aging 148 (2006) 23–31www.elsevier.com/locate/psychresns

Psychiatry Research: Neuroim

Dysfunction of dorsolateral prefrontal cortex in antipsychotic-naïveschizophreniform psychosis

Ben J. Harrisona,⁎, Murat Yücela,b,⁎, Marnie Shawc, Warrick J. Brewera,b,Pradeep J. Nathand, Stephen C. Strothere, James S. Olverf, Gary F. Egang,

Dennis Velakoulisa, Patrick D. McGorryb, Christos Pantelisa,g

aMelbourne Neuropsychiatry Centre, Department of Psychiatry, The University of Melbourne, AustraliabORYGEN Research Centre and the Early Psychosis Prevention and Intervention Centre (EPPIC), Department of Psychiatry,

The University of Melbourne, and NorthWestern Mental Health Program, Melbourne Health, AustraliacCognitive Neuroscience Laboratory, School of Psychology, Flinders University, Australia

dBehavioural Neuroscience Laboratory, Department of Physiology, Monash Centrefor Brain and Behavior, Monash University, AustraliaeThe Rotman Research Institute, University of Toronto, Ontario, Canada

fCentre for Positron Emission Tomography, Austin Hospital, The University of Melbourne, AustraliagHoward Florey Institute, The University of Melbourne, Australia

Received 6 December 2005; received in revised form 31 January 2006; accepted 13 February 2006

Abstract

Reports of abnormal activation of the dorsolateral prefrontal cortex (dlPFC) are common in functional neuroimaging studies ofschizophrenia, although very few have examined brain activity in patients close to the onset of illness. In this H2

15O PET study,eight young male patients with first-episode schizophreniform psychosis and age-matched control subjects performed a version ofthe Stroop task that we have previously shown to engage the middle-frontal gyrus. At the time of testing, patients wereantipsychotic-naïve and were scanned within 1 week of initial contact with our clinical program. All patients received a laterdiagnosis of schizophrenia 6 months after participating in the study. Whole-brain (within-group) and region-of-interest (between-group) analyses were carried out and data underwent spatial reproducibility testing. Compared with healthy subjects, patientsshowed significantly greater reaction-time (RT) interference but normal RT accuracy on the Stroop task. This pattern correlatedwith significant under-activation of the posterior left middle-frontal gyri in the patient versus control group. These findings supportan emerging model of impaired cognitive control in schizophrenia and suggest that there is significant dysfunction of the dlPFCclose to the onset of illness that may coincide with, or be modulated by, the transition-to-illness phase.© 2006 Elsevier Ireland Ltd. All rights reserved.

Keywords: Schizophrenia; Dorsolateral prefrontal cortex; Cognitive control; First-episode psychosis

⁎ Corresponding authors. Ben Harrison (electronic). Murat Yücel(postal) Melbourne Neuropsychiatry Centre, Department of Psychia-try, The University of Melbourne, Level 3, National NeuroscienceFacility, 161 Barry Street, Carlton, Melbourne, Australia. Tel.: +61 38344 1877; fax: +61 3 8345 0599.

E-mail address: habj@unimelb.edu.au (B.J. Harrison).

0925-4927/$ - see front matter © 2006 Elsevier Ireland Ltd. All rights resedoi:10.1016/j.pscychresns.2006.02.006

1. Introduction

Extending early observations of reduced frontal lobemetabolism and blood flow changes in schizophrenia(i.e. resting “hypofrontality”; Ingvar and Franzén, 1974;Buchsbaum et al., 1982), the most reproducible finding

rved.

24 B.J. Harrison et al. / Psychiatry Research: Neuroimaging 148 (2006) 23–31

from positron emission tomography (PET) and functionalmagnetic resonance imaging (fMRI) studies of thisdisorder has been that patients show decreased task-related activation of the dorsolateral prefrontal cortex(dlPFC) (Andreasen et al., 1992; Weinberger et al., 1986;see also meta-analyses by Davidson and Heinrichs, 2003;Glahn et al., 2005). Although a more complex picture ofthis abnormality has emerged with studies using fMRI(Callicott et al., 2003; Manoach, 2003), the relationshipbetween impaired dlPFC activity and impaired highercognition in schizophrenia remains compelling. Recently,for example, studies have focused on a specific role for themiddle-frontal gyrus (Brodmann's area 9/46) in mediat-ing cognitive control (CC) deficits in patients withschizophrenia — the ability to coordinate one'sthoughts/actions in line with specific goals or task-oriented behaviors (Barch et al., 2001; MacDonald andCarter, 2003; MacDonald et al., 2005; Perlstein et al.,2003). This pattern has been observed most readily infMRI studies of stimulus-response compatibility para-digms, where patients show a reduced capacity to sustaintask-relevant (i.e. correct) responses, in the face ofdistraction from task-irrelevant (i.e. incorrect) items —a phenomenon that has been linked to impaired contextprocessing in schizophrenia (Cohen et al., 1999).

