 1 risperdone + olanzapine 1 asenapine + aripiprazole 1 Second and first generation antipsychotics clozapine + chlorpromazine 1 clozapine + haloperidol 1 zuclopenthizol + quetiapine 1 zuclopenthixol + olanzapine 1 Notes: Standard deviation in parentheses. WAIS-III FSIQ = Wechsler Adult Intelligence Scale 3rd Edition Full Scale Intelligence Quotient, WTAR = Wechsler Test of Adult Reading, CPZ = chlorpromazine, PANSS = Positive and Negative Syndrome Scale. Prolactin-raising antipsychotics: amisulpride, chlorpromazine, haloperidol, flupentixol, paliperidone, risperidone, zuclopenthixol; prolactin-sparing antipsychotics: aripiprazole, asenapine clozapine, quetiapine fumerate, olanzapine, ziprasidone. Total number of participants receiving any prolactin-raising medication is 11. The other 5 participants were receiving one or a combination of prolactin-sparing antipsychotics. 95

Hormone levels

Serum testosterone, estradiol, and prolactin levels did not differ significantly between men with schizophrenia and healthy men (see Table 5.4).

Relationship of testosterone to brain activity

Correlation analyses between fROI activation in the IFG and circulating testosterone levels during angry versus non-threat emotional face processing conditions revealed a moderately strong, significant positive correlation (r = .63, p = .01) in men with schizophrenia but only a weak, non-significant inverse relationship in the healthy men (r

= -.21, p = .45; see Figure 5.2b). Fisher’s r-to-z transformation confirmed the difference between the correlations was significant, z = 2.42, p = .02, demonstrating that the relationship between testosterone and activity in the fROI was significantly greater in men with schizophrenia relative to healthy men.

5.5 Discussion

Our first hypothesis was that men and women with schizophrenia would show abnormal neural activity in frontal-limbic areas in comparison to healthy men and women. In accordance with our hypothesis and findings from previous studies, processing of angry versus non-threat facial expressions was associated with significant activation within limbic (hippocampus, amygdala, parahippocampal gyrus), occipital (fusiform gyrus, inferior and middle occipital gyri, lingual gyrus), temporal (superior temporal gyrus), parietal (left superior parietal lobule,) and frontal areas (inferior, medial and superior frontal gyri, insula) in healthy men and women, while people with schizophrenia 96

generally showed similar regions of activity but failed to show increased activity in the limbic areas (Kohler et al., 2003, Mandal et al., 1998; Phillips et al., 1999). Between group analyses showed that healthy men and women demonstrated significantly greater activation during recognition of threat-related faces than men and women with schizophrenia in the bilateral IFG, predominantly in the left hemisphere. Other groups have shown abnormalities in the left IFG in schizophrenia during working memory, language and emotion processing tasks (Meyer-Lindenberg et al., 2005; Phillips et al.,

1999). Functional and anatomical connectivity abnormalities in the left IFG, which have been reported in schizophrenia (Jeong et al., 2009), may underlie abnormalities in neural activation observed in our schizophrenia group. Underactivation of the left IFG has also been reported during various language tasks in people with schizophrenia

(Ragland et al., 2004). People with schizophrenia also had significantly lower neural activity in the left IFG when presented with emotionally ambivalent stimuli relative to healthy controls (Lee et al., 2014). Further, in processing happy faces, people with schizophrenia showed significantly decreased activation in the IFG relative to nonpsychotic siblings (Li et al., 2012). Our finding of significantly decreased activation in this region supports work showing that dysfunctional left IFG activity is implicated in emotion processing in people with schizophrenia.

Despite growing interest in the role of sex steroids in schizophrenia, few studies have examined the relationship between testosterone and neural activity using imaging methods. To our knowledge, this is the first study to report a relationship between circulating testosterone levels and facial emotion processing in men with schizophrenia.

We found a significant positive relationship between serum testosterone levels and activation of the IFG during processing of angry faces in men with schizophrenia. 97

Although previous studies have typically reported a positive relationship between neural activity during processing angry faces and endogenous testosterone levels in healthy men and women, we did not find a significant relationship in our healthy male sample.

However, while these previous studies examined the relationship between testosterone and activity in the amygdala and orbitofrontal cortex, while the present study focused on the left IFG using a fROI approach. As such, this may point to the existence of differential regional effects of endogenous testosterone. Differential explicit (as in the current study) versus implicit (as in previous studies of emotional faces) task demands between studies may also account for the differential relationships with testosterone.

Taken together, these findings suggest that differences in normal serum testosterone levels may modulate emotion processing deficits in the frontal cortex of men with schizophrenia when engaged in the processing of emotional (angry) faces.

Our results indicate that having low normal testosterone levels is associated with increasingly aberrant neural processing of angry faces in men with schizophrenia. The positive correlation between testosterone and brain activation in a region where people with schizophrenia show significantly less activation relative to healthy controls suggests that higher testosterone levels within a normal range may have beneficial effects for emotion processing in men with schizophrenia. Using an emotional go/no-go paradigm integrating cognitive and emotion processing, Vercammen et al., (2013) demonstrated that higher levels of testosterone were associated with better task performance and decreased activity in the prefrontal cortex, insula and precuneus in schizophrenia, suggesting that higher endogenous testosterone levels within a normal range may play an enhanced modulatory role in cognitive-emotional processing in men with schizophrenia. There was no significant difference in testosterone levels between 98

men with schizophrenia and healthy men in our sample, suggesting that the correlation between testosterone levels and brain activation in the fROI are due to hormonal- neuronal interactions and are not based on abnormal hormone levels. However, other studies have shown a reduction of circulating testosterone levels in men with schizophrenia relative to healthy men (Kaneda & Fujii, 2000) while others report no difference in circulating testosterone levels (Ferrier et al., 1982; Oades & Schepker,

1994). Hence, evidence for circulating testosterone differences in men with schizophrenia remains controversial and it may be that only a subset of males with schizophrenia have abnormally low testosterone levels.

Testosterone has been shown to have neuroprotective effects in neurodegenerative diseases such as Alzheimer’s disease and mild cognitive dysfunction (Bialek et al.,

2004). Furthermore, in healthy older men, addition of exogenous testosterone has been shown to improve spatial and verbal memory (Cherrier et al., 2007). One study to date has examined adjunctive treatment with testosterone in men with schizophrenia and found that those who were treated with testosterone showed significantly greater reduction of negative, but not positive symptoms (Ko et al., 2008). Further investigation of the efficacy of testosterone as an adjunctive therapy to treat emotion processing deficits and social impairment in schizophrenia will need to be conducted.

Although we found a significant correlation between testosterone levels and brain activity in emotion processing of angry faces in schizophrenia, it remains unclear whether the modulatory effect of testosterone is a consequence of the disease process or whether changes in sex steroid signalling is part of the pathogenesis. Some studies report abnormalities in hypothalamic-pituitary-gonadal (HPA) axis function in 99

schizophrenia that could offer possible explanations for the psychobiological role of testosterone in emotion processing deficits in schizophrenia (Bradley & Dinan, 2010).

Receptors of sex steroids are located throughout the human brain including in the cortical and subcortical areas (Perlman et al., 2004, 2005; Montague et al., 2008;

Beyenburg et al., 2000). Testosterone effects are mediated directly through the androgen receptor or indirectly through the estrogen receptor by conversion of testosterone to estrogen. In its free form, testosterone crosses the blood-brain barrier (Iqbal et al., 1983) and modulates the action of neuronal cells via genomic and nongenomic mechanisms

(Bialek et al., 2004). Thus, the modulatory effect of testosterone in schizophrenia could either play a role in the neurodevelopment of the disease or be a result of the disease process itself.

Limitations

It is important to point out some limitations of the present study. Prolactin-elevating antipsychotics have been shown to decrease testosterone levels, thus, the numerically increased prolactin levels in our sample may represent a potential confound. However, we did not find any statistically significant differences in serum prolactin or testosterone levels between healthy men and men with schizophrenia; thus, testosterone levels were not reduced below normal levels overall. Although the sample size of our study was consistent with other imaging studies of people with schizophrenia, future research examining the relationship between serum testosterone and brain activation in men with schizophrenia should examine larger samples and further assess potentially influencing factors (e.g., prolactin levels). For recruitment, we restricted the maximum age to 50 years as testosterone levels are known to decrease with age (Harman et al., 2001). There was a significant difference in age between groups although there were no significant 100

difference in testosterone levels and when age was entered as a covariate, we obtained a similar significant although smaller region within the same ROI. Further, previous studies have reported sex differences with greater activation in limbic areas (amygdala and parahippocampal gyrus) and prefrontal cortices in healthy male participants relative to healthy female participants during implicit emotional processing (Fusar-Poli et al.,

2009). However, in the present study, we found no sex differences in brain activity within the healthy control group or people with schizophrenia during emotional face recognition. The absence of prominent sex differences may be due to the relatively smaller sample size when analysing males and females separately, thus, a larger sample may also be useful to examine the potential modulatory effect of sex on brain activation in people with schizophrenia.

Most circulating testosterone is bound to sex hormone binding protein, which leaves only a small proportion to enter the cells and bind to receptors. Although we measured circulating total testosterone, which is an indirect measure of free testosterone, free testosterone correlates well with total testosterone and both have been used as an index of bioavailable testosterone (Vermeulen et al., 1999). Lastly, given the discrepant findings regarding testosterone in schizophrenia research, it is possible that the modulatory effect of testosterone in schizophrenia is present in a subgroup of the disease population. Thus, future studies with larger samples could assess differential effects on the basis of androgen and/or estrogen receptor genotype.

Conclusions

This preliminary finding provides the first evidence for a link between circulating testosterone levels and brain activity in a region characterised by hypoactivity during 101

negative emotion face processing in men with schizophrenia, suggesting that an increase in normal levels of testosterone may have beneficial effects for emotion processing in men with schizophrenia.

102

6.0 Adjunctive selective estrogen receptor modulator increases neural activity in the hippocampus and inferior frontal gyrus during emotional face recognition in schizophrenia

6.1 Abstract

Estrogen has been implicated in the development and course of schizophrenia with most evidence suggesting a neuroprotective effect. Treatment with raloxifene, a selective estrogen receptor modulator, can reduce symptom severity, improve cognition and normalize brain activity during learning in schizophrenia. People with schizophrenia are especially impaired in the identification of negative facial emotions. The present study was designed to determine the extent to which adjunctive raloxifene treatment would alter abnormal neural activity during angry facial emotion recognition in schizophrenia.

Twenty people with schizophrenia (12 men, 8 women) participated in a thirteen-week, randomized, double-blind, placebo-controlled, crossover trial of adjunctive raloxifene treatment (120 mg/day orally) and preformed a facial emotion recognition task during fMRI after each treatment phase. Two-sample t-tests in regions of interest selected a priori were performed to assess activation differences between raloxifene and placebo conditions during recognition of angry faces. Adjunctive raloxifene significantly increased activation in the right hippocampus and left inferior frontal gyrus compared to the placebo condition (FWE p < 0.05). There was no significant difference in performance accuracy or reaction time between active and placebo conditions. This study provides the first evidence suggesting that adjunctive raloxifene treatment changes neural activity in brain regions associated with facial emotion recognition in schizophrenia. These findings support the hypothesis that estrogen plays a modifying 103

role in schizophrenia and shows that adjunctive raloxifene treatment may reverse abnormal neural activity during facial emotion recognition, which is relevant to impaired social functioning in men and women with schizophrenia.

104

6.2 Introduction

Schizophrenia is a disabling psychiatric disorder, with a 70-80% unemployment rate

(Rosenheck et al., 2006), associated with multi-faceted deficits in cognitive function

(Gur et al., 2007; Weickert et al., 2000) and emotion processing (Kohler et al., 2000;

Schneider et al., 2006). Although antipsychotics are the first line of treatment for schizophrenia, these medications are often limited in their effectiveness and leave many patients with residual symptoms while producing unwanted side effects (Levine et al.,

2011). There is growing evidence that sex hormones may influence the course and symptoms of schizophrenia. The onset of the disease typically occurs during adolescence (Burke et al., 1990) and the clinical presentation, response to treatment and symptom severity can differ between men and women (Goldstein & Link, 1988; Gur et al., 1996; Loranger, 1984). Schizophrenia occurs less frequently and has a later average age of onset in women (Häfner, 2003; Ochoa et al., 2012). Furthermore, women tend to experience a less severe course of the disease compared to men (Gur et al., 1996).

Studies have found lower estrogen levels in women with schizophrenia relative to healthy women, relapses are more frequent when estrogen levels are low, such as during the early follicular phase of the menstrual cycle, postpartum and after menopause when there is a second peak of illness onset and a more severe course of illness (Gogos et al.,

2015; Hallonquist et al., 1993; Huber et al., 2001; Riecher-Rössler et al., 1992; Riecher-

Rossler et al., 1994; Riecher-Rossler & Kulkarni, 2011). These findings support the estrogen hypothesis of schizophrenia that posits that estrogen may have a neuroprotective effect against the disease (Seeman & Lang, 1990).

105

Alterations in dopaminergic and serotonergic systems are also key components of schizophrenia pathogenesis and animal research has shown a modulatory effect of estrogen upon dopamine and serotonin neurotransmitter systems in the brain (Bethea et al.; Biegon & McEwen, 1982; Di Paolo, 1994; Hafner et al., 1991). The neurobiological benefit of estrogen (Resnick & Maki, 2001) has led to an increase in the number of studies investigating estrogen as a potential therapeutic treatment in schizophrenia

(Akhondzadeh et al., 2003; Begemann et al., 2012; Kulkarni et al., 2008; Kulkarni et al.,

2011; Kulkarni et al., 1996; Kulkarni et al., 2014). Despite growing positive evidence that estrogen may reduce symptom severity in schizophrenia and may benefit cognition during aging (Jacobs et al., 1998; Tang et al., 1996; Zandi et al., 2002), little is known regarding the therapeutic effects of estrogen on aspects of schizophrenia that are relatively unresponsive to standard therapeutic intervention, such as cognitive dysfunction and social impairment.

Selective estrogen receptor modulators (SERMs) specifically activate estrogen receptors and no other nuclear receptors (Kian Tee et al., 2004; Lonard & Smith, 2002). Unlike estrogen, SERMs do not stimulate estrogen receptors in breast or uterine tissue

(therefore, avoiding any adverse effects in these tissues). Raloxifene is a first generation

SERM that is used to treat osteoporosis in postmenopausal women (MacGregor &

Jordan, 1998) and has demonstrated benefits in preventing age-related decreases in neural activity in healthy older men (Goekoop et al., 2005). More recently, adjunctive treatment with raloxifene in postmenopausal women with schizophrenia has demonstrated reductions in positive and negative symptom severity and general psychopathology (Kulkarni et al., 2010; Usall et al., 2011). A recent clinical trial has shown that adjunctive raloxifene administered at 60 mg/day improved memory and 106

verbal fluency in post-menopausal women with schizophrenia (Huerta-Ramos et al.,

2014). Additionally, we have shown that daily, oral adjunctive raloxifene treatment at

120 mg/day improved verbal memory and attention and increased brain activity during learning in both men and women with schizophrenia (Kindler et al., 2015; Weickert et al., 2015). However, no clinical trial to date has examined the extent to which adjunctive raloxifene treatment may influence facial emotion recognition, which is a critical skill linked to social function in men and women with schizophrenia (Hooker &

Park, 2002).

Poor social functioning is a core feature of schizophrenia (American Psychiatric

Association, 2000) that is associated with deficits in emotion processing in which people with schizophrenia often have difficulty identifying and discriminating among different facial expressions (Edwards et al., 2001). Most evidence suggests that people with schizophrenia display hypoactivation in frontal and limbic regions and perform worse relative to healthy individuals during facial emotion identification tasks (Gur et al., 2002; Ji et al., 2015; Phillips et al., 1999). Regions of particular interest in relation to facial emotion recognition are the bilateral amygdala, hippocampus and inferior frontal gyrus (IFG). Amygdala damage has been linked to impaired recognition of fearful and angry facial expressions (Adolphs et al., 1994; Broks et al., 1998; Calder,

1996). Further, amygdala dysfunction has been implicated in facial emotion processing tasks in people with schizophrenia, with several studies reporting both hyper or hypo activation (Hempel et al., 2003; Li et al., 2010; Takahashi et al., 2004; Williams et al.,

2004). Thus, although functional abnormality of the amygdala in schizophrenia is well established, the exact direction of the disruption remains unclear.

107

Neural activity in the hippocampus is associated with amygdala activity and emotional memory. The hippocampus represents a key interface between sensory systems and the limbic system, and is necessary to form and elicit long-term memories (Squire et al.,

1992). It has been suggested that the hippocampus encodes the emotional sense of experiences so that they may be recalled at a later time in association with an emotional valence (Bellace et al., 2013). Together, the hippocampus and amygdala interact during emotion processing and emotional memory retrieval (Buchanan, 2007; Phelps, 2004).

Regarding the cortical IFG, a meta-analysis of 105 fMRI studies in healthy participants reported significantly increased IFG activation during processing of angry faces relative to a baseline control condition (Fusar-Poli et al., 2009). In schizophrenia, fMRI studies report abnormal IFG activation during semantic processing (Jeong et al., 2009) and emotion processing (Lee et al., 2014). We recently reported lower levels of neural activity in the left IFG in schizophrenia relative to healthy controls during recognition of negative facial emotions (Ji et al., 2015).

The aim of the present study was to determine the extent to which a hormone intervention therapy using the SERM raloxifene will influence neural activity underlying recognition of facial emotions in men and women with schizophrenia. For the fMRI analysis we focused on cortical and subcortical regions of interest where there is evidence of functional abnormalities during facial emotion recognition in people with schizophrenia compared to healthy controls: the amygdala, hippocampus, and IFG. We focused on angry versus neutral faces because people with schizophrenia are particularly impaired in the identification of faces displaying negative emotions (Kohler et al., 2003; Mandal et al., 1998). Further, there is evidence of a relationship between 108

the magnitude of brain activity during processing of direct social threat (anger) and symptom improvement in people with schizophrenia while this relationship was not found for indirect threat (fear) (Kumari et al., 2011). We predicted that administration of raloxifene will increase brain activity in the hippocampus and IFG in men and women with schizophrenia, compared to the placebo. Due to contrasting reports regarding the direction of abnormal amygdala activation in schizophrenia we had a non-directional hypothesis that raloxifene would have a significant influence on neural activity relative to placebo.

6.3 Materials and Methods

Participants

The study sample consisted of 20 people with schizophrenia or schizoaffective disorder

(12 male, 8 female). Patients were recruited via a national television documentary, the outpatient mental health unit at the Prince of Wales Hospital and community mental health clinics in the South Eastern Sydney and Illawarra Area Health Service. All patients were between 22 and 51 years of age and were receiving antipsychotic medication for at least one year prior to taking part in the study. Clinical diagnostic interviews using the Structured Clinical Interview for DSM-IV (SCID) (First et al.,

2007) were performed by a trained psychologist or psychiatrist. Symptom severity was assessed using the Positive and Negative Syndrome Scale (PANSS) (Kay et al., 1987).

Duration of illness was defined as the difference between the age at first hospitalization and age at the time of scanning. Exclusion criteria included a comorbid Axis I DSM-IV disorder, substance abuse or dependency within the past 5 years, seizures, central nervous system infection, uncontrolled diabetes or hypertension, a history of 109

neurological illness, head injury with loss of consciousness, structural brain abnormalities as assessed by MRI scan, intellectual disability (current IQ < 70) or contraindications to the administration of raloxifene. Women were excluded if they were currently pregnant or were receiving hormone therapy and refused alternate forms of birth control. See Table 6.1 for demographic and clinical characteristics of the sample.

Fasting peripheral blood samples were collected between 9am and 11am to control for alterations in hormone levels due to diurnal variations. Clotted and heparinised blood were delivered on ice to the Prince of Wales Hospital South Eastern Area Laboratory

Services Pathology Unit immediately following collection. Prolactin, follicle- stimulating hormone and luteinizing hormone were assayed using a chemiluminescent immunometric assay (Siemens Immulite 2000).

Participants were assessed with the Wechsler Test of Adult Reading (WTAR)

(Wechsler, 2001) to obtain an estimate of premorbid intellectual functioning and a four subtest version of the Wechsler Adult Intelligence scale, 3rd edition (WAIS-III)

(Wechsler, 1997) comprised of the Arithmetic, Digit Symbol, Similarities and Picture

Completion subtests to assess current intellectual functioning. The study procedures were approved by the University of New South Wales and the South Eastern Sydney and Illawarra Area Health Service Ethic Committees. All participants provided written informed consent prior to participation in the study.

Study Design

In a thirteen week, randomized, double-blind, placebo-controlled, crossover trial patients received 120mg/day of encapsulated raloxifene and placebo (encapsulated 110

lactose) as an adjunctive treatment to their currently prescribed antipsychotic medication (see Supplementary Figure 6.1). All quality assessment/control testing of encapsulated raloxifene was performed by IDT Australia Ltd, Victoria, Australia.

Following a baseline assessment, participants were randomly assigned to receive raloxifene (10 patients) or placebo (10 patients) for 6 weeks using a computer generated randomization schedule provided by the Prince of Wales Hospital Pharmacy Clinical

Trials Unit. Following the first 6-week period of the trial, there was a one week

"washout" period followed by the second 6-week period when patients received the alternate treatment (raloxifene or placebo). Patients were monitored throughout the trial to assess potential adverse events and compliance was determined by returned pill counts and hormonal blood assays. Functional Magnetic Resonance Imaging (fMRI) was used to measure blood oxygen level-dependent (BOLD) signal changes as patients performed a facial emotion recognition task at week 6 (end of trial period 1) and at week 13 (end of trial period 2).

Facial emotion recognition task

During each test session, participants completed a facial emotion recognition task in the scanning environment (see Supplementary Figure 6.2). Stimuli consisted of 60 unique colour pictures of human faces representing equal numbers of the following emotions: anger, fear, happy, sad and neutral (Gur et al., 2002). Each of the five emotional expressions was presented a total of 12 times. Stimuli were presented on an inverted computer screen via a set of mirrors in the scanner for 5.5 s each and for each presentation, individuals were asked to identify the affect displayed using a button response box.

111

Demographic and task statistical analyses

Data analysis was performed using IBM SPSS 22 for Windows. Antipsychotic dose, hormone levels, PANSS scores and task performance measures were compared between treatment conditions using t-tests or Wilcoxon rank-sum tests as appropriate. Task performance included measures of accuracy (% correct responses) and reaction times

(RTs) for each facial expression. Due to technical problems, behavioural data was limited to 17 patients during the active raloxifene phase and 11 patients during the placebo phase.

Image acquisition and processing

Echoplanar MR brain images were acquired using a 3 Tesla Phillips Achieva MRI scanner with an 8 channel bird-cage type head coil at Neuroscience Research Australia,

Randwick, NSW, Australia. Each participant received a T1-weighted high-resolution anatomical scan in order to screen for structural abnormalities and for co-registration

(TR: 5.4 ms; TE: 2.4 ms; FOV: 256 mm; matrix: 256 x 256; sagittal plane; slice thickness: 1 mm, no gap; 180 slices). During the facial emotion recognition task, 210

T2*-weighted MR images providing BOLD contrast (TR/TE = 3000/30; 21 interleaved slices, slice thickness = 3.0 mm, gap = 1.0 mm, voxel size = 3 x 3x 3 mm; flip angle =

90°; field of view = 24 cm) were acquired. Three dummy scans were obtained before each fMRI data acquisition to allow for the equilibration of the MRI signal.

BOLD fMRI data were preprocessed and analysed using Statistical Parametric Mapping software (SPM8; http://www.fil.ion.ucl.ac.uk/spm) running under MATLAB version

2012b. For each participant, the 210 volume functional time series images were realigned to the first image in the sequence and coregistered to the T1 anatomical scan. 112

Images were transformed into stereotactic space (Montreal Neurological Institute, MNI) and smoothed with a 10 mm FWHM Gaussian filter. All data sets were screened for artefacts, excessive movement exceeding 3mm translation on x, y or z axes and successful normalization. Motion parameters were included as regressors in the first level analysis to further control for motion effects.

fMRI analyses

At the first level of analysis, a contrast was created for subject-level time series to assess the difference in BOLD signal between conditions of interest: angry faces versus neutral faces. At the second level, we first constructed single sample t-test models for the active and placebo conditions separately at the whole brain level to assess the main task effect for the recognition of angry faces. Next, to assess activation differences between raloxifene and placebo conditions in specific regions selected a priori which have been previously shown to have a lower level of activation during emotional face recognition in schizophrenia (Gur et al., 2002; Ji et al., 2015; Phillips et al., 1999), we performed paired t-tests between raloxifene and placebo conditions in the regions of interest (ROI).

