Anatomy of Attentional Networks

Home , Attentional control, Attentional shift, Stroop effect

THE ANATOMICAL RECORD (PART B: NEW ANAT.) 281B:21–36, 2004

FEATURE ARTICLE

AMIR RAZ

Attention is a central theme in psychological science. As with other biological systems, attention has a distinct anatomy that carries out basic psychological functions. Disparate attentional networks correlate with discrete neural circuitry and can be influenced by specific brain injuries, states, and drugs. Accordingly, thinking about attention as an organ system is advantageous for understanding the details of this complex cognitive process. In the context of an influential model of attention, this article introduces the broad notion of attention, then addresses its prominent characteristics, mechanisms, and theories. The presentation emphasizes the role of recent neuroimaging data in outlining the functional neuroanatomy subserving distinct attentional networks. A discussion of pertinent results connects attentional networks with self-regulation, development, and rehabilitation training. Anat Rec (Part B: New Anat) 281B:21–36, 2004. © 2004 Wiley-Liss, Inc.

KEY WORDS: attentional networks; imaging; neuroimaging; cognitive control; neuroscience; hemispheres

INTRODUCTION one out of what seem several simulta- plying formal information theory, neous objects or trains of thought.” Broadbent likened attention to a filter. One of the oldest and most central James’ account strongly joins atten- He proposed that attention was issues in psychological science, attention with subjective experience. More- bounded by the amount of information is the process of selecting for ac- over, James’s addressing of attention tive processing ideas stored in mem- tion located between parallel sensory both to objects and to “trains of ory in our minds, or aspects of our systems and a limited-capacity per- thought” is important for understand- physical environment, such as ob- ceptual system (Broadbent, 1958). ing current approaches to sensory ori- jects. The study of attention has be- This view facilitated objective studies enting and executive control (Driver come a huge enterprise; last year of the limitations of the human ability and Frackowiak, 2001; Fernandez- alone about 300 articles were pub- to deal with multiple signals at a time Duque and Posner, 2001; Fan et al., lished on attention (Fig. 1). Through- in a variety of practical tasks. 2003a,b). However, as outlined below, out history, many great minds have As psychology moved toward the wrestled with the definition of atten- study of cognitive mechanisms, new tion. In 370 B.C., Aristotle regarded objective methods allowed for inves- attention as a narrowing of the senses. tigation into the processes of selec- Centuries later, William James (1890) Attention is not, as the tion. For example, studies showed contended that “everyone knows what quote from William that words could activate their se- attention is. It is the taking possession mantic associates without aware- of the mind in clear and vivid form of James implies, the same ness of the word’s identity (i.e., as being aware. priming). The parallel organization of sensory information extends to se- Dr. Raz is on the faculty of the MRI Unit mantic processing. The act of select- in the Department of Psychiatry, Divi- ing a word meaning for active atten- sion of Child and Adolescent Psychiatry, at Columbia University College of Phy- attention in the sense of orienting to tion appears to suppress the sicians and Surgeons and New York sensory objects can actually be invol- availability of other meanings of the State Psychiatric Institute. untary and can occur unconsciously. selected item and of competing *Correspondence to: Amir Raz, MRI Unit, Department of Psychiatry, Division So attention is not, as the quote from items. Consequently, attention has of Child and Adolescent Psychiatry, Co- William James implies, the same as since been viewed more as a mecha- lumbia University College of Physicians and Surgeons and New York State Psy- being aware (see Table 1 for a list of nism for providing priority for mo- chiatric Institute, 1051 Riverside Drive, important terms and definitions in tor acts than as a filter or bottleneck. Box 74, New York, NY 10032. Fax: 212- this field of study). For example, the premotor theory 543-6660; E-mail: [email protected] Following a lull in the field during suggests that attention is simply a DOI 10.1002/ar.b.20035 the early 1900s, Donald Broadbent re- preparation for response, or a select- Published online in Wiley InterScience (www.interscience.wiley.com). sumed the quest after World War II to ing of the goal of an intended action discover attentional mechanisms. Ap- (Rizzolatti et al., 1987). The focus on

tion. These similar and demanding GROSS CHARACTERISTICS OF tasks can at ﬁrst only be done in alter- ATTENTION nation, but after months of practice can be efﬁciently combined (Spelke et al., Visual attention often serves as a con- 1976). Given these examples, theories venient lens to examine the character- of attentional limits must also account istics of attention. Researchers and for how these limits can be eliminated clinicians have investigated the optics, (e.g., following practice) when the bot- anatomy, development, pathology, tleneck is seemingly bypassed and the and underlying neural processes of vi- correct action is effortlessly performed. sion, making it the most widely stud-

TABLE 1. Terms and deﬁnitions related to attentional network research

Figure 1. Research into Attention. Relative Foveate To look at directly so the image being viewed numbers of publications for each of five key falls on the fovea of the retina words, normalized to their relative frequen- Attention The mental ability to select stimuli, responses, cies in the 1960-1969 decade. The 2000 memories, and thoughts that are data span only the years 2000 to 2003. The behaviorally relevant among a host of others relative frequency values were computed by adding up all the papers that used each that are behaviorally irrelevant of these keywords in 1960–1969 in PubMed Attentional networks Neural circuits subserving attentional and finding the proportion using each key- processing, which preserve a degree of word. These were the base proportions. The anatomical and functional independence relative proportions were then the propor- but interact in many practical situations tions in each subsequent decade divided Executive network The mechanism for monitoring and resolving by the original proportions from 1960–1969. conflict among thoughts, feelings, and From Cavanagh (2003, 2004). responses; an attentional system concerned with such tasks as working memory, response may explain why atten- planning, switching, and inhibitory control tional systems seem to differ for Orienting The process of selecting information from sensory input space that is within reach and space Alerting The process of achieving and maintaining a that lies beyond. However, intended state of high sensitivity to incoming stimuli actions are probably not the only Top-down A downstream effect such as cognitive sources of capacity limits and atten- modulation control, as opposed to a bottom-up effect tional selection. After all, even when Cognitive control Processes such as conflict resolution, error we have no intention to act and pas- correction, inhibitory control, planning, and sively watch scenes go by (e.g., at the resource allocation movies), we still select only a subset Self-regulation A key mediator between genetic of the information that reaches the predisposition, early experience, and adult senses. Attention, therefore, was fur- functioning (e.g., in controlling the reaction to stress, the capacity to maintain focused ther operationalized as facilitating attention or the ability to interpret mental some kinds of memory and even states both internally and in others) consciousness. Neuroimaging Technologically advanced, often noninvasive, Other views construe attention in tools for tapping neurophysiological aspects the context of resource limitation/ of the behaving brain selection. Attentional limits are typ- Mismatch negativity An electrophysiological manifestation of ically found in unpracticed tasks. (MMN) involuntary preattentive processing occurring However, with practice, most atten- when some regularity in the auditory stream tion-demanding tasks can be ren- is violated by a change (deviant stimulus); dered attention-independent, or au- the MMN is advantageous with regard to the neural processing of unattended input tomatized. One example of this Anterior cingulate Part of the brain’s limbic system classically phenomenon involves visually cortex (ACC) related to affect but also active in many searching for a set of letters among studies of cognition; the ACC might be the other random letters. This task ini- brain’s error detection and correction tially demands attention and gives device part of a circuit involved in a form of steep slopes of latencies against dis- attention that serves to regulate both play size. However, after weeks of cognitive and emotional processing practice, the slopes become flat Functional magnetic A noninvasive neuroimaging technique that (Spelke et al., 1976; Schneider and resonance registers blood flow to functioning areas of Shiffrin, 1977). Another example imaging (fMRI) the brain in high magnetic fields Positron emission A neuroimaging technique measuring the flow comes from a demonstration involv- tomography (PET) of blood containing radioactive atoms that ing two verbal tasks, reading for emit positrons comprehension and writing to dicta- FEATURE ARTICLE THE ANATOMICAL RECORD (PART B: NEW ANAT.) 23

