R

ARCHIVAL REPORT

General and Specific Functional ConnectivityDisturbances in First-Episode Schizophrenia DuringCognitive Control PerformanceAlex Fornito, Jong Yoon, Andrew Zalesky, Edward T. Bullmore, and Cameron S. Carter

Background: Cognitive control impairments in schizophrenia are thought to arise from dysfunction of interconnected networks of brainregions, but interrogating the functional dynamics of large-scale brain networks during cognitive task performance has proved difficult. Weused functional magnetic resonance imaging to generate event-related whole-brain functional connectivity networks in participants withfirst-episode schizophrenia and healthy control subjects performing a cognitive control task.

Methods: Functional connectivity during cognitive control performance was assessed between each pair of 78 brain regions in 23 patientsand 25 control subjects. Network properties examined were region-wise connectivity, edge-wise connectivity, global path length, cluster-ing, small-worldness, global efficiency, and local efficiency.

Results: Patients showed widespread functional connectivity deficits in a large-scale network of brain regions, which primarily affectedconnectivity between frontal cortex and posterior regions and occurred irrespective of task context. A more circumscribed and task-specificconnectivity impairment in frontoparietal systems related to cognitive control was also apparent. Global properties of network topology inpatients were relatively intact.

Conclusions: The first episode of schizophrenia is associated with a generalized connectivity impairment affecting most brain regions butthat is particularly pronounced for frontal cortex. Superimposed on this generalized deficit, patients show more specific cognitive-control-related functional connectivity reductions in frontoparietal regions. These connectivity deficits occur in the context of relatively preserved

global network organization.

nescaetkib

(thcsp(oswfmc

M

P

f(wsI

Key Words: Complex, executive function, fMRI, graph, psychosis,small world

C ognitive deficits are a core characteristic of schizophrenia(1–3). Although patients typically perform poorly across arange of cognitive domains (4), impairments of cognitive

control—the coordination of cognitive resources in support of goaldirected behavior—feature prominently (5). These impairments arestable (6,7), present before illness onset (8), and manifest in pa-tients’ unaffected relatives (9). Characterizing their neural basis istherefore critical for understanding the pathogenesis of the disor-der.

Cognitive control deficits in schizophrenia are associated withabnormal activation of frontal, temporal, parietal, and subcorticalregions (10 –12). The widespread spatial distribution of these ab-normalities suggests that a network-based understanding will pro-vide a useful model of neural dysfunction in schizophrenia. Indeed,most contemporary pathophysiological theories characterize theillness as emerging from disturbed interregional brain connectivity,rather than isolated dysfunction in specific loci (13–16). Severalstudies have reported disturbances of functional connectivity, de-fined as the temporal correlation between spatially distributed

From the Brain Mapping Unit (AF, ETB), Department of Psychiatry, Universityof Cambridge, Cambridge, United Kingdom; Melbourne Neuropsychia-try Centre (AF, AZ), Department of Psychiatry, University of Melbourneand Melbourne Health, Parkville, Australia; Translational Cognitive andAffective Neuroscience Laboratory (JY, CSC), University of California,Davis, California; GlaxoSmithKline Clinical Unit Cambridge (ETB), Adden-brooke’s Hospital, Cambridge, United Kingdom.

Address correspondence to Alex Fornito, Ph.D., Melbourne NeuropsychiatryCentre, University of Melbourne; Levels 1 and 2, 161 Barry Street, CarltonSouth, Victoria 3053, Australia; E-mail: fornitoa@unimelb.edu.au.

ceceived Oct 5, 2010; revised Jan 11, 2011; accepted Feb 10, 2011.

0006-3223/$36.00doi:10.1016/j.biopsych.2011.02.019

europhysiologic signals (17), in schizophrenia across a range ofxperimental paradigms, with frontotemporal and frontoparietalystems being the most affected (18 –24). However, it is often un-lear whether such disturbances are task-specific or reflect a gener-lized, context-independent dysfunction. Some studies providedvidence for both (23,25,26), but these analyses have only beenractable for networks comprising a select few brain regions. To ournowledge, no distinction between general and specific connectiv-

ty deficits in schizophrenia has been made at the level of whole-rain networks.

In this study, we adapted a beta series correlation technique27,28) to construct whole-brain, event-related functional connec-ivity networks in people with first-episode schizophrenia andealthy volunteers performing a cognitive control task (29 –31). Weharacterized general and cognitive control-specific effects ofchizophrenia on functional brain connectivity, as well as globalroperties of network organization quantified using graph analysis

32–34). Following evidence that cognitive control impairmentsccur against a background of generalized cognitive deficit inchizophrenia (4,5), we hypothesized that patients would show aidespread reduction of functional connectivity, particularly in

rontotemporal regions, regardless of task demands, in addition toore circumscribed connectivity reductions specific to cognitive

ontrol processes.

ethods and Materials

articipantsTwenty-three patients with schizophrenia (two schizophreni-

orm) and 25 healthy subjects were recruited from the communitydemographic and clinical data are shown in Table 1). Volunteers

ith schizophrenia were outpatients within the first year of psycho-is onset. Participants were evaluated with the Structured Clinicalnterview for DSM-IV-TR to confirm diagnosis and exclude illness in

ontrol subjects. Patients aged under 16 years underwent the

BIOL PSYCHIATRY 2011;xx:xxx© 2011 Society of Biological Psychiatry.