Importantly, and supporting the specificity of suchfindings in schizophrenia, CC-related hypofunction of thedlPFC has been characterized in chronically ill, medicatedand medication-naïve patients (Javitt et al., 2000; MacDo-nald and Carter, 2003; Perlstein et al., 2003) and has beendistinguished quantitatively from other psychiatric dis-orders (Holmes et al., 2005). However, questions stillremain about the timing and stability of dlPFC-CC deficitsin schizophrenia andwhether theremay be variation in theirexpression across different stages of illness. For instance,no studies have examined this in patients at the very earlieststage of illness where there is evidence for significant tem-poral flux in the anatomy and physiological integrity of thedlPFC (Pantelis et al., 2005). To this end, we have reportedsignificant progressive changes (atrophy) in dlPFC greymatter over the initial few years of illness that are apparentbefore illness onset (Pantelis et al., 2003a), while a recentfMRI study suggested a progressive worsening of dlPFChypofunction on a putative CC task between first-episodeand chronically ill patients versus relative normality inpeople identified at ultra-high risk (UHR) for psychosis(Morey et al., 2005). Hence, these findings suggest thatinferences of a static or trait-invariant dlPFC-CC dysfunc-tion in schizophrenia may be misleading.

In this H215O PET study, we examined dlPFC-CC

activity in a sample of young, antipsychotic-naïve patientswith schizophreniform psychosis, who each later transi-

tioned to schizophrenia and who, at the time of testing,were within a week of initial contact with our early inter-vention program after experiencing a first episode ofpsychotic symptoms. The critical difference between thisgroup of first-episode patients and those of previousstudies (Barch et al., 2001; MacDonald et al., 2005)relates to the duration of untreated psychosis (DUP). Inearlier studies, patients were potentially untreated for upto 5 years longer than the current sample and may havebeen a more severely affected group given the negativecorrelation of DUP and cognitive outcome in schizophre-nia (Amminger et al., 2002; Harrigan et al., 2003).Therefore, we considered it useful to study a sample ofpatients as close as possible to the onset of illness where itis more appropriate to test the null hypothesis of “no traitdeficit” of dlPFC-CC in schizophrenia.

In this study, subjects performed a classic stimulus-response compatibility paradigm, the Stroop task, thatin previous studies including our own has been shownto engage the middle-frontal gyrus in healthy subjects(Banich et al., 2000; Erickson et al., 2004; Harrison et al.,2005; Kerns et al., 2004; MacDonald et al., 2000; Milhamet al., 2003b). Under this paradigm, CC is estimated fromthe behavioral performance of subjects asked to name theprinted color of incongruent color word nouns (i.e. RED)where the word noun serves as a potent distraction to task-irrelevant word reading. For a dlPFC-CC deficit to exist,we considered that patientswould need to show significantunder-activity of the middle-frontal gyrus compared withhealthy subjects as well as a corresponding impairment ofreaction time (RT) and/or task accuracy (error) scores.

2. Methods

2.1. Subjects

Eight male patients with first-episode schizophreni-form psychosis (mean age 21.2±3 years) were recruitedduring initial treatment at the Early Psychosis Preven-tion and Intervention Centre (EPPIC), a program ofORYGEN Youth Health, Melbourne. All patients held acurrent diagnosis based on the Structured DiagnosticInterview for DSM-IV (SCID-I; First et al., 1998) and aconfirmed diagnosis of schizophrenia assessed at least 6months after their participation in the study. Patientswere scanned within 1 week following contact with theEPPIC; program admission criteria, as described else-where (McGorry et al., 1996), were age of onset be-tween 16 and 30 years and the presence of activepsychosis as reflected by at least one of the following: (i)delusions; (ii) hallucinations; (iii) disorder of thinking/speech, other than simple acceleration or retardation;

25B.J. Harrison et al. / Psychiatry Research: Neuroimaging 148 (2006) 23–31

and (iv) disorganized, bizarre, or markedly inappropri-ate behavior. Scores on the Positive and NegativeSyndrome Scale (PANSS; Kay et al., 1991) weretotal=50.3±7.5; positive symptoms=29.7±2.5; andnegative symptoms=20.5±6.9. Exclusion criteria in-cluded a significant current history of alcohol or illicitdrug dependence and a recent history of psychoactivemedication use, including steroids, or any other contra-indication to PET scanning.

Eight control subjects (7 male; 1 female) were alsorecruited by approaching ancillary hospital staff andtheir families or via local advertisements (mean age22.6±2 yrs). Control subjects were matched with thepatient group for age (t(14)=0.95,Pb0.36) and estimatedpremorbid IQ (t(1, 15)=1.16, Pb0.12; controls 108.5±9;patients 101.1±7) with the National Adult Reading Test(NART; Nelson and O'Connell, 1978). All participantsthat entered the study were screened for co-morbidmedical and psychiatric conditions by clinical assessment,and by physical and neurological examination. Allparticipants spoke English as a first language and pre-sented with adequate visual and auditory functioning.Five subjects in each group were smokers and three werenon-smokers. All participants gave written informedconsent to participation in this study, which was approvedby the Behavioral Research and Ethics Committees forthe North Western Mental Health Care Network,Melbourne and the Austin Hospital Human ResearchEthics Committee.