Box 1. Temporal Aspects of Attention: The Attentional Blink

Temporal allocation of visual attention is often explored using the attentional blink (AB) paradigm (Raymond et al., 1992). AB isolates one aspect of attention allocation and avoids the problem of interpreting results tapping concurrent higher cognitive functions. In the AB, visual attention must be directed to two items in close temporal and spatial proximity. The blink refers to the brief period in which processing of the first target item interferes with processing of a second target (or probe). In a typical AB task, a stream of letters appears in a rapid serial visual presentation while participants identify a target and also determine whether a probe appears later in the same stimulus stream. A classic example of AB, following Raymond et al. (1992), is to present one blue letter in a rapid visual stream of black letters at the same spatial location (Box 1 Fig. 1). The task is to identify the blue target letter and then detect the presence or absence of a probe letter. Probe detection in this dual-task condition is compared against a baseline condition in which only the probe letter must be de- Box 1 Figure 1. The attentional blink paradigm tests attention by overloading it; a list of tected. Processing of the probe is im- stimuli is presented very rapidly in succession at the same location on a computer paired for up to 600 msec after the screen. Each item overwrites the last while participants monitor the list using two target in the dual-task condition (Ray- criteria. For example, the criteria might be to detect the blue letter target and to mond et al., 1992; Shapiro et al., identify the “B” probe. If the interval of appearance between the target and the probe is greater than about 3/4 of a second, then both are usually reported correctly. 1994). This impairment of probe pro- However, when the interval between the target and the probe is shorter (e.g., 200–500 cessing disappears after about 700 msec), the ability to report the probe declines. This decline constitutes the attentional msec (i.e., probes appearing later blink, an interval of time when attention is supposedly switching from the first criterion than 700 msec following targets can to the second. Adapted with permission from Hollingsworth et al. (2001). be detected as well as probes appearing in streams requiring no target through an attentional “gate” opened advantages of the AB lie in the fact identification). Notably, detection of by the processing of the target (Ray- that it avoids the problems of ceil- probes immediately following the tar- mond et al., 1992). One benefit of this ing effects, continuous-performance get (i.e., in the T ϩ 1 position) is task is that it isolates temporal allo- tests, response inhibition or motor spared relative to detection of probes cation of attention without introduc- control, and can be effectively used in subsequent positions. It is as if the ing the spatial switching component with pathological populations (Holl- item in this position is able to slip of other visual attentional tasks. Other ingsworth et al., 2001). ied perceptual system. Studying visual to single characters permits us to visual attention: “spotlight”, “zoom attention allows us to explore how we catch typos, glean information about lens,” “gating,” and “gradient,” along move the attentional “beam” around punctuation marks, and even spot with the common usage of terms such to various areas of the visual field and minute imperfections on the physical as “attentional gaze” or “attentional change the detail with which we look paper. However, at this level of detail, focus” (Shalev and Algom, 2000). The at any given area. For example, it is we may miss the bigger idea conveyed notion of a spotlight, which started possible to look at this page and pay in a paragraph. As we shift our focus, out as a crude metaphor and has since attention to its setup as a whole, or we can change the target location of been taken up as a serious viewpoint concentrate on specific words and cer- our attention or the size of our atten- to be tested, is a natural derivative of tain letters therein. Paying attention tional field. Many metaphors describe the spatial milieu in which attention is 24 THE ANATOMICAL RECORD (PART B: NEW ANAT.) FEATURE ARTICLE assumed to operate and relates to the select one out of these streams of con- fluences priority or processing prefer- common experience concerning the versation to follow in detail. We usu- ence, but also modulates sensory kind of attention needed for reading ally do that based on the location of processing and top-down control versus proofreading. The bulk of the the person: we may visually orient to- (Pessoa et al., 2003; Raz, 2004) and evidence suggests that attention can ward the person and/or hone in on may even affect accommodation (Raz have an influence quite early in the their frequency of voice. For example, et al., 2004a). visual system and scientists are still we can typically separate a male voice Compared to unattended stimuli, learning what visual operations atten- from a female voice more easily than behavioral findings show that reaction could potentially influence (Rob- we can separate two male voices. Al- tion time to attended stimuli is usually ternatively, we may pay attention to ertson and Garavan, in press). How- faster. Neurophysiological data reveal the content of the information by fol- ever, it is important to complement enhancement of electrical activity the spatial aspects of attention with its lowing a story line. When we attend to over extrastriate visual areas by about temporal limitations (see Box 1). one stream, the other information 90 msec following visual presentation. Given a large visual array of individ- coming from conversations around us Although performance may improve ual features, one can choose either to goes into the background: it is on increased attentional investment, examine it globally or to investigate its present, but does not reach focal anal- specific features. Using compound ysis. Data suggest that much of this great controversy has existed over stimuli for studying global and local unattended information is actually what orienting attention to a sensory processing, it is possible to study how processed in subtle and complicated stimulus actually does. The psycho- one can shift back and forth between ways (Kihlstrom, 1996). Moreover, physics literature provides reliable ac- them by changing the attentional fo- unattended information can suddenly counts of how visual thresholds corre- cus (Navon, 2003). For example, cer- become interesting (e.g., when our late with attentional investment. tain patients have difficulty examining Studies show that improvement in vi- the local features. These patients usu- sual acuity is not synonymous with ally have damage to the left tem- altered thresholds for detection, bet- poroparietal lobe. Other patients may Researchers in the field ter performance, or faster reaction do well with the local features, but fail agree that attention is times. Whereas acuity requires the to appreciate the overall contour; they resolution of detail, detection thresh- usually have damage to the right tem- not a unitary term. olds and reaction time can involve the poroparietal lobe. Indeed, the parietal Rather, we can summation of luminance, which lobe tends to emphasize the shifting fractionate attention into might obscure detail (Raz et al., between local and global stimuli, 2004b). while the temporal lobe seems to de- subsystems of more Although investing attention is fre- termine whether one can actually ex- quently associated with looking di- amine a local or global feature of the circumscribed function rectly at the scene of interest, covert stimulus. and anatomy. attention is the ability to grant such We usually foveate, or look at, ex- information priority in processing actly the thing in which we are inter- without eye movement. Researchers ested, and that process generally re- name is mentioned), whereby we ori- have shown that the performance im- lates our attention to where we fixate. ent to the new information. These at- provement at attended locations re- However, it is easy to dissociate atten- tentional phenomena have been stud- sults somewhat from an enhanced tion from the fovea. We can cue peo- ied experimentally in some detail and spatial resolution at the cued location ple to attend to some location in space will be described next. other than the center of gaze and then (Yeshurun and Carrasco, 1998, 1999, show that they are sensitive to infor- 2000; Carrasco et al., 2002). Studies ATTENTION AND PERCEPTION mation that occurs at the cued loca- exploring the relationship between vi- tion and relatively slow or insensitive Attention is not a panacea to percep- sual attention and contrast sensitivity to information at the fovea, as mea- tion because there is a great deal at- show that covert attention not only sured by a low threshold or fast re- tention cannot do (Raz et al., 2004b; improves discrimination in a wide va- sponse time (Posner, 1980). It is be- Raz, 2004a,b). For example, while ori- riety of visual tasks, but could also lieved that these covert shifts in enting to a location, attention can give increase the rate at which information attention are used to select the part of priority to that location (i.e., targets is processed. Some findings do indi- the visual field to which one usually that appear there will be perceived cate that contrast sensitivity is greater wants to move the eyes because in more rapidly and with lower thresh- in the lower versus higher visual me- everyday life, one usually follows co- olds). However, attention cannot sub- ridian. However, most evidence sets vert shifts of attention with an eye stitute for the acuity provided by the limits to the effects of attention on movement. Attention to visual ele- fovea. While the fovea is critical for spatial resolution and specifies that ments can also apply to other modal- acuity, the costs in reaction time for certain visual, not attentional, con- ities, such as the auditory system. an unexpected foveal stimulus are just straints determine aspects of spatial When multiple people talk simulta- as great as for an unexpected periph- resolution (Talgar and Carrasco, neously, it is sometimes necessary to eral event. Thus, visual attention in- 2002). FEATURE ARTICLE THE ANATOMICAL RECORD (PART B: NEW ANAT.) 25