The following bilateral structural ROIs were selected from the Anatomical Automatic

Labelling (AAL) Atlas in the SPM8 toolbox (Brett et al., 2002): amygdala, hippocampus and IFG. We corrected for multiple comparisons across the whole brain one-sample t tests using false discovery rate (FDR) corrections (p < 0.05). Small volume corrections (SVC) were applied for ROIs (FWE p < 0.05).

113

6.4 Results

Demographics, symptom measures, compliance and blood analyses

Demographic and clinical characteristics of the study sample are presented in Table 6.1.

The men and women with schizophrenia in this sample were chronically ill, treated primarily with second generation antipsychotics, and displayed mild to moderate symptom severity based on PANSS scores. Based on returned pill counts, compliance for period 1 of the trial was 97.6% and 97.3% for period 2 for this fMRI study. There were no severe adverse events that were attributed to the study medication. There was no statistically significant difference between PANSS-positive and –negative symptoms scores during raloxifene and placebo conditions; however, there was a trend towards a statistically significant decrease in PANSS positive symptom severity scores with raloxifene treatment. There were no clinically relevant differences on the hormone panel measures between the raloxifene and placebo conditions although there was a statistically significant increase in follicle-stimulating hormone levels with raloxifene treatment.

114

Table 6.1: Demographics, clinical characteristics and blood analyses of patients (n=20)

Placebo Raloxifene t/Z Variable Baseline treatment treatment df value p Age (years) 36.5 (8.5) Education (years) 13.4 (2.1) WAIS-III IQ Estimated full scale IQ 92.4 (10.0) WTAR Estimated premorbid IQ 104.0 (6.5)

Sex (M/F) 12/8 Race (total) Caucasian 15 Asian 1 Caucasian/Asian 3 Other 1 Handedness (Right/Left) 19/1 Diagnosis Schizophrenia 14 Schizoaffective 6 Age of onset 23.7 (6.3) Illness duration 12.9 (6.8) Number of hospitalisations 0-5 13 >5 7 Antipsychotic dose (CPZ 699.0 680.7 equivalent) 703.0 (590.0) (592.0) (574.7) 19 1.34 0.18 % on prolactin-raising antipsychotics 55% Hormone assays 500.3 484.2 Prolactin (ml U/ml) 619.8 (654.6) (457.7) (474.4) 19 1.45 0.15 Luteinizing Hormone 5.2 (5.0) 6.6 (6.8) 5.1 (4.6) 19 <.001 1.0 Follicle Stimulating Hormone 9.2 (23.3) 9.1 (22.9) 9.8 (20.2) 19 2.07 0.04 PANSS Positive 14.5 (4.9) 14.4 (5.7) 13.4 (4.4) 19 2.03 0.06 Negative 14.8 (6.4) 14.5 (4.7) 14.1 (5.7) 19 0.54 0.6 General 30.7 (7.9) 28.4 (6.6) 28.5 (6.9) 19 0.14 0.89 Total 60.0 (16.6) 57.3 (13.9) 56.0 (13.8) 19 0.76 0.46 Second generation antipsychotics Olanzapine 4 Clozapine 1 Amisulpride 2 Risperidone 2 Aripiprazole 1 Ziprasidone 1 Quetiapine 1 Clozapine + paliperidone 1 Clozapine + amisulpride 2 115

Clozapine + aripiprazole 1 Clozapine + chlorpromazine 1 Risperidone + quetiapine fumerate 1 Second and first generation antipsychotics Clozapine + haloperidol 1 Zuclopenthixol + olanzipine 1 Standard deviation in parentheses; WAIS-III FSIQ, Wechsler Adult Intelligence Scale 3rd edition Full Scale Intelligence Quotient; WTAR, Wechsler Test of Adult Reading; CPZ, chlorpromazine; PANSS, Positive and Negative Syndrome Scale. Prolactin-raising antipsychotics: amisulpride, chlorpromazine, haloperidol, flupentixol, paliperidone, risperidone, zuclopenthixol; prolactin-sparing antipsychotics: aripiprazole, asenapine, clozapine, quetiapine fumerate, olanzapine, ziprasidone. Total number of participants receiving any prolactin-raising medication is 11. The other 9 participants were receiving one or a combination of prolactin-sparing antipsychotics.

116

Behavioural results

Performance measures of participants on the facial emotion recognition task included reaction times (RTs) and response accuracy (% correct) presented in Supplementary

Table 6.1. The Wilcoxon signed-rank tests indicated that there was no significant difference in accuracy and RT between adjunctive raloxifene and placebo conditions.

Imaging

Whole brain one-sample t-tests for angry versus neutral face recognition revealed significant activation for the raloxifene and placebo conditions (see Figure 6.1, Table

6.2). During raloxifene treatment, angry face recognition elicited bilateral activation in a widespread network including inferior, middle, medial and superior frontal gyrus, middle temporal gyrus, inferior parietal lobe, insula, lingual gyrus, fusiform gyrus, parahippocampal gyrus and cingulate gyrus (FDR p < 0.05). The placebo condition elicited activation in a similar yet more restricted network including inferior, middle, medial and superior frontal gyrus and precentral gyrus (FDR p < 0.05).

117

Figure 6.1: Task-related neural activity. Coronal slices depicting areas of significant neural activation while viewing angry versus neutral faces during raloxifene condition (a) and during the placebo condition (b).

118

Table 6.2: Whole brain analysis showing regions of activation during anger processing in people with schizophrenia during raloxifene treatment and placebo condition

Brodmann Peak coordinate Brain region L/R CS T area X Y Z Raloxifene treatment: Main effect Lingual gyrus (extends to middle occipital and inferior occipital 18 R 1321 10 -86 -10 6.30 gyrus)

Lingual gyrus (extends to fusiform gyrus, cerebellum and 18 L 1110 -18 -88 -8 6.29 parahippocampal gyrus)

Fusiform gyrus 37 R 166 50 -50 -10 4.46 Inferior frontal gyrus 44 L 2900 -44 24 24 7.66 (extends to middle frontal gyrus and insula) Inferior frontal gyrus (extends to insula) 47 R 119 30 34 6 4.82

Middle temporal gyrus 19 R 48 36 -82 16 4.31

Middle occipital gyrus 19 L 104 -32 -84 16 6.33

Middle frontal gyrus 46 R 373 50 28 24 4.65

Inferior parietal lobe 7 L 242 -32 -48 44 5.37

Medial frontal gyrus (extends to medial and superior frontal gyrus, 8 L 475 -8 16 50 5.99 limbic lobe and cingulate gyrus)

Placebo: Main effect Middle frontal gyrus 46 L 535 -48 28 24 6.77 (extends to inferior frontal gyrus) Middle frontal gyrus 9 R 81 42 12 32 4.68 (extends to inferior frontal gyrus) Middle frontal gyrus 6 L 32 -50 8 46 4.59

Precentral gyrus 6 L 42 -42 -4 44 5.47 Medial frontal gyrus (extends to superior frontal 6 R 222 4 14 52 5.66 gyrus) CS = cluster size (voxels). Data are MNI coordinates for activation significant at FDR p < 0.05.

119

When comparing the 3 brain regions selected a priori by ROI (amygdala, hippocampus

and IFG bilaterally) patients showed greater activation only within the bilateral inferior

frontal gyrus and hippocampus with raloxifene treatment relative to the placebo

condition; however, only the left inferior frontal gyrus and right hippocampus reached

the strict statistical significance level (see Figure 6.2, Table 6.3). The placebo condition

did not elicit greater activation than the raloxifene treatment in any region during

recognition of angry faces. There were no significant activation differences between

raloxifene and placebo conditions in relation to the amygdala ROI.

Table 6.3: Region of interest analyses showing significant treatment effects

Contrast L/R CS x y z T Z Brain region

Raloxifene > placebo L 86 -44 -48 4 5.80* 4.35 Inferior frontal gyrus

R 36 34 -14 -24 4.71* 3.79 Hippocampus

Placebo> raloxifene none

Significant activation differences in the hippocampus and inferior frontal gyrus following raloxifene treatment compared to placebo. Centers of activation clusters are given by MNI stereotactic coordinates (x, y, z). CS = cluster size, T= t-value, Z= z-value, l/r= left/right hemisphere, *small volume FWE corrected, p < 0.05.

120

Figure 6.2: Effects of raloxifene on BOLD activity in ROIs. Patients showed significantly greater activation within the left inferior frontal gyrus and right hippocampus (small volume FWE corrected, p < 0.05) during raloxifene treatment relative to the placebo condition.

6.5 Discussion

The main aim of the present study was to determine the extent to which raloxifene would influence neural activity associated with recognition of faces with a negative valence in men and women with schizophrenia. In accordance with our hypothesis, angry face recognition elicited significantly greater activation in the left IFG and right hippocampus during raloxifene treatment relative to placebo of the same individuals with schizophrenia studied over time. However, we did not detect any change in performance accuracy over the course of the study with raloxifene administration during facial emotion recognition in people with schizophrenia.

The hippocampus is involved in emotion processing and has been shown to be hypoactive in schizophrenia during facial emotion processing tasks. Using the same emotion-recognition paradigm as in the present study, Gur et al. (Gur et al., 2002) found that people with schizophrenia showed significantly less activation in the bilateral 121

hippocampus relative to healthy controls during exposure to positive and negative facial expressions. While we did not find significant increases in hippocampal activity in both hemispheres, we were able to detect an increase in activity of the right hippocampus.

Our finding of increased activation in ROIs that are hemisphere-specific supports previous work showing evidence for the laterization of emotion processing (Killgore &

Yurgelun-Todd, 2007). Papanicolaou et al. (Papanicolaou et al., 2002) demonstrated specialization of the hippocampus in healthy individuals where the left hippocampus activates during mnemonic processing of verbal items and the right side activates for processing visual stimuli that are difficult to encode verbally. Similarly, Bellace et al.

(Bellace et al., 2013) reported greater activation in the left hippocampus for emotional words and the right hippocampus for emotional pictures. Another fMRI study reported that encoding faces was associated primarily with activation in the right hippocampus

(Kelley et al., 1998). Thus, our finding of significantly greater activation in the right hippocampus during raloxifene compared to placebo may reflect the ability of raloxifene to enhance neural processing of emotional visuospatial stimuli.

Several fMRI studies of emotion processing in schizophrenia have found abnormal activity in cortical regions, particularly the IFG. We previously reported hypoactivation of the left IFG in people with schizophrenia relative to healthy controls during angry face processing where matching the emotional face to the correct emotional word is required (Ji et al., 2015). Similarly, Russell et al. (Russell et al., 2000) showed less

BOLD signal in the left IFG during a socioemotional task in men with schizophrenia compared to healthy controls. Furthermore, decreased functional connectivity of the

IFG, which may contribute to abnormalities during emotion processing, has been reported during semantic processing (Jeong et al., 2009) and cognitive processing 122

(Peeters et al., 2015). Goekoop et al. (Goekoop et al., 2005) reported increased activation in the left IFG following raloxifene treatment compared to placebo in healthy elderly males during a face recognition task. Our finding of increased activation in the left IFG with adjunctive raloxifene supports findings of Goekoop et al. (Goekoop et al.,

2005) on the effect of raloxifene on brain activity. Further, the left IFG includes Broca’s area which plays a role in verbal processing (Broca, 1861). The paradigm in the present study requires subjects to match the facial emotion with the written emotion displayed, thus marked increase in activation in the left IFG during raloxifene may reflect increased verbal processing to identify the presented visuospatial stimuli.

We did not observe any significant changes in brain activation in the amygdala during recognition of angry faces between raloxifene and placebo conditions. Some neuroimaging studies report hypoactivation of the amygdala during processing of facial emotions, while others report normal or enhanced activation of this region in people with schizophrenia (Hempel et al., 2003; Li et al., 2010; Takahashi et al., 2004;

Williams et al., 2004). One study sought to determine whether variability in amygdala activation may be related to the time-course of the experiment. People with schizophrenia initially showed hyper-responsivity of the amygdala to facial expressions relative to healthy controls; while in the latter phase of the experiment patients exhibited decreased bilateral amygdala response (Suslow et al., 2013). It has been well established that amygdala response to facial expressions habituates with repeated presentations

(Britton et al., 2008; Wright et al., 2001); however, Suslow et al. (Suslow et al., 2013) demonstrated a more rapid decrease in amygdala response over time in schizophrenia relative to healthy people. Therefore, heterogeneous findings regarding amygdala activation during emotional face processing may result from differences in length of the 123

experimental paradigms, such that longer tasks are associated with habituation of amygdala activity and therefore less activation when averaged across time. Our study did not show a significant difference in the level of amygdala activation during placebo and raloxifene conditions in schizophrenia. It is possible that raloxifene does not affect activation in this region or that any increase in activation was limited to the initial stimuli presentations and thus, was not significant when we assessed across the whole task.

The exact mechanism by which raloxifene exerts its effects is not entirely known, but may involve neurotransmitter systems or other neuromodulators. The action of raloxifene on mood may be related to mimicking estrogenic effects on pre- and postsynaptic modulation of serotonergic and/or dopaminergic neurotransmission (Cyr et al.,

2000; Landry et al., 2002; Sánchez et al., 2010). A large body of evidence has shown that dopamine neurotransmission influences hippocampal plasticity and function (Kulla

& Manahan-Vaughan, 2000; Rossato et al., 2009; Sajikumar & Frey, 2004), and there is an increase of dopaminergic activity in the left inferior frontal gyrus during human emotion processing (Badgaiyan et al., 2009). Raloxifene has also been shown to decrease the inflammatory response in mouse and rat microglia cells in vitro (Suuronen et al., 2005). Our group recently found increased inflammatory mRNA expression in a subgroup of people diagnosed with schizophrenia (approximately 40%) suggesting that anti-inflammatory therapies may be of benefit (Fillman et al., 2013; Fillman et al.,

2015). Thus, it is possible that the ability of raloxifene to reverse abnormalities in brain activation in patients may in part be due to its ability to suppress inflammation. Further, a number of studies have shown that raloxifene has antioxidant properties (Armagan et al., 2009; Konyalioglu et al., 2007), therefore its beneficial effects in schizophrenia may 124

be related to the control of oxidative stress; which has been shown to be elevated in people with schizophrenia (Emiliani et al., 2014).

The hippocampus is one of the primary sites of estrogen receptors in the brain and in animal studies oestrogen administration has been linked to increases in synaptic spine density (Woolley & McEwen, 1994). Raloxifene has also been shown to reduce neuronal loss in the hippocampus (Ciriza et al., 2004). Our group previously demonstrated increased activation of the right hippocampus during probabilistic association learning with adjunctive raloxifene (Kindler et al., 2015). Thus, our present finding of increased neural activation in the hippocampus during raloxifene administration is consistent with these previous findings of the role of oestrogen receptor modulation in the hippocampus. We also find increased activity in the prefrontal cortex of people with schizophrenia, a brain region also expressing estrogen receptors (Montague et al., 2008; Perlman et al., 2005). Thus, the neural substrate of

SERM action may include direct effects in both the hippocampus and IFG.

There are a number of limitations to our study. The limited sample size did not allow us to evaluate sex differences. It is possible that raloxifene affects neural activation related to emotion processing differentially in male and female patients. Additionally, while the crossover design has the strength of comparing the same person in different conditions, the use of a crossover design can also be a limitation due to the potential for any effects of treatment in the first period carrying over into the second placebo period for those who received raloxifene first. In that instance, we would be less likely to detect changes associated with raloxifene; however, the carryover effect did not appear to negatively influence our results given the effects of raloxifene on brain activity reported. Although 125

raloxifene has a relatively short half-life, raloxifene may have prolonged effects on brain cells; therefore, future studies should use a parallel group design in order to assess possible long lasting effects. Additionally, our sample was not large enough to separate out individuals taking different antipsychotic medications; a larger study would allow determination of whether raloxifene may have different effects in the presence of particular antipsychotics. Our sample consisted of patients displaying mild to moderate symptom severity on the positive and negative syndrome scale, and it is possible that raloxifene may have a more beneficial effect on symptoms in patients with more severe symptomatology. Lastly, we were unable to obtain some task behavioural data for several participants due to technical problems, which may have impacted our ability to detect whether raloxifene had a statistically significant effect on emotional face recognition accuracy. However, others (Bookheimer et al., 2000; Egan et al., 2001;

Karuza et al., 2014; Turk-Browne et al., 2009) have demonstrated that BOLD activity can be a more sensitive measure of change than measures of behaviour and can predict behavioural decline. These studies illustrate the ability of fMRI to reveal significant neurobiological effects between groups in the absence of behavioural differences, which may explain our findings of activation differences in ROI regions while performance measures did not vary.

Notwithstanding the limitations, our study provides the first evidence of significant effects of raloxifene treatment on neural activation during a key component of social processing. Due to the clinical importance of social impairment in schizophrenia and its relationship to poor functional outcomes, development of new treatments to improve this core deficit is one of the most pressing challenges in the field. Our findings of increased activity in the hippocampus and IFG during the recognition of emotional 126

faces with raloxifene suggest that adjunctive raloxifene treatment facilitates hippocampal and IFG activity in men and women with schizophrenia. These findings indicate that future research on the use of raloxifene to treat social impairment in schizophrenia is warranted. Given the emerging evidence regarding the efficacy of psychosocial treatments such as cognitive behavioural therapy and social skills training for improving social functioning in schizophrenia (Heinssen et al., 2000; Li et al., 2015), optimal outcomes may arise from a combination of pharmacologic therapy such as raloxifene with behavioural interventions.

In conclusion, our double-blind, randomized, placebo-controlled, crossover trial found that raloxifene treatment at 120 mg/day enhanced brain activation in key regions involved in facial emotion recognition deficits in schizophrenia and thus, there is potential for using adjunctive raloxifene treatment in order to treat social impairment in schizophrenia.

127

Supplementary Figure 6.1: Trial design Outcome measures for crossover analyses were assessed at week 6 and 13.

Supplementary Figure 6.2: Emotional face recognition paradigm Participants had to identify the emotion displayed. The present study only assessed brain activation for angry and neutral face recognition.

128

Supplementary Table 6.1: Mean reaction times and performance accuracy

Raloxifene Placebo Z p Angry RT 2624 (460) 2537 (489) 1.48 .14 Angry Accuracy 87% (10%) 90% (10%) .62 .53 Neutral RT 2502 (651) 2342 (511) .82 .41 Neutral Accuracy 84% (17%) 80% (17%) .90 .93 Abbrevations: RT = reaction time. RT displayed in milliseconds. Performance accuracy displayed as % correct. SD in parentheses.

129

7.0 General Discussion

The studies within this thesis support that hormones (estrogen, testosterone and DHEA) are associated with specific pathological features of schizophrenia by demonstrating novel relationships with structural and functional brain abnormalities and illustrating the potential for reversing such abnormalities. In Chapter 3, I compared basal levels of sex hormones from a cohort of patients and healthy controls, separated by gender. While some of my findings were in line with my hypotheses and in accordance with previous studies (e.g., increased DHEA in all patients), I also made somewhat unexpected discoveries including no diagnostic difference in estrogen levels in either men or women with schizophrenia compared to controls. Subsequent chapters in this thesis include several novel findings regarding the potential relationship of brain volume and increased neural activity to sex steroid levels. In Chapter 4, I showed that the molar ratio of cortisol to DHEA, which can be viewed as a peripheral marker of increased sensitivity to stressors, is associated with grey matter volume in the prefrontal cortex and hippocampus, brain areas critical for cognitive functioning and stress regulation.

Further, in Chapter 5, I showed that circulating testosterone levels in male patients are associated with increased neural activity during identification of the emotion of anger on faces. Finally, in Chapter 6, I found that 6 weeks of raloxifene administration in male and female patients enhances neural activity significantly more than a placebo during identification of angry faces in typically hypoactive brain regions, suggesting that even when hormone (estrogen) levels are within the normal range, estrogen receptors may still be targets that are modifiable with treatment.

130

7.1 Basal hormone levels in schizophrenia – implications and mechanisms

In the quest to discover better treatment and prevention strategies, we must first understand and untangle the underlying processes and mechanisms contributing to the onset of schizophrenia. Epidemiological data provided clues to look at sex hormones; the age of onset of schizophrenia typically occurs around adolescence to early adulthood with a further interesting phenomenon where a small subset of women experience onset around menopause. This second peak later in women is completely unmatched in men.

The event may be explained by hormonal changes because the later onset of schizophrenia in mature women is associated with the rapid decline in estrogen levels.

The neuroprotective nature of hormones in humans led myself and others to ask a logical question: If sex hormones protect against cognitive decline (Holland et al., 2011) which we see in schizophrenia (Weickert et al., 2000) and if sex hormones have other neuroprotective properties, do affected people with schizophrenia have abnormal levels of sex hormones? While previous studies have attempted to address this question, there have been mixed reports and generally testosterone levels have been measured in males while estrogen levels have been mainly examined in females with schizophrenia. In

Chapter 3, I showed that testosterone displayed abnormal levels diagnostically

(significantly decreased in male patients and increased in female patients), suggesting that sex is a modulatory factor of circulating testosterone in people with schizophrenia.

Furthermore, although I covaried for age in the comparative analyses, the wide age range of participants (18-55 years) does not allow us to understand if increased circulating testosterone in female patients exists in the early stages of the illness

(suggesting that increased testosterone is an important contributing factor of the illness) or if it increases with illness duration (suggesting that increased testosterone is a 131

consequence of the illness). In order to understand if testosterone is a risk factor for schizophrenia in females, it will be important for future studies to examine whether testosterone levels are changed in ultra-high risk or first-episode female patients.

Moreover, longitudinal studies measuring testosterone levels over time will help discern whether/how testosterone levels may be related to the course of the illness. In Chapter 3,

I also found that DHEA is significantly upregulated in both sexes of patients, suggesting that, unlike testosterone levels, this phenomenon is not modulated by sex. Moreover, my finding of increased DHEA in chronic patients has also been found in first-episode schizophrenia (Strous et al., 2004), suggesting that elevated DHEA may be both a risk factor and a possible consequence of the illness that is not resolved by treatment.

My findings of abnormal hormone levels led me to my next series of questions: What are the implications of abnormal hormone levels and can hormone levels within a normal range also have modulatory effects on brain function? My results along with those from other studies suggest that abnormal hormone (estrogen, testosterone, DHEA) levels exist in at least some people with schizophrenia and that discrepancies in reports are likely due to heterogeneous sampling of participants. Some studies report hypoestrogenism in female patients; however, our cohort consisted of mildly/moderately ill patients that may explain the lack of a diagnostic difference in females’ estrogen levels. There have been reports of low testosterone in male patients

(Akhondzadeh et al., 2006) which I was able to replicate in a larger sample, and, further, previous work in our study sample has demonstrated that lower testosterone levels in men with schizophrenia may be associated with poorer cognitive functioning (Moore et al., 2013). Other groups have shown that low testosterone is a risk factor for dementia in healthy elderly men (Carcaillon et al., 2014), therefore the negative relationship 132

between testosterone levels and cognition may not be unique to men with schizophrenia.

I also found increased testosterone levels in female patients relative to healthy females; however, the implications of this finding are uncertain as there was no association with symptom severity or cognitive functioning and there is no evidence from animal or human studies focusing on the female sex to suggest whether this may have positive or deleterious consequences. Prolactin-raising antipsychotics are generally associated with decreases in testosterone levels (Konarzewska et al., 2009), suggesting that increased testosterone in female patients is not a consequence of antipsychotic medication.

Contrary to this, amenorrhea, which is frequently reported in females with schizophrenia, can be induced by prolactin-raising antipsychotics and is associated with increased testosterone levels, suggesting that our finding of increased testosterone in female patients may be related to antipsychotic use. Moreover, one study demonstrated that risperidone treatment (a typical antipsychotic that raises prolactin) decreases testosterone in males and increases testosterone in females (Kinon et al., 2003), suggesting that prolactin-raising antipsychotics may affect major sex hormones differently in men and women. Our findings of lower testosterone levels in male patients and higher testosterone levels in female patients relative to controls are consistent with this possibility. Therefore, future studies examining antipsychotic-naïve patients are needed to understand whether increased testosterone in female patients is related to antipsychotic use. Still, in the case that it is unrelated to antipsychotics, it is also uncertain whether elevated testosterone levels in female patients at a relatively low concentration level compared with men has any relevance to the illness based on our negative findings of a relationship between testosterone levels and symptom severity and cognitive functioning.