2.2. Stroop task and behavioral analysis

Subjects completed a version of the Stroop color-word paradigm (Stroop, 1935) that has been previouslyreported by our group (Harrison et al., 2005; Yücel et al.,2002). It consisted of sequential congruent (A) andincongruent (B) trials where each trial corresponded to acontinuous 6-s PET scan. Eight trials were presented in a4AB design on a computer monitor located approxi-mately 6 cm from the subject in the PET scanner. Foreach trial, 36 stimulus words were presented consecu-tively 3 mm above a fixation point (white cross) for1300 ms with an inter-stimulus interval (ISI) of 350 ms.Instructions specified that subjects attend to and name asquickly as possible the color of the print in which theword was written, without reading the word.

Voice onset latencies were recorded with a micro-phone that was fixed to the subject's mask, although itwas not visible to the subject. We determined the meanlatency of responses for each of the four congruent andfour incongruent conditions. Responses that were notclearly recorded, were abnormally fast (b100 ms) or that

were abnormally slow (N1200 ms) were excluded fromanalysis. The rate of excluded responses was comparablefor both groups and accounted for less than approxi-mately 10% of all responses made. We also calculated thenumber of errors made during the eight Stroop scan trials.Thesewere defined as errors due tomisses (omissions) anderrors due to incorrect verbalizations (commissions).‘Task’ (congruent or incongruent trial) by ‘Group’ (controlor patient) differences in vocalized reaction times (RTs)and commission error scores were examined usingrepeated measures analyses of variance (ANOVAs) andpost hoc comparisons in the Statistical Package for theSocial Sciences (SPSS) Version 11.

2.3. Image acquisition, preprocessing and analysis

Image acquisition and preprocessing parameters wereidentical to those previously published (Harrison et al.,2005; Yücel et al., 2002). For each subject, eight H2

15OPET scans (i.e. 4AB task pairs) were acquired using aSiemens/CT1 951R ECAT PET scanner, which gen-erates 31 transaxial slices across an axial field of view of10.8 cm. PET images were reconstructed resulting indata volumes with 128*128*31 voxels (each of2.43*2.43*3.375 mm3). A high-resolution T1-weight-ed MRI was also acquired for each subject (GE Signa1.5T scanner, voxel size 0.9*0.9*1.4 mm3). Spatialrealignment of the individual PET images was performedin SPM2. Data were smoothed with a 12-mm FWHMGaussian filter. Normalization to standard space wasperformed using FSL (http://www.fmrib.ox.ac.uk/fsl/index.html).

Functional data were analyzed using NPAIRS 1.0. Fordetails of the NPAIRS program, visit http://www.neurovia.umn.edu/incweb/npairs_info.html or see Strother et al.(2002). Initial preprocessing involved volume mean nor-malization of scans (i.e. proportional scaling) and dimen-sionality reduction with principal components analysis(PCA), retaining 20 principal components (Harrison et al.,2005; Shaw et al., 2002). To characterize within-groupdifferences in task-related rCBF, each scan was classifiedas either congruent or incongruent, and the primarydifference in task-related variance between them wasdetermined. To do so, NPAIRS combines analyses withsplit-half resampling, which takes the specified data andrandomly divides it into two disjoint halves. Each half isthen separately analyzedwith a chosen statistical model; inthis case, canonical variate analysis (CVA) with 35 splits,and the two results are compared. Among other metrics,NPAIRS estimates the spatial reproducibility of covariancepatterns produced by the split analyses. Reproducibility isderived from the Pearson product correlation coefficient

26 B.J. Harrison et al. / Psychiatry Research: Neuroimaging 148 (2006) 23–31

(r) of the scatter plots of resulting pairs of independentstatistical parametric maps (i.e. 35 pairs). The r-valuesfrom the data splits are then displayed as a histogram,which is further summarized by its median, avoidingpotential outlying r-values from influential subjects inindividual splits. These summarized patterns are thenexpressed as canonical variates (CVs) and associatedcanonical eigenimages (CEs) and the latter converted tomultivariate statistical parametric maps (for a detaileddescription of NPAIRS spatial reproducibility testing, seeStrother et al., 2002). Probability values corresponding toCEs are equivalent to an empirical correction for randomsubject effects. We classified results as significant ifreaching peak height probability of Puncorrectedb0.001 andN20 contiguous voxels.

Table 1Within-group pattern of brain activity associated with performance of the St

Region Brodmannarea

Patients

Clustersize

Voxel co-ordinat

x y

ActivationCerebellum 639 16 −52

647 −13 −36

Thalamus 53 −22 −34Superior occipital gyrus 18/19 119 24 −94Parahippocampal gyrus 28 73 22 −10Anterior-cingulate cortex 32 211 −8 12

4 12Superior frontal gyrus 6Middle-frontal gyrus 9/46 166 −46 10

−45 2Pre/post central gyrus 4/6 193 0 −26Brainstem

Orbital frontal gyrus

DeactivationSuperior temporal gyrus 38Parahippocampal gyrus 36Medial temporal gyrus 20Inferior frontal gyrus 11 155 −42 26Orbital frontal gyrus 47 84 −30 36Lingual gyrus 18 173 −6 −88Fusiform gyrus 20 62 −44 −70Superior frontal gyrus 8 122 −28 22Inferior temporal gyrus 20 116 46 −12Gyrus cuneus 31/7