ATYPICAL ATTENTION magnitude. The MMN presumably ATTENTION AND indicates a mechanism that com- NEUROIMAGING Biological rhythms affect attentional pares the current auditory input to performance. Diurnal reductions in the memory traces formed by previ- Two recently developed tools have attention normally occur during the ous auditory inputs and signals the been particularly valuable in studying hours of maximum sleepiness, 2:00 occurrence of a mismatch (Naa- attention as an organ system: brain a.m. to 7:00 a.m., which coincides tanen, 2003). imaging and genomics. Neuroimaging with the period of lowest body tem- In adults, MMN tends to decline has illuminated the anatomical areas perature. Attentional performance is during drowsiness and its persis- involved in attention, while the Hu- usually enhanced in the evening, tence into adult human sleep is still man Genome Project is identifying when body temperature peaks (Raz, debated (Atienza et al., 2002). Other genes involved in these networks. This 1999; Manly et al., 2002). During the EEG components do reflect the article focuses on data concerning the states of sleep deprivation and sleep, brain’s reaction to novelty. Although neuroimaging of attention. The genet- voluntary attention is often markedly active midbrain inhibition blocks ics of attention and the interaction be- attenuated or even absent. However, cortical activity in the developed tween the two methodologies are ad- evidence suggests that certain atten- brain, the sleeping infant brain is dressed elsewhere (Fossella et al., tional as well as preattentional mech- probably not as capable of blocking 2002b; Fan et al., 2003c). anisms remain intact albeit unstable and inhibiting information effi- The past 2 decades have ushered in (Raz, 1999; Raz and Posner, 2000; Do- ciently. Indeed, MMN is obtainable a new era of methodological advances ran et al., 2001; Raz et al., 2001). from newborns and young infants in tools for noninvasive imaging of the Dreaming is usually divorced from a and experimental data show that living brain. The information gleaned sense of controlled awareness. How- during those early developmental from such advances has identified ever, purported accounts of lucid stages, the brain can learn, even dur- both anatomical and functional as- dreaming, in which persons dream ing sleep (Cheour et al., 2002). pects of the neural processes underly- while knowing that they are dreaming attention (Kastner and Ungerlei- ing, suggest that some control mech- der, 2000, 2001; Driver and anisms may be available during sleep. Two recently developed Frackowiak, 2001; Pessoa et al., 2003). Other common anecdotes in support Functional neuroimaging methods al- of this concept include the incorpora- tools have been low researchers to measure changes tion of ambient sound into the dream particularly valuable in in brain activity associated with si- content as well as the idea of sensitiv- multaneous changes in behavior, or in ity to one’s own name. studying attention as an response to a wide variety of stimuli. To investigate information process- organ system: brain ERP, functional magnetic resonance ing in sleep, researchers measure elec- imaging (fMRI), magnetoencephalog- trical recordings from the scalp using imaging and genomics. raphy (MEG), near-infrared spectros- electroencephalography (EEG). By copy (NIRS), positron emission to- averaging the brain’s electrical re- mography (PET), and single photon sponse potentials to stimuli using a Attention is sometimes associated emission computed tomography technique called event-related poten- with hypnosis, a special cognitive (SPECT) all measure aspects of phys- tials (ERPs), it is possible to examine state sometimes confused with sleep. iological change in brain activity. the processing capability of the sleep- Hypnosis has been used clinically for Early neuroimaging techniques of ing brain. One such component to ex- hundreds of years and is primarily a the 1970s allowed researchers to amine is the mismatch negativity phenomenon involving attentive re- record from cells in alert monkeys (MMN). The MMN is an electrophys- ceptive concentration. Following par- performing an experimental cognitive iological manifestation of involuntary ticular suggestions, highly hypnotiz- task. Data from these assays showed preattentive processing in response to able individuals may experience increased firing rates for cells in a oddball stimuli. In a typical MMN attentional and perceptual changes, number of brain regions when the paradigm, a deviant auditory stimulus which may not typically occur during monkeys paid attention compared to is infrequently interspersed within a common awareness. In a responsive when they did not (Wurtz and Gold- sequence of standard auditory stim- subject, suggestions for perceptual al- berg, 1972). These efforts directly re- uli. The MMN is evident in the dif- teration are accompanied by repro- lated to studies of other aspects of at- ference waveform resulting from the ducible changes in brain action. For tention, such as research carried out subtraction of the ERP elicited by example, the activity of the anterior in the 1960s to decipher the mecha- the standard stimulus from that elic- cingulate cortex (ACC) to painful nism of brain information processing ited by the novel auditory stimuli stimuli or conflict resolution can be (e.g., Broadbent). Since the 1980s, the (the deviants). The difference wave- modulated by hypnotic suggestion study of patients with focal brain le- form, occurring even without atten- (Rainville et al., 1997; Raz and Sha- sions and ERPs has established links tion, normally peaks between 100 piro, 2002; Raz, 2004). Interestingly, between the brain areas showing se- and 250 msec from the onset of the most children are highly hypnotizable lective enhancement and specific as- deviant event. This value depends on and more easily inducted into hypno- pects of information processing. Re- the dimension of deviance and its sis than are adults. cently, neuroimaging methods were 26 THE ANATOMICAL RECORD (PART B: NEW ANAT.) FEATURE ARTICLE developed that expanded the types of 2004b). Neuroimaging data have con- that attention enables the exercise of attention that could be studied physi- firmed earlier notions that attention is voluntary control over thoughts, feel- ologically. Some of these findings be- not a unitary concept and that the hu- ings, and actions as a means of self- gan to influence computational mod- man brain likely entertains several at- regulation (Raz, 2004) in adulthood els of attention (LaBerge, 1994). tentional systems of different though and throughout development (Posner Psychological studies have eluci- interrelated functions (Posner and Pe- and Rothbart, 1998, 2000; Bronson, dated the pertinent questions about tersen, 1990). Indeed, for the last 2000; Rothbart et al., 2004). Varia- attention. These studies created a sub- 15–20 years, brain imaging has been tions in the operational efficiency of stantial number of experimental par- examining questions of attention and these attentional systems explain dif- adigms, including dichotic listening, many other cognitive tasks. Major re- ferences in self-regulation and emo- visual search, dual-task performance, sults of this research can be summa- tional control and promise to help de- flanker tasks, rapid serial visual pre- rized in three statements (Posner, scribe mechanisms of volition and sentation, and negative priming. 2004c). One, networks of neural areas, sustained effort (Rueda et al., 2004b, Through these models, researchers often widely scattered over the brain, in press). In a way, this network ap- hoped to capture the range of phe- but not involving the whole brain, proach owes a lot to the work carried nomena encompassed by attention compute cognitive and emotional out about 50 years ago by Donald O. under the controlled conditions re- tasks. Each node of the network com- Hebb, who outlined in his cell-assem- quired by scientific scrutiny. Converg- putes a different aspect and together bly theory the ensemble of areas in the ing evidence is vital to such analysis as the networks orchestrate the whole brain that might be involved in pro- scientists test hypotheses in a variety task. Two, some networks are in- cessing cognitive and emotional tasks of different ways so that artifacts of volved in the control of other net- (Posner and Rothbart, in press). one method are circumvented by an- works. These attentional networks are other. Confidence in the conclusions involved in the selection and control DIFFERENT TYPES OF ATTENTION increases with replications and the of networks that process sensory in- systematic dismissal of potential and Attention does not imply a singular lingering caveats. Since inferences are mechanism; rather, it is a complex necessarily indirect (e.g., observation Attention does not imply system presiding over a number of of behavior or measurement of brain distinct neuronal circuits. This under- activations), by putting them to- a singular mechanism; standing comes from combining data gether, one further constrains the pro- rather, it is a complex gathered using various neuroimaging posed theories. techniques. Early PET data show that While the psychology of attention system presiding over a different attributes such as color, mo- has furnished a number of interesting number of distinct tion, and shape are processed by at results regarding the limits of perfor- least partially separate systems (Cor- mance and of unconscious process- neuronal circuits. betta et al., 1991a, 1991b). Addition- ing, there was no agreement on ally, a recent review of fMRI results whether attention involved separate reveals a distributed system of brain mechanisms from those used to pro- formation and information from regions that control attention (Fan et cess data, nor did psychological stud- memory. Three, these networks al., 2003a; Pessoa et al., 2003). Ac- ies analyze the neural mechanisms of change with development, learning, cordingly, researchers have formu- attention. Furthermore, because ob- brain injury, and pathology. lated experimental paradigms that servable responses are needed as evi- Given the increased description of considerably shape our understand- dence, purely psychological studies attentional networks in recent years, ing of attention. More recently, the are limited in their ability to deter- we can now use these networks as notion of attention has been extended mine how much information is pro- model systems for the exploration of to the control of action, thought, and cessed concerning unattended mes- symptoms arising from various forms affect both in adults and across devel- sages. Fortunately, brain imaging of pathology (Berger and Posner, opment (Posner and Rothbart, 1998, allows monitoring of incidental and 2000). Current research has demon- 2000; Rueda et al., 2002; Rothbart et involuntary processing of unattended strated that distinct brain areas in- al., 2003, 2004; Rueda et al., 2004a, stimuli. Insights into the neural sys- deed mediate different attentional 2004b, in press). tems underlying attention, therefore, processes and thus it is now possible Three largely disparate attentional markedly benefited from experimen- to examine selective attention as an control systems—select, orient, and tal paradigms involving selection of organ system with its own functional alert—have been identified. Although sensory information coupled with the anatomy, circuitry, and cellular struc- the three systems interact in many technological innovation of imaging ture (Posner and Fan, 2004). Although practical contexts, the operations of the living brain. still incomplete, this information has these attentional networks have a cer- Brain imaging, including data from illuminated important questions in tain degree of functional and anatom- implanted electrodes in both human cognitive science and has provided in- ical independence. Whereas the select and nonhuman primates, has forged sights into neurological and psychiat- system involves choosing among con- an impressive link between psychol- ric disorders of both children and flicting actions, orient refers to a point ogy and neuroscience (Posner, adults. Converging data now point out of reference to sensory objects, and FEATURE ARTICLE THE ANATOMICAL RECORD (PART B: NEW ANAT.) 27 alert involves changes in the internal tion either by brain damage or by re- (Posner and Fan, 2004), we have re- state in preparation for perceiving a stricting transmitter input. Likewise, cently devised a simple Attention Net- stimulus. Insult to each of these dis- frontal attentional networks have pro- work Test (ANT) that can be per- tinct attentional circuits, irrespective vided similar understanding of pa- formed by adults, children, patients, of the source, produces dissimilar thologies at higher levels of cognition. and even nonhuman animals (Fan et neuropsychological deficits. Correlat- For example, evidence exists relat- al., 2002). The ANT takes about 30 ing the etiologies of such injuries with ing frontal networks to attention def- min to administer and provides three subsequent changes in behavior and icit hyperactivity disorder (ADHD) numbers that indicate the efficiency cognition allows us to extend the role (Berger and Posner, 2000). Recent re- of the networks that perform the alert, of attentional neuroimaging possibly search has linked the cholinergic sys- orient, and conflict-resolution func- into rehabilitation (Robertson, 1999; tem to orienting (Beane and Mar- tions. Previous work with this test Fan et al., 2003b; Posner and Fan, rocco, 2004). provided evidence on its reliability, 2004; Rueda et al., in press). Re- The alert, or vigilant, system func- heritability, and the independence of searchers have illuminated the neuro- tions to achieve and maintain a state the results (Fan et al., 2001a, 2001b; chemical substrates of these atten- of high sensitivity to incoming stim- Fossella et al., 2002a). Recent neuro- tional networks and detailed analyses uli. The process of acquiring and imaging findings have also used the are now emerging for these orthogo- maintaining alertness in the face of ANT to probe and assess these distinct nal sensory systems (Raz and Shapiro, prolonged dull tasks has been an area attentional networks (Fan et al., 2002; Fan et al., 2003b; Posner, of vigorous research. Recent reports 2001a). These behavioral and imaging 2004a). have suggested some of the anatomy data suggest that the networks are Neuroimaging findings have shown involved in how arousal, alertness, largely independent and further per- activation of frontal areas associated and vigilance affect performance mit identification of the neuroanat- with paying attention in the context of omy subserving these attentional net- strong mental effort. For example, works (Fig. 2). when experimental tasks appear to in- Although previous studies have ex- volve significant exertion during tar- Alerting involves a amined areas of the brain involved in get detection or conflict resolution, change in the internal various ANT components (Corbetta et the area of the ACC is consistently ac- al., 2000; Hopfinger et al., 2000), neu- tivated. Thus, the select network is a state in preparation for roimaging the ANT has illuminated higher-level metacognitive attentional perceiving a stimulus. areas involved in carrying out these system related to the subjective im- attentional networks as a whole. fMRI pression of mental effort (Fernandez- The alert state is critical data from the ANT indicate that three Duque et al., 2000a, 2000b). This form for optimal performance largely orthogonal networks relate to of supervisory or executive attention components of attention (Fan et al., comprises the mechanisms for moni- in tasks involving higher 2001a). Pharmacological studies have toring and resolving conflict among cognitive functions. related each of these networks with thoughts, feelings, and responses. specific chemical neuromodulators. Therefore, this attentional system is First, cholinergic systems arising in concerned with such tasks as working (Robertson and Garavan, in press). the basal forebrain play an important memory, planning, switching, and in- They also highlight the importance of role in orienting (Beane and Mar- hibitory control. Since frontal pa- norepinephrine modulation in the rocco, 2004). Second, the norepineph- tients are less able to implement a set alerting that follows warning signals erine system arising in the locus coer- of instructed goals, the orbital frontal (Witte and Marrocco, 1997; Witte et uleus of the midbrain functions in area is probably important for this ex- al., 1997). alerting (Witte and Marrocco, 1997; ecutive faculty (Duncan et al., 1996). Witte et al., 1997; Coull et al., 2001). Visual orienting is the process of SUBSTRATES OF ATTENTIONAL Third, the ACC and lateral prefrontal selecting information from sensory cortex are target areas of the mesocor- NETWORKS input. Orienting of attention toward tical dopamine system involved in ex- visual objects can be either triggered Attention can enhance neural process- ecutive attention (Marrocco and Da- by the stimuli (e.g., by brief presenta- ing at multiple levels. As a selective vidson, 1998; Deth et al., 2004; tions) or shifted as a result of volun- aspect of information processing, at- Robbins et al., 2004). tary control. Orienting has thus be- tention allows some things to be priv- Alerting involves a change in the in- come a vehicle for separating ileged while others are ignored. Vigor- ternal state in preparation for perceiv- automatic from voluntary mecha- ous studies into the modulation of ing a stimulus. The alert state is criti- nisms. Early PET results have under- neural responses by attention have re- cal for optimal performance in tasks lined the importance of the parietal vealed that attention can be viewed as involving higher cognitive functions. area together with thalamic and mid- a system of anatomical areas carrying Neuroimaging studies have shown ac- brain areas for visual orienting. The out the functions of alerting, orient- tivity in the frontal and parietal re- orienting network has been used to ing, and executive control (Posner and gions particularly of the right hemi- understand the effects of lesions that Raichle, 1996; Posner, 2004a). In line sphere when people are required to produce neglect of sensory informa- with an attentional research agenda achieve and maintain the alert state 28 THE ANATOMICAL RECORD (PART B: NEW ANAT.) FEATURE ARTICLE