133

It is important to note that significant associations have been reported between hormones (testosterone and estrogen) within a normal range with symptom severity and cognition in people with schizophrenia. Negative correlations between endogenous estrogen levels and positive symptoms have been reported in both sexes (Bergemann et al., 2007; Halari et al., 2004), while a positive correlation between endogenous estrogen levels and cognition has been found in female patients (Hoff et al., 2001; Ko et al.,

2006). This suggests that there may be an optimal range for circulating estrogen levels and that the range may very between individuals and sexes. More studies examining circulating estrogen in male patients are needed to understand whether it may be beneficial for their cognitive functioning. Men with schizophrenia showing testosterone levels within the normal range according to respective age displayed positive associations between endogenous testosterone levels and cognitive functioning (Moore et al., 2013). Importantly, in Chapter 5 I found that circulating testosterone levels relates to the ability of the human frontal cortex to increase cortical blood flow when required by task demands. Since this effect is correlative, it is possible that increased testosterone allows for greater neural activity or other tissue changes resulting in more perfusion or that a healthier prefrontal physiological response results in the ability for a person to maintain higher levels of blood testosterone. These findings along with previous studies demonstrate that increases in androgens and estrogens within the normal, healthy range may benefit cognitive abilities and be informative of the degree of abnormal activation underlying such abilities. However, these studies are limited in number and sample size, therefore a more comprehensive examination of testosterone on measures of symptomatology, cognition and neural activity in larger samples is required to better understand the possible role of testosterone in modulating cortical activity and function in schizophrenia. 134

7.2 Can the cortisol/DHEA ratio be used as a peripheral biomarker of some aspect of schizophrenia?

Reliable objective biomarkers, measurable indicators of biological state/condition, are desirable in order to supplement the current subjective assessments used to diagnose and better treat schizophrenia. Given the heterogeneous nature of the disease, no single biomarker has been identified that could be applied toward the entire population of patients. Traditionally, biomarkers included any measurable biochemical or molecular alterations found in the human body (Hulka & Wilcosky, 1988); however, more recently the definition has expanded to include any measures (including clinical endophenotypes such as brain imaging data) which can indicate normal or pathogenic processes (Weickert et al., 2013). Biomarkers would be particularly useful tools for risk assessment, intervention, treatment response prediction and diagnosis for schizophrenia.

Biomarkers in the blood have great potential as they are easily accessible and as evidenced by my findings in Chapter 3 and other studies showing elevated peripheral cytokines in a subset of people with schizophrenia (Fillman et al., 2015), at least some patients display a unique molecular signature.

There is abundant evidence for alterations in HPA regulatory mechanisms and disturbances in HPA activation in schizophrenia, manifesting in elevated peripheral cortisol levels in a subpopulation of individuals with schizophrenia (Girshkin et al.,

2014). This subsequently led to an interest in the involvement of DHEA as a functional antagonist of cortisol activity to potentially reverse stress-related pathology. Therefore, the HPA axis has been considered a source of change when in searching for peripheral 135

biomarkers. The cortisol/DHEA ratio is a balance of anabolic to catabolic hormones and has been examined and implicated in depression (Kahl et al., 2006), post-traumatic stress disorder (Van Voorhees et al., 2014) and other diseases of stress (o Hartaigh et al.,

2012) generally showing increased ratios. My finding of decreased cortisol/DHEA ratios in schizophrenia is novel and may represent a subgroup of patients as other groups have reported increased ratios (Ritsner et al., 2004) or no diagnostic difference in ratios (Gallagher et al., 2007; Garner et al., 2011) compared with healthy people.

Although it is counter to the hypothesis that the cortisol/DHEA ratio may be elevated in diseases of stress, my finding of increased cortisol/DHEA ratios in people with schizophrenia was robust and may be related to the blunted stress response observed in some people with schizophrenia, which is supported by previous findings of lower GR in the brain of people with psychosis (Sinclair et al., 2012; Sinclair et al., 2011; Sinclair et al., 2012). Varied reports of increased or no diagnostic elevation of cortisol suggest that hypercortisolemia only exists in a subgroup of patients. It will be important to identify what contributes to this phenomenon (genetic factors and/or early life stressors) and, subsequently, discern whether the cortisol/DHEA ratio is related to pathology in one or more of these subgroups. Similarly, reports of increased or no diagnostic difference in DHEA levels warrant further investigation before the ratio can be used effectively as a biomarker. To date, no study, due to limited sample sizing, has identified whether increased DHEA reflects exceedingly high DHEA levels from a particular subgroup of patients (perhaps who exhibit worse pathology) or if most patients on average have increased DHEA.

The direct effects of glucocorticoids are mediated by glucocorticoid receptors (GR). It has been shown that some people with schizophrenia have downregulated levels of GR 136

mRNA in frontal and temporal cortices and in the hippocampus as compared with healthy people, with variability among regions (Sinclair et al., 2012; Webster et al.,

2002). These findings suggest that GR dysregulation may contribute to differences in susceptibility to stress which may explain the clinical heterogeneity seen in schizophrenia. It is therefore possible that the cortisol/DHEA ratio may be a more predictive biomarker in a subgroup of patients who display altered GR mRNA expression. However, our sample of patients did not express abnormal cortisol levels as compared with healthy controls at the one time point sampled, suggesting that GR dysregulation was not significant or that it is not a strong predictor of circulating cortisol levels.

The fact that increased cortisol/DHEA ratios have been reported in multiple diseases of stress and that our finding is counter to what was expected suggests that the cortisol/DHEA ratio may not be a suitable diagnostic marker for schizophrenia. Here, I review the utility of the cortisol/DHEA as a peripheral biomarker of abnormal brain volume and behavioural phenotypes in schizophrenia.

7.2.1 Brain volume, cognition and symptoms

I reported, for the first time, that cortisol/DHEA ratios were inversely related to brain volume in the hippocampus and PFC in people with schizophrenia. Reduced hippocampal volume in schizophrenia and other disorders of stress (post-traumatic stress disorder (Gurvits et al., 1996), Cushing’s syndrome (Starkman et al., 1992) and depression (Sheline et al., 1999)) implicates glucocorticoid-induced brain atrophy.

Therefore, my finding of an inverse relationship of cortisol/DHEA to brain volume 137

suggests that the ratio may be a useful blood biomarker of increased sensitivity to brain insults and may also support findings that cortisol/DHEA ratios are inversely related to behavioural phenotypes in at least some people with schizophrenia (Garner et al., 2011).

However, contrary to what was expected, cortisol/DHEA ratios were not associated with cognitive functioning in schizophrenia. This may be explained by the presence of distinct cognitive subtypes in people with schizophrenia, where some patients show marked cognitive decline and others show fairly preserved intellect (Weickert et al.,

2000) and that these subtypes are associated with meaningfully distinct differences in grey matter volume (Geisler et al., 2015). It is therefore possible that cortisol/DHEA ratios may be associated with cognitive functioning in more homogenous subgroups based on brain morphometry or other clinical features. Furthermore, in contrast to findings from Garner et al. showing that negative symptoms improve as cortisol/DHEA decreases in schizophrenia (Garner et al., 2011), cortisol/DHEA ratios were not associated with symptom severity in our patient sample, which may be attributed to the relatively mild illness severity of patients in our study. Future studies using larger sample sizes and wider symptom range will enable researchers to categorise patients into more homozygous groups according to symptom severity in order to determine whether cortisol/DHEA ratios can be used as a biomarker of clinical phenotypes.

7.2.2 Possible mediating factors between cortisol/DHEA and brain volume

My finding of an inverse relationship between cortisol/DHEA ratios and brain volume was only found in people with schizophrenia while no such relationship was found in healthy people. While this suggests a role for DHEA and HPA axis dysfunction in contributing to the pronounced grey matter reductions in the disease, we must also 138

consider other pathological factors that may contribute to our finding. Brain-derived neurotrophic factor (BDNF) can directly influence the HPA-axis regulation through alterations of corticotropin-releasing hormone levels (Jeanneteau et al., 2012) and is critical for neuronal development, differentiation and maintenance and contributes to regulation of hippocampal volume (Mondelli et al., 2011; Stoop & Poo, 1996). There are several reports of reduced BDNF in the brain (most notably in the DLPFC) (Durany et al., 2001; Hashimoto et al., 2005; Weickert et al., 2003) and periphery of people with schizophrenia (Jindal et al., 2010); therefore, deficits in BDNF signalling may increase susceptibility to brain insults by interfering with the negative feedback inhibition of the

HPA axis. Treatment with DHEA increases BDNF expression in cortical neurons

(Rahmani et al., 2013) and the hippocampus (Sakr et al., 2014). Further, Miodownik et al. (Miodownik et al., 2011) reported evidence for the co-involvement of peripheral

BDNF levels and cortisol/DHEAS ratios in predicting the beneficial clinical response to

L-theanine treatment in people with schizophrenia. Taken together with our findings, these studies suggest that the combined effect of neurotrophin and HPA axis systems may contribute to brain alterations in schizophrenia through activation of common molecular pathways.

Inflammation is associated with structural brain alterations in schizophrenia (Ellman et al., 2010; Fillman et al., 2015; Meisenzahl et al., 2001). Our group has previously found that individuals with schizophrenia displaying elevated pro-inflammatory cytokine mRNA levels have reduced volume in Broca's area compared with patients with lower cytokines (Fillman et al., 2015). Moreover, IL-10 is an anti-inflammatory cytokine that induces synapse formation in the hippocampus (Lim et al., 2013) and is decreased in patients with schizophrenia (Xiu et al., 2014). Chronic stress is associated with immune 139

response (Frank et al., 2007) and treatment with DHEA increases serum levels of IL-10

(Cheng & Tseng, 2000). It is possible that the findings from this thesis may represent stress-induced immune activation leading to brain volume reductions. It is also possible that the collective effect of stress, BDNF and inflammation determine hippocampal and

DLPFC volume, through shared biological pathways. In support of this theory, findings from Mondelli et al. (Mondelli et al., 2011) suggest a synergic effect of BDNF, IL-6 and cortisol in determining hippocampal volume. Further investigation should involve a cluster analysis including BDNF levels, cytokines, cortisol/DHEA ratios and brain volumes in order to understand how the combined effects of growth factors, inflammation and hormones may determine volumetric changes.

7.3 Hormone therapy for schizophrenia: where we stand now and potential for continued use

The fact that circulating peripheral hormones can penetrate through the blood-brain barrier (Banks, 2012) and readily diffuse into neurons and glial to act on intracellular receptors and are the source of at least some hormones within CNS deems investigation of hormones in the blood extremely important, and supports the use of effective hormone therapies to treat brain disorders such as schizophrenia.

Androgen and estrogen supplementation have been used to treat various symptoms of schizophrenia, and researchers are still assessing their efficacy while understanding their mechanisms of action in the brain. The benefits from circulating sex hormones and adjunctive hormone therapy are likely mediated by their effects on neurotransmitters

(Bergeron et al., 1996; Compagnone & Mellon, 1998; Cyr et al., 2001; Di Paolo, 1994; 140

Herbison & Fenelon, 1995; Kranz et al., 2014; Kugaya et al., 2003; Moses et al., 2000;

Pasqualini et al., 1995; Sanchez et al., 2010). Overall, as compared with testosterone and DHEA, evidence demonstrates that estrogen has the greatest positive modulatory effect and on a more diverse number of neurotransmitters, which may explain why estrogen-like drug compounds as compared with other hormone-based compounds are leading in potential for clinical utility to treat schizophrenia. The findings from this thesis fuel an important discussion regarding whether all patients or only a certain subgroup of patients can benefit from hormone treatment in addition to important considerations that should be reviewed for future applications. Findings from our group

(Weickert et al., 2015) where only 40% of patients were responders to raloxifene

(showing clinically reliable change on one or more cognitive measures from baseline) suggest that a subgroup of male and female patients benefit from estrogen-based hormone treatment. It will be important for future clinical trials with adequate sample size to perform similar responder analyses and subsequent analyses (e.g. genotyping) in order to help understand the factors that may contribute to hormone treatment response.

7.3.1 Is there still therapeutic potential for DHEA augmentation in schizophrenia?

The neuroprotective effects of DHEA in the brain and rapidly diminishing DHEA levels with age led to its clinical use in healthy populations. DHEA treatment has been found to improve cognitive functioning in healthy older adults (Hirshman et al., 2003; Stangl et al., 2011; Yamada et al., 2010) but there have also been trials showing no significant improvement (Kritz-Silverstein et al., 2008; Merritt et al., 2012; van Niekerk et al.,

2001). Despite these mixed findings, adjunctive DHEA has also been trialled repeatedly in people with schizophrenia with the idea that its modulatory actions on 141

neurotransmitter systems and potential for restoring aberrant HPA axis signalling may ameliorate symptom severity and cognitive deficits. Somewhat unsurprisingly, these studies have produced mixed results regarding efficacy. While there are some positive reports of reduced symptoms and improved cognitive functioning with adjunctive

DHEA (Ritsner et al., 2006; Strous et al., 2003), studies also show non-significant effects compared to placebo (Strous et al., 2007) and to date, relatively small sample sizes have been trialled. Mixed reports from these clinical trials may suggest that DHEA was administered to patients prior to defining appropriate subgroups according to symptom severity, cognitive functioning or baseline DHEA levels. Whereas testosterone augmentation for schizophrenia was used based on clinical observations of hypogonadism in some male patients (Akhondzadeh et al., 2006; Taherianfard &

Shariaty, 2004) and, similarly, estrogen-like compounds were used adjunctively based on hypoestrogenism in postmenopausal women (Hoff et al., 2001), the majority of studies suggest DHEA is increased or not different in people with schizophrenia compared with healthy people. If people with schizophrenia produce more DHEA as a compensatory mechanism and it is currently uncertain to what extent it is helping patients normalise and recover, it is unclear why researchers believe DHEA administration will add any additional benefit. Moreover, there is evidence that high concentrations of DHEA can be neurotoxic in mice and in vitro (Kimonides et al., 1999;

Safiulina et al., 2006). Considering my finding of increased DHEA in patients along with other similar reports, the efficacy DHEA augmentation may depend on a patient’s basal DHEA and cortisol levels. Perhaps adjunctive DHEA should be confined to a subgroup of patients with low DHEA and high cortisol who are unable to compensate for increased vulnerability to stressors, and that the dose should be carefully optimised to avoid becoming exceedingly high. Future studies should consider administering the 142

dexamethasone suppression test in order to test HPA axis function. People with schizophrenia who are non-suppressors (associated with poor outcome) (Tandon et al.,

1991) and/or who also have low DHEA levels may respond better to DHEA administration. Identification of biomarkers, such as glucocorticoid receptor genetic polymorphisms, may also be helpful in predicting who may benefit from adjunctive

DHEA. While the therapeutic benefit of DHEA as an adjunctive treatment for schizophrenia is not yet well established, our finding of increased DHEA in patients warrants further investigation as to whether people with schizophrenia (as a whole or certain subgroups) may benefit from add-on treatment when their DHEA levels already exceed those of healthy controls on average.

7.3.2 Implications of circulating testosterone levels when considering the development of new treatments

To date, there has been considerably less focus on testosterone’s role in schizophrenia as compared with estrogen’s role, therefore it may be unsurprising that testosterone augmentation has only been evaluated in a single pilot investigation. Ko et al. (Ko et al.,

2008) conducted a placebo-controlled, double-blind trial where 30 men with schizophrenia were randomised to receive either adjunctive testosterone or a placebo for

4 weeks. While the authors found that testosterone-treated patients showed significantly greater improvement in negative symptoms compared with placebo-treated patients, the study did not include a healthy comparison group at baseline to determine whether the patients displayed hypogonadism or lower than optimal testosterone levels prior to hormone treatment. Thus, the unanswered question was: did testosterone exert its beneficial effects by restoring low circulating testosterone levels to normal levels or 143

does an increase in testosterone have positive effects regardless of a patient’s basal levels? Testosterone-induced psychotic episodes in steroid users (Ruiz Feliu & Campos

Mangas, 2012) suggests that testosterone levels exceeding a normal range may have detrimental effects. Moreover, testosterone can enhance release of glutamate and dopamine which may contribute to heightened symptomatology (Gogos et al., 2012;

Purves-Tyson et al., 2012). These findings suggest that the beneficial effects of testosterone are dose dependent such that exceedingly low and high testosterone levels may be deleterious while there exists an optimal mid-range of testosterone levels, representing an inverted U-shaped curve. Therefore, while selective estrogen receptor modulators (SERMs) have been safely and effectively used in both men and women with schizophrenia, adjunctive treatment with testosterone-like compounds in male patients without hypogonadism and in female patients should be approached with caution because the implications of elevated testosterone are still unclear.

7.3.3 The efficacy of raloxifene for treating schizophrenia

In contrast to the limited evidence that adjunctive testosterone may benefit patients, estrogen and estrogen-like compounds (SERMs such as raloxifene) have been more widely tested as adjunctive treatment for schizophrenia (Akhondzadeh et al., 2003;

Kulkarni et al., 1996; Kulkarni et al., 2015; Kulkarni et al., 2010; Weickert et al., 2015).

It is noteworthy that raloxifene appears to be an effective treatment in patients who do not display hypoestrogenism (as evidenced by our study sample), therefore the beneficial effects of add-on treatment are not acting by merely restoring hormone levels/action.

144

Taken together with clinical findings and one other published study of raloxifene effects on neural activity of probabilistic association learning (Kindler et al., 2015), my findings from Chapter 6 support a role for estrogen receptor modulation in treating individuals with schizophrenia. While the direction of abnormal neural activity

(hypoactive versus hyperactive) seen in schizophrenia is region and task-dependent, my findings of increased BOLD activity by raloxifene were specific to typical hypoactive regions during emotional face identification, demonstrating raloxifene’s ability to restore BOLD signalling. Importantly, I have demonstrated that raloxifene facilitates neural activity in both men and women with schizophrenia in brain regions (frontal cortex and hippocampus) where estrogen receptors are localised in both sexes. However, previous work by our group has demonstrated gender specific effects in verbal fluency where raloxifene treatment as compared with a placebo significantly improved verbal fluency in female patients but not male patients (Weickert et al., 2015). Verbal fluency relies on frontal cortical regions that overlap with the IFG, therefore we may expect that if raloxifene can increase activity in the IFG in men it would also improve cognitive performance. Together, these findings are not easily interpreted and suggest that there are interactions between sex and raloxifene treatment on certain domains of cognition and emotion processing. People with schizophrenia have altered estrogen receptors

(lower mRNA levels and failure to express the fully functional wild-type) (Perlman et al., 2005; Weickert et al., 2008), suggesting that raloxifene’s ability to restore neural activity in schizophrenia reflects the benefit of stimulating the estrogen signalling pathway. This is supported by findings from animal studies that report increased ER mRNA following raloxifene administration (Zhou et al., 2002). However, the authors from the above study also demonstrate that the agonist/antagonist profile for SERMs may differ among brain areas, which may explain why raloxifene improves some 145

cognitive domains in people with schizophrenia while demonstrating no significant effects on other cognitive domains (Weickert et al., 2015).

Among the hormone therapies discussed, evidence from previous studies demonstrates that estrogen-like compounds appear to yield the most promising clinical benefit.

Further, its tolerability and lack of cancerous adverse effects adds to its strengths. This thesis provides the first evidence of a modulatory effect of raloxifene on emotion processing, which is related to social functioning outcomes, and may further support the use of raloxifene as adjunctive treatment in men and women with schizophrenia.

7.4 Methodological considerations

In considering the relationship of hormones to clinical symptoms and how hormones of peripheral origin can reflect the amount found in the brain, I combined biological measurements in the blood with s/fMRI to observe how hormones are involved functionally and structurally in vivo, an idea that has only recently been explored in schizophrenia. The use of multiple methods (neuroimaging, endocrine, clinical and behavioural assessment) to understand the relationship of hormones to brain abnormalities in schizophrenia has provided much insight, but there are undoubtedly strengths and weaknesses associated with each method.

7.4.1 Quantifying hormone levels

A potential weakness of assessing serum levels of hormones is that they are peripheral measures and not direct measures of hormone levels in the brain. However, due to the 146

correlative relationship between hormone levels in the CNS and periphery, there is a strong case for measuring peripheral hormone levels. Moreover, methods of quantifying circulating hormone levels have become more reliable and sensitive over time.

Chemiluminescent immunometric assays were used to measure hormones (testosterone, estrogen and cortisol), which is a widely used and validated method for measuring analytes in the blood. Further, GC-MS analysis was used to measure DHEA and quantifies minute amounts of steroidal compounds with high accuracy. The studies in this thesis quantified hormone levels in participants from blood serum, however there are other methods available for the assessment of basal hormone levels in humans.

Saliva sampling may have certain advantages over blood sampling, particularly for participants who may exhibit sensitivity/fear toward venepuncture (Kirschbaum &

Hellhammer, 1989, 1994; Lac, 2001). A proportion of patients have a blunted cortisol response following acute stress (Brenner et al., 2009; Jansen et al., 1998); therefore, such sensitivity to venepuncture may introduce a confounding factor. These patients may thus have higher circulating cortisol that is not reflected when they have blood drawn at the clinic because the environment is perceived to be more stressful for them than the average control participant. This may explain variable reports of circulating cortisol levels in studies of people with schizophrenia and should be taken into consideration. Moreover, while the majority of schizophrenia studies extract information from blood measurements at a single time point, this does not allow us to observe any fluctuations that might occur within the day or over longer periods of time.

Relatively small saliva samples are required to obtain measurements of hormones; therefore sampling across multiple time points may be performed quite easily in order to obtain a more accurate reflection of steroid levels over a time course which may help address the problem of a single steroid measurement. Both cortisol and DHEA 147

measurements from saliva correspond to levels found in the blood as steroids in the blood enter the oral cavity via passive diffusion, which is independent of saliva flow- rate and other transport mechanisms (Gallagher et al., 2006; Granger et al., 1999;

Shirtcliff et al., 2001).

Another important issue to consider is the measurement of free versus total hormones.

Sex hormone-binding globulin (SHBG) is a glycoprotein that binds to circulating androgen and estrogen. Testosterone has the highest affinity for SHBG, followed by estrogen and DHEA (Dunn et al., 1981; Rosner, 1991). Hormones only in their unbound free form are biologically active and able to bind to their receptors. Therefore, the amount of SHBG influences sex steroid bioavailability by inhibiting their function. In this thesis, I only reported total levels of hormones while their interactions with SHBG were not determined. However, there is evidence that free testosterone and total testosterone are highly correlated (Akhondzadeh et al., 2006; Ko et al., 2007); therefore, the total hormone levels measured should reflect the bioavailable and active hormones.

However, we must also consider previous reports that antipsychotics can decrease

SHBG levels (Birkenaes et al., 2009; Smith et al., 2002), which may lead to increased unbound hormones. This suggests that unbound testosterone may be even lower in drug- naïve patients, and may support the use for testosterone administration in this sample of people. A decrease in SHBG levels by antipsychotics may also contribute to my finding of increased testosterone in female patients. It will be important for future studies to compare levels of SHBG and sex hormones in drug-naïve patients and chronic patients who have been taking antipsychotics.

148

7.4.2 Strengths and weaknesses of MRI fMRI is a powerful technology that helps researchers answer questions within psychology and cognitive neuroscience: How or what brain regions are involved when people process emotions in various context? What is the impact of psychiatric therapies on neural activity? While fMRI is often useful in complementing clinical behavioural data there are also limitations to consider and there has been an unsurprising amount of debate among scientists for the use of fMRI. The different views may reflect the critical need to thoroughly understand the actual capacities and limitations of fMRI technology.

The principal advantages of fMRI include its non-invasive nature and high spatial resolution. However, fMRI suffers from poor temporal resolution; thus, it does not allow inferences about the exact timing in which activations occur (Logothetis, 2008).

In order to address this issue in the future, multimodal data fusion of imaging techniques that have complementary strengths and weaknesses may be a preferred and more effective strategy. For example, simultaneous recording and analysis of fMRI and electroencephalography (EEG) (which has high temporal resolution but lacks spatial resolution) (Michel et al., 2004) has gained increasing attention, allowing the achievement of both high temporal and spatial resolution of neural activity.

Despite the ability of multimodal imaging to improve our understanding of brain function and activity, the majority of limitations that arise from fMRI are the result of a complex circuitry and functional organization of the human brain. There are also a number of limitations of the methodology and potential flaws in modelling the principles of fMRI, which are discussed briefly here:

149

- The BOLD signal measured is susceptible to artefacts including head movement

of the participant during scan acquisition. Head movement is an important

practical challenge to overcome and may produce signal changes that are falsely

interpreted as increases in BOLD activity. In order to address this, the studies in

this thesis excluded participant scans that showed movement greater than 3mm

in any direction. Further, the standard pre-processing steps performed in SPM

are also designed to help filter and reduce movement artefacts.

- Repetition of stimuli eventually results in reduced BOLD signal (Müller et al.,

2012) (habituation), with some regions adapting more quickly than others. In

this case, the correlative nature between BOLD response and neuronal firing

decreases and the BOLD activity is no longer an accurate measure of neural

activity. This could potentially be problematic for studies in Chapters 5 and 6

where I measured BOLD activity during identification of emotional faces,

because the emotions displayed were repeated throughout the paradigm. It is

possible that this habituation also varies between individuals and groups, which

is not often or easily addressed. While EEG-fMRI coupling may address this

issue, it is also important to assess the strengths (improved accuracy of

determining true neuronal activity) and weaknesses (MRI scanner artefacts in

EEG signal and vice versa; time consuming) before utilising it.