Mid-cingulate gyrus 31Superior parietal cortexSuperior occipital gyrus 18Middle-frontal gyrus 10 255 −32 60

Activities are reported if exceeding a minimum cluster extent of at least 30 cox, y, z co-ordinates are reported in MNI-to-Talairach space. The NPAIRcorresponded to canonical correlations of 0.93 and 0.92 for patients and con

To test for group differences in functional activationbetween controls and patients, we also performed aconfirmatory region-of-interest (ROI) analysis of themiddle-frontal gyrus. This ROI approach was chosenbecause our primary hypothesis involved this brain re-gion, but also to reduce the risk of false-positive activa-tion (Type 1 error) when comparing small groups ofsubjects across whole-brain volumes. The actual se-lection of the ROI was based on within-group results(Table 1), which indicated that both patients and controlsengaged an overlapping area of the middle-frontal gyrus.The inclusive ROI dimensions were; x=−40/−56 mm;y=−6/+14 mm; z=+16/+46 mm). The analysis itselfinvolved identical procedures to that described above,with the addition of a dependent variable distinguishing

roop task

Healthy subjects

es Z-score

Clustersize

Voxel co-ordinates Z-score

z x y z

−20 5.02 906 −11 −54 −34 4.16−36 4.53 242 38 −58 −48 3.99

356 26 −58 −30 3.784 3.04

28 3.64−34 3.3142 3.2336 3.00

354 −8 −4 56 3.8320 3.32 335 −42 10 22 3.4132 3.0456 3.17 275 −24 −26 58 3.59

269 0 −28 −18 3.564 −36 −18 3.47

101 48 32 −16 3.44

564 24 −8 −32 4.1428 −12 −32 3.79

1065 −48 8 −24 3.80−8 3.60 −26 16 −26 3.72−36 3.39 282 2 44 −26 4.04

6 3.40−20 3.02 142 −42 −70 6 3.2256 3.40 109 26 24 50 3.18

−40 3.38 −48 −10 −28 3.67135 −10 −76 20 3.28375 −4 −50 38 3.73120 6 −26 42 3.24226 38 −52 54 3.41121 44 −66 28 3.10

4 3.27

ntiguous voxels at a probability thresholding of Puncorrectedb0.001. TheS (split-half) reproducibility values of these brain activity patternstrols, respectively.

27B.J. Harrison et al. / Psychiatry Research: Neuroimaging 148 (2006) 23–31

the two study groups. This tested for any significantdifference in (Stroop) task-related activation of thedlPFC between controls and patients (Puncorrectedb0.001,and N20 contiguous voxels).

3. Results

3.1. Behavioral

Mean reaction-time (RT) scores for the congruent andincongruent conditions were 585/775 ms (S.D.=86/62 ms) for controls and 571/871 ms (S.D.=75/102 ms)for patients (Fig. 1). Repeated measures ANOVArevealed a significant main effect of task condition(F(1, 14)=93.6,Pb0.001) and a task-by-group interactionfor RT performance (F(1, 14)=4.77, Pb0.04). Patientsshowed proportionally greater slowing only in theincongruent condition, signifying a greater RT interferenceeffect (190 versus 300 ms; F(1, 14)=5.36, Pb0.03). Therewere no main effects of task condition (F(1, 14)=3.0,Pb0.17) or task-by-group interaction (F(1, 13)=1.81,Pb0.20) for task accuracy/error scores. However, it shouldbe noted that rates of error were very low across the 4ABtrials, accounting for less than 2% of the total responsesmade for patients and controls, respectively (controls 1.0%;patients 1.1%).

Fig. 1. Mean and standard deviation of reaction-time performance onthe Stroop task.

3.2. PET rCBF

Task-related rCBF activations corresponding to with-in-group canonical eigenimage (CE) results for patientsand control subjects are given in Table 1. These CEscorresponded to canonical correlations of 0.93 and 0.92for patients and controls, respectively, indicating within-group reproducibility. For patients, significant rCBF acti-vation during Stroop interferencewas observed bilaterallyin the cerebellum, thalamus, dorsal anterior-cingulatecortex, right extrastriate cortex and parahippocampal gyri,left primary motor area and left middle-frontal gyri. Forcontrol subjects, significant rCBF activation was ob-served bilaterally in the cerebellum, left primary andsupplementary motor areas, brainstem and middle-frontalgyri and right orbital prefrontal cortex.

Confirmatory ROI analysis of task-related rCBFactivity of left middle-frontal gyri between patients andcontrols revealed significantly greater activity in the con-trol group (Z=3.77, Pb0.001; x, y, z=−48, 13, 32; BA 9/46) (Fig. 2).