rior colliculus, superior parietal lobe, and frontal eye fields are often activated in studies of the orienting network. Orienting can be reflexive, as when a sudden target event directs attention to its location, or it can be voluntary, as when a person searches the visual field for a target. Orienting typically involves head and/or eye movements toward the target, as in overt orienting. However, the process can also be covert. A few dorsal brain areas, including the superior parietal lobe and temporoparietal junction, serve as common sources of attention to sensory stimuli. These regions produce effects within a network of areas that depend on modality, such as ventral visual areas in the case of visual input. The strongest evidence for lo- calization of mental operations stems from the area of attentional orienting toward sensory stimuli. A confluence of methods and experimental sophis- tication has demonstrated how separate brain areas can be invoked to or- ganize a simple attentional shift. There is agreement that orienting of attention to a visual stimulus produces amplification in prestriate regions. This activation affects processing in all subsequent regions, which feeds back to influence processing in the primary visual cortex (V1) and perhaps in the lateral geniculate nu- cleus of the thalamus. Executive control of attention involves more complex mental operations both to monitor and resolve conflict between computations occurring in different brain areas. Executive Figure 2. Anatomy of attentional networks. fMRI images collected from 16 healthy adults control is most needed in situations performing the ANT in a 3 Tesla magnet (Fan et al., 2001). Cross-sectional views of the involving planning or decision-mak- three attentional networks; the alerting network shows thalamic activation, the orienting network shows parietal activations, and the conflict network shows anterior cingulate cortex activation. First presented at the 2001 Annual Meeting of the Society for Neuro- science, this figure outlines some of the functional anatomy subserving these distinct attentional networks. The color bar shows fMRI signal level (Z-scores) above the 0.05 significance threshold. The novelty of the ANT is that it illuminates the separate attentional networks described in recent literature, all within one relatively short task that can be run with children and animals, as well as deficit populations and adults. for even a brief period. Thus, lesions deficits produce a profound neglect in of these areas reduce the ability to the visual field opposite the lesion. maintain alertness. Right frontal le- Alerting is thought to involve the cor- sions impair one’s ability to sustain tical distribution of the brain’s nore- Figure 3. The Stroop effect: an example of attention voluntarily, which produces pinepherine system arising in the lo- selective attention. In responding to the ink more errors over time than are found cus coeruleus of the midbrain. color of an incompatible color word, partic- for left frontal patients in tasks involv- The orienting network concerns the ipants are usually much slower and less ac- ing continuous performance. Right selection of information from sensory curate than in identifying the ink color of a neutral (or congruent) item. This is called the parietal patients have difficulties in input. Cholinergic systems arising in Stroop interference effect and is one of the both maintaining the alert state and in the basal forebrain play an important most robust and well-studied phenomena in attentional orienting. Together, these role in orienting. The pulvinar, supe- attentional research FEATURE ARTICLE THE ANATOMICAL RECORD (PART B: NEW ANAT.) 29