- Certain drugs can stimulate the autonomic nervous system controlling vascular

dilation, which can affect blood flow and volume at baseline resulting in an

altered BOLD response. Caffeine (an adenosine antagonist that constricts

cerebral vessels) can reduce baseline blood flow and increase deoxyhemoglobin 150

levels, which results in an increase BOLD response (Mulderink et al., 2002). All

participants in the studies performed in this thesis were asked to fast for 9 hours

prior to testing, limiting the possible effects of caffeine on the BOLD signal.

However, smoking was not an exclusion criteria and nicotine has been shown to

increase the BOLD response in a region-specific and task-dependent manner

(Bruijnzeel et al., 2015), possibly introducing a confound.

- The great anatomical variability between the brains of people with schizophrenia

and healthy people may introduce a confound in the BOLD signal. While spatial

normalisation during the pre-processing process attempts to account for

differences in brain size among participants by aligning each participant's

anatomy to a standard template, the more a participant's brain deviates from the

template, the more likely the warping process will results in errors, which can

lead to inaccurate magnitude estimation and mislocalization (Crinion et al.,

2007). In order to address this to the best of our ability, raw scans were screened

for structural artefacts and cases removed if too aberrant.

7.4.3 Clinical difficulties: recruitment, confounding variables, heterogeneity

Understanding the mechanisms behind the symptoms and deficits of schizophrenia is an extraordinarily difficult task to achieve due to the heterogeneous nature of the disease and variability in past/ongoing treatments. Recruitment of people with schizophrenia is a lengthy, expensive and challenging process due to the time commitment required and the extensive exclusion criteria used for most rigorous studies. Moreover, it is difficult and often impossible to entirely account for the effect that chronic use of antipsychotics 151

can have on the variables measured in the study. There is substantial evidence that antipsychotics can induce structural remodelling (Ho et al., 2011). In fact, Tost et al.

(Tost et al., 2010) found a rapid decrease in striatal grey matter volume within 2 hours of haloperidol treatment in healthy people. For each of the studies in this thesis, participants with schizophrenia were required to be receiving stable doses of antipsychotics for at least one year prior to entry of the study in order to help ensure that their symptoms were relatively stable. However, there is considerable variability in that participants received different types of antipsychotics and at different dosages. While we calculated chlorpromazine equivalent dose in order to determine whether there are acute effects of medication on the brain, this measurement is limited to the extent that the conversion factor is valid.

Other variable aspects of the presented studies (and the majority of studies in schizophrenia) are the precise diagnoses and clinical presentation of patients. Although the patients who participated in these studies were diagnosed with schizophrenia

(approximately 1/3 of the sample) or schizoaffective disorder (approximately 2/3 of the sample) by their psychiatrist with a confirmed diagnosis by our study clinician using the

SCID, these diagnoses lump heterogeneous sub-phenotypes

(schizoaffective/schizophrenia, depressed/bipolar schizoaffective, paranoid/disorganised/residual/undifferentiated schizophrenia) together. Moreover, there is evidence that schizophrenia patients may be more impaired in their ability to discriminate facial emotions as compared with people with schizoaffective disorder

(Chen et al., 2012), which should be considered for treatments (pharmacological and behavioural) targeting social emotional behaviour in patients. Furthermore, the presentation of clinical symptoms varies according to severity and type of symptom 152

present such that some patients experience visual and/or auditory hallucinations while other patients may never experience hallucinations. Due to the limited sample pool of our studies (and the majority of studies), categorisation into more homogeneous subtypes is not always a viable option. Moreover, the complex gene-environment interactions contributing to the onset of psychosis differs within the diagnosis, such that different at-risk genes and environmental stressors may converge for different people.

This makes understanding the causal factors of schizophrenia an exceptionally difficult task and should be carefully considered when discerning whether circulating hormones play an influential role in the onset/course of schizophrenia for all people with schizophrenia or only a subpopulation of patients.

7.5 Future directions

In completing this series of studies, I have contemplated the following: how influential is one hormone or a group of related hormones in the larger context of the disease?

While there is no distinct, definite answer to this, I have shown that an estrogen-based compound targeting both ESR1 and ESR2 receptors, directly or indirectly, can result in changes in the brain. However, in order to assess the relative impact and mechanisms of action of raloxifene, we must engage in further scrutiny and there is enormous potential to do so.

It will be important and useful for future studies to determine the mechanisms by which raloxifene exerts its effects upon binding to the estrogen receptor. Since raloxifene may act to decrease chemokines and suppress cytokines in the brain, it is possible that raloxifene dampens chronic neuroinflammation (Li et al., 2014). In order to test this 153

hypothesis, future clinical trials using raloxifene administration should examine circulating cytokines in the blood before and after treatment. One study showed that raloxifene decreases the inflammatory response caused by lipopolysaccharide in rodent microglia cells (Arevalo et al., 2012). The same study found that treatment with raloxifene reduced the number of reactive astrocytes in the hippocampus following injury. Together, these findings indicate that raloxifene counteracts brain inflammation by reducing reactive gliosis and suggest that raloxifene may help treat schizophrenia by reducing inflammation in the subpopulation of patients who exhibit an upregulated immune response (Fillman et al., 2015). Inflammation is also thought to underlie white matter dysconnectivity found in schizophrenia (Prasad et al., 2015). While I and others

(Kindler et al., 2015) have demonstrated that raloxifene can increase/restore neural activity, it is well known that estrogen also has modulatory effects on glia and more specifically white matter. Future clinical trials of raloxifene should use diffusion tensor imaging to measure fractional anisotropy (associated with increased myelination, fiber coherence and/or number of axons) (Dong et al., 2004; Neil et al., 1998; Song et al.,

2003; Takahashi et al., 2002; Takahashi et al., 2000) before and following treatment. If fractional anisotropy increases with raloxifene administration and if the amount of increase is correlated with improvements in cognitive functioning and/or symptom severity, this may suggest that raloxifene has consequences of repairing inflammation- related myelin dysfunction, which contributes to symptoms of schizophrenia (d’Albis &

Houenou, 2015).

154

7.6 Conclusions

I performed a variety of tests and detailed analyses using a combination of endocrinological, clinical and neuroimaging markers in order to investigate the relationship between circulating hormones and brain pathology in schizophrenia. My work provides supporting evidence for the role of various hormones in the manifestation of abnormal neural activity and morphometry in people with schizophrenia. The major conclusions of the studies arising from this thesis are:

1. Because of their ability to cross the blood brain barrier, abnormal hormone

levels of peripheral origin may be particularly relevant to brain pathology

and are likely related to alterations in major neurotransmitter systems

involved in schizophrenia. Further studies are required to determine how

manipulation of sex may modulate the effects on brain volume and activity.

2. Variable reports on the presence/lack of diagnostic differences in hormone

levels and/or presence/lack of correlational relationships between hormones

and symptom severity/cognition suggest that other factors (such as genetic

polymorphisms and environmental stressors) may further interact with

hormones in their effects on clinical presentation. Moreover, the associations

between hormones and behavioural phenotypes may be subtle but still

important.

3. At least some people with schizophrenia have upregulated DHEA, which

may reflect a compensatory and/or regulatory mechanism; however, more

studies are required to determine the effective nature of this response. 155

4. Cortisol/DHEA ratios in people with schizophrenia may be a biological

marker of stress-induced reductions in cortical and subcortical grey matter

volume; however, this relationship may be indirect and additional studies are

required to confirm this finding and to determine whether other regulatory

factors (such as BDNF and inflammation) are also involved.

5. In men with schizophrenia but not healthy men, there is a link between

circulating testosterone levels and neural activity underlying the

identification of angry faces, suggesting that testosterone may influence

emotion processing.

6. Along with its demonstrated benefits of reducing symptomatology and

enhancing cognitive functioning, adjunctive raloxifene increases neural

activity in important cortical and limbic regions during emotion processing

in schizophrenia. Further work is needed to determine whether it has

potential to improve social impairment in schizophrenia.

The current scarcity of effective treatment options for the wide range of symptoms in schizophrenia reinforces the gravity with which improved options must be thoroughly considered, tested and established. Collectively, the studies in this thesis have further expanded the current knowledge of the role of hormones in terms of their association with brain abnormalities and potential therapeutic use in people with schizophrenia.

Therefore, it is hoped that the findings from this thesis corroborate current efforts and demonstrate how future work can be developed regarding effective hormone-based interventions, which may ultimately help prevent and treat at least some impairments in individuals with schizophrenia. 156

8.0 References

Abel, K. M., Drake, R., & Goldstein, J. M. (2010). Sex differences in schizophrenia. Int Rev Psychiatry, 22(5), 417-428.

Adams, M. M., Fink, S. E., Janssen, W. G., Shah, R. A., & Morrison, J. H. (2004). Estrogen modulates synaptic N-methyl-D-aspartate receptor subunit distribution in the aged hippocampus. J Comp Neurol, 474(3), 419-426.

Adolphs, R., Tranel, D., Damasio, H., & Damasio, A. (1994). Impaired recognition of emotion in facial expressions following bilateral damage to the human amygdala. Nature, 372(6507), 669-672.

Agid, O., Shapira, B., Zislin, J., Ritsner, M., Hanin, B., Murad, H., . . . Lerer, B. (1999). Environment and vulnerability to major psychiatric illness: a case control study of early parental loss in major depression, bipolar disorder and schizophrenia. Mol Psychiatry, 4(2), 163-172.

Akhondzadeh, S., Nejatisafa, A. A., Amini, H., Mohammadi, M. R., Larijani, B., Kashani, L., . . . Kamalipour, A. (2003). Adjunctive estrogen treatment in women with chronic schizophrenia: a double-blind, randomized, and placebo- controlled trial. Prog Neuropsychopharmacol Biol Psychiatry, 27(6), 1007-1012.

Akhondzadeh, S., Rezaei, F., Larijani, B., Nejatisafa, A.-A., Kashani, L., & Abbasi, S. H. (2006). Correlation between testosterone, gonadotropins and prolactin and severity of negative symptoms in male patients with chronic schizophrenia. Schizophr Res, 84(2–3), 405-410.

Aleman, A., Kahn, R. S., & Selten, J. (2003). Sex differences in the risk of schizophrenia: Evidence from meta-analysis. Arch Gen Psychiatry, 60(6), 565- 571.

American Psychiatric Association. (1980). Diagnostic and Statistical Manual of Mental Disorders (3rd ed.). Washington, DC.

American Psychiatric Association. (2000). Diagnostic and Statistical Manual of Mental Disorders (4th ed.). Washington, DC.

American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders (5th ed.). Washington, DC.

Andreasen, N. C., Liu, D., Ziebell, S., Vora, A., & Ho, B.-C. (2013). Relapse Duration, Treatment Intensity, and Brain Tissue Loss in Schizophrenia: A Prospective Longitudinal MRI Study. Am J Psychiatry, 170(6), 10.1176/appi.ajp.2013.12050674.

Angermeyer, M. C., Kühn, L., & Goldstein, J. M. (1990). Gender and the course of schizophrenia: Differences in treated outcomes. Schizophrenia Bulletin, 16(2), 293-307. 157

Arabadzisz, D., Diaz-Heijtz, R., Knuesel, I., Weber, E., Pilloud, S., Dettling, A. C., . . . Pryce, C. R. (2010). Primate early life stress leads to long-term mild hippocampal decreases in corticosteroid receptor expression. Biol Psychiatry, 67(11), 1106-1109.

Arevalo, M. A., Diz-Chaves, Y., Santos-Galindo, M., Bellini, M. J., & Garcia-Segura, L. M. (2012). Selective oestrogen receptor modulators decrease the inflammatory response of glial cells. J Neuroendocrinol, 24(1), 183-190.

Armagan, G., Kanit, L., Terek, C. M., Sozmen, E. Y., & Yalcin, A. (2009). The levels of glutathione and nitrite-nitrate and the expression of Bcl-2 mRNA in ovariectomized rats treated by raloxifene against kainic acid. Int J Neurosci, 119(2), 227-239.

Arseneault, L., Cannon, M., Fisher, H. L., Polanczyk, G., Moffitt, T. E., & Caspi, A. (2011). Childhood trauma and children's emerging psychotic symptoms: A genetically sensitive longitudinal cohort study. Am J Psychiatry, 168(1), 65-72.

Asgharpour, M., Tehrani-Doost, M., Ahmadi, M., & Moshki, H. (2015). Visual attention to emotional face in schizophrenia: an eye tracking study. Iran J Psychiatry, 10(1), 13-18.

Ashburner, J. (2007). A fast diffeomorphic image registration algorithm. Neuroimage, 38(1), 95-113.

Ashburner, J., & Friston, K. J. (2000). Voxel-based morphometry--the methods. Neuroimage, 11(6 Pt 1), 805-821.

Asthana, S., Baker, L. D., Craft, S., Stanczyk, F. Z., Veith, R. C., Raskind, M. A., & Plymate, S. R. (2001). High-dose estradiol improves cognition for women with AD: results of a randomized study. Neurology, 57(4), 605-612.

Asthana, S., Craft, S., Baker, L. D., Raskind, M. A., Birnbaum, R. S., Lofgreen, C. P., . . . Plymate, S. R. (1999). Cognitive and neuroendocrine response to transdermal estrogen in postmenopausal women with Alzheimer's disease: results of a placebo-controlled, double-blind, pilot study. Psychoneuroendocrinology, 24(6), 657-677.

Badgaiyan, R. D., Fischman, A. J., & Alpert, N. M. (2009). Dopamine Release During Human Emotional Processing. Neuroimage, 47(4), 2041-2045.

Baig, B. J., Whalley, H. C., Hall, J., McIntosh, A. M., Job, D. E., Cunningham-Owens, D. G., . . . Lawrie, S. M. (2010). Functional magnetic resonance imaging of BDNF val66met polymorphism in unmedicated subjects at high genetic risk of schizophrenia performing a verbal memory task. Psychiatry Res, 183(3), 195- 201.

Banks, W. A. (2012). Brain Meets Body: The Blood-Brain Barrier as an Endocrine Interface. Endocrinology, 153(9), 4111-4119.

Barrett-Connor, E., & Stuenkel, C. A. (2001). Hormone replacement therapy (HRT)— risks and benefits. International Journal of Epidemiology, 30(3), 423-426. 158

Bastianetto, S., Ramassamy, C., Poirier, J., & Quirion, R. (1999). Dehydroepiandrosterone (DHEA) protects hippocampal cells from oxidative stress-induced damage. Brain Res Mol Brain Res, 66(1-2), 35-41.

Baulieu, E. E. (1997). Neurosteroids: of the nervous system, by the nervous system, for the nervous system. Recent Prog Horm Res, 52, 1-32.

Baulieu, E. E. (1998). Neurosteroids: a novel function of the brain. Psychoneuroendocrinology, 23(8), 963-987.

Baulieu, E. E., & Robel, P. (1998). Dehydroepiandrosterone (DHEA) and dehydroepiandrosterone sulfate (DHEAS) as neuroactive neurosteroids. Proc Natl Acad Sci U S A, 95(8), 4089-4091.

Begemann, M. J., Dekker, C. F., van Lunenburg, M., & Sommer, I. E. (2012). Estrogen augmentation in schizophrenia: a quantitative review of current evidence. Schizophr Res, 141(2-3), 179-184.

Bellace, M., Williams, J. M., Mohamed, F. B., & Faro, S. H. (2013). An fMRI study of the activation of the hippocampus by emotional memory. Int J Neurosci, 123(2), 121-127.

Bergemann, N., Mundt, C., Parzer, P., Jannakos, I., Nagl, I., Salbach, B., . . . Resch, F. (2005). Plasma concentrations of estradiol in women suffering from schizophrenia treated with conventional versus atypical antipsychotics. Schizophr Res, 73(2-3), 357-366.

Bergemann, N., Mundt, C., Parzer, P., Pakrasi, M., Eckstein-Mannsperger, U., Haisch, S., . . . Resch, F. (2005). Estrogen as an adjuvant therapy to antipsychotics does not prevent relapse in women suffering from schizophrenia: results of a placebo- controlled double-blind study. Schizophr Res, 74(2-3), 125-134.

Bergemann, N., Parzer, P., Runnebaum, B., Resch, F., & Mundt, C. (2007). Estrogen, menstrual cycle phases, and psychopathology in women suffering from schizophrenia. Psychol Med, 37(10), 1427-1436.

Bergeron, R., de Montigny, C., & Debonnel, G. (1996). Potentiation of neuronal NMDA response induced by dehydroepiandrosterone and its suppression by progesterone: effects mediated via sigma receptors. J Neurosci, 16(3), 1193- 1202.

Berman, K. F., Schmidt, P. J., Rubinow, D. R., Danaceau, M. A., Van Horn, J. D., Esposito, G., . . . Weinberger, D. R. (1997). Modulation of cognition-specific cortical activity by gonadal steroids: A positron-emission tomography study in women. Proceedings of the National Academy of Sciences, 94(16), 8836-8841.

Berrios, G. E., & Hauser, R. (1988). The early development of Kraepelin's ideas on classification: a conceptual history. Psychol Med, 18(4), 813-821. 159

Bethea, C. L., Mirkes, S. J., Su, A., & Michelson, D. Effects of oral estrogen, raloxifene and arzoxifene on gene expression in serotonin neurons of macaques. Psychoneuroendocrinology, 27(4), 431-445.

Beyazyüz, M., Albayrak, Y., Beyazyüz, E., Ünsal, C., & Göka, E. (2014). Increased serum dehydroepiandrosterone sulfate in the first episode but not in subsequent episodes in male patients with schizophrenia. Neuropsychiatric Disease and Treatment, 10, 687-693.

Beyer, C., Green, S. J., & Hutchison, J. B. (1994). Androgens influence sexual differentiation of embryonic mouse hypothalamic aromatase neurons in vitro. Endocrinology, 135(3), 1220-1226.

Beyer, C., & Hutchison, J. B. (1997). Androgens stimulate the morphological maturation of embryonic hypothalamic aromatase-immunoreactive neurons in the mouse. Brain Res Dev Brain Res, 98(1), 74-81.

Biegon, A., & McEwen, B. S. (1982). Modulation by estradiol of serotonin receptors in brain. J Neurosci, 2(2), 199-205.

Birkenaes, A. B., Birkeland, K. I., Friis, S., Opjordsmoen, S., & Andreassen, O. A. (2009). Hormonal markers of metabolic dysregulation in patients with severe mental disorders after olanzapine treatment under real-life conditions. J Clin Psychopharmacol, 29(2), 109-116.

Blakemore, S.-J., Burnett, S., & Dahl, R. E. (2010). The Role of Puberty in the Developing Adolescent Brain. Human Brain Mapping, 31(6), 926-933.

Bookheimer, S. Y., Strojwas, M. H., Cohen, M. S., Saunders, A. M., Pericak-Vance, M. A., Mazziotta, J. C., & Small, G. W. (2000). Patterns of Brain Activation in People at Risk for Alzheimer's Disease. New England Journal of Medicine, 343(7), 450-456.

Bradley, A. J., & Dinan, T. G. (2010). A systematic review of hypothalamic-pituitary- adrenal axis function in schizophrenia: implications for mortality. J Psychopharmacol, 24(4 Suppl), 91-118.

Brenner, K., Liu, A., Laplante, D. P., Lupien, S., Pruessner, J. C., Ciampi, A., . . . King, S. (2009). Cortisol response to a psychosocial stressor in schizophrenia: blunted, delayed, or normal? Psychoneuroendocrinology, 34(6), 859-868.

Brett, M., Anton, J., Valabregue, R., & Poline, J. (2002). Region of interest analysis using an SPM toolbox. Neuroimage.

Britton, J. C., Shin, L. M., Barrett, L. F., Rauch, S. L., & Wright, C. I. (2008). Amygdala and fusiform gyrus temporal dynamics: responses to negative facial expressions. BMC Neurosci, 9, 44.

Broca, P. (1861). {Remarque sur le siege de la facult\\'{e} du language articul\\'{e}, suivie d\'une observation d\'aph\\'{e} mie (perte de la parole)}. {Bulletin de la soci\\'{e} t\\'{e} anatomique de Paris}, 36, 330-356. 160

Broks, P., Young, A. W., Maratos, E. J., Coffey, P. J., Calder, A. J., Isaac, C. L., . . . Hadley, D. (1998). Face processing impairments after encephalitis: amygdala damage and recognition of fear. Neuropsychologia, 36(1), 59-70.

Brown, E. S., D, J. W., Frol, A., Bobadilla, L., Khan, D. A., Hanczyc, M., . . . Cullum, C. M. (2004). Hippocampal volume, spectroscopy, cognition, and mood in patients receiving corticosteroid therapy. Biol Psychiatry, 55(5), 538-545.

Brown, R. C., Han, Z., Cascio, C., & Papadopoulos, V. (2003). Oxidative stress- mediated DHEA formation in Alzheimer's disease pathology. Neurobiol Aging, 24(1), 57-65.

Bruijnzeel, A. W., Alexander, J. C., Perez, P. D., Bauzo-Rodriguez, R., Hall, G., Klausner, R., . . . Febo, M. (2015). Acute Nicotine Administration Increases BOLD fMRI Signal in Brain Regions Involved in Reward Signaling and Compulsive Drug Intake in Rats. International Journal of Neuropsychopharmacology, 18(2), pyu011.

Buchanan, T. W. (2007). Retrieval of Emotional Memories. Psychological bulletin, 133(5), 761-779.

Buckley, P. F., Miller, B. J., Lehrer, D. S., & Castle, D. J. (2009). Psychiatric Comorbidities and Schizophrenia. Schizophrenia Bulletin, 35(2), 383-402.

Burke, K. C., Burke, J. D., Jr., Regier, D. A., & Rae, D. S. (1990). Age at onset of selected mental disorders in five community populations. Arch Gen Psychiatry, 47(6), 511-518.

Calder, A. J. (1996). Facial Emotion Recognition after Bilateral Amygdala Damage: Differentially Severe Impairment of Fear. Cognitive Neuropsychology, 13(5), 699-745.

Cannon, T. D., Chung, Y., He, G., Sun, D., Jacobson, A., van Erp, T. G. M., . . . Heinssen, R. (2015). Progressive Reduction in Cortical Thickness as Psychosis Develops: A Multisite Longitudinal Neuroimaging Study of Youth at Elevated Clinical Risk. Biological Psychiatry, 77(2), 147-157.

Cannon, T. D., Thompson, P. M., van Erp, T. G. M., Toga, A. W., Poutanen, V.-P., Huttunen, M., . . . Kaprio, J. (2002). Cortex mapping reveals regionally specific patterns of genetic and disease-specific gray-matter deficits in twins discordant for schizophrenia. Proceedings of the National Academy of Sciences, 99(5), 3228-3233.

Canuso, C. M., & Pandina, G. (2007). Gender and schizophrenia. Psychopharmacol Bull, 40(4), 178-190.

Carcaillon, L., Brailly-Tabard, S., Ancelin, M. L., Tzourio, C., Foubert-Samier, A., Dartigues, J. F., . . . Scarabin, P. Y. (2014). Low testosterone and the risk of dementia in elderly men: Impact of age and education. Alzheimers Dement, 10(5 Suppl), S306-314. 161

Cardounel, A., Regelson, W., & Kalimi, M. (1999). Dehydroepiandrosterone protects hippocampal neurons against neurotoxin-induced cell death: mechanism of action. Proc Soc Exp Biol Med, 222(2), 145-149.

Carr, V. J., Neil, A. L., Halpin, S. A., Holmes, S., & Lewin, T. J. (2003). Costs of schizophrenia and other psychoses in urban Australia: findings from the Low Prevalence (Psychotic) Disorders Study. Aust N Z J Psychiatry, 37(1), 31-40.

Carrion, V. G., Weems, C. F., Richert, K., Hoffman, B. C., & Reiss, A. L. (2010). Decreased Prefrontal Cortical Volume Associated With Increased Bedtime Cortisol in Traumatized Youth. Biological Psychiatry, 68(5), 491-493.

Cascio, C., Brown, R. C., Liu, Y., Han, Z., Hales, D. B., & Papadopoulos, V. (2000). Pathways of dehydroepiandrosterone formation in rat brain glia. The Journal of Steroid Biochemistry and Molecular Biology, 75(2–3), 177-186.

Chabungbam, G., Avasthi, A., & Sharan, P. (2007). Sociodemographic and clinical factors associated with relapse in schizophrenia. Psychiatry Clin Neurosci, 61(6), 587-593.

Chapman, C., Tubridy, N., Cook, M. J., Mitchell, P. J., MacGregor, L. R., Lovelock, C., . . . Butzkueven, H. (2006). Short-term effects of methylprednisolone on cerebral volume in multiple sclerosis relapses. Journal of Clinical Neuroscience, 13(6), 636-638.

Chavez, C., Hollaus, M., Scarr, E., Pavey, G., Gogos, A., & van den Buuse, M. (2010). The effect of estrogen on dopamine and serotonin receptor and transporter levels in the brain: an autoradiography study. Brain Res, 1321, 51-59.