3.3. Omnibus brain–behavioral correlation

Pearson's product-moment correlations (one-tailed,simple regression) were carried out between subjects'within-group CV scores (i.e. omnibus whole-brain acti-vity estimate) and RT performance. Because of the subjectnumbers in this study, additional correlations betweensubjects' functional, behavioral and demographic/clinicalvariables were limited due to the issue of multiple com-parisons. For both groups, there were significant overallpositive correlations between the CV and RT measures(controls, r=0.74, Pb0.001; patients, r=0.78, Pb0.001).These correlations reflect a certain degree of functional-specificity with the current paradigm, with higher CVscores covarying with higher RT scores (i.e. indexingincongruent task activity) and vice versa, thus reflectingthe relative cognitive demands of the two Stroopconditions on rCBF activity.

4. Discussion

Disturbance of cognitive control (CC), the ability tocoordinate one's thoughts and actions in line withspecific goals or task-oriented behaviors, has been linkedto a range of phenomenological features of schi-zophrenia, including disorganization symptoms andworking memory deficits and, in recent fMRI studies,has been ascribed to a primary dysfunction of the dlPFC(Barch et al., 2001; MacDonald and Carter, 2003;Perlstein et al., 2001). In this study, we tested the dlPFC-

Fig. 2. Posterior left middle-frontal gyri activation during Stroop task performance in antipsychotic-naïve patients with schizophreniform psychosis(left) and age/IQ matched healthy control subjects (right). Reproducible Z-score activations are displayed at a range 2.33 to 5.0 to aid visualization ofclusters within dlPFC region exceeding a probability threshold of Puncorrectedb0.001.

28 B.J. Harrison et al. / Psychiatry Research: Neuroimaging 148 (2006) 23–31

CCmodel by examining a group of young antipsychotic-naïve patients with schizophreniform psychosis whowere later diagnosed with schizophrenia and, at the timeof testing, had only recently experienced their firstepisode of psychotic symptoms. Specifically, we soughtto test whether these patients would show a less severepattern of dlPFC-CC dysfunction compared with exist-ing recent studies of first episode patients withpresumably longer durations of untreated psychosis(Barch et al., 2001; MacDonald et al., 2005). However,contrary to the null hypothesis of ‘no trait deficit’ ofdlPFC-CC in patients at an early stage of illness, patientsin this study did show significant under-activation of theleft middle-frontal gyri relative to healthy subjects aswell as evidence for a behavioral CC deficit. Together,these findings support the CC model of schizophreniaand suggest that there is significant physiologicaldysfunction of the dlPFC close to the onset of illness.

Consistent with our previous findings (Harrison etal., 2005) and also other functional imaging studies ofhealthy subjects (Erickson et al., 2004; Milham et al.,2003b), both groups showed significant task-relatedactivation of the dlPFC during Stroop task performance.These activations, which occurred in a region of the leftposterior middle-frontal gyrus, were almost identical forboth groups and, as reflected by a brain-wide activityestimate (i.e. CV scores), showed a positive correlationwith subjects' RT performance. Overall, these findingsappear to be consistent with multiple studies now thatimplicate this brain region as responsible for generatingCC on the Stroop task (Banich et al., 2000; Erickson etal., 2004; Kerns et al., 2004; MacDonald et al., 2000;Milham et al., 2003b). Specifically, the dlPFC maycontribute to Stroop performance by maintaining acontext for the task in working memory over time (i.e.

color naming versus word reading), and/or by biasingthe top-down processing of stimuli towards a correctcolor-naming response (Miller and Cohen, 2001).However, the primary implication of this study is thatdespite activating an almost identical region of thedlPFC during task performance, patients also showedsignificant under-activity of this region when comparedto healthy subjects.

Importantly, the pattern of reduced dlPFC activity infirst-episode patients was accompanied by a relativeimpairment in the speed of RT performance on theStroop task, where patients showed significantly greaterRT interference than control subjects, but showed normaltask accuracy. Although we suggest that this latterfinding should be considered with regards to moresophisticated studies of this paradigm in schizophrenia(Barch et al., 2004), it nevertheless seems reasonable thatpatients' reduced dlPFC activity contributed to theirreduced capacity for CC on the Stroop task, given thatboth dimensions were found to be correlated in thisstudy. It is also worth noting here that in studies of theStroop task in healthy subjects, the strength or efficiencyof dlPFC-CC has been inferred from the pattern ofdecreased activity in posterior brain regions (e.g.fusiform gyrus) due to their role in processing wordstimulus features, i.e., more deactivation, more inhibi-tion of task-irrelevant processing (Banich et al., 2001;Carter et al., 1995; Milham et al., 2003a,b, 2002). Whilenot a specific focus of our study, patients' did show lessextensive deactivation of posterior regions, which in turnmay also support our interpretation of dlPFC dysfunctionand impaired CC at this early stage of illness.