Box 2. Posnerian Trinity of Attention

By following a trinity proposed by Michael I. Posner (Posner and Pe- tersen, 1990; Posner and Raichle, 1996; Fan et al., 2003b; Posner and Fan, 2004), attentional networks can be construed in terms of, one, obtain- ing and maintaining the alert state; two, orienting to sensory information; and three, the executive functions involved in resolving conflict between competing areas of the brain that might be simultaneously active. The orienting network relies heavily on parietal systems, including the superior parietal lobe and the temporal parietal junction. It is involved in both orienting to visual information and stimuli in other modalities. The alerting network relies heavily on thalamic areas, locus coeruleus, and cortical areas. The executive attention network relies on the anterior cingulate and lateral areas of the prefrontal cortex (Box 2 Fig. 1). In a set of influential experiments investigating the neuropharmacology of attention in alert monkeys, Mar- rocco and Davidson (1998) correlated Box 2 Figure 1. A sketch of the functional anatomy of the attentional networks. The each brain network with different neu- pulvinar, superior colliculus, superior parietal lobe, and frontal eye fields are often activated in studies of the orienting network. The temporoparietal junction is active romodulators. These data identified when a target occurs at a novel location. The anterior cingulate gyrus is an important that the orienting network is modu- part of the executive network. Right frontal and parietal areas are active when people lated by the cholinergic system, the maintain the alert state. Adapted with permission from Raz and Shapiro (2002). alerting network by the norepinepherine system, and the executive network by mainly dopamine systems. Together with Box 2 Figure 2, Box 2 2002) or even pleasant thoughts opposite direction. If the fish in ques- Table 1 depicts the structures in- (Beauregard et al., 2001). The dorsal tion is swimming in the same direc- volved as the sources of the Posne- part of the ACC, which is involved in tion as the surrounding school of fish, rian trinity of attention, the sites on cognitive tasks, is distinguishable the two are congruent; if they face which these structures operate, and from the more ventral part, which opposite directions, they are incon- the neuromodulators they use. functions in emotional tasks (Bush et gruent. Although the sources of attentional al., 2000; Bush, 2004). Thus, it is rea- Recently, Jin Fan and his col- effects are limited to networks, atten- sonable to consider the ACC an im- leagues showed that three separate tion can influence any part of the portant node in the monitoring and conflict-resolution tasks activate a brain, including the brain’s primary resolution of conflict that occurs in common conjunction area involving sensory areas and circuits governing emotional and cognitive regulation the ACC, together with a conjunction emotions. This influence was nicely (Posner, 2004c). area corresponding to the prefrontal demonstrated in a summary of many Many of the tasks used to activate cortex (Fan et al., 2003a). These data studies looking at the role of the ACC the ACC are Stroop-like tasks that highlight the brain network involved in in the monitoring and resolution of require language. However, research- aspects of self-regulation. conflict (Bush et al., 2000). The exact ers have developed some language- It was in this context that the atten- mental operations performed by the independent tasks suitable for young tion network task (ANT) was devel- ACC remain in dispute (Bush, 2004). children. For example, an adaptation oped (Fan et al., 2002). The ANT, a However, overall the ACC is involved of the Flanker task (Eriksen and Erik- variation of the Eriksen Flanker task, in self-regulation, in which subjects sen, 1974) asks a child the direction requires a subject to determine are required to damp down or ward of a swimming fish that is surrounded whether the central arrow points off negative thoughts (Ochsner et al., by other fish swimming in the same or rightward or leftward while the sur- 30 THE ANATOMICAL RECORD (PART B: NEW ANAT.) FEATURE ARTICLE

Box 2. (Continued)

rounding arrows may be either congruent or incongruent (Box 2 Fig. 2). This design permits researchers to measure the subject’s ability to resolve conﬂict, or cognitive regulation, by subtracting the congruent reaction time from the incongruent reaction time. However, before the task starts, the subject is given a cue, which may be no cue at all, a double cue (alerting the subject to get ready because the target is about to occur), or it may be a cue orienting the subject to the target’s location. This design permits two additional subtractions in order to extract both a measure of alerting and a measure of orienting. Analysis of reaction time data reveals little correlation between these three tasks. Although small interactions do occur, such as when the orienting cue occurs at the target location, there appears to be relative independence between the congruency condition and the alerting and orienting conditions. Using fMRI, Jin Fan has also shown activation of unique neuroana- tomical sites for each attentional network. The alerting network elicits a strong thalamic activation, while orienting activates superior and inferior areas of the parietal lobe and the con- ﬂict or executive network induces activation at the ACC and lateral prefrontal cortex. Whereas many other studies have measured these attentional networks separately, the ANT allows measurement of all attentional networks within one relatively short task that can be run with children and animals, as well as adults and patients. Box 2 Figure 2. Experimental procedure for the ANT. a: The four cue conditions. b: The six stimuli used in the experiment. c: The task as it unfolds in time. Adapted with permission from Fan et al. (2002).

Box 2. TABLE 1. Attentional networks: brain regions and neuromodulators

Attentional network Neuroanatomy Neuromodulator Select (executive, conﬂict, supervisory, Anterior cingulate cortex; lateral ventral prefrontal Dopamine focal, metacognitive attention) cortex; basal ganglia Orient Superioparietal; temporoparietal; frontal eye ﬁelds; Acetylcholine superior colliculus Alert (arousal, vigilant attention) Locus coeruleus: right frontal and parietal cortex Norepinepherine FEATURE ARTICLE THE ANATOMICAL RECORD (PART B: NEW ANAT.) 31

Box 3. Development and Training of Attentional Networks

The ANT (see Box 2) has been carefully crafted for use with children by replacing the symbolic flanker arrows with age-appropriate fish stimuli, adding the motivating story of feeding the central fish, and providing brief audiovisual feedback (Box 3 Fig. 1). Behavioral results from two studies in healthy pediatric populations show that the three scores of the fish ANT (FANT) do not correlate and interactions do not exist between them (Rueda et al., 2002). (FANT was coined as a tribute to Jin Fan, who, together with Michael I. Posner and colleagues, developed the ANT and its versions. A combination of “Fan” and “ANT,” the FANT refers to the child ANT.) Whereas there is stable improvement in latencies and errors between 6- and 10-year-old subjects, FANT scores do not show reliable Box 3 Figure 1. Sketch of the sequence of (F)ANT trials. Scores are obtained using the changes at these ages. Direct com- following subtractions: select ϭ (RT incongruent) Ϫ (RT congruent); orient ϭ (RT central parison of children and adults sug- cue) Ϫ (RT spatial cue); alert ϭ (RT no cue) Ϫ (RT double cue). See BOX 2 for more gests that the alerting and conflict information. Adapted with permission from Rueda et al. (2004a). networks, but not the orienting network, continue to improve with age. The development of attentional 7, alertness can change up to and level functions. The alerting network, networks must involve both genes beyond age 10. Although alerting is however, continues to develop and specific experience (Fossella et steady during middle childhood, effi- throughout adolescence and into al., 2002a,b; Fan et al., 2003c). The ciency scores for children are much adulthood. Clearly, children’s diffi- executive attention network as in- higher than in adults. Perhaps higher culty maintaining a task set in the ab- dexed by the ANT shows a strong alerting scores reflect a child’s diffi- sence of a cue telling them to get development from 4 to 7 years of age, culty in maintaining the alert state ready is an important aspect of later but does not seem to change from without a cue. Orienting scores, how- childhood and relates to the ability to age 7 to adulthood (Rueda et al., ever, do not vary by age in this study; maintain task sets without specific in- 2002, 2004a,b). During this period, orienting to the correct location pro- structions. the efficiency of executive control as duces a similar improvement in reac- These collective findings indicate measured by the ANT also predicts a tion time across all ages. Further- the possibility of tracing differences in number of behavioral and question- more, ANT data collected from 40 the development of these networks naire measures of attentional regula- 7-year-old children suggest that, as for children with different genotypes tion among normal individuals (Pos- with adults, efficiency scores for the and attentional capacities. Research- ner and Rothbart, 2000; Rothbart et three attentional networks are inde- ers have developed training exercises al., 2004). pendent under some conditions. designed to influence the rate of de- Complementing earlier develop- Lack of correlation may also reflect velopment of attentional networks mental assays (Akhtar and Enns, unreliability rather than indepen- (Berger et al., 2000). Using these ex- 1989), one recent study involves a dence. ercises on children with different ge- cross-sectional experiment with four Whereas executive attention pla- netic and attentional backgrounds age groups ranging from six through teaus after 7 years of development, presents the opportunity for specific nine years of age, along with compar- the orienting network seems to be studies of genetic-environmental in- isons of children (age ten) and adult formed as early as age 4. In more teraction and individual differences performance on both child and adult complex situations, where orienting is (Fan et al., 2003; Posner, 2004c; versions of the ANT (Rueda et al., combined with different dimensions, Sommer et al., in press). 2002, in press). The findings reveal there may be a later development. Understanding how attentional net- that latency and accuracy improve at The same might be true of many con- works function in the everyday life of each age interval and positive values flict tasks that rely not only on the a child is a key reason for studying exist for the average efficiency of network that has been discussed but the development of these networks. each of the networks. Whereas exec- also on other networks such as those Potential effects on learning and utive attention seems stable after age involved in planning and other higher- training are of particular conse- 32 THE ANATOMICAL RECORD (PART B: NEW ANAT.) FEATURE ARTICLE