Che, Y., Wang, J.-F., Shao, L., & Young, L. T. (2010). Oxidative damage to RNA but not DNA in the hippocampus of patients with major mental illness. Journal of Psychiatry & Neuroscience : JPN, 35(5), 296-302.

Chen, F., Knecht, K., Birzin, E., Fisher, J., Wilkinson, H., Mojena, M., . . . Reszka, A. A. (2005). Direct agonist/antagonist functions of dehydroepiandrosterone. Endocrinology, 146(11), 4568-4576.

Chen, Y., Cataldo, A., Norton, D. J., & Ongur, D. (2012). Distinct facial processing in schizophrenia and schizoaffective disorders. Schizophr Res, 134(1), 95-100.

Cheng, G. F., & Tseng, J. (2000). Regulation of murine interleukin-10 production by dehydroepiandrosterone. J Interferon Cytokine Res, 20(5), 471-478.

Cherrier, M. M., Asthana, S., Plymate, S., Baker, L., Matsumoto, A. M., Peskind, E., . . . Craft, S. (2001). Testosterone supplementation improves spatial and verbal memory in healthy older men. Neurology, 57(1), 80-88.

Chin, E. H., Shah, A. H., Schmidt, K. L., Sheldon, L. D., Love, O. P., & Soma, K. K. (2008). Sex differences in DHEA and estradiol during development in a wild songbird: Jugular versus brachial plasma. Horm Behav, 54(1), 194-202. 162

Ciriza, I., Carrero, P., Azcoitia, I., Lundeen, S. G., & Garcia-Segura, L. M. (2004). Selective estrogen receptor modulators protect hippocampal neurons from kainic acid excitotoxicity: differences with the effect of estradiol. J Neurobiol, 61(2), 209-221.

Ciufolini, S., Dazzan, P., Kempton, M. J., Pariante, C., & Mondelli, V. (2014). HPA axis response to social stress is attenuated in schizophrenia but normal in depression: Evidence from a meta-analysis of existing studies. Neuroscience & Biobehavioral Reviews, 47, 359-368.

Compagnone, N. A., & Mellon, S. H. (1998). Dehydroepiandrosterone: a potential signalling molecule for neocortical organization during development. Proc Natl Acad Sci U S A, 95(8), 4678-4683.

Compagnone, N. A., & Mellon, S. H. (2000). Neurosteroids: biosynthesis and function of these novel neuromodulators. Front Neuroendocrinol, 21(1), 1-56.

Corigliano, V., De Carolis, A., Trovini, G., Dehning, J., Di Pietro, S., Curto, M., . . . Comparelli, A. (2014). Neurocognition in schizophrenia: from prodrome to multi-episode illness. Psychiatry Res, 220(1-2), 129-134.

Couture, S. M., Penn, D. L., & Roberts, D. L. (2006). The Functional Significance of Social Cognition in Schizophrenia: A Review. Schizophrenia Bulletin, 32(suppl 1), S44-S63.

Creese, I., Burt, D., & Snyder, S. (1976). Dopamine receptor binding predicts clinical and pharmacological potencies of antischizophrenic drugs. Science, 192(4238), 481-483.

Crinion, J., Ashburner, J., Leff, A., Brett, M., Price, C., & Friston, K. (2007). Spatial normalization of lesioned brains: Performance evaluation and impact on fMRI analyses. Neuroimage, 37(3), 866-875.

Cyr, M., Ghribi, O., Thibault, C., Morissette, M., Landry, M., & Di Paolo, T. (2001). Ovarian steroids and selective estrogen receptor modulators activity on rat brain NMDA and AMPA receptors. Brain Res Brain Res Rev, 37(1-3), 153-161.

Cyr, M., Landry, M., & Di Paolo, T. (2000). Modulation by Estrogen-Receptor Directed Drugs of 5-Hydroxytryptamine-2A Receptors in Rat Brain. Neuropsychopharmacology, 23(1), 69-78. d’Albis, M.-A., & Houenou, J. (2015). The Kraepelinian Dichotomy Viewed by Neuroimaging. Schizophrenia Bulletin, 41(2), 330-335.

Day, R., Nielsen, J. A., Korten, A., Ernberg, G., Dube, K. C., Gebhart, J., . . . et al. (1987). Stressful life events preceding the acute onset of schizophrenia: a cross- national study from the World Health Organization. Cult Med Psychiatry, 11(2), 123-205. de Kloet, E. R., Joels, M., & Holsboer, F. (2005). Stress and the brain: from adaptation to disease. Nat Rev Neurosci, 6(6), 463-475. 163

Depp, C. A., Moore, D. J., Sitzer, D., Palmer, B. W., Eyler, L. T., Roesch, S., . . . Jeste, D. V. (2007). Neurocognitive Impairment in Middle-Aged and Older Adults with Bipolar Disorder: Comparison to Schizophrenia and Normal Comparison Subjects. Journal of affective disorders, 101(1-3), 201-209. di Michele, F., Caltagirone, C., Bonaviri, G., Romeo, E., & Spalletta, G. (2005). Plasma dehydroepiandrosterone levels are strongly increased in schizophrenia. J Psychiatr Res, 39(3), 267-273.

Di Paolo, T. (1994). Modulation of brain dopamine transmission by sex steroids. Rev Neurosci, 5(1), 27-41.

Dickson, H., Cullen, A. E., Reichenberg, A., Hodgins, S., Campbell, D. D., Morris, R. G., & Laurens, K. R. (2014). Cognitive impairment among children at-risk for schizophrenia. J Psychiatr Res, 50, 92-99.

Diorio, D., Viau, V., & Meaney, M. J. (1993). The role of the medial prefrontal cortex (cingulate gyrus) in the regulation of hypothalamic-pituitary-adrenal responses to stress. J Neurosci, 13(9), 3839-3847.

Done, D. J., Crow, T. J., Johnstone, E. C., & Sacker, A. (1994). Childhood antecedents of schizophrenia and affective illness: social adjustment at ages 7 and 11. BMJ, 309(6956), 699-703.

Dong, Q., Welsh, R. C., Chenevert, T. L., Carlos, R. C., Maly-Sundgren, P., Gomez- Hassan, D. M., & Mukherji, S. K. (2004). Clinical applications of diffusion tensor imaging. J Magn Reson Imaging, 19(1), 6-18.

Dunn, J. F., Nisula, B. C., & Rodbard, D. (1981). Transport of steroid hormones: binding of 21 endogenous steroids to both testosterone-binding globulin and corticosteroid-binding globulin in human plasma. J Clin Endocrinol Metab, 53(1), 58-68.

Durany, N., Michel, T., Zochling, R., Boissl, K. W., Cruz-Sanchez, F. F., Riederer, P., & Thome, J. (2001). Brain-derived neurotrophic factor and neurotrophin 3 in schizophrenic psychoses. Schizophr Res, 52(1-2), 79-86.

Eaton, W. W., Thara, R., Federman, B., Melton, B., & Liang, K. Y. (1995). Structure and course of positive and negative symptoms in schizophrenia. Arch Gen Psychiatry, 52(2), 127-134.

Eberhard, J., Riley, F., & Levander, S. (2003). Premorbid IQ and schizophrenia. Increasing cognitive reduction by episodes. Eur Arch Psychiatry Clin Neurosci, 253(2), 84-88.

Edwards, J., Pattison, P. E., Jackson, H. J., & Wales, R. J. (2001). Facial affect and affective prosody recognition in first-episode schizophrenia. Schizophr Res, 48(2–3), 235-253.

Egan, M. F., Goldberg, T. E., Kolachana, B. S., Callicott, J. H., Mazzanti, C. M., Straub, R. E., . . . Weinberger, D. R. (2001). Effect of COMT Val108/158 Met genotype 164

on frontal lobe function and risk for schizophrenia. Proceedings of the National Academy of Sciences, 98(12), 6917-6922.

Ellman, L. M., Deicken, R. F., Vinogradov, S., Kremen, W. S., Poole, J. H., Kern, D. M., . . . Brown, A. S. (2010). Structural brain alterations in schizophrenia following fetal exposure to the inflammatory cytokine interleukin-8. Schizophr Res, 121(1-3), 46-54.

Emiliani, F. E., Sedlak, T. W., & Sawa, A. (2014). Oxidative stress and schizophrenia: recent breakthroughs from an old story. Curr Opin Psychiatry, 27(3), 185-190.

Eranti, S. V., MacCabe, J. H., Bundy, H., & Murray, R. M. (2013). Gender difference in age at onset of schizophrenia: a meta-analysis. Psychol Med, 43(1), 155-167.

Fang, H., Tong, W., Branham, W. S., Moland, C. L., Dial, S. L., Hong, H., . . . Sheehan, D. M. (2003). Study of 202 natural, synthetic, and environmental chemicals for binding to the androgen receptor. Chem Res Toxicol, 16(10), 1338-1358.

Fenton, W. S., & McGlashan, T. H. (1994). Antecedents, symptom progression, and long-term outcome of the deficit syndrome in schizophrenia. Am J Psychiatry, 151(3), 351-356.

Ferrier, I. N., Cotes, P. M., Crow, T. J., & Johnstone, E. C. (1982). Gonadotropin secretion abnormalities in chronic schizophrenia. Psychol Med, 12(2), 263-273.

Fett, A. K., Viechtbauer, W., Dominguez, M. D., Penn, D. L., van Os, J., & Krabbendam, L. (2011). The relationship between neurocognition and social cognition with functional outcomes in schizophrenia: a meta-analysis. Neurosci Biobehav Rev, 35(3), 573-588.

Fillman, S. G., Cloonan, N., Catts, V. S., Miller, L. C., Wong, J., McCrossin, T., . . . Weickert, C. S. (2013). Increased inflammatory markers identified in the dorsolateral prefrontal cortex of individuals with schizophrenia. Mol Psychiatry, 18(2), 206-214.

Fillman, S. G., Weickert, T. W., Lenroot, R. K., Catts, S. V., Bruggemann, J. M., Catts, V. S., & Weickert, C. S. (2015). Elevated peripheral cytokines characterize a subgroup of people with schizophrenia displaying poor verbal fluency and reduced Broca/'s area volume. Mol Psychiatry.

Fiore, C., Inman, D. M., Hirose, S., Noble, L. J., Igarashi, T., & Compagnone, N. A. (2004). Treatment with the neurosteroid dehydroepiandrosterone promotes recovery of motor behavior after moderate contusive spinal cord injury in the mouse. J Neurosci Res, 75(3), 391-400.

First, M., Williams, J., Spitzer, R., & Gibbon, M. (2007). Structured Clinical Interview for DSM-IV-TR Axis I Disorders, Patient Edition, 1st revision edn., New York, Biometrics Research Department, New York State Psychiatric Institute

165

Fraguas, D., Gonzalez-Pinto, A., Mico, J. A., Reig, S., Parellada, M., Martinez- Cengotitabengoa, M., . . . Arango, C. (2012). Decreased glutathione levels predict loss of brain volume in children and adolescents with first-episode psychosis in a two-year longitudinal study. Schizophr Res, 137(1-3), 58-65.

Frank, M. G., Baratta, M. V., Sprunger, D. B., Watkins, L. R., & Maier, S. F. (2007). Microglia serve as a neuroimmune substrate for stress-induced potentiation of CNS pro-inflammatory cytokine responses. Brain Behav Immun, 21(1), 47-59.

Friess, E., Schiffelholz, T., Steckler, T., & Steiger, A. (2000). Dehydroepiandrosterone- -a neurosteroid. Eur J Clin Invest, 30 Suppl 3, 46-50.

Fuller, R., Nopoulos, P., Arndt, S., O'Leary, D., Ho, B. C., & Andreasen, N. C. (2002). Longitudinal assessment of premorbid cognitive functioning in patients with schizophrenia through examination of standardized scholastic test performance. Am J Psychiatry, 159(7), 1183-1189.

Fusar-Poli, P., Placentino, A., Carletti, F., Landi, P., Allen, P., Surguladze, S., . . . Politi, P. (2009). Functional atlas of emotional faces processing: a voxel-based meta- analysis of 105 functional magnetic resonance imaging studies. J Psychiatry Neurosci, 34(6), 418-432.

Fusar-Poli, P., Smieskova, R., Kempton, M. J., Ho, B. C., Andreasen, N. C., & Borgwardt, S. (2013). Progressive brain changes in schizophrenia related to antipsychotic treatment? A meta-analysis of longitudinal MRI studies. Neurosci Biobehav Rev, 37(8), 1680-1691.

Gallagher, P., Leitch, M. M., Massey, A. E., McAllister-Williams, R. H., & Young, A. H. (2006). Assessing cortisol and dehydroepiandrosterone (DHEA) in saliva: effects of collection method. J Psychopharmacol, 20(5), 643-649.

Gallagher, P., Watson, S., Smith, M. S., Young, A. H., & Ferrier, I. N. (2007). Plasma cortisol-dehydroepiandrosterone (DHEA) ratios in schizophrenia and bipolar disorder. Schizophr Res, 90(1–3), 258-265.

Garner, B., Phassouliotis, C., Phillips, L. J., Markulev, C., Butselaar, F., Bendall, S., . . . McGorry, P. D. (2011). Cortisol and dehydroepiandrosterone-sulphate levels correlate with symptom severity in first-episode psychosis. J Psychiatr Res, 45(2), 249-255.

Gattaz, W. F., Vogel, P., Riecher-Rössler, A., & Soddu, G. (1994). Influence of the menstrual cycle phase on the therapeutic response in schizophrenia. Biological Psychiatry, 36(2), 137-139.

Gawryluk, J. W., Wang, J. F., Andreazza, A. C., Shao, L., & Young, L. T. (2011). Decreased levels of glutathione, the major brain antioxidant, in post-mortem prefrontal cortex from patients with psychiatric disorders. Int J Neuropsychopharmacol, 14(1), 123-130.

Geisler, D., Walton, E., Naylor, M., Roessner, V., Lim, K. O., Charles Schulz, S., . . . Ehrlich, S. (2015). Brain structure and function correlates of cognitive subtypes in schizophrenia. Psychiatry Res, 234(1), 74-83. 166

Girshkin, L., Matheson, S. L., Shepherd, A. M., & Green, M. J. (2014). Morning cortisol levels in schizophrenia and bipolar disorder: A meta-analysis. Psychoneuroendocrinology, 49, 187-206.

Goekoop, R., Barkhof, F., Duschek, E. J. J., Netelenbos, C., Knol, D. L., Scheltens, P., & Rombouts, S. A. R. B. (2005). Raloxifene Treatment Enhances Brain Activation during Recognition of Familiar Items: a Pharmacological fMRI Study in Healthy Elderly Males. Neuropsychopharmacology, 31(7), 1508-1518.

Goekoop, R., Duschek, E. J., Knol, D. L., Barkhof, F., Netelenbos, C., Scheltens, P., & Rombouts, S. A. (2005). Raloxifene exposure enhances brain activation during memory performance in healthy elderly males; its possible relevance to behavior. Neuroimage, 25(1), 63-75.

Gogos, A., Kwek, P., & van den Buuse, M. (2012). The role of estrogen and testosterone in female rats in behavioral models of relevance to schizophrenia. Psychopharmacology (Berl), 219(1), 213-224.

Gogos, A., Sbisa, A. M., Sun, J., Gibbons, A., Udawela, M., & Dean, B. (2015). A Role for Estrogen in Schizophrenia: Clinical and Preclinical Findings. Int J Endocrinol, 2015, 615356.

Goldberg, T. E., Karson, C. N., Leleszi, J. P., & Weinberger, D. R. (1988). Intellectual impairment in adolescent psychosis. A controlled psychometric study. Schizophr Res, 1(4), 261-266.

Goldberg, T. E., Torrey, E. F., Gold, J. M., Bigelow, L. B., Ragland, R. D., Taylor, E., & Weinberger, D. R. (1995). Genetic risk of neuropsychological impairment in schizophrenia: a study of monozygotic twins discordant and concordant for the disorder. Schizophr Res, 17(1), 77-84.

Goldman-Rakic, P. S., & Selemon, L. D. (1997). Functional and anatomical aspects of prefrontal pathology in schizophrenia. Schizophr Bull, 23(3), 437-458.

Goldstein, J. M., & Link, B. G. (1988). Gender and the expression of schizophrenia. J Psychiatr Res, 22(2), 141-155.

Gonzalez-Alvear, G. M., & Werling, L. L. (1994). Regulation of [3H]dopamine release from rat striatal slices by sigma receptor ligands. J Pharmacol Exp Ther, 271(1), 212-219.

Gore, J. C. (2003). Principles and practice of functional MRI of the human brain. Journal of Clinical Investigation, 112(1), 4-9.

Grace, A. A. (2016). Dysregulation of the dopamine system in the pathophysiology of schizophrenia and depression. Nat Rev Neurosci, advance online publication.

Granger, D. A., Schwartz, E. B., Booth, A., Curran, M., & Zakaria, D. (1999). Assessing dehydroepiandrosterone in saliva: a simple radioimmunoassay for use in studies of children, adolescents and adults. Psychoneuroendocrinology, 24(5), 567-579. 167

Green, M. F. (1996). What are the functional consequences of neurocognitive deficits in schizophrenia? Am J Psychiatry, 153(3), 321-330.

Green, M. F., Horan, W. P., & Lee, J. (2015). Social cognition in schizophrenia. Nat Rev Neurosci, 16(10), 620-631.

Guastella, A. J., Ward, P. B., Hickie, I. B., Shahrestani, S., Hodge, M. A. R., Scott, E. M., & Langdon, R. (2015). A single dose of oxytocin nasal spray improves higher-order social cognition in schizophrenia. Schizophr Res, 168(3), 628-633.

Guazzo, E. P., Kirkpatrick, P. J., Goodyer, I. M., Shiers, H. M., & Herbert, J. (1996). Cortisol, dehydroepiandrosterone (DHEA), and DHEA sulfate in the cerebrospinal fluid of man: relation to blood levels and the effects of age. J Clin Endocrinol Metab, 81(11), 3951-3960.

Guo, J. Y., Huhtaniska, S., Miettunen, J., Jääskeläinen, E., Kiviniemi, V., Nikkinen, J., . . . Murray, G. K. (2015). Longitudinal regional brain volume loss in schizophrenia: Relationship to antipsychotic medication and change in social function. Schizophr Res, 168(1–2), 297-304.

Gur, R. E., Calkins, M. E., Gur, R. C., Horan, W. P., Nuechterlein, K. H., Seidman, L. J., & Stone, W. S. (2007). The Consortium on the Genetics of Schizophrenia: neurocognitive endophenotypes. Schizophr Bull, 33(1), 49-68.

Gur, R. E., McGrath, C., Chan, R. M., Schroeder, L., Turner, T., Turetsky, B. I., . . . Gur, R. C. (2002). An fMRI study of facial emotion processing in patients with schizophrenia. Am J Psychiatry, 159(12), 1992-1999.

Gur, R. E., Petty, R. G., Turetsky, B. I., & Gur, R. C. (1996). Schizophrenia throughout life: sex differences in severity and profile of symptoms. Schizophr Res, 21(1), 1-12.

Gurvits, T. V., Shenton, M. E., Hokama, H., Ohta, H., Lasko, N. B., Gilbertson, M. W., . . . Pitman, R. K. (1996). Magnetic resonance imaging study of hippocampal volume in chronic, combat-related posttraumatic stress disorder. Biol Psychiatry, 40(11), 1091-1099.

Häfner, H. (2003). Gender differences in schizophrenia. Psychoneuroendocrinology, 28, Supplement 2, 17-54.

Hafner, H., Behrens, S., De Vry, J., & Gattaz, W. F. (1991). An animal model for the effects of estradiol on dopamine-mediated behavior: implications for sex differences in schizophrenia. Psychiatry Res, 38(2), 125-134.

Hafner, H., Riecher-Rossler, A., An Der Heiden, W., Maurer, K., Fatkenheuer, B., & Loffler, W. (1993). Generating and testing a causal explanation of the gender difference in age at first onset of schizophrenia. Psychol Med, 23(4), 925-940.

Haijma, S. V., Van Haren, N., Cahn, W., Koolschijn, P. C. M. P., Hulshoff Pol, H. E., & Kahn, R. S. (2012). Brain Volumes in Schizophrenia: A Meta-Analysis in Over 18 000 Subjects. Schizophrenia Bulletin. 168

Hajszan, T., MacLusky, N. J., & Leranth, C. (2004). Dehydroepiandrosterone increases hippocampal spine synapse density in ovariectomized female rats. Endocrinology, 145(3), 1042-1045.

Halari, R., Kumari, V., Mehrotra, R., Wheeler, M., Hines, M., & Sharma, T. (2004). The relationship of sex hormones and cortisol with cognitive functioning in Schizophrenia. J Psychopharmacol, 18(3), 366-374.

Halbreich, U., & Kahn, L. S. (2001). Role of estrogen in the aetiology and treatment of mood disorders. CNS Drugs, 15(10), 797-817.

Halbreich, U., Rojansky, N., Palter, S., Tworek, H., Hissin, P., & Wang, K. (1995). Estrogen augments serotonergic activity in postmenopausal women. Biol Psychiatry, 37(7), 434-441.

Hallonquist, J. D., Seeman, M. V., Lang, M., & Rector, N. A. (1993). Variation in symptom severity over the menstrual cycle of schizophrenics. Biol Psychiatry, 33(3), 207-209.

Hambrecht, M., Maurer, K., Hafner, H., & Sartorius, N. (1992). Transnational stability of gender differences in schizophrenia? An analysis based on the WHO study on determinants of outcome of severe mental disorders. Eur Arch Psychiatry Clin Neurosci, 242(1), 6-12.

Harris, D. S., Wolkowitz, O. M., & Reus, V. I. (2001). Movement disorder, memory, psychiatric symptoms and serum DHEA levels in schizophrenic and schizoaffective patients. World J Biol Psychiatry, 2(2), 99-102.

Hashimoto, T., Bergen, S. E., Nguyen, Q. L., Xu, B., Monteggia, L. M., Pierri, J. N., . . . Lewis, D. A. (2005). Relationship of brain-derived neurotrophic factor and its receptor TrkB to altered inhibitory prefrontal circuitry in schizophrenia. J Neurosci, 25(2), 372-383.

Haslam, J. (1810). Illustrations of Madness.

Havelock, J. C., Auchus, R. J., & Rainey, W. E. (2004). The rise in adrenal androgen biosynthesis: adrenarche. Semin Reprod Med, 22(4), 337-347.

Hayes, L. N., Severance, E. G., Leek, J. T., Gressitt, K. L., Rohleder, C., Coughlin, J. M., . . . Sawa, A. (2014). Inflammatory molecular signature associated with infectious agents in psychosis. Schizophr Bull, 40(5), 963-972.

Heeger, D. J., & Ress, D. (2002). What does fMRI tell us about neuronal activity? Nat Rev Neurosci, 3(2), 142-151.

Heinssen, R. K., Liberman, R. P., & Kopelowicz, A. (2000). Psychosocial skills training for schizophrenia: lessons from the laboratory. Schizophr Bull, 26(1), 21-46.

Heldring, N., Pike, A., Andersson, S., Matthews, J., Cheng, G., Hartman, J., . . . Gustafsson, J.-Å. (2007). Estrogen Receptors: How Do They Signal and What Are Their Targets. Physiological Reviews, 87(3), 905-931. 169

Hempel, A., Hempel, E., Schonknecht, P., Stippich, C., & Schroder, J. (2003). Impairment in basal limbic function in schizophrenia during affect recognition. Psychiatry Res, 122(2), 115-124.

Herbison, A. E., & Fenelon, V. S. (1995). Estrogen regulation of GABAA receptor subunit mRNA expression in preoptic area and bed nucleus of the stria terminalis of female rat brain. J Neurosci, 15(3 Pt 2), 2328-2337.

Hirshman, E., Wells, E., Wierman, M. E., Anderson, B., Butler, A., Senholzi, M., & Fisher, J. (2003). The effect of dehydroepiandrosterone (DHEA) on recognition memory decision processes and discrimination in postmenopausal women. Psychonomic Bulletin & Review, 10(1), 125-134.

Ho, B.-C., Andreasen, N. C., Ziebell, S., Pierson, R., & Magnotta, V. (2011). Long-term Antipsychotic Treatment and Brain Volumes: A Longitudinal Study of First- Episode Schizophrenia. Arch Gen Psychiatry, 68(2), 128-137.

Hoff, A. L., Kremen, W. S., Wieneke, M. H., Lauriello, J., Blankfeld, H. M., Faustman, W. O., . . . Nordahl, T. E. (2001). Association of estrogen levels with neuropsychological performance in women with schizophrenia. Am J Psychiatry, 158(7), 1134-1139.

Holland, J., Bandelow, S., & Hogervorst, E. (2011). Testosterone levels and cognition in elderly men: A review. Maturitas, 69(4), 322-337.

Hooker, C., & Park, S. (2002). Emotion processing and its relationship to social functioning in schizophrenia patients. Psychiatry Res, 112(1), 41-50.