A broader implication of the current study's findingsbeyond disturbances of CC is that physiological dysfunc-tion of the dlPFC appears to exist soon after the onset of

29B.J. Harrison et al. / Psychiatry Research: Neuroimaging 148 (2006) 23–31

schizophrenia. This finding is supported by a handful ofother studies that have also characterized significant in-activation of the dlPFC during higher cognitive perfor-mance in first-episode patients (Boksman et al., 2005;MacDonald et al., 2005; Morey et al., 2005). However, aunique feature of our patient group compared with priorstudies is that they were likely to have been recruited atcloser proximity to illness onset given the specific earlyintervention focus of the EPPIC/ORYGEN program(McGorry et al., 1996). Such patients typically have amean duration of illness of 180 days (approx. 6months) atfirst contact, with a median of 49 days of untreated psy-chosis (Harrigan et al., 2003). Compared with otherstudies (e.g. Morey et al., 2005), this is reflected by thelowermean age of our patient group (early twenties versusmid-twenties) and shorter duration of illness (6 versus20 months). Nevertheless, the study by Morey et al.(2005) remains particularly relevant to our findingsbecause, in addition, they reported relatively intactdlPFC function in a group of individuals classified atultra-high risk (UHR) for psychosis. If it can be assumedthat this UHR group included a percentage of casesultimately to develop schizophrenia, then the dlPFCdysfunction that has been characterized in first-episodepatients would seem to coincide with or be modulated bythe transition-to-illness phase. While this requires furtherclarification, a primary insult to higher cognition aroundthe time of illness onset fits with our previous notion thatthose behavioral processes, namely executive functions,which normally optimize developmentally during the lateadolescent/early adulthood stage (i.e. the time of maximalrisk for psychosis), are accordingly, the most oftencompromised in a patient with schizophrenia (Panteliset al., 2003b). It would therefore seem important for futurestudies to focus on the functional pathophysiology of thisstage of illness and its links to normal brain maturation,and moreover, whether functional correlates during thisstage can be used as a guide to prognosis, treatmentresponse or prediction of outcome.

In closing, there are some limitations to this study thatdeserve consideration. This study involved the use of PETto examine brain activity in patient and control subjects.Although we argue that this method was sufficientlysuited to address our study aims, the use of fMRI,particularly event-related fMRI, offers better spatio-temporal resolution and seems the method of choice forfuture studies. While studies of antipsychotic-naïve first-episode patients often and understandably involve smallpatient numbers, this may limit their generalisability andideally should be extended. On the other hand, and incontrast to most functional imaging studies of schizo-phrenia, our results were derived through testing the spa-

tial reproducibility of data, which strengthens thegeneralisability of findings. This issue of reproducibilityin brain-activation studies of schizophrenia is particularlyrelevant for the dlPFC, where large differences in thespatial locations of activity have been reported amongindividual patients (see Manoach, 2003). However, ourfindings indicate that in a group of young, schizophreni-form patients there was a spatially reliable reduction ofdlPFC activity. Lastly, we consider dlPFC dysfunction torepresent only one, albeit a crucial aspect of impairedcognitive control in schizophrenia. Other studies of suchpatients that focus on more dynamic aspects of cognitivecontrol, in particular involving anterior-cingulate cortex(e.g. Kerns et al., 2005), may lead to a broaderunderstanding of schizophrenia's neural basis andspecific implications for this stage of illness.

Acknowledgments

This work was supported by National Health andMedical Research Council (NHMRC) grant 970599, theNHMRC Brain Research Network and Janssen-Cilag. Itwas undertaken to fulfill part of the requirement of aDoctor of Philosophy (PhD) to BJH funded by anAustralian Post-graduate Award (APA). The authorsthank colleagues and staff from the Department ofNuclear Medicine, Centre for PET, Austin Hospital.Murat Yücel is supported by an NH&MRC ProgramGrant (ID: 350241) and Melbourne NeuropsychiatryCentre is supported by the Department of Psychiatry,University of Melbourne and Melbourne Health.

References

Amminger, G.P., Edwards, J., Brewer, W.J., Harrigan, S., McGorry,P.D., 2002. Duration of untreated psychosis and cognitive dete-rioration in first-episode schizophrenia. Schizophrenia Research54, 223–230.

Andreasen, N.C., Rezai, K., Alliger, R., Swayze II, V.W., Flaum, M.,Kirchner, P., Cohen, G., O'Leary, D.S., 1992. Hypofrontality inneuroleptic-naive patients and in patients with chronic schizophrenia.Assessment with xenon 133 single-photon emission computed to-mography and the Tower of London. Archives of General Psychiatry49, 943–958.

Banich, M.T., Milham, M.P., Atchley, R.A., Cohen, N.J., Webb, A.,Wszalek, T., Kramer, A.F., Liang, Z., Barad, V., Gullett, D., Shah, C.,Brown, C., 2000. Prefrontal regions play a predominant role inimposing an attentional ‘set’: evidence from fMRI. Cognitive BrainResearch 10, 1–9.

Banich, M.T., Milham, M.P., Jacobson, B.L., Webb, A., Wszalek, T.,Cohen, N.J., Kramer, A.F., 2001. Attentional selection and theprocessing of task-irrelevant information: insights from fMRIexaminations of the Stroop task. Progress in Brain Research 134,459–470.

Barch, D.M., Carter, C.S., Braver, T.S., Sabb, F.W., MacDonald III, A.,Noll, D.C., Cohen, J.D., 2001. Selective deficits in prefrontal

30 B.J. Harrison et al. / Psychiatry Research: Neuroimaging 148 (2006) 23–31

cortex function in medication-naive patients with schizophrenia.Archives of General Psychiatry 58, 280–288.