Box 3. (Continued)

quence. Whereas training in non- adult-likepattern. Additionally, the data suggest that working memory human animals can persist for hun- data also show a striking increase in training potentially could be of clin- dreds of thousands of trials, as- IQ, suggesting the generalization of ical use for ameliorating the symp- sessing the effects of training in these effects. Specifically, the exper- toms in ADHD. Training efforts human and especially in children imental group improves significantly similar to the above-mentioned at- can be difficult. Training for only a in the matrix portion of the child ver- tentional assays have been re- few days would be unlikely to have sion of the IQ test. Thus, attentional ported in adults and children with a large effect on behavior, but may training appears to improves the un- ADHD (Posner, 2004c). As neuro- cause minimal changes that would derlying neural networks involved in imaging begins to unravel the ef- indicate greater potential benefits conflict resolution and may even gen- fects of practice on brain sub- of more extensive attentional train- eralize to different tasks (Posner, strates (Garavan et al., 2004; ing. Given this premise, Posner and 2004c). Landau et al., 2004), cumulative his colleagues conducted a 5-day Similar findings have been reported data suggest that these attentional attentional training pilot. On the in 8-year-old children diagnosed with networks can be modified. Intro- first and last day, they examined ADHD. After completing a training ducing attentional training in pre- the children’s EEG during the ANT program to strengthen working mem- schools or childcare centers may task, as well as their behavioral ory, participants have significantly be a valuable innovation. Atten- data, IQ, and temperament mea- improved motor performances and IQ tional training might be of particular sures. Preliminary analyses indi- results on the nonverbal complex rea- importance because it forms spe- cate that even a brief 5-day atten- soning task of the Raven’s progres- cific neural networks involved in tional training shifts the attentional sive matrices (Klingberg et al., 2002, multiple domains of behavior (Pos- networks of children toward a more in press; Olesen et al., 2004). These ner, 2004c). ing, error detection, novel or not well- dorsal ACC, and parts of the basal the visual word-form area, which ar- learned responses, difficult or danger- ganglia. These experimental tasks ranges visual letters into words. Oth- ous conditions, and in overcoming provide a means of fractionating the ers similarly report that in highly habitual actions. The ACC and lateral functional contributions of areas suggestible individuals, hypnotic in- frontal cortex are target areas of the within the executive attention net- structions to view a color image in ventral tegmental dopamine system. work. Chiefly, the ACC is more active gray scale prevents normal activation Brain imaging data have repeatedly on incongruent than on congruent tri- of the color areas of the prestriate cor- shown that the ACC is an important als. This difference could reflect the tex (Kosslyn et al., 2000). These data node in this network. More specifi- general finding that lateral areas are imply that, at least in highly suggest- cally, neuroimaging studies have involved in representing specific in- ible individuals, attentional manipu- shown activation of the dorsal ACC in formation over time, while medial ar- lations can influence aspects of self- tasks requiring people to respond ei- eas are more related to the detection regulation by affecting neural activity ther to a prepotent response or to a of conflict. in specific brain areas. These findings rather strong conflicting dimension. Patients with focal brain lesions of support the use of special attentional For example, in the classic Stroop the ACC initially display deficits of interventions such as hypnosis to task, experienced readers name the voluntary behavior. The notion of study related brain processes (Raz ink color of a displayed word (Stroop, ACC involvement in cognitive control and Shapiro, 2002). 1935). Subjects are usually slower and and volition has been a topic of much less accurate in responding to the ink interest lately. Based on behavioral, ANT AND LANT color of an incompatible color word, optical, and neuroimaging data, we such as the word “RED” displayed in recently reported that effective post- Aspects of laterality and development blue ink, than in identifying the ink hypnotic suggestion, or verbal exhor- may operate within attentional net- color of a control item (e.g., “LOT” tation, to construe words as meaning- works. For example, human vigilance inked in red; Fig. 3). Another such less symbols can modulate focal brain networks tend to be right-lateralized, task involves subjects responding to activity in highly suggestible individ- while orienting appears to be specific the direction of a central arrow when uals (Raz et al., 2002, 2003; Raz, to both modality and hemisphere. flanking arrows could either point in 2004). This top-down influence is both PET data collected from healthy indi- the same (congruent) or opposite (in- potent and selective: it removes viduals show that the right parietal congruent) direction (see Box 2 Fig. Stroop conflict and reduces neuroim- lobe is involved in attentional shifts 2). The ANT uses this flanker task to aging signals in both the ACC and ex- on either side, whereas the left pari- measure conflict. Such conflict tasks trastriate cortex (Raz, 2004). Whereas etal lobe was involved in attentional have been shown by neuroimaging a drop in ACC activity is likely related shifts on the right side only. Patients studies to activate the lateral prefron- to the abrogation of conflict, the oc- with right parietal damage exhibit im- tal cortex, frontal areas such as the cipital-parietal area may be related to paired leftward orienting for both vi- FEATURE ARTICLE THE ANATOMICAL RECORD (PART B: NEW ANAT.) 33

Figure 4. LANT vs. ANT. A lateralized version of the ANT, the LANT assesses the attentional capacity of each hemisphere and has been used to compare hemispheric attentional networks in children and in adults. C, conflict network; O, orienting network; A, alerting network. sual fields. When the corpus callosum pending on the training side involved. research with the LANT will likely fur- is intact and the two cerebral hemi- On the LANT, conflict and alerting ther elucidate issues related to atten- spheres are in normal communica- were unaffected, but neurofeedback tion and laterality. tion, a single focus of orienting is enhanced orienting in the left visual maintained. However, when the cal- field (Barnea et al., 2004b,c, in press lossal commissures are disconnected, c). Overall, the neurofeedback train- CONCLUSION as in split-brain patients, each hemi- ing reduced conflict and increased Researchers in the field agree that at- sphere has a partly separate focus for alerting in the ANT but not in the orienting (Zaidel, 1995). From a mo- tention is not a unitary term. Rather, LANT. Therefore, ANT stimuli likely we can fractionate attention into sub- dality perspective, visual, tactile, and engage complex interhemispheric in- auditory neglect seem to involve dif- systems of more circumscribed func- teractions as each hemisphere orches- ferent areas of the parietal lobe. None- tion and anatomy. Neuroimaging of trates different bilateral control cir- theless, the phenomenon of cross- the human brain permits us to view cuits. modal integration, in which visual attention as an organ system with its Another study examined the role of information influences the location own unique anatomy. This attitude fa- from which we hear sounds, suggests developmental stage on L/ANT perfor- cilitates detailed examination of the that these areas do communicate. mance (Barnea et al., 2004a). Twenty- synaptic, cellular, and genetic basis of Eran Zaidel and colleagues recently five young Israeli adults (25–30 years normal attentional networks as well addressed these issues by using a lat- of age) and fourteen Israeli children as pathological processes. Moreover, eralized ANT (LANT) to examine (10–12 years of age) performed both we can then link attention to the study hemispheric asymmetries associated the ANT and the LANT. The results of brain states that change both as a with the development and perfor- show that in both adults and children, function of arousal level (from wake- mance of attentional networks (Bar- the LANT generally yields larger, fulness through hypnosis to deep nea et al., 2004a). A recent study ex- more reliable estimates of the three sleep) and across development, from plored the potential effects of networks. In adults given the L/ANT, infancy to adulthood. Studies of atten- neurofeedback, an operant condition- conflict and orienting in the ANT se- tion suggest that our subjective expe- ing protocol for self-regulating brain lectively correlate with the LANT in rience and cognitive control are asso- activation patterns by rewarding en- different brain hemispheres: ANT ciated with these brain networks. The hancement or reduction of specific conflict correlates with conflict in the characteristics of these networks and EEG frequencies measured over se- left hemisphere (LH) LANT conflict, lected scalp locations. In the study, and orienting in the ANT correlates their development in phylogeny and both ANT and LANT were adminis- with orienting in right hemisphere ontogeny thus become central issues tered to ten 11-year-old girls both be- LANT. Whereas conflict is equal and of brain research of particular impor- fore and after 20 half-hour neurofeed- significant in both hemispheres, the tance to psychology. Pioneered by Mi- back sessions. Preliminary analysis of RH specializes in orienting. Com- chael I. Posner, this approach pro- the data revealed that neurofeedback pared to adults, children have a larger vides a basis for considering the many improved overall performance. On the conflict in both the ANT (latency mea- pathologies of attention due to insults ANT, training decreased conflict and sures) and the LANT (sensitivity mea- to the adult brain, developmental dis- increased orienting and alerting, and sures). Orienting is significant only for orders, and gene-environment inter- neurofeedback had specific effects de- adults within the RH (Fig. 4). Further actions. 34 THE ANATOMICAL RECORD (PART B: NEW ANAT.) FEATURE ARTICLE