Hor, K., & Taylor, M. (2010). Suicide and schizophrenia: a systematic review of rates and risk factors. Journal of Psychopharmacology (Oxford, England), 24(4_supplement), 81-90.

Horrobin, D. F. (1998). Schizophrenia: the illness that made us human. Medical Hypotheses, 50(4), 269-288.

Howes, O. D., & Kapur, S. (2009). The Dopamine Hypothesis of Schizophrenia: Version III—The Final Common Pathway. Schizophrenia Bulletin, 35(3), 549- 562.

Huber, T. J., Rollnik, J., Wilhelms, J., von zur Muhlen, A., Emrich, H. M., & Schneider, U. (2001). Estradiol levels in psychotic disorders. Psychoneuroendocrinology, 26(1), 27-35.

Huerta-Ramos, E., Iniesta, R., Ochoa, S., Cobo, J., Miquel, E., Roca, M., . . . Usall, J. (2014). Effects of raloxifene on cognition in postmenopausal women with schizophrenia: a double-blind, randomized, placebo-controlled trial. Eur Neuropsychopharmacol, 24(2), 223-231.

Hulka, B. S., & Wilcosky, T. (1988). Biological markers in epidemiologic research. Arch Environ Health, 43(2), 83-89. 170

Ihnen, G. H., Penn, D. L., Corrigan, P. W., & Martin, J. (1998). Social perception and social skill in schizophrenia. Psychiatry Res, 80(3), 275-286.

Jacob, M. H., da, R. J. D., Jahn, M. P., Kucharski, L. C., Bello-Klein, A., & Ribeiro, M. F. (2010). Age-related effects of DHEA on peripheral markers of oxidative stress. Cell Biochem Funct, 28(1), 52-57.

Jacobs, D. M., Tang, M. X., Stern, Y., Sano, M., Marder, K., Bell, K. L., . . . Mayeux, R. (1998). Cognitive function in nondemented older women who took estrogen after menopause. Neurology, 50(2), 368-373.

Jacobs, E., & D'Esposito, M. (2011). Estrogen shapes dopamine-dependent cognitive processes: implications for women's health. J Neurosci, 31(14), 5286-5293.

Jansen, L. M., Gispen-de Wied, C. C., Gademan, P. J., De Jonge, R. C., van der Linden, J. A., & Kahn, R. S. (1998). Blunted cortisol response to a psychosocial stressor in schizophrenia. Schizophr Res, 33(1-2), 87-94.

Jeanneteau, F. D., Lambert, W. M., Ismaili, N., Bath, K. G., Lee, F. S., Garabedian, M. J., & Chao, M. V. (2012). BDNF and glucocorticoids regulate corticotrophin- releasing hormone (CRH) homeostasis in the hypothalamus. Proc Natl Acad Sci U S A, 109(4), 1305-1310.

Jeong, B., Wible, C. G., Hashimoto, R., & Kubicki, M. (2009). Functional and anatomical connectivity abnormalities in left inferior frontal gyrus in schizophrenia. Hum Brain Mapp, 30(12), 4138-4151.

Ji, E., Weickert, C. S., Lenroot, R., Catts, S. V., Vercammen, A., White, C., . . . Weickert, T. W. (2015). Endogenous testosterone levels are associated with neural activity in men with schizophrenia during facial emotion processing. Behavioural Brain Research, 286, 338-346.

Jindal, R. D., Pillai, A. K., Mahadik, S. P., Eklund, K., Montrose, D. M., & Keshavan, M. S. (2010). Decreased BDNF in patients with antipsychotic naive first episode schizophrenia. Schizophr Res, 119(1-3), 47-51.

Johnstone, E. C., Crow, T. J., Frith, C. D., Carney, M. W., & Price, J. S. (1978). Mechanism of the antipsychotic effect in the treatment of acute schizophrenia. Lancet, 1(8069), 848-851.

Johnstone, E. C., Crow, T. J., Frith, C. D., Husband, J., & Kreel, L. (1976). Cerebral ventricular size and cognitive impairment in chronic schizophrenia. Lancet, 2(7992), 924-926.

Jones, P., & Cannon, M. (1998). The new epidemiology of schizophrenia. Psychiatr Clin North Am, 21(1), 1-25.

Kahl, K. G., Bens, S., Ziegler, K., Rudolf, S., Dibbelt, L., Kordon, A., & Schweiger, U. (2006). Cortisol, the cortisol-dehydroepiandrosterone ratio, and pro- inflammatory cytokines in patients with current major depressive disorder comorbid with borderline personality disorder. Biol Psychiatry, 59(7), 667-671. 171

Kalimi, M., Shafagoj, Y., Loria, R., Padgett, D., & Regelson, W. (1994). Anti- glucocorticoid effects of dehydroepiandrosterone (DHEA). Molecular and Cellular Biochemistry, 131(2), 99-104.

Kalsbeek, A., van der Spek, R., Lei, J., Endert, E., Buijs, R. M., & Fliers, E. (2012). Circadian rhythms in the hypothalamo-pituitary-adrenal (HPA) axis. Mol Cell Endocrinol, 349(1), 20-29.

Kancheva, R., Hill, M., Novak, Z., Chrastina, J., Kancheva, L., & Starka, L. (2011). Neuroactive steroids in periphery and cerebrospinal fluid. Neuroscience, 191, 22-27.

Kapur, S., & Mamo, D. (2003). Half a century of antipsychotics and still a central role for dopamine D2 receptors. Prog Neuropsychopharmacol Biol Psychiatry, 27(7), 1081-1090.

Karbasforoushan, H., & Woodward, N. D. (2012). Resting-state networks in schizophrenia. Curr Top Med Chem, 12(21), 2404-2414.

Karishma, K. K., & Herbert, J. (2002). Dehydroepiandrosterone (DHEA) stimulates neurogenesis in the hippocampus of the rat, promotes survival of newly formed neurons and prevents corticosterone-induced suppression. Eur J Neurosci, 16(3), 445-453.

Karuza, E. A., Emberson, L. L., & Aslin, R. N. (2014). Combining fMRI and behavioral measures to examine the process of human learning. Neurobiol Learn Mem, 109, 193-206.

Kay, S. R., Fiszbein, A., & Opler, L. A. (1987). The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr Bull, 13(2), 261-276.

Kelley, W. M., Miezin, F. M., McDermott, K. B., Buckner, R. L., Raichle, M. E., Cohen, N. J., . . . Petersen, S. E. (1998). Hemispheric specialization in human dorsal frontal cortex and medial temporal lobe for verbal and nonverbal memory encoding. Neuron, 20(5), 927-936.

Keshavan, M. S., & Cornblatt, B. A. (2010). Developmental models and hypotheses- driven early interventions in schizophrenia. In C. U. Press (Ed.), Neurodevelopment and Schizophrenia. Cambridge.

Khashan, A. S., Abel, K. M., McNamee, R., Pedersen, M. G., Webb, R. T., Baker, P. N., . . . Mortensen, P. B. (2008). Higher risk of offspring schizophrenia following antenatal maternal exposure to severe adverse life events. Arch Gen Psychiatry, 65(2), 146-152.

Kian Tee, M., Rogatsky, I., Tzagarakis-Foster, C., Cvoro, A., An, J., Christy, R. J., . . . Leitman, D. C. (2004). Estradiol and Selective Estrogen Receptor Modulators Differentially Regulate Target Genes with Estrogen Receptors α and β. Molecular Biology of the Cell, 15(3), 1262-1272. 172

Killgore, W. D. S., & Yurgelun-Todd, D. A. (2007). The right-hemisphere and valence hypotheses: could they both be right (and sometimes left)? Social Cognitive and Affective Neuroscience, 2(3), 240-250.

Kimonides, V. G., Khatibi, N. H., Svendsen, C. N., Sofroniew, M. V., & Herbert, J. (1998). Dehydroepiandrosterone (DHEA) and DHEA-sulfate (DHEAS) protect hippocampal neurons against excitatory amino acid-induced neurotoxicity. Proc Natl Acad Sci U S A, 95(4), 1852-1857.

Kimonides, V. G., Spillantini, M. G., Sofroniew, M. V., Fawcett, J. W., & Herbert, J. (1999). Dehydroepiandrosterone antagonizes the neurotoxic effects of corticosterone and translocation of stress-activated protein kinase 3 in hippocampal primary cultures. Neuroscience, 89(2), 429-436.

Kimura, M., Tanaka, S., Yamada, Y., Kiuchi, Y., Yamakawa, T., & Sekihara, H. (1998). Dehydroepiandrosterone decreases serum tumor necrosis factor-alpha and restores insulin sensitivity: independent effect from secondary weight reduction in genetically obese Zucker fatty rats. Endocrinology, 139(7), 3249-3253.

Kindler, J., Weickert, C. S., Skilleter, A. J., Catts, S. V., Lenroot, R., & Weickert, T. W. (2015). Selective Estrogen Receptor Modulation Increases Hippocampal Activity during Probabilistic Association Learning in Schizophrenia. Neuropsychopharmacology, 40(10), 2388-2397.

Kinon, B. J., Gilmore, J. A., Liu, H., & Halbreich, U. M. (2003). Prevalence of hyperprolactinemia in schizophrenic patients treated with conventional antipsychotic medications or risperidone1. Psychoneuroendocrinology, 28, Supplement 2, 55-68.

Kirschbaum, C., & Hellhammer, D. H. (1989). Salivary cortisol in psychobiological research: an overview. Neuropsychobiology, 22(3), 150-169.

Kirschbaum, C., & Hellhammer, D. H. (1994). Salivary cortisol in psychoneuroendocrine research: recent developments and applications. Psychoneuroendocrinology, 19(4), 313-333.

Ko, Y. H., Joe, S. H., Cho, W., Park, J. H., Lee, J. J., Jung, I. K., . . . Kim, S. H. (2006). Estrogen, Cognitive Function and Negative Symptoms in Female Schizophrenia. Neuropsychobiology, 53(4), 169-175.

Ko, Y. H., Jung, S. W., Joe, S. H., Lee, C. H., Jung, H. G., Jung, I. K., . . . Lee, M. S. (2007). Association between serum testosterone levels and the severity of negative symptoms in male patients with chronic schizophrenia. Psychoneuroendocrinology, 32(4), 385-391.

Ko, Y. H., Lew, Y. M., Jung, S. W., Joe, S. H., Lee, C. H., Jung, H. G., & Lee, M. S. (2008). Short-term testosterone augmentation in male schizophrenics: a randomized, double-blind, placebo-controlled trial. J Clin Psychopharmacol, 28(4), 375-383. 173

Kohler, C. G., Bilker, W., Hagendoorn, M., Gur, R. E., & Gur, R. C. (2000). Emotion recognition deficit in schizophrenia: association with symptomatology and cognition. Biol Psychiatry, 48(2), 127-136.

Kohler, C. G., Turner, T. H., Bilker, W. B., Brensinger, C. M., Siegel, S. J., Kanes, S. J., . . . Gur, R. C. (2003). Facial emotion recognition in schizophrenia: intensity effects and error pattern. Am J Psychiatry, 160(10), 1768-1774.

Kohler, C. G., Walker, J. B., Martin, E. A., Healey, K. M., & Moberg, P. J. (2010). Facial emotion perception in schizophrenia: a meta-analytic review. Schizophr Bull, 36(5), 1009-1019.

Konarzewska, B., Wolczynski, S., Szulc, A., Galinska, B., Poplawska, R., & Waszkiewicz, N. (2009). Effect of risperidone and olanzapine on reproductive hormones, psychopathology and sexual functioning in male patients with schizophrenia. Psychoneuroendocrinology, 34(1), 129-139.

Konyalioglu, S., Durmaz, G., & Yalcin, A. (2007). The potential antioxidant effect of raloxifene treatment: a study on heart, liver and brain cortex of ovariectomized female rats. Cell Biochemistry and Function, 25(3), 259-266.

Kranz, G. S., Rami-Mark, C., Kaufmann, U., Baldinger, P., Hahn, A., Hoflich, A., . . . Kasper, S. (2014). Effects of hormone replacement therapy on cerebral serotonin-1A receptor binding in postmenopausal women examined with [carbonyl-(1)(1)C]WAY-100635. Psychoneuroendocrinology, 45, 1-10.

Kritz-Silverstein, D., von Mühlen, D., Laughlin, G. A., & Bettencourt, R. (2008). Effects of Dehydroepiandrosterone Supplementation on Cognitive Function and Quality of Life: The DHEA and Well-Ness (DAWN) Trial. Journal of the American Geriatrics Society, 56(7), 1292-1298.

Kritzer, M. F., & Kohama, S. G. (1999). Ovarian hormones differentially influence immunoreactivity for dopamine beta- hydroxylase, choline acetyltransferase, and serotonin in the dorsolateral prefrontal cortex of adult rhesus monkeys. J Comp Neurol, 409(3), 438-451.

Kroboth, P. D., Amico, J. A., Stone, R. A., Folan, M., Frye, R. F., Kroboth, F. J., . . . Hakala, C. (2003). Influence of DHEA administration on 24-hour cortisol concentrations. J Clin Psychopharmacol, 23(1), 96-99.

Kugaya, A., Epperson, C. N., Zoghbi, S., van Dyck, C. H., Hou, Y., Fujita, M., . . . Innis, R. B. (2003). Increase in prefrontal cortex serotonin 2A receptors following estrogen treatment in postmenopausal women. Am J Psychiatry, 160(8), 1522- 1524.

Kulkarni, J., de Castella, A., Fitzgerald, P. B., & et al. (2008). Estrogen in severe mental illness: A potential new treatment approach. Arch Gen Psychiatry, 65(8), 955- 960.

Kulkarni, J., de Castella, A., Headey, B., Marston, N., Sinclair, K., Lee, S., . . . Burger, H. (2011). Estrogens and men with schizophrenia: is there a case for adjunctive therapy? Schizophr Res, 125(2-3), 278-283. 174

Kulkarni, J., de Castella, A., Smith, D., Taffe, J., Keks, N., & Copolov, D. (1996). A clinical trial of the effects of estrogen in acutely psychotic women. Schizophr Res, 20(3), 247-252.

Kulkarni, J., Gavrilidis, E., Wang, W., Worsley, R., Fitzgerald, P. B., Gurvich, C., . . . Burger, H. (2014). Estradiol for treatment-resistant schizophrenia: a large-scale randomized-controlled trial in women of child-bearing age. Mol Psychiatry.

Kulkarni, J., Gavrilidis, E., Wang, W., Worsley, R., Fitzgerald, P. B., Gurvich, C., . . . Burger, H. (2015). Estradiol for treatment-resistant schizophrenia: a large-scale randomized-controlled trial in women of child-bearing age. Mol Psychiatry, 20(6), 695-702.

Kulkarni, J., Gurvich, C., Lee, S. J., Gilbert, H., Gavrilidis, E., de Castella, A., . . . Davis, S. R. (2010). Piloting the effective therapeutic dose of adjunctive selective estrogen receptor modulator treatment in postmenopausal women with schizophrenia. Psychoneuroendocrinology, 35(8), 1142-1147.

Kulkarni, J., Riedel, A., de Castella, A. R., Fitzgerald, P. B., Rolfe, T. J., Taffe, J., & Burger, H. (2001). Estrogen — a potential treatment for schizophrenia. Schizophr Res, 48(1), 137-144.

Kulla, A., & Manahan-Vaughan, D. (2000). Depotentiation in the dentate gyrus of freely moving rats is modulated by D1/D5 dopamine receptors. Cereb Cortex, 10(6), 614-620.

Kumari, V., Fannon, D., Peters, E. R., ffytche, D. H., Sumich, A. L., Premkumar, P., . . . Kuipers, E. (2011). Neural changes following cognitive behaviour therapy for psychosis: a longitudinal study. Brain, 134(8), 2396-2407.

Labrie, F. (2004). Adrenal androgens and intracrinology. Semin Reprod Med, 22(4), 299-309.

Lac, G. (2001). Saliva assays in clinical and research biology. Pathol Biol (Paris), 49(8), 660-667.

Lamont, E. W., Coutu, D. L., Cermakian, N., & Boivin, D. B. (2010). Circadian rhythms and clock genes in psychotic disorders. Isr J Psychiatry Relat Sci, 47(1), 27-35.

Landry, M., Levesque, D., & Di Paolo, T. (2002). Estrogenic properties of raloxifene, but not tamoxifen, on D2 and D3 dopamine receptors in the rat forebrain. Neuroendocrinology, 76(4), 214-222.

Laursen, T. M., Munk-Olsen, T., & Vestergaard, M. (2012). Life expectancy and cardiovascular mortality in persons with schizophrenia. Curr Opin Psychiatry, 25(2), 83-88.

Lee, J., & Park, S. (2005). Working Memory Impairments in Schizophrenia: A Meta- Analysis. J Abnorm Psychol, 114(4), 599-611. 175

Lee, S.-K., Chun, J. W., Lee, J. S., Park, H.-J., Jung, Y.-C., Seok, J.-H., & Kim, J.-J. (2014). Abnormal Neural Processing during Emotional Salience Attribution of Affective Asymmetry in Patients with Schizophrenia. PLoS ONE, 9(3), e90792.

Lee, S. K., Chun, J. W., Lee, J. S., Park, H. J., Jung, Y. C., Seok, J. H., & Kim, J. J. (2014). Abnormal neural processing during emotional salience attribution of affective asymmetry in patients with schizophrenia. PLoS ONE, 9(3), e90792.

Leeson, V. C., Sharma, P., Harrison, M., Ron, M. A., Barnes, T. R. E., & Joyce, E. M. (2011). IQ Trajectory, Cognitive Reserve, and Clinical Outcome Following a First Episode of Psychosis: A 3-Year Longitudinal Study. Schizophrenia Bulletin, 37(4), 768-777.

Leucht, S., Wahlbeck, K., Hamann, J., & Kissling, W. (2003). New generation antipsychotics versus low-potency conventional antipsychotics: a systematic review and meta-analysis. Lancet, 361(9369), 1581-1589.

Levine, S. Z., Rabinowitz, J., Ascher-Svanum, H., Faries, D. E., & Lawson, A. H. (2011). Extent of attaining and maintaining symptom remission by antipsychotic medication in the treatment of chronic schizophrenia: Evidence from the CATIE study. Schizophr Res, 133(1–3), 42-46.

Lezak, M. (2004). Neuropsychological Assessment, 4th Edition. Oxford University Press. New York, NY.

Li, H., Chan, R. C. K., McAlonan, G. M., & Gong, Q.-y. (2010). Facial Emotion Processing in Schizophrenia: A Meta-analysis of Functional Neuroimaging Data. Schizophrenia Bulletin, 36(5), 1029-1039.

Li, H., Klein, G., Sun, P., & Buchan, A. M. (2001). Dehydroepiandrosterone (DHEA) reduces neuronal injury in a rat model of global cerebral ischemia. Brain Res, 888(2), 263-266.

Li, R., Xu, W., Chen, Y., Qiu, W., Shu, Y., Wu, A., . . . Hu, X. (2014). Raloxifene suppresses experimental autoimmune encephalomyelitis and NF-kappaB- dependent CCL20 expression in reactive astrocytes. PLoS ONE, 9(4), e94320.

Li, Z. J., Guo, Z. H., Wang, N., Xu, Z. Y., Qu, Y., Wang, X. Q., . . . Kingdon, D. (2015). Cognitive-behavioural therapy for patients with schizophrenia: a multicentre randomized controlled trial in Beijing, China. Psychol Med, 45(9), 1893-1905.

Lieberman, J. A., Perkins, D., Belger, A., Chakos, M., Jarskog, F., Boteva, K., & Gilmore, J. (2001). The early stages of schizophrenia: speculations on pathogenesis, pathophysiology, and therapeutic approaches. Biol Psychiatry, 50(11), 884-897.

Lightman, S. L., & Conway-Campbell, B. L. (2010). The crucial role of pulsatile activity of the HPA axis for continuous dynamic equilibration. Nat Rev Neurosci, 11(10), 710-718. 176

Lim, D., Flicker, L., Dharamarajan, A., & Martins, R. N. (2003). Can testosterone replacement decrease the memory problem of old age? Med Hypotheses, 60(6), 893-896.

Lim, S.-H., Park, E., You, B., Jung, Y., Park, A. R., Park, S. G., & Lee, J.-R. (2013). Neuronal Synapse Formation Induced by Microglia and Interleukin 10. PLoS ONE, 8(11), e81218.

Lindamer, L. A., Buse, D. C., Lohr, J. B., & Jeste, D. V. (2001). Hormone replacement therapy in postmenopausal women with schizophrenia: positive effect on negative symptoms? Biol Psychiatry, 49(1), 47-51.

Logothetis, N. K. (2008). What we can do and what we cannot do with fMRI. Nature, 453(7197), 869-878.

Lonard, D. M., & Smith, C. L. (2002). Molecular perspectives on selective estrogen receptor modulators (SERMs): progress in understanding their tissue-specific agonist and antagonist actions. Steroids, 67(1), 15-24.

Loranger, A. W. (1984). SEx difference in age at onset of schizophrenia. Arch Gen Psychiatry, 41(2), 157-161.

Lu, S. F., Mo, Q., Hu, S., Garippa, C., & Simon, N. G. (2003). Dehydroepiandrosterone upregulates neural androgen receptor level and transcriptional activity. J Neurobiol, 57(2), 163-171.

Lubin, A., Gieseking, C. F., & Williams, H. L. (1962). Direct measurement of cognitive deficit in schizophrenia. J Consult Psychol, 26, 139-143.

Lustig, R. H. (1994). Sex hormone modulation of neural development in vitro. Horm Behav, 28(4), 383-395.

MacGregor, J. I., & Jordan, V. C. (1998). Basic guide to the mechanisms of antiestrogen action. Pharmacol Rev, 50(2), 151-196.

Majewska, M. D. (1992). Neurosteroids: endogenous bimodal modulators of the GABAA receptor. Mechanism of action and physiological significance. Prog Neurobiol, 38(4), 379-395.

Mamah, D., Wang, L., de Erausquin, D. B. G. A., Gado, M., & Csernansky, J. G. (2007). Structural Analysis of the Basal Ganglia in Schizophrenia. Schizophr Res, 89(1- 3), 59-71.

Mandal, M. K., Pandey, R., & Prasad, A. B. (1998). Facial expressions of emotions and schizophrenia: A review. Schizophrenia Bulletin, 24(3), 399-412.

Maninger, N., Wolkowitz, O. M., Reus, V. I., Epel, E. S., & Mellon, S. H. (2009). Neurobiological and neuropsychiatric effects of dehydroepiandrosterone (DHEA) and DHEA sulfate (DHEAS). Frontiers in Neuroendocrinology, 30(1), 65-91.

Mao, X., & Barger, S. W. (1998). Neuroprotection by dehydroepiandrosterone-sulfate: role of an NFkappaB-like factor. Neuroreport, 9(4), 759-763. 177

Markham, J. A. (2012). Sex steroids and schizophrenia. Rev Endocr Metab Disord, 13(3), 187-207.

Marx, C. E., Jarskog, L. F., Lauder, J. M., Gilmore, J. H., Lieberman, J. A., & Morrow, A. L. (2000). Neurosteroid modulation of embryonic neuronal survival in vitro following anoxia. Brain Res, 871(1), 104-112.

Mathis, C., Paul, S. M., & Crawley, J. N. (1994). The neurosteroid pregnenolone sulfate blocks NMDA antagonist-induced deficits in a passive avoidance memory task. Psychopharmacology (Berl), 116(2), 201-206.

Matsuda, Y., Hirano, H., & Watanabe, Y. (2002). Effects of estrogen on acetylcholine release in frontal cortex of female rats: involvement of serotonergic neuronal systems. Brain Res, 937(1–2), 58-65.

Matsumoto, A., Arai, Y., & Osanai, M. (1985). Estrogen stimulates neuronal plasticity in the deafferented hypothalamic arcuate nucleus in aged female rats. Neurosci Res, 2(5), 412-418.

Maximov, P. Y., Lee, T. M., & Jordan, V. C. (2013). The discovery and development of selective estrogen receptor modulators (SERMs) for clinical practice. Curr Clin Pharmacol, 8(2), 135-155.

McEwen, B. S. (2006). Protective and damaging effects of stress mediators: central role of the brain. Dialogues in Clinical Neuroscience, 8(4), 367-381.

McEwen, B. S., Alves, S. E., Bulloch, K., & Weiland, N. G. (1997). Ovarian steroids and the brain: implications for cognition and aging. Neurology, 48(5 Suppl 7), S8-15.

McGlashan, T. H. (2011). Eugen Bleuler: centennial anniversary of his 1911 publication of Dementia Praecox or the group of schizophrenias. Schizophr Bull, 37(6), 1101-1103.

McGrath, J., Saha, S., Welham, J., El Saadi, O., MacCauley, C., & Chant, D. (2004). A systematic review of the incidence of schizophrenia: the distribution of rates and the influence of sex, urbanicity, migrant status and methodology. BMC Med, 2, 13.