Barch, D.M., Carter, C.S., Cohen, J.D., 2004. Factors influencingStroop performance in schizophrenia. Neuropsychology 18,477–484.

Boksman, K., Théberge, J., Williamson, P., Drost, D.J., Malla, A.,Densmore, M., Takhar, J., Pavlosky, W., Menon, R.S., Neufeld,R.W., 2005. A 4.0-T fMRI study of brain connectivity duringword fluency in first-episode schizophrenia. SchizophreniaResearch 75, 247–263.

Buchsbaum, M.S., Ingvar, D.H., Kessler, R., Waters, R.N., Cappelletti,J., van Kammen, D.P., King, A.C., Johnson, J.L., Manning, R.G.,Flynn, R.W., Mann, L.S., Bunney Jr., W.E., Sokoloff, L., 1982.Cerebral glucography with positron tomography. Use in normalsubjects and in patients with schizophrenia. Archives of GeneralPsychiatry 39, 251–259.

Callicott, J.H., Mattay, V.S., Verchinski, B.A., Marenco, S., Egan, M.F.,Weinberger, D.R., 2003. Complexity of prefrontal cortical dysfunc-tion in schizophrenia: more than up or down. American Journal ofPsychiatry 160, 2209–2215.

Carter, C.S., Mintun, M., Cohen, J.D., 1995. Interference and facilitationeffects during selective attention: an H2

15O PET study of Stroop taskperformance. Neuroimage 2, 264–272.

Cohen, J.D., Barch, D.M., Carter, C., Servan-Schreiber, D., 1999.Context-processing deficits in schizophrenia: converging evidencefrom three theoretically motivated cognitive tasks. Journal ofAbnormal Psychology 108, 120–133.

Davidson, L.L., Heinrichs, R.W., 2003. Quantification of frontal andtemporal lobe brain-imaging findings in schizophrenia: a meta-analysis. Psychiatry Research: Neuroimaging 122, 69–87.

Erickson, K.I., Milham, M.P., Colcombe, S.J., Kramer, A.F., Banich,M.T., Webb, A., Cohen, N.J., 2004. Behavioral conflict, anteriorcingulate cortex, and experiment duration: implications of diver-ging data. Human Brain Mapping 21, 98–107.

First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B., 1998. StructuredClinical Interview for DSM-IV Axis 1 Disorders. AmericanPsychiatric Press, Washington, DC.

Glahn, D.C., Ragland, J.D., Abramoff, A., Barrett, J., Laird, A.R.,Bearden, C.E., Velligan, D.I., 2005. Beyond hypofrontality: aquantitative meta-analysis of functional neuroimaging studies ofworking memory in schizophrenia. Human Brain Mapping 25,60–69.

Harrigan, S.M.,McGorry, P.D., Krstev, H., 2003.Does treatment delay infirst-episode psychosis really matter? Psychological Medicine 33,97–110.

Harrison, B.J., Shaw, M., Yücel, M., Purcell, R., Brewer, W.J., Strother,S.C., Egan, G.F., Olver, J.S., Nathan, P.J., Pantelis, C., 2005.Functional connectivity during Stroop task performance. Neuro-image 24, 181–191.

Holmes, A.J., MacDonald III, A., Carter, C.S., Barch, D.M., AndrewStenger, V., Cohen, J.D., 2005. Prefrontal functioning duringcontext processing in schizophrenia and major depression: anevent-related fMRI study. Schizophrenia Research 76, 199–206.

Ingvar, D.H., Franzén, G., 1974. Distribution of cerebral activity inchronic schizophrenia. Lancet 2, 1484–1486.

Javitt, D.C., Shelley, A.M., Silipo, G., Lieberman, J.A., 2000. Deficitsin auditory and visual context-dependent processing in schizo-phrenia: defining the pattern. Archives of General Psychiatry 57,1131–1137.

Kay, S.R., Opler, L.A., Spitzer, R.L., Williams, J.B., Fiszbein, A.,Gorelick, A., 1991. SCID-PANSS: two-tier diagnostic system forpsychotic disorders. Comprehensive Psychiatry 32, 355–361.

Kerns, J.G., Cohen, J.D., MacDonald III, A.W., Cho, R.Y., Stenger, V.A.,Carter, C.S., 2004. Anterior cingulate conflict monitoring andadjustments in control. Science 303, 1023–1026.

Kerns, J.G., Cohen, J.D., MacDonald III, A.W., Johnson, M.K.,Stenger, V.A., Aizenstein, H., Carter, C.S., 2005. Decreased con-flict- and error-related activity in the anterior cingulate cortex insubjects with schizophrenia. American Journal of Psychiatry 162,1833–1839.

MacDonald III, A.W., Cohen, J.D., Stenger, V.A., Carter, C.S., 2000.Dissociating the role of the dorsolateral prefrontal and anteriorcingulate cortex in cognitive control. Science 288, 1835–1838.

MacDonald III, A.W., Carter, C.S., 2003. Event-related FMRI study ofcontext processing in dorsolateral prefrontal cortex of patients withschizophrenia. Journal of Abnormal Psychology 112, 689–697.