ACKNOWLEDGMENTS nitive neuroscience of attention. New Fan J, Wu Y, Fossella JA, Posner MI. I thank Michael I. Posner for his York: Guilford Press. p 207–218. 2001b. Assessing the heritability of attentional networks. BMC Neurosci 2:14. graceful mentorship, professional and Carrasco M, Williams PE, Yeshurun Y. 2002. Covert attention increases spatial Fan J, McCandliss BD, Sommer T, Raz A, personal inspiration, and continued resolution with or without masks: sup- Posner MI. 2002. Testing the efficiency friendship and support. I am also in- port for signal enhancement. J Vis 2:467– and independence of attentional net- debted to Jin Fan and Mindy Tanzola 479. works. J Cogn Neurosci 14:340–347. Fan J, Flombaum JI, McCandliss BD, for their consistent collaboration con- Cavanagh P. 2003. Attention routines. In: Thomas KM, Posner MI. 2003a. Cogni- cerning aspects of the content and International workshop on visual attentive and brain consequences of conflict. presentation, respectively. tion. San Miniato, Italy. Neuroimage 18:42–57. Cavanagh P. 2004. Attention Routines and Fan J, Raz A, Posner MI. 2003b. Atten- the Architecture of Selection. In: Posner tional mechanisms. In: Aminoff MJ, Da- MI, editor. Cognitive Neuroscience of At- roff RB, editors. Encyclopedia of neuro- LITERATURE CITED tention. New York: Guilford Press. p 13– logical sciences. New York: Elsevier Akhtar N, Enns JT. 1989. Relations be- 28. Science. p 292–299. tween covert orienting and filtering in Cheour M, Martynova O, Naatanen R, Fan J, Fossella J, Sommer T, Wu Y, Posner the development of visual attention. J Erkkola R, Sillanpaa M, Kero P, Raz A, MI. 2003c. Mapping the genetic varia- Exp Child Psychol 48:315–334. Kaipio ML, Hiltunen J, Aaltonen O, tion of executive attention onto brain activity. Proc Natl Acad Sci USA 100:7406– Atienza M, Cantero JL, Dominguez-Marin Savela J, Hamalainen H. 2002. Speech 7411. E. 2002. Mismatch negativity (MMN): an sounds learned by sleeping newborns. Fernandez-Duque D, Baird JA, Posner MI. objective measure of sensory memory Nature 415:599–600. 2000a. Awareness and metacognition. and long-lasting memories during sleep. Corbetta M, Miezin FM, Dobmeyer S, Conscious Cogn 9:324–326. Int J Psychophysiol 46:215–225. Shulman GL, Petersen SE. 1991a. Selec- Fernandez-Duque D, Baird JA, Posner MI. tive and divided attention during visual Barnea A, Rassis A, Neta M, Raz A, Zaidel 2000b. Executive attention and metacog- discriminations of shape, color, and E. 2004a. The lateralized attention net- nitive regulation. Conscious Cogn 9:288– speed: functional anatomy by positron work test (LANT) in children and adults. 307. emission tomography. J Neurosci 11: San Francisco: Cognitive Neuroscience Fernandez-Duque D, Posner MI. 2001. 2383–2402. Society. Brain imaging of attentional networks in Barnea A, Rassis A, Raz A, Othmer S, Corbetta M, Miezin FM, Shulman GL, Pe- normal and pathological states. J Clin Zaidel E. 2004b. The effect of neurofeed- tersen SE. 1991b. Selective attention Exp Neuropsychol 23:74–93. back on attention in children. San Fran- modulates extrastriate visual regions in Fossella J, Posner MI, Fan J, Swanson JM, cisco: Cognitive Neuroscience Society. humans during visual feature discrimi- Pfaff DW. 2002a. Attentional phenotypes Barnea A, Rassis A, Raz A, Othmer S, nation and recognition. Ciba Found for the analysis of higher mental func- Zaidel E. 2004c. Effects of neurofeed- Symp 163:165–175. tion. Sci World 2:217–223. back on hemispheric attention networks. Corbetta M, Kincade JM, Ollinger JM, Fossella J, Sommer T, Fan J, Wu Y, Swan- Tennet: an annual meeting of theoretical McAvoy MP, Shulman GL. 2000. Volun- son JM, Pfaff DW, Posner MI. 2002b. and clinical neuropsychology. Montreal: tary orienting is dissociated from target Assessing the molecular genetics of at- Canada. detection in human posterior parietal tention networks. BMC Neurosci 3:14. Barnea A, Rassis A, Raz A, Othmer S, cortex. Nat Neurosci 3:292–297. Garavan H, Kelley D, Rosen A, Rao SM, Zaidel E. Effects of neurofeedback on Coull JT, Nobre AC, Frith CD. 2001. The Stein EA. 2004. Practice-related func- hemispheric attentional networks. Brain noradrenergic alpha2 agonist clonidine tional activation changes in a working Cogn (in press c). modulates behavioural and neuroana- memory task. Microsc Res Tech 51:54– Beane M, Marrocco RT. 2004. Cholinergic tomical correlates of human attentional 63. and noradrenergic inputs to the poste- orienting and alerting. Cereb Cortex 11: Hollingsworth DE, McAuliffe SP, Knowl- rior parietal cortex modulate the compo- 73–84. ton BJ. 2001. Temporal allocation of vi- nents of exogenous attention. In: Posner Deth RC, Kuznetsova A, Waly M. 2004. sual attention in adult attention deficit MI, editor. Cognitive neuroscience of at- Attention-related signaling activities of hyperactivity disorder. J Cogn Neurosci tention. New York: Guilford Press. p the d4 dopamine receptor. In: Posner 13:298–305. 313–325. MI, editor. Cognitive neuroscience of at- Hopfinger JB, Buonocore MH, Mangun Beauregard M, Levesque J, Bourgouin P. tention. New York: Guilford Press. p GR. 2000. The neural mechanisms of 2001. Neural correlates of conscious self- 269–282. top-down attentional control. Nat Neu- regulation of emotion. J Neurosci 21: Doran SM, Van Dongen HP, Dinges DF. rosci 3:284–291. RC165. 2001. Sustained attention performance James W. 1890. The principles of psychol- Berger A, Posner MI. 2000. Pathologies of during sleep deprivation: evidence of ogy. New York: Dover. brain attentional networks. Neurosci state instability. Arch Ital Biol 139:253– Kastner S, Ungerleider LG. 2000. Mecha- Biobehav Rev 24:3–5. 267. nisms of visual attention in the human Berger A, Jones L, Rothbart MK, Posner Driver J, Frackowiak RSJ, editors. 2001. cortex. Annu Rev Neurosci 23:315–341. MI. 2000. Computerized games to study Imaging selective attention in the human Kastner S, Ungerleider LG. 2001. The neu- the development of attention in child- brain. New York: Elsevier. ral basis of biased competition in human hood. Behav Res Methods Instrum Com- Duncan J, Emslie H, Williams P, Johnson visualcortex.Neuropsychologia39:1263– put 32:297–303. R, Freer C. 1996. Intelligence and the 1276. Broadbent DE. 1958. Perception and com- frontal lobe: the organization of goal- Kihlstrom JF. 1996. Perception without munication. New York: Pergamon Press. directed behavior. Cogn Psychol 30:257– awareness of what is perceived, learning Bronson MB. 2000. Self-regulation in early 303. without awareness of what is learned. In: childhood: nature and nurture. New Eriksen BA, Eriksen CW. 1974. Effects of Velmans S, editor. The science of con- York: Guilford Press. noise letters upon the identification of a sciousness: psychological, neuropsycho- Bush G, Luu P, Posner MI. 2000. Cognitive target letter in a nonsearch task. Percept logical and clinical reviews. London: and emotional influences in anterior cin- Psychophys 16:143–149. Routledge. p 23–46. gulate cortex. Trends Cogn Sci 4:215– Fan J, McCandliss BD, Flombaum JI, Pos- Klingberg T, Forssberg H, Westerberg H. 222. ner MI. 2001a. Imaging attentional net- 2002. Training of working memory in Bush G. 2004. Multimodal studies of cin- works. San Diego, CA: Annual Meeting children with ADHD. J Clin Exp Neuro- gulate cortex. In: Posner MI, editor. Cog- of the Society for Neuroscience. psychol 24:781–791. FEATURE ARTICLE THE ANATOMICAL RECORD (PART B: NEW ANAT.) 35