McNelis, J. C., Manolopoulos, K. N., Gathercole, L. L., Bujalska, I. J., Stewart, P. M., Tomlinson, J. W., & Arlt, W. (2013). Dehydroepiandrosterone exerts antiglucocorticoid action on human preadipocyte proliferation, differentiation, and glucose uptake. Am J Physiol Endocrinol Metab, 305(9), E1134-1144.

Meaney, M. J., Diorio, J., Francis, D., Widdowson, J., LaPlante, P., Caldji, C., . . . Plotsky, P. M. (1996). Early environmental regulation of forebrain glucocorticoid receptor gene expression: implications for adrenocortical responses to stress. Dev Neurosci, 18(1-2), 49-72.

Meisenzahl, E. M., Rujescu, D., Kirner, A., Giegling, I., Kathmann, N., Leinsinger, G., . . . Moller, H. J. (2001). Association of an interleukin-1beta genetic 178

polymorphism with altered brain structure in patients with schizophrenia. Am J Psychiatry, 158(8), 1316-1319.

Meltzer, H. Y., Lee, M. A., & Jayathilake, K. (2001). The blunted plasma cortisol response to apomorphine and its relationship to treatment response in patients with schizophrenia. Neuropsychopharmacology, 24(3), 278-290.

Mendrek, A., Bourque, J., Dubé, A., Lakis, N., & Champagne, J. (2012). Emotion Processing in Women with Schizophrenia Is Menstrual Cycle Phase and Affective Valence Dependent: An fMRI Study. ISRN Psychiatry, 2012, 13.

Mendrek, A., Lakis, N., & Jimenez, J. (2011). Associations of sex steroid hormones with cerebral activations during mental rotation in men and women with schizophrenia. Psychoneuroendocrinology, 36(9), 1422-1426.

Menozzi, R., Cagnacci, A., Zanni, A. L., Bondi, M., Volpe, A., & Del Rio, G. (2000). Sympathoadrenal response of postmenopausal women prior and during prolonged administration of estradiol. Maturitas, 34(3), 275-281.

Merritt, P., Stangl, B., Hirshman, E., & Verbalis, J. (2012). Administration of dehydroepiandrosterone (DHEA) increases serum levels of androgens and estrogens but does not enhance short-term memory in post-menopausal women. Brain Res, 1483, 54-62.

Mesiano, S., & Jaffe, R. B. (1997). Developmental and functional biology of the primate fetal adrenal cortex. Endocr Rev, 18(3), 378-403.

Michel, C. M., Murray, M. M., Lantz, G., Gonzalez, S., Spinelli, L., & Grave de Peralta, R. (2004). EEG source imaging. Clin Neurophysiol, 115(10), 2195-2222.

Miodownik, C., Maayan, R., Ratner, Y., Lerner, V., Pintov, L., Mar, M., . . . Ritsner, M. S. (2011). Serum levels of brain-derived neurotrophic factor and cortisol to sulfate of dehydroepiandrosterone molar ratio associated with clinical response to L-theanine as augmentation of antipsychotic therapy in schizophrenia and schizoaffective disorder patients. Clin Neuropharmacol, 34(4), 155-160.

Mizoguchi, K., Ishige, A., Aburada, M., & Tabira, T. (2003). Chronic stress attenuates glucocorticoid negative feedback: involvement of the prefrontal cortex and hippocampus. Neuroscience, 119(3), 887-897.

Mocking, R. J., Pellikaan, C. M., Lok, A., Assies, J., Ruhe, H. G., Koeter, M. W., . . . Schene, A. H. (2015). DHEAS and cortisol/DHEAS-ratio in recurrent depression: State, or trait predicting 10-year recurrence? Psychoneuroendocrinology, 59, 91-101.

Mondelli, V., Cattaneo, A., Belvederi Murri, M., Di Forti, M., Handley, R., Hepgul, N., . . . Pariante, C. M. (2011). Stress and inflammation reduce brain-derived neurotrophic factor expression in first-episode psychosis: a pathway to smaller hippocampal volume. J Clin Psychiatry, 72(12), 1677-1684. 179

Mondelli, V., Pariante, C. M., Navari, S., Aas, M., D'Albenzio, A., Di Forti, M., . . . Dazzan, P. (2010). Higher cortisol levels are associated with smaller left hippocampal volume in first-episode psychosis. Schizophr Res, 119(1-3), 75-78.

Montague, D., Weickert, C. S., Tomaskovic-Crook, E., Rothmond, D. A., Kleinman, J. E., & Rubinow, D. R. (2008). Oestrogen receptor alpha localisation in the prefrontal cortex of three mammalian species. J Neuroendocrinol, 20(7), 893- 903.

Moore, L., Kyaw, M., Vercammen, A., Lenroot, R., Kulkarni, J., Curtis, J., . . . Weickert, T. W. (2013). Serum testosterone levels are related to cognitive function in men with schizophrenia. Psychoneuroendocrinology, 38(9), 1717- 1728.

Morales, A. J., Nolan, J. J., Nelson, J. C., & Yen, S. S. (1994). Effects of replacement dose of dehydroepiandrosterone in men and women of advancing age. J Clin Endocrinol Metab, 78(6), 1360-1367.

Morris, R. W., Purves-Tyson, T. D., Weickert, C. S., Rothmond, D., Lenroot, R., & Weickert, T. W. (2015). Testosterone and reward prediction-errors in healthy men and men with schizophrenia. Schizophr Res, 168(3), 649-660.

Moses, E. L., Drevets, W. C., Smith, G., Mathis, C. A., Kalro, B. N., Butters, M. A., . . . Berga, S. L. (2000). Effects of estradiol and progesterone administration on human serotonin 2A receptor binding: a PET study. Biol Psychiatry, 48(8), 854- 860.

Mothersill, O., & Donohoe, G. (2016). Neural effects of social environmental stress - an activation likelihood estimation meta-analysis. Psychol Med, 1-9.

Muck-Seler, D., Pivac, N., Jakovljevic, M., & Brzovic, Z. (1999). Platelet serotonin, plasma cortisol, and dexamethasone suppression test in schizophrenic patients. Biol Psychiatry, 45(11), 1433-1439.

Mulderink, T. A., Gitelman, D. R., Mesulam, M. M., & Parrish, T. B. (2002). On the use of caffeine as a contrast booster for BOLD fMRI studies. Neuroimage, 15(1), 37-44.

Muller, M., den Tonkelaar, I., Thijssen, J. H., Grobbee, D. E., & van der Schouw, Y. T. (2003). Endogenous sex hormones in men aged 40-80 years. Eur J Endocrinol, 149(6), 583-589.

Müller, N. G., Strumpf, H., Scholz, M., Baier, B., & Melloni, L. (2012). Repetition Suppression versus Enhancement—It's Quantity That Matters. Cerebral Cortex.

Myers, R. E., Anderson, L. I., & Dluzen, D. E. (2003). Estrogen, but not testosterone, attenuates methamphetamine-evoked dopamine output from superfused striatal tissue of female and male mice. Neuropharmacology, 44(5), 624-632.

Myin-Germeys, I., & van Os, J. (2007). Stress-reactivity in psychosis: evidence for an affective pathway to psychosis. Clin Psychol Rev, 27(4), 409-424. 180

Nagatsu, T., Levitt, M., & Udenfriend, S. (1964). Tyrosine Hydroxylase. The Initial Step in Norepinephrine Biosynthesis. J Biol Chem, 239, 2910-2917.

Nechmad, A., Maayan, R., Ramadan, E., Morad, O., Poyurovsky, M., & Weizman, A. (2003). Clozapine decreases rat brain dehydroepiandrosterone and dehydroepiandrosterone sulfate levels. Eur Neuropsychopharmacol, 13(1), 29- 31.

Neckelmann, G., Specht, K., Lund, A., Ersland, L., Smievoll, A. I., Neckelmann, D., & Hugdahl, K. (2006). Mr morphometry analysis of grey matter volume reduction in schizophrenia: association with hallucinations. Int J Neurosci, 116(1), 9-23.

Neil, J. J., Shiran, S. I., McKinstry, R. C., Schefft, G. L., Snyder, A. Z., Almli, C. R., . . . Conturo, T. E. (1998). Normal brain in human newborns: apparent diffusion coefficient and diffusion anisotropy measured by using diffusion tensor MR imaging. Radiology, 209(1), 57-66.

Nelson, M. D., Saykin, A. J., Flashman, L. A., & Riordan, H. J. (1998). Hippocampal volume reduction in schizophrenia as assessed by magnetic resonance imaging: a meta-analytic study. Arch Gen Psychiatry, 55(5), 433-440.

Nguyen, T.-V., McCracken, J. T., Ducharme, S., Cropp, B. F., Botteron, K. N., Evans, A. C., & Karama, S. (2013). Interactive Effects of Dehydroepiandrosterone and Testosterone on Cortical Thickness during Early Brain Development. The Journal of Neuroscience, 33(26), 10840-10848.

Nishioka, N., & Arnold, S. E. (2004). Evidence for oxidative DNA damage in the hippocampus of elderly patients with chronic schizophrenia. Am J Geriatr Psychiatry, 12(2), 167-175.

Nuechterlein, K. H., Dawson, M. E., Ventura, J., Gitlin, M., Subotnik, K. L., Snyder, K. S., . . . Bartzokis, G. (1994). The vulnerability/stress model of schizophrenic relapse: a longitudinal study. Acta Psychiatr Scand Suppl, 382, 58-64. o Hartaigh, B., Loerbroks, A., Thomas, G. N., Engeland, C. G., Hollands, M. A., Fischer, J. E., & Bosch, J. A. (2012). Age-dependent and -independent associations between depression, anxiety, DHEAS, and cortisol: from the MIPH Industrial Cohort Studies (MICS). Psychoneuroendocrinology, 37(7), 929-936.

Ochoa, S., Usall, J., Cobo, J., Labad, X., & Kulkarni, J. (2012). Gender differences in schizophrenia and first-episode psychosis: a comprehensive literature review. Schizophr Res Treatment, 2012, 916198.

Okasha, A., & Okasha, T. (2000). Notes on mental disorders in Pharaonic Egypt. History of Psychiatry, 11(44), 413-424.

Olney, J. W., Newcomer, J. W., & Farber, N. B. (1999). NMDA receptor hypofunction model of schizophrenia. J Psychiatr Res, 33(6), 523-533.

Olsen, L., Rasmussen, H. B., Hansen, T., Bagger, Y. Z., Tanko, L. B., Qin, G., . . . Werge, T. (2006). Estrogen receptor alpha and risk for cognitive impairment in postmenopausal women. Psychiatr Genet, 16(2), 85-88. 181

Osterlund, M. K., Halldin, C., & Hurd, Y. L. (2000). Effects of chronic 17beta-estradiol treatment on the serotonin 5-HT(1A) receptor mRNA and binding levels in the rat brain. Synapse, 35(1), 39-44.

Palmer, B. W., Heaton, R. K., Paulsen, J. S., Kuck, J., Braff, D., Harris, M. J., . . . Jeste, D. V. (1997). Is it possible to be schizophrenic yet neuropsychologically normal? Neuropsychology, 11(3), 437-446.

Papanicolaou, A. C., Simos, P. G., Castillo, E. M., Breier, J. I., Katz, J. S., & Wright, A. A. (2002). The Hippocampus and Memory of Verbal and Pictorial Material. Learning & Memory, 9(3), 99-104.

Papasozomenos, S. C. (1997). The heat shock-induced hyperphosphorylation of tau is estrogen-independent and prevented by androgens: implications for Alzheimer disease. Proc Natl Acad Sci U S A, 94(13), 6612-6617.

Parker, L. N. (1991). Control of adrenal androgen secretion. Endocrinol Metab Clin North Am, 20(2), 401-421.

Pasqualini, C., Olivier, V., Guibert, B., Frain, O., & Leviel, V. (1995). Acute stimulatory effect of estradiol on striatal dopamine synthesis. J Neurochem, 65(4), 1651-1657.

Patel, P. D., Katz, M., Karssen, A. M., & Lyons, D. M. (2008). Stress-induced changes in corticosteroid receptor expression in primate hippocampus and prefrontal cortex. Psychoneuroendocrinology, 33(3), 360-367.

Peeters, S. C. T., van Bronswijk, S., van de Ven, V., Gronenschild, E. H. B. M., Goebel, R., van Os, J., & Marcelis, M. (2015). Cognitive correlates of frontoparietal network connectivity ‘at rest’ in individuals with differential risk for psychotic disorder. European Neuropsychopharmacology.

Penn, D., Roberts, D. L., Munt, E. D., Silverstein, E., Jones, N., & Sheitman, B. (2005). A pilot study of social cognition and interaction training (SCIT) for schizophrenia. Schizophr Res, 80(2), 357-359.

Perälä, J., Suvisaari, J., Saarni, S. I., & et al. (2007). LIfetime prevalence of psychotic and bipolar i disorders in a general population. Arch Gen Psychiatry, 64(1), 19- 28.

Peralta, V., & Cuesta, M. J. (2001). How many and which are the psychopathological dimensions in schizophrenia? Issues influencing their ascertainment. Schizophr Res, 49(3), 269-285.

Pérez-Neri, I., Montes, S., Ojeda-López, C., Ramírez-Bermúdez, J., & Ríos, C. (2008). Modulation of neurotransmitter systems by dehydroepiandrosterone and dehydroepiandrosterone sulfate: Mechanism of action and relevance to psychiatric disorders. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 32(5), 1118-1130.

Perlman, W. R., Matsumoto, M., Beltaifa, S., Hyde, T. M., Saunders, R. C., Webster, M. J., . . . Weickert, C. S. (2005). Expression of estrogen receptor alpha exon- 182

deleted mRNA variants in the human and non-human primate frontal cortex. Neuroscience, 134(1), 81-95.

Perlman, W. R., Tomaskovic-Crook, E., Montague, D. M., Webster, M. J., Rubinow, D. R., Kleinman, J. E., & Weickert, C. S. (2005). Alteration in estrogen receptor alpha mRNA levels in frontal cortex and hippocampus of patients with major mental illness. Biol Psychiatry, 58(10), 812-824.

Perlman, W. R., Webster, M. J., Kleinman, J. E., & Weickert, C. S. (2004). Reduced glucocorticoid and estrogen receptor alpha messenger ribonucleic acid levels in the amygdala of patients with major mental illness. Biol Psychiatry, 56(11), 844- 852.

Petrides, M. (1995). Impairments on nonspatial self-ordered and externally ordered working memory tasks after lesions of the mid-dorsal part of the lateral frontal cortex in the monkey. J Neurosci, 15(1 Pt 1), 359-375.

Peyroux, E., & Franck, N. (2014). RC2S: A Cognitive Remediation Program to Improve Social Cognition in Schizophrenia and Related Disorders. Frontiers in Human Neuroscience, 8, 400.

Phelps, E. A. (2004). Human emotion and memory: interactions of the amygdala and hippocampal complex. Curr Opin Neurobiol, 14(2), 198-202.

Phillips, M. L., Williams, L., Senior, C., Bullmore, E. T., Brammer, M. J., Andrew, C., . . . David, A. S. (1999). A differential neural response to threatening and non-threatening negative facial expressions in paranoid and non-paranoid schizophrenics. Psychiatry Research: Neuroimaging, 92(1), 11-31.

Pope, H. G., Jr., & Katz, D. L. (1988). Affective and psychotic symptoms associated with anabolic steroid use. Am J Psychiatry, 145(4), 487-490.

Prasad, K. M., Upton, C. H., Nimgaonkar, V. L., & Keshavan, M. S. (2015). Differential susceptibility of white matter tracts to inflammatory mediators in schizophrenia: an integrated DTI study. Schizophr Res, 161(1), 119-125.

Purves-Tyson, T. D., Boerrigter, D., Allen, K., Zavitsanou, K., Karl, T., Djunaidi, V., . . . Weickert, C. S. (2015). Testosterone attenuates and the selective estrogen receptor modulator, raloxifene, potentiates amphetamine-induced locomotion in male rats. Horm Behav, 70, 73-84.

Purves-Tyson, T. D., Handelsman, D. J., Double, K. L., Owens, S. J., Bustamante, S., & Weickert, C. S. (2012). Testosterone regulation of sex steroid-related mRNAs and dopamine-related mRNAs in adolescent male rat substantia nigra. BMC Neuroscience, 13(1), 1-12.

Purves-Tyson, T. D., Owens, S. J., Double, K. L., Desai, R., Handelsman, D. J., & Weickert, C. S. (2014). Testosterone induces molecular changes in dopamine signaling pathway molecules in the adolescent male rat nigrostriatal pathway. PLoS ONE, 9(3), e91151. 183

Rahmani, A., Shoae-Hassani, A., Keyhanvar, P., Kheradmand, D., & Darbandi-Azar, A. (2013). Dehydroepiandrosterone Stimulates Nerve Growth Factor and Brain Derived Neurotrophic Factor in Cortical Neurons. Advances in Pharmacological Sciences, 2013, 7.

Ramsey, J. M., Schwarz, E., Guest, P. C., van Beveren, N. J. M., Leweke, F. M., Rothermundt, M., . . . Bahn, S. (2013). Distinct Molecular Phenotypes in Male and Female Schizophrenia Patients. PLoS ONE, 8(11), e78729.

Reichenberg, A., Harvey, P. D., Bowie, C. R., Mojtabai, R., Rabinowitz, J., Heaton, R. K., & Bromet, E. (2009). Neuropsychological Function and Dysfunction in Schizophrenia and Psychotic Affective Disorders. Schizophrenia Bulletin, 35(5), 1022-1029.

Resnick, S. M., & Maki, P. M. (2001). Effects of Hormone Replacement Therapy on Cognitive and Brain Aging. Ann N Y Acad Sci, 949(1), 203-214.

Riecher-Rössler, A., Häfner, H., Maurer, K., Stummbaum, M., & Schmidt, R. (1992). Schizophrenic symptomatology varies with serum estradiol levels during menstrual cycle. Schizophr Res, 6(2), 114-115.

Riecher-Rossler, A., Hafner, H., Stumbaum, M., Maurer, K., & Schmidt, R. (1994). Can estradiol modulate schizophrenic symptomatology? Schizophr Bull, 20(1), 203- 214.

Riecher-Rossler, A., & Kulkarni, J. (2011). Estrogens and gonadal function in schizophrenia and related psychoses. Curr Top Behav Neurosci, 8, 155-171.

Ritsner, M., Gibel, A., Ram, E., Maayan, R., & Weizman, A. (2006). Alterations in DHEA metabolism in schizophrenia: two-month case-control study. Eur Neuropsychopharmacol, 16(2), 137-146.

Ritsner, M., Maayan, R., Gibel, A., Strous, R. D., Modai, I., & Weizman, A. (2004). Elevation of the cortisol/dehydroepiandrosterone ratio in schizophrenia patients. European Neuropsychopharmacology, 14(4), 267-273.

Ritsner, M. S., Gibel, A., Ratner, Y., Tsinovoy, G., & Strous, R. D. (2006). Improvement of sustained attention and visual and movement skills, but not clinical symptoms, after dehydroepiandrosterone augmentation in schizophrenia: a randomized, double-blind, placebo-controlled, crossover trial. J Clin Psychopharmacol, 26(5), 495-499.

Robinson, D., Woerner, M. G., Alvir, J. M., Bilder, R., Goldman, R., Geisler, S., . . . Lieberman, J. A. (1999). Predictors of relapse following response from a first episode of schizophrenia or schizoaffective disorder. Arch Gen Psychiatry, 56(3), 241-247.

Romeo, E., Cheney, D. L., Zivkovic, I., Costa, E., & Guidotti, A. (1994). Mitochondrial diazepam-binding inhibitor receptor complex agonists antagonize dizocilpine amnesia: putative role for allopregnanolone. J Pharmacol Exp Ther, 270(1), 89- 96. 184

Rosenheck, R., Leslie, D., Keefe, R., McEvoy, J., Swartz, M., Perkins, D., . . . Lieberman, J. (2006). Barriers to employment for people with schizophrenia. Am J Psychiatry, 163(3), 411-417.

Rosner, W. (1991). Plasma steroid-binding proteins. Endocrinol Metab Clin North Am, 20(4), 697-720.

Rossato, J. I., Bevilaqua, L. R. M., Izquierdo, I., Medina, J. H., & Cammarota, M. (2009). Dopamine Controls Persistence of Long-Term Memory Storage. Science, 325(5943), 1017-1020.

Roy, C. S., & Sherrington, C. S. (1890). On the Regulation of the Blood-supply of the Brain. The Journal of Physiology, 11(1-2), 85-158.117.

Ruiz Feliu, M., & Campos Mangas, M. C. (2012). P-890 - Acute psychosis induced by anabolic steroids. European Psychiatry, 27, Supplement 1, 1.

Russell, T. A., Rubia, K., Bullmore, E. T., Soni, W., Suckling, J., Brammer, M. J., . . . Sharma, T. (2000). Exploring the social brain in schizophrenia: left prefrontal underactivation during mental state attribution. Am J Psychiatry, 157(12), 2040- 2042.

Safiulina, D., Peet, N., Seppet, E., Zharkovsky, A., & Kaasik, A. (2006). Dehydroepiandrosterone inhibits complex I of the mitochondrial respiratory chain and is neurotoxic in vitro and in vivo at high concentrations. Toxicol Sci, 93(2), 348-356.

Saha, S., Chant, D., & McGrath, J. (2007). A systematic review of mortality in schizophrenia: is the differential mortality gap worsening over time? Arch Gen Psychiatry, 64(10), 1123-1131.

Saha, S., Chant, D., Welham, J., & McGrath, J. (2005). A systematic review of the prevalence of schizophrenia. PLoS Med, 2(5), e141.

Sajikumar, S., & Frey, J. U. (2004). Late-associativity, synaptic tagging, and the role of dopamine during LTP and LTD. Neurobiol Learn Mem, 82(1), 12-25.

Sakr, H. F., Khalil, K. I., Hussein, A. M., Zaki, M. S., Eid, R. A., & Alkhateeb, M. (2014). Effect of dehydroepiandrosterone (DHEA) on memory and brain derived neurotrophic factor (BDNF) in a rat model of vascular dementia. J Physiol Pharmacol, 65(1), 41-53.

Salokangas, R. K. (1983). Prognostic implications of the sex of schizophrenic patients. The British Journal of Psychiatry, 142(2), 145-151.

Salvatore, P., Ghidini, S., Zita, G., De Panfilis, C., Lambertino, S., Maggini, C., & Baldessarini, R. J. (2008). Circadian activity rhythm abnormalities in ill and recovered bipolar I disorder patients. Bipolar Disord, 10(2), 256-265.

Sanchez, M. G., Bourque, M., Morissette, M., & Di Paolo, T. (2010). Steroids- dopamine interactions in the pathophysiology and treatment of CNS disorders. CNS Neurosci Ther, 16(3), e43-71. 185

Sánchez, M. G., Bourque, M., Morissette, M., & Di Paolo, T. (2010). Steroids- Dopamine Interactions in the Pathophysiology and Treatment of CNS Disorders. CNS Neurosci Ther, 16(3), e43-e71.

Sapolsky, R. M. (1985). A mechanism for glucocorticoid toxicity in the hippocampus: increased neuronal vulnerability to metabolic insults. J Neurosci, 5(5), 1228- 1232.

Sapolsky, R. M., Krey, L. C., & McEwen, B. S. (1984). Stress down-regulates corticosterone receptors in a site-specific manner in the brain. Endocrinology, 114(1), 287-292.

Sapolsky, R. M., Krey, L. C., & McEwen, B. S. (1985). Prolonged glucocorticoid exposure reduces hippocampal neuron number: implications for aging. J Neurosci, 5(5), 1222-1227.

Sapolsky, R. M., Uno, H., Rebert, C. S., & Finch, C. E. (1990). Hippocampal damage associated with prolonged glucocorticoid exposure in primates. J Neurosci, 10(9), 2897-2902.

Savla, G. N., Vella, L., Armstrong, C. C., Penn, D. L., & Twamley, E. W. (2012). Deficits in Domains of Social Cognition in Schizophrenia: A Meta-Analysis of the Empirical Evidence. Schizophrenia Bulletin.

Sawada, H., Ibi, M., Kihara, T., Urushitani, M., Akaike, A., & Shimohama, S. (1998). Estradiol protects mesencephalic dopaminergic neurons from oxidative stress- induced neuronal death. J Neurosci Res, 54(5), 707-719.

Saykin, A. J., Gur, R. C., Gur, R. E., Mozley, P. D., Mozley, L. H., Resnick, S. M., . . . Stafiniak, P. (1991). Neuropsychological function in schizophrenia. Selective impairment in memory and learning. Arch Gen Psychiatry, 48(7), 618-624.

Schneider, F., Gur, R. C., Koch, K., Backes, V., Amunts, K., Shah, N. J., . . . Habel, U. (2006). Impairment in the specificity of emotion processing in schizophrenia. Am J Psychiatry, 163(3), 442-447.