MacDonald III, A.W., Carter, C.S., Kerns, J.G., Ursu, S., Barch, D.M.,Holmes, A.J., Stenger, V.A., Cohen, J.D., 2005. Specificity ofprefrontal dysfunction and context processing deficits to schizo-phrenia in never-medicated patients with first-episode psychosis.American Journal of Psychiatry 162, 475–484.

Manoach, D.S., 2003. Prefrontal cortex dysfunction during workingmemory performance in schizophrenia: reconciling discrepantfindings. Schizophrenia Research 60, 285–298.

McGorry, P.D., Edwards, J., Mihalopoulos, C., Harrigan, S.M., Jackson,H.J., 1996. EPPIC: an evolving systemof early detection and optimalmanagement. Schizophrenia Bulletin 22, 305–326.

Milham, M.P., Erickson, K.I., Banich, M.T., Kramer, A.F., Webb, A.,Wszalek, T., Cohen, N.J., 2002. Attentional control in the agingbrain: insights from an fMRI study of the Stroop task. Brain andCognition 49, 277–296.

Milham, M.P., Banich, M.T., Barad, V., 2003a. Competition for priorityin processing increases prefrontal cortex's involvement in top-downcontrol: an event-related fMRI study of the Stroop task. CognitiveBrain Research 17, 212–222.

Milham, M.P., Banich, M.T., Claus, E.D., Cohen, N.J., 2003b. Practice-related effects demonstrate complementary roles of anterior cingulateand prefrontal cortices in attentional control. Neuroimage 18,483–493.

Miller, E.K., Cohen, J.D., 2001. An integrative theory of prefrontalcortex function. Annual Review of Neuroscience 24, 167–202.

Morey, R.A., Inan, S., Mitchell, T.V., Perkins, D.O., Lieberman, J.A.,Belger, A., 2005. Imaging frontostriatal function in ultra-high-risk,early, and chronic schizophrenia during executive processing.Archives of General Psychiatry 62, 254–262.

Nelson, H.E., O'Connell, A., 1978. Dementia: the estimation ofpremorbid intelligence levels using the New Adult Reading Test.Cerebral Cortex 14, 234–244.

Pantelis, C., Velakoulis, D., McGorry, P.D., Wood, S.J., Suckling, J.,Phillips, L.J., Yung, A.R., Bullmore, E.T., Brewer, W., Soulsby, B.,Desmond, P., McGuire, P.K., 2003a. Neuroanatomical abnormal-ities before and after onset of psychosis: a cross-sectional andlongitudinal MRI comparison. Lancet 361, 281–288.

Pantelis, C., Yücel, M., Wood, S.J., McGorry, P.D., Velakoulis, D.,2003b. Early and late neurodevelopmental disturbances inschizophrenia and their functional consequences. Australian andNew Zealand Journal of Psychiatry 37, 399–406.

Pantelis, C., Yücel, M., Wood, S.J., Velakoulis, D., Sun, D., Berger, G.,Stuart, G.W., Yung, A., Phillips, L., McGorry, P.D., 2005. Structuralbrain imaging evidence for multiple pathological processes atdifferent stages of brain development in schizophrenia. Schizophre-nia Bulletin 31, 672–696.

Perlstein, W.M., Carter, C.S., Noll, D.C., Cohen, J.D., 2001. Relationof prefrontal cortex dysfunction to working memory and

31B.J. Harrison et al. / Psychiatry Research: Neuroimaging 148 (2006) 23–31

symptoms in schizophrenia. American Journal of Psychiatry 158,1105–1113.

Perlstein, W.M., Dixit, N.K., Carter, C.S., Noll, D.C., Cohen, J.D.,2003. Prefrontal cortex dysfunction mediates deficits in workingmemory and prepotent responding in schizophrenia. BiologicalPsychiatry 53, 25–38.

Shaw, M.E., Strother, S.C., McFarlane, A.C., Morris, P., Anderson, J.,Clark, C.R., Egan, G.F., 2002. Abnormal functional connectivity inposttraumatic stress disorder. Neuroimage 15, 661–674.

Stroop, J.R., 1935. Studies of interference in serial verbal reactions.Journal of Experimental Psychology 18, 643–662.

Strother, S.C., Anderson, J., Hansen, L.K., Kjems, U., Kustra, R.,Sidtis, J., Frutiger, S., Muley, S., LaConte, S., Rottenberg, D.,

2002. The quantitative evaluation of functional neuroimagingexperiments: the NPAIRS data analysis framework. Neuroimage15, 747–771.

Weinberger, D.R., Berman, K.F., Zec, R.F., 1986. Physiologic dys-function of dorsolateral prefrontal cortex in schizophrenia. I.Regional cerebral blood flow evidence. Archives of GeneralPsychiatry 43, 114–124.

Yücel, M., Pantelis, C., Stuart, G.W., Wood, S.J., Maruff, P.,Velakoulis, D., Pipingas, A., Crowe, S.F., Tochon-Danguy, H.J.,Egan, G.F., 2002. Anterior cingulate activation during Stroop taskperformance: a PET to MRI coregistration study of individualpatients with schizophrenia. American Journal of Psychiatry 159,251–254.