Klingberg T, Fernell E, Olesen PJ, Johnson Posner MI. 2004a. Cognitive neuroscience motor acts in inferior area 6 of the ma- M, Gustafsson P, Dahlstro¨m K, Gillberg of attention. New York: Guilford Press. caque monkey. Exp Brain Res 67:220– CG, Forssberg H, Westerberg H. Com- Posner MI. 2004b. The achievement of 224. puterized training of working memory in brain imaging: past and future. In: Kan- Robbins TW, Milstein JA, Dalley JW. 2004. children with ADHD: a aontrolled, ran- wisher N, Duncan J, editors. Attention Neuropharmacology of attention. In: domized, double-blind trial. J Am Assoc and performance XX. New York: Oxford Posner MI, editor. Cognitive neuro- Child Adolesc Psych (in press). University Press. p 505–528. science of attention. New York: Guilford Posner MI. 2004c. Tenth annual George A. Kosslyn SM, Thompson WL, Costantini- Press. p 269–282. Ferrando MF, Alpert NM, Spiegel D. Miller distinguished lecture at the an- 2000. Hypnotic visual illusion alters nual meeting of the Cognitive Neuro- Robertson IH. 1999. The rehabilitation of color processing in the brain. Am J Psy- science Society. San Francisco: Cogni- attention. In: Stuss DT, Winocur G, Rob- chiatry 157:1279–1284. tive Neuroscience Society. ertson IH, editors. Cognitive rehabilita- LaBerge D. 1994. Quantitative models of Rainville P, Duncan GH, Price DD, Carrier tion. Cambridge: Cambridge University attention and response processes in B, Bushnell MC. 1997. Pain affect en- Press. p 302–313. shape identification tasks. J Math Psy- coded in human anterior cingulate but Robertson IH, Garavan H. Vigilant atten- chol 38:198–243. not somatosensory cortex. Science 277: tion. In: Gazzaniga MS, editor. The new Landau SM, Schumacher EH, Garavan H, 968–971. cognitive neurosciences. New York: MIT Druzgal TJ, D’Esposito M. 2004. A func- Raymond JE, Shapiro KL, Arnell KM. Press (in press). tional MRI study of the influence of 1992. Temporary suppression of visual Rothbart MK, Ellis LK, Rueda MR, Posner practice on component processes of processing in an RSVP task: an atten- MI. 2003. Developing mechanisms of working memory. Neuroimage 22:211– tional blink? J Exp Psychol Hum Percept temperamental effortful control. J Pers 221. Perform 18:849–860. 71:1113–1143. Raz A. 1999. The effects of total sleep de- Manly T, Lewis GH, Robertson IH, Watson Rothbart MK, Ellis LK, Posner MI. 2004. privation on the spotlight of visual atten- PC, Datta AK. 2002. Coffee in the corn- Temperament and self regulation. In: tion and on pre-attentional processing. flakes: time-of-day as a modulator of ex- Baumeister RF, Vohs KD, editors. Hand- Jerusalem: Hebrew University. ecutive response control. Neuropsycho- book of self regulation. New York: Guil- Raz A, Posner MI. 2000. Attention under logia 40:1–6. ford Press. p 357–370. sleep deprivation: mismatch negativity Marrocco RT, Davidson MC. 1998. Neuro- Rueda MR, Fan J, Halparin J, Gruber D, chemistry of attention. In: Parasuraman altered by attentional state. J Cogn Neu- rosci 85C:81. Posner MI. 2002. Assaying the develop- R, editor. The attentional brain. Cam- ment of attentional networks in six to ten bridge, MA: MIT Press. p 35–50. Raz A, Deouell LY, Bentin S. 2001. Is preattentive processing compromised by year old children. San Francisco: Annual Naatanen R. 2003. Mismatch negativity: Meeting of the Cognitive Neuroscience clinical research and possible applica- prolonged wakefulness? effects of total sleep deprivation on the mismatch neg- Society. tions. Int J Psychophysiol 48:179–188. Rueda MR, Fan J, McCandliss BD, Halpa- Navon D. 2003. What does a compound ativity. Psychophysiology 38:787–795. Raz A, Shapiro T. 2002. Hypnosis and neu- rin JD, Gruber DB, Lercari LP, Posner letter tell the psychologist’s mind? Acta MI. 2004a. Development of attentional Psychol (Amst) 114:273–309. roscience: a cross talk between clinical and cognitive research. Arch Gen Psychi- networks in childhood. Neuropsycholo- Ochsner KN, Bunge SA, Gross JJ, Gabrieli gia 42:1029–1040. JD. 2002. Rethinking feelings: an FMRI atry 59:85–90. Raz A, Shapiro T, Fan J, Posner MI. 2002. Rueda MR, Posner MI, Rothbart MK. study of the cognitive regulation of emo- 2004b. Attentional control and self tion. J Cogn Neurosci 14:1215–1229. Hypnotic suggestion and the modulation of Stroop interference. Arch Gen Psychi- regulation. In: Baumeister RF, Vohs Olesen PJ, Westerberg H, Klingberg T. atry 59:1155–1161. KD, editors. Handbook of self regula- 2004. Increased prefrontal and parietal Raz A, Landzberg KS, Schweizer HR, Ze- tion. New York: Guilford Press. p 283– activity after training of working mem- 300. ory. Nat Neurosci 7:75–79. phrani ZR, Shapiro T, Fan J, Posner MI. 2003. Posthypnotic suggestion and the Rueda MR, Fan J, McCandliss BD, Halpa- Pessoa L, Kastner S, Ungerleider LG. 2003. modulation of Stroop interference under rin JD, Gruber DB, Lercari LP, Posner Neuroimaging studies of attention: from cycloplegia. Conscious Cogn 12:332– MI. Development of attentional net- modulation of sensory processing to top- 346. works in childhood. Neuropsychologia down control. J Neurosci 23:3990–3998. Raz A. 2004a. Atypical attention: hypnosis (in press). Posner MI, Fan J. 2004. Attention as an and conflict reduction. In: Posner MI, Schneider W, Shiffrin RM. 1977. Con- organ system. In: Pomerantz JR, Crai editor. Cognitive neuroscience of atten- trolled and automatic human informa- MC, editors. Topics in integrative neuro- tion. New York: Guilford Press. p 420– tion processing: I, detection, search and science: from cells to cognition. Cam- 429. attention. Psychol Rev 84:1–66. bridge: Cambridge University Press. Raz A. 2004b. Attention. In: Spielbereger Shalev L, Algom D. 2000. Stroop and Posner MI, Petersen SE. 1990. The atten- C, editor. Encyclopedia of applied psy- Garner effects in and out of Posner’s tion system of the human brain. Annu chology. San Diego, CA: Elsevier Science beam: reconciling two conceptions of Rev Neurosci 13:25–42. Academic Press. p 203–208. selective attention. J Exp Psychol 26: Posner MI, Raichle ME. 1996. Images of Raz A. 2004c. Neural substrates of atten- 997–1017. mind, rev. Washington, DC: Scientific tion. In: Adelman G, Smith BH, editors. Shapiro KL, Raymond JE, Arnell KM. American Books. Encyclopedia of neuroscience. New 1994. Attention to visual pattern infor- Posner MI, Rothbart MK. 1998. Attention, York: Elsevier Science Academic Press. mation produces the attentional blink in self-regulation and consciousness. Phi- Raz A, Marinoff GP, Landzberg KS, Guy- rapid serial visual presentation. J Exp los Trans R Soc Lond B Biol Sci 353: ton D. 2004a. Substrates of negative ac- Psychol Hum Percept Perform 20:357– 1915–1927. commodation. Binocul Vis Strabismus 371. Posner MI, Rothbart MK. 2000. Develop- Q 19:71–74. Sommer T, Fossella JA, Fan J, Posner MI. ing mechanisms of self-regulation. Dev Raz A, Marinoff GP, Zephrani ZR, Inhibitory control: cognitive subfunc- Psychopathol 12:427–441. Schweizer HR, Posner MI. 2004b. See tions, individual differences and varia- Posner MI, Rothbart MK. Hebb’s neural clearly: suggestion, hypnosis, attention, tion in dopaminergic genes. Proccedings networks support the integration of psy- and visual acuity. Int J Clin Exp Hypn of the Hanse Institute Frankfurt, Ger- chological science. Canadian Psycholo- 52:159–187. many. (in press). gist (in press). Rizzolatti G, Gentilucci M, Fogassi L, Lup- Spelke E, Hirst W, Neisser U. 1976. Skills Posner MI. 1980. Orienting of attention. Q pino G, Matelli M, Ponzoni-Maggi S. of divided attention. Cognition 4:215– J Exp Psychol 32:3–25. 1987. Neurons related to goal-directed 230. 36 THE ANATOMICAL RECORD (PART B: NEW ANAT.) FEATURE ARTICLE

Stroop JR. 1935. Studies of interference in Witte EA, Davidson MC, Marrocco RT. 1997. Yeshurun Y, Carrasco M. 1999. Spatial at- serial verbal reactions. J Exp Psychol 18: Effects of altering brain cholinergic activity tention improves performance in spatial 643–661. on covert orienting of attention: comparison resolution tasks. Vision Res 39:293–306. Talgar CP, Carrasco M. 2002. Vertical me- of monkey and human performance. Psy- Yeshurun Y, Carrasco M. 2000. The locus ridian asymmetry in spatial resolution: chopharmacology (Berl) 132:324–334. of attentional effects in texture segmen- visual and attentional factors. Psychon Wurtz RH, Goldberg ME. 1972. The primate tation. Nat Neurosci 3:622–627. Bull Rev 9:714–722. superior colliculus and the shift of visual Zaidel E. 1995. Interhemispheric transfer Witte EA, Marrocco RT. 1997. Alteration of attention. Invest Ophthalmol 11:441–450. in the split brain: long-term status fol- brain noradrenergic activity in rhesus Yeshurun Y, Carrasco M. 1998. Attention lowing complete cerebral commissurot- monkeys affects the alerting component improves or impairs visual performance omy. In: Davidson RH, Hugdahl, K, edi- of covert orienting. Psychopharmacol- by enhancing spatial resolution. Nature tors. Brain asymmetry. Cambridge, MA: ogy (Berl) 132:315–323. 396:72–75. MIT Press. p 491–532.