Schreier, A., Wolke, D., Thomas, K., Horwood, J., Hollis, C., Gunnell, D., . . . Harrison, G. (2009). Prospective study of peer victimization in childhood and psychotic symptoms in a nonclinical population at age 12 years. Arch Gen Psychiatry, 66(5), 527-536.

Schwartzman, A. E., & Douglas, V. I. (1962). Intellectual loss in schizophrenia. I. Can J Psychol, 16, 1-10.

Seeman, M. V., & Lang, M. (1990). The role of estrogens in schizophrenia gender differences. Schizophrenia Bulletin, 16(2), 185-194.

Seeman, P., Lee, T., Chau-Wong, M., & Wong, K. (1976). Antipsychotic drug doses and neuroleptic/dopamine receptors. Nature, 261(5562), 717-719.

Seidman, L. J., Faraone, S. V., Goldstein, J. M., Goodman, J. M., Kremen, W. S., Matsuda, G., . . . Tsuang, M. T. (1997). Reduced subcortical brain volumes in 186

nonpsychotic siblings of schizophrenic patients: A pilot magnetic resonance imaging study. American Journal of Medical Genetics, 74(5), 507-514.

Selemon, L. D., Kleinman, J. E., Herman, M. M., & Goldman-Rakic, P. S. (2002). Smaller frontal gray matter volume in postmortem schizophrenic brains. Am J Psychiatry, 159(12), 1983-1991.

Selemon, L. D., Mrzljak, J., Kleinman, J. E., Herman, M. M., & Goldman-Rakic, P. S. (2003). Regional specificity in the neuropathologic substrates of schizophrenia: a morphometric analysis of Broca's area 44 and area 9. Arch Gen Psychiatry, 60(1), 69-77.

Selemon, L. D., Rajkowska, G., & Goldman-Rakic, P. S. (1998). Elevated neuronal density in prefrontal area 46 in brains from schizophrenic patients: application of a three-dimensional, stereologic counting method. J Comp Neurol, 392(3), 402- 412.

Senanarong, V., Vannasaeng, S., Poungvarin, N., Ploybutr, S., Udompunthurak, S., Jamjumras, P., . . . Cummings, J. L. (2002). Endogenous estradiol in elderly individuals: cognitive and noncognitive associations. Arch Neurol, 59(3), 385- 389.

Shapleske, J., Rossell, S. L., Chitnis, X. A., Suckling, J., Simmons, A., Bullmore, E. T., . . . David, A. S. (2002). A computational morphometric MRI study of schizophrenia: effects of hallucinations. Cereb Cortex, 12(12), 1331-1341.

Shaywitz, S. E., Shaywitz, B. A., Pugh, K. R., Fulbright, R. K., Skudlarski, P., Mencl, W. E., . . . Gore, J. C. (1999). Effect of estrogen on brain activation patterns in postmenopausal women during working memory tasks. Jama, 281(13), 1197- 1202.

Sheline, Y. I., Sanghavi, M., Mintun, M. A., & Gado, M. H. (1999). Depression Duration But Not Age Predicts Hippocampal Volume Loss in Medically Healthy Women with Recurrent Major Depression. The Journal of Neuroscience, 19(12), 5034-5043.

Shirtcliff, E. A., Granger, D. A., Schwartz, E., & Curran, M. J. (2001). Use of salivary biomarkers in biobehavioral research: cotton-based sample collection methods can interfere with salivary immunoassay results. Psychoneuroendocrinology, 26(2), 165-173.

Shtasel, D. L., Gur, R. E., Gallacher, F., Heimberg, C., & Gur, R. C. (1992). Gender differences in the clinical expression of schizophrenia. Schizophr Res, 7(3), 225- 231.

Sigmundsson, T., Suckling, J., Maier, M., Williams, S., Bullmore, E., Greenwood, K., . . . Toone, B. (2001). Structural abnormalities in frontal, temporal, and limbic regions and interconnecting white matter tracts in schizophrenic patients with prominent negative symptoms. Am J Psychiatry, 158(2), 234-243. 187

Silver, H., Knoll, G., Isakov, V., Goodman, C., & Finkelstein, Y. (2005). Blood DHEAS concentrations correlate with cognitive function in chronic schizophrenia patients: a pilot study. J Psychiatr Res, 39(6), 569-575.

Sinclair, D., Fullerton, J. M., Webster, M. J., & Shannon Weickert, C. (2012). Glucocorticoid receptor 1B and 1C mRNA transcript alterations in schizophrenia and bipolar disorder, and their possible regulation by GR gene variants. PLoS ONE, 7(3), e31720.

Sinclair, D., Purves-Tyson, T. D., Allen, K. M., & Weickert, C. S. (2014). Impacts of stress and sex hormones on dopamine neurotransmission in the adolescent brain. Psychopharmacology (Berl), 231(8), 1581-1599.

Sinclair, D., Tsai, S. Y., Woon, H. G., & Weickert, C. S. (2011). Abnormal glucocorticoid receptor mRNA and protein isoform expression in the prefrontal cortex in psychiatric illness. Neuropsychopharmacology, 36(13), 2698-2709.

Sinclair, D., Webster, M. J., Fullerton, J. M., & Weickert, C. S. (2012). Glucocorticoid receptor mRNA and protein isoform alterations in the orbitofrontal cortex in schizophrenia and bipolar disorder. BMC Psychiatry, 12(1), 1-13.

Smith, S., Wheeler, M. J., Murray, R., & O'Keane, V. (2002). The effects of antipsychotic-induced hyperprolactinaemia on the hypothalamic-pituitary- gonadal axis. J Clin Psychopharmacol, 22(2), 109-114.

Song, S. K., Sun, S. W., Ju, W. K., Lin, S. J., Cross, A. H., & Neufeld, A. H. (2003). Diffusion tensor imaging detects and differentiates axon and myelin degeneration in mouse optic nerve after retinal ischemia. Neuroimage, 20(3), 1714-1722.

Sousa, A., & Ticku, M. K. (1997). Interactions of the neurosteroid dehydroepiandrosterone sulfate with the GABA(A) receptor complex reveals that it may act via the picrotoxin site. J Pharmacol Exp Ther, 282(2), 827-833.

Squire, L. R., Ojemann, J. G., Miezin, F. M., Petersen, S. E., Videen, T. O., & Raichle, M. E. (1992). Activation of the hippocampus in normal humans: a functional anatomical study of memory. Proc Natl Acad Sci U S A, 89(5), 1837-1841.

Stangl, B., Hirshman, E., & Verbalis, J. (2011). Administration of Dehydroepiandrosterone (DHEA) Enhances Visual-Spatial Performance in Post- Menopausal Women. Behav Neurosci, 125(5), 742-752.

Stanley D. Rosenberg , P. D., Weili Lu , P. D., Kim T. Mueser , P. D., Mary Kay Jankowski , P. D., & Francine Cournos , M. D. (2007). Correlates of Adverse Childhood Events Among Adults With Schizophrenia Spectrum Disorders. Psychiatric Services, 58(2), 245-253.

Starkman, M. N., Gebarski, S. S., Berent, S., & Schteingart, D. E. (1992). Hippocampal formation volume, memory dysfunction, and cortisol levels in patients with Cushing's syndrome. Biological Psychiatry, 32(9), 756-765. 188

Stevens, A. A., Goldman-Rakic, P. S., Gore, J. C., Fulbright, R. K., & Wexler, B. E. (1998). Cortical dysfunction in schizophrenia during auditory word and tone working memory demonstrated by functional magnetic resonance imaging. Arch Gen Psychiatry, 55(12), 1097-1103.

Stoop, R., & Poo, M. M. (1996). Synaptic modulation by neurotrophic factors: differential and synergistic effects of brain-derived neurotrophic factor and ciliary neurotrophic factor. J Neurosci, 16(10), 3256-3264.

Straub, R. H., Konecna, L., Hrach, S., Rothe, G., Kreutz, M., Scholmerich, J., . . . Lang, B. (1998). Serum dehydroepiandrosterone (DHEA) and DHEA sulfate are negatively correlated with serum interleukin-6 (IL-6), and DHEA inhibits IL-6 secretion from mononuclear cells in man in vitro: possible link between endocrinosenescence and immunosenescence. J Clin Endocrinol Metab, 83(6), 2012-2017.

Strous, R. D., Maayan, R., Lapidus, R., Goredetsky, L., Zeldich, E., Kotler, M., & Weizman, A. (2004). Increased circulatory dehydroepiandrosterone and dehydroepiandrosterone-sulphate in first-episode schizophrenia: relationship to gender, aggression and symptomatology. Schizophr Res, 71(2-3), 427-434.

Strous, R. D., Maayan, R., Lapidus, R., Stryjer, R., Lustig, M., Kotler, M., & Weizman, A. (2003). Dehydroepiandrosterone augmentation in the management of negative, depressive, and anxiety symptoms in schizophrenia. Arch Gen Psychiatry, 60(2), 133-141.

Strous, R. D., Stryjer, R., Maayan, R., Gal, G., Viglin, D., Katz, E., . . . Weizman, A. (2007). Analysis of clinical symptomatology, extrapyramidal symptoms and neurocognitive dysfunction following dehydroepiandrosterone (DHEA) administration in olanzapine treated schizophrenia patients: a randomized, double-blind placebo controlled trial. Psychoneuroendocrinology, 32(2), 96-105.

Styne, D. M. (1994). Physiology of puberty. Horm Res, 41 Suppl 2, 3-6.

Sulcova, J., Hill, M., Hampl, R., & Starka, L. (1997). Age and sex related differences in serum levels of unconjugated dehydroepiandrosterone and its sulphate in normal subjects. J Endocrinol, 154(1), 57-62.

Sumner, B. E. H., & Fink, G. (1998). Testosterone as well as estrogen increases serotonin2A receptor mRNA and binding site densities in the male rat brain. Molecular Brain Research, 59(2), 205-214.

Suslow, T., Lindner, C., Dannlowski, U., Walhofer, K., Rodiger, M., Maisch, B., . . . Kugel, H. (2013). Automatic amygdala response to facial expression in schizophrenia: initial hyperresponsivity followed by hyporesponsivity. BMC Neurosci, 14, 140.

Suuronen, T., Nuutinen, T., Huuskonen, J., Ojala, J., Thornell, A., & Salminen, A. (2005). Anti-inflammatory effect of selective estrogen receptor modulators (SERMs) in microglial cells. Inflammation Research, 54(5), 194-203. 189

Taherianfard, M., & Shariaty, M. (2004). Evaluation of serum steroid hormones in schizophrenic patients. Indian J Med Sci, 58(1), 3-9.

Takahashi, H., Koeda, M., Oda, K., Matsuda, T., Matsushima, E., Matsuura, M., . . . Okubo, Y. (2004). An fMRI study of differential neural response to affective pictures in schizophrenia. Neuroimage, 22(3), 1247-1254.

Takahashi, M., Hackney, D. B., Zhang, G., Wehrli, S. L., Wright, A. C., O'Brien, W. T., . . . Selzer, M. E. (2002). Magnetic resonance microimaging of intraaxonal water diffusion in live excised lamprey spinal cord. Proc Natl Acad Sci U S A, 99(25), 16192-16196.

Takahashi, M., Ono, J., Harada, K., Maeda, M., & Hackney, D. B. (2000). Diffusional anisotropy in cranial nerves with maturation: quantitative evaluation with diffusion MR imaging in rats. Radiology, 216(3), 881-885.

Tandon, R., Mazzara, C., DeQuardo, J., Craig, K. A., Meador-Woodruff, J. H., Goldman, R., & Greden, J. F. (1991). Dexamethasone suppression test in schizophrenia: relationship to symptomatology, ventricular enlargement, and outcome. Biol Psychiatry, 29(10), 953-964.

Tang, M. X., Jacobs, D., Stern, Y., Marder, K., Schofield, P., Gurland, B., . . . Mayeux, R. (1996). Effect of oestrogen during menopause on risk and age at onset of Alzheimer's disease. Lancet, 348(9025), 429-432.

Tannenbaum, C., Barrett-Connor, E., Laughlin, G. A., & Platt, R. W. (2004). A longitudinal study of dehydroepiandrosterone sulphate (DHEAS) change in older men and women: the Rancho Bernardo Study. Eur J Endocrinol, 151(6), 717-725.

Thompson, K. N., Phillips, L. J., Komesaroff, P., Yuen, H. P., Wood, S. J., Pantelis, C., . . . McGorry, P. D. (2007). Stress and HPA-axis functioning in young people at ultra high risk for psychosis. J Psychiatr Res, 41(7), 561-569.

Tost, H., Braus, D. F., Hakimi, S., Ruf, M., Vollmert, C., Hohn, F., & Meyer- Lindenberg, A. (2010). Acute D2 receptor blockade induces rapid, reversible remodeling in human cortical-striatal circuits. Nat Neurosci, 13(8), 920-922.

Tourney, G., & Hatfield, L. (1972). Plasma androgens in male schizophrenics. Arch Gen Psychiatry, 27(6), 753-755.

Turk-Browne, N. B., Scholl, B. J., Chun, M. M., & Johnson, M. K. (2009). Neural evidence of statistical learning: efficient detection of visual regularities without awareness. J Cogn Neurosci, 21(10), 1934-1945.

Usall, J., Huerta-Ramos, E., Iniesta, R., Cobo, J., Araya, S., Roca, M., . . . Ochoa, S. (2011). Raloxifene as an adjunctive treatment for postmenopausal women with schizophrenia: a double-blind, randomized, placebo-controlled trial. J Clin Psychiatry, 72(11), 1552-1557.

Valenti, G., Ferrucci, L., Lauretani, F., Ceresini, G., Bandinelli, S., Luci, M., . . . Schwartz, R. S. (2009). Dehydroepiandrosterone sulfate and cognitive function 190

in the elderly: The InCHIANTI Study. Journal of Endocrinological Investigation, 32(9), 766-772. van Erp, T. G. M., Hibar, D. P., Rasmussen, J. M., Glahn, D. C., Pearlson, G. D., Andreassen, O. A., . . . Turner, J. A. (2015). Subcortical brain volume abnormalities in 2028 individuals with schizophrenia and 2540 healthy controls via the ENIGMA consortium. Mol Psychiatry. van Haren, N. E., Cahn, W., Hulshoff Pol, H. E., & Kahn, R. S. (2008). Schizophrenia as a progressive brain disease. Eur Psychiatry, 23(4), 245-254. van Niekerk, J. K., Huppert, F. A., & Herbert, J. (2001). Salivary cortisol and DHEA: association with measures of cognition and well-being in normal older men, and effects of three months of DHEA supplementation. Psychoneuroendocrinology, 26(6), 591-612. van Os, J., & Kapur, S. (2009). Schizophrenia. Lancet, 374(9690), 635-645.

Van Voorhees, E. E., Dennis, M. F., Calhoun, P. S., & Beckham, J. C. (2014). Association of DHEA, DHEAS, and cortisol with childhood trauma exposure and post-traumatic stress disorder. Int Clin Psychopharmacol, 29(1), 56-62.

Veldhuis, J. D., Iranmanesh, A., Lizarralde, G., & Johnson, M. L. (1989). Amplitude modulation of a burstlike mode of cortisol secretion subserves the circadian glucocorticoid rhythm. Am J Physiol, 257(1 Pt 1), E6-14.

Venkatasubramanian, G., Chittiprol, S., Neelakantachar, N., Shetty, T., & Gangadhar, B. N. (2010). Effect of antipsychotic treatment on Insulin-like Growth Factor-1 and cortisol in schizophrenia: a longitudinal study. Schizophr Res, 119(1-3), 131-137.

Ventura, J., Nuechterlein, K. H., Hardesty, J. P., & Gitlin, M. (1992). Life events and schizophrenic relapse after withdrawal of medication. Br J Psychiatry, 161, 615- 620.

Vercammen, A., Skilleter, A. J., Lenroot, R., Catts, S. V., Weickert, C. S., & Weickert, T. W. (2013). Testosterone Is Inversely Related to Brain Activity during Emotional Inhibition in Schizophrenia. PLoS ONE, 8(10), e77496.

Vermes, I., Varszegi, M., Toth, E. K., & Telegdy, G. (1979). Action of androgenic steroids on brain neurotransmitters in rats. Neuroendocrinology, 28(6), 386-393.

Vita, A., De Peri, L., Deste, G., Barlati, S., & Sacchetti, E. (2015). The Effect of Antipsychotic Treatment on Cortical Gray Matter Changes in Schizophrenia: Does the Class Matter? A Meta-analysis and Meta-regression of Longitudinal Magnetic Resonance Imaging Studies. Biol Psychiatry, 78(6), 403-412.

Walder, D. J., Walker, E. F., & Lewine, R. J. (2000). Cognitive functioning, cortisol release, and symptom severity in patients with schizophrenia. Biol Psychiatry, 48(12), 1121-1132. 191

Walker, E. F., Brennan, P. A., Esterberg, M., Brasfield, J., Pearce, B., & Compton, M. T. (2010). Longitudinal changes in cortisol secretion and conversion to psychosis in at-risk youth. J Abnorm Psychol, 119(2), 401-408.

Walker, E. F., & Diforio, D. (1997). Schizophrenia: a neural diathesis-stress model. Psychol Rev, 104(4), 667-685.

Walker, E. F., Walder, D. J., & Reynolds, F. (2001). Developmental changes in cortisol secretion in normal and at-risk youth. Dev Psychopathol, 13(3), 721-732.

Webb, S. J., Geoghegan, T. E., Prough, R. A., & Michael Miller, K. K. (2006). The biological actions of dehydroepiandrosterone involves multiple receptors. Drug Metab Rev, 38(1-2), 89-116.

Webster, M. J., Knable, M. B., O'Grady, J., Orthmann, J., & Weickert, C. S. (2002). Regional specificity of brain glucocorticoid receptor mRNA alterations in subjects with schizophrenia and mood disorders. Mol Psychiatry, 7(9), 985-994, 924.

Wechsler, D. (1997). Wechsler Adult Intelligence Scale- Third Edition. Psychological Corporation, San Antonio.

Wechsler, D. (2001). Wechsler Test of Adult Reading. The Psychological Corporation, San Antonio.

Weickert, C. S., Hyde, T. M., Lipska, B. K., Herman, M. M., Weinberger, D. R., & Kleinman, J. E. (2003). Reduced brain-derived neurotrophic factor in prefrontal cortex of patients with schizophrenia. Mol Psychiatry, 8(6), 592-610.

Weickert, C. S., Miranda-Angulo, A. L., Wong, J., Perlman, W. R., Ward, S. E., Radhakrishna, V., . . . Kleinman, J. E. (2008). Variants in the estrogen receptor alpha gene and its mRNA contribute to risk for schizophrenia. Hum Mol Genet, 17(15), 2293-2309.

Weickert, C. S., Weickert, T. W., Pillai, A., & Buckley, P. F. (2013). Biomarkers in Schizophrenia: A Brief Conceptual Consideration. Disease Markers, 35(1), 7.

Weickert, T. W., Goldberg, T. E., Gold, J. M., Bigelow, L. B., Egan, M. F., & Weinberger, D. R. (2000). Cognitive impairments in patients with schizophrenia displaying preserved and compromised intellect. Arch Gen Psychiatry, 57(9), 907-913.

Weickert, T. W., Weinberg, D., Lenroot, R., Catts, S. V., Wells, R., Vercammen, A., . . . Weickert, C. S. (2015). Adjunctive raloxifene treatment improves attention and memory in men and women with schizophrenia. Mol Psychiatry, 20(6), 685-694. 192

Weinberger, D. R. (1987). Implications of normal brain development for the pathogenesis of schizophrenia. Arch Gen Psychiatry, 44(7), 660-669.

Weinberger, D. R., Mattay, V., Callicott, J., Kotrla, K., Santha, A., van Gelderen, P., . . . Frank, J. (1996). fMRI applications in schizophrenia research. Neuroimage, 4(3 Pt 3), S118-126.

Weiser, M. J., Osterlund, C., & Spencer, R. L. (2011). Inhibitory effects of corticosterone in the hypothalamic paraventricular nucleus (PVN) on stress- induced ACTH secretion and gene expression in the PVN and anterior pituitary. Journal of neuroendocrinology, 23(12), 1231-1240.

Wellman, C. L. (2001). Dendritic reorganization in pyramidal neurons in medial prefrontal cortex after chronic corticosterone administration. Journal of Neurobiology, 49(3), 245-253.

Widstrom, R. L., & Dillon, J. S. (2004). Is there a receptor for dehydroepiandrosterone or dehydroepiandrosterone sulfate? Semin Reprod Med, 22(4), 289-298.

Wieck, A., & Haddad, P. M. (2003). Antipsychotic-induced hyperprolactinaemia in women: pathophysiology, severity and consequences. Selective literature review, 182(3), 199-204.

Williams, L. M., Das, P., Harris, A. W., Liddell, B. B., Brammer, M. J., Olivieri, G., . . . Gordon, E. (2004). Dysregulation of arousal and amygdala-prefrontal systems in paranoid schizophrenia. Am J Psychiatry, 161(3), 480-489.

Wilson, M. (1993). DSM-III and the transformation of American psychiatry: a history. Am J Psychiatry, 150(3), 399-410.

Wolkowitz, O. M., Epel, E. S., & Reus, V. I. (2001). Stress hormone-related psychopathology: pathophysiological and treatment implications. World J Biol Psychiatry, 2(3), 115-143.

Woods, S. W. (2003). Chlorpromazine equivalent doses for the newer atypical antipsychotics. J Clin Psychiatry, 64(6), 663-667.

Woolley, C. S., & McEwen, B. S. (1994). Estradiol regulates hippocampal dendritic spine density via an N-methyl-D-aspartate receptor-dependent mechanism. J Neurosci, 14(12), 7680-7687.

Woolley, J. D., Chuang, B., Lam, O., Lai, W., O’Donovan, A., Rankin, K. P., . . . Vinogradov, S. (2014). Oxytocin administration enhances controlled social cognition in patients with schizophrenia. Psychoneuroendocrinology, 47, 116- 125.

Wright, C. I., Fischer, H., Whalen, P. J., McInerney, S. C., Shin, L. M., & Rauch, S. L. (2001). Differential prefrontal cortex and amygdala habituation to repeatedly presented emotional stimuli. Neuroreport, 12(2), 379-383.

Wright, I. C., McGuire, P. K., Poline, J. B., Travere, J. M., Murray, R. M., Frith, C. D., . . . Friston, K. J. (1995). A voxel-based method for the statistical analysis of 193

gray and white matter density applied to schizophrenia. Neuroimage, 2(4), 244- 252.

Xiu, M. H., Yang, G. G., Tan, Y. L., Chen da, C., Tan, S. P., Wang, Z. R., . . . Zhang, X. Y. (2014). Decreased interleukin-10 serum levels in first-episode drug-naive schizophrenia: relationship to psychopathology. Schizophr Res, 156(1), 9-14.

Yamada, S., Akishita, M., Fukai, S., Ogawa, S., Yamaguchi, K., Matsuyama, J., . . . Ouchi, Y. (2010). Effects of dehydroepiandrosterone supplementation on cognitive function and activities of daily living in older women with mild to moderate cognitive impairment. Geriatr Gerontol Int, 10(4), 280-287.

Yildirim, O., Dogan, O., Semiz, M., & Kilicli, F. (2011). Serum cortisol and dehydroepiandrosterone-sulfate levels in schizophrenic patients and their first- degree relatives. Psychiatry Clin Neurosci, 65(6), 584-591.

Young, E. A., Abelson, J., & Lightman, S. L. (2004). Cortisol pulsatility and its role in stress regulation and health. Front Neuroendocrinol, 25(2), 69-76.

Yung, A. R., & McGorry, P. D. (1996). The initial prodrome in psychosis: descriptive and qualitative aspects. Aust N Z J Psychiatry, 30(5), 587-599.

Yung, A. R., Nelson, B., Stanford, C., Simmons, M. B., Cosgrave, E. M., Killackey, E., . . . McGorry, P. D. (2008). Validation of "prodromal" criteria to detect individuals at ultra high risk of psychosis: 2 year follow-up. Schizophr Res, 105(1-3), 10-17.

Zandi, P. P., Carlson, M. C., Plassman, B. L., Welsh-Bohmer, K. A., Mayer, L. S., Steffens, D. C., & Breitner, J. C. (2002). Hormone replacement therapy and incidence of Alzheimer disease in older women: the Cache County Study. Jama, 288(17), 2123-2129.

Zdanowicz, N., Mees, L., Jacques, D., Tordeurs, D., & Reynaert, C. (2014). Assessment and treatment of the risk of psychosis in adolescents--a review. Psychiatr Danub, 26(2), 115-121.

Zhou, W., Koldzic-Zivanovic, N., Clarke, C. H., de Beun, R., Wassermann, K., Bury, P. S., . . . Thomas, M. L. (2002). Selective estrogen receptor modulator effects in the rat brain. Neuroendocrinology, 75(1), 24-33.