Published by Elsevier Inc. All rights reserved.

hsnb

C

3mt“qi(pmdtta

N

a

tsi

(ftcsi3acnbcptR7cgitdof

F

atcuflnte

N

r pat

2 BIOL PSYCHIATRY 2011;xx:xxx A. Fornito et al.

w

Schedule for Affective Disorders and Schizophrenia for School-AgeChildren—Present and Lifetime Version. Board certified PhD andMD-level clinicians conducted diagnostic evaluations. Diagnoseswere confirmed by consensus conference. Psychosis onset was de-termined by establishing the time at which DSM-IV criteria A and Bfor schizophrenia were met.

Control subjects were excluded if they had a lifetime Axis Idiagnosis or a first-degree relative with a psychotic disorder. Exclu-sion criteria for all subjects were IQ � 70, drug/alcohol dependence

istory or abuse in the previous 3 months, positive urine drugcreen before testing, significant head trauma, or any known mag-etic resonance imaging contraindication. The study was approvedy the internal review board at the University of California—Davis.

ognitive TaskSubjects performed the AX-Continuous Performance Task (35–

7). A cue letter was presented for 500 msec, followed by a 3500-sec delay period, and a 500-msec probe letter. Subjects made a

arget response (right index finger button press) to the probe letterX” only when it followed the cue letter “A.” All other stimuli re-uired a nontarget response (right middle finger button press),

ncluding trials in which X was preceded by any letter other than acollectively referred to as “B”). Trials with target (AX) cue–probeairings occurred with 75% frequency, establishing a tendency toake a target response to the X probe. BX trials placed the highest

emand on cognitive control because subjects had to overcomehe tendency to make a target response to X. Trials in which eitherhe A or B cue was followed by a non-X letter (collectively referred tos Y) were also included.

euroimaging MethodsAcquisition. Functional scans (T2*-weighted echo-planar im-

ging, repetition time 2000 msec, echo time � 40 msec, flip angle �90°, field of view � 22 cm) were acquired using a 1.5-T GE (Wauke-sha, Wisconsin) scanner. Twenty-four contiguous 4.0-mm axialslices with 3.4 mm2 in-plane resolution from 80.0 mm above to 16mm below the anterior commissure–posterior commissure linewere obtained.

fMRI Processing. Initial preprocessing steps, implemented inAutomated Image Registration software (http://bishopw.loni.ucla.edu/AIR), were slice-time and motion correction, and removal of

Table 1. Demographic and Clinical Characteristics

Male/Female (n)Age (years)a

Wechsler Abbreviated Scale of Intelligence—Intelligence Quotient

Education (years)Parent Education (years)Scale for the Assessment of Negative Symptoms TotalScale for the Assessment of Positive Symptoms TotalBrief Psychiatric Rating Scale TotalStrauss Carpenter Outcome Scale TotalGlobal Assessment Scale TotalMedication

aAge range for control subjects was 16 to 32 years; fo

linear trends and significant time series outliers. The images were w

ww.sobp.org/journal

hen spatially normalized using a nonlinear algorithm and spatiallymoothed (8-mm full-width at half-maximum Gaussian kernel) us-ng SPM5 (http://www.fil.ion.ucl.ac.uk/spm/software).

To generate measures of event-related functional connectivity27,28), we modeled every cue and probe event with a unique deltaunction, convolved with a canonical hemodynamic response func-ion, using SPM5. All events were modeled, but only cue events fororrect trials were included in the analysis, because these repre-ented trial periods in which cognitive control demands were max-mized. Subjects who committed no errors had 160 trials (128 A cue;2 B cue). Separate regressors modeling each event were defined ingeneral linear model to yield 128 unique A cue and 32 unique B

ue beta values for every voxel. Each beta value reflected the mag-itude of the hemodynamic response evoked by each event. Theeta values were then sorted by condition (i.e., A cue or B cue) andoncatenated to generate a beta series for each condition. We thenarcellated each participant’s brain into discrete regions of interest

o represent network nodes (19,38,39) using an anatomic atlas (40).egions with less than 25% brain coverage were excluded, yielding8 anatomic regions (Table S1 in Supplement 1). Pairwise Pearsonorrelations between regional mean beta series were computed toenerate a {78 � 78} functional connectivity matrix for each partic-

pant and each task condition (A or B cue). These correlations be-ween regional beta series reflected correlated variations in hemo-ynamic responses evoked by each task condition. An overview ofur method is provided in Figure 1 (see Section S1 in Supplement 1

or further details).

unctional Connectivity AnalysisNetwork functional connectivity was examined at a region-wise

nd edge-wise basis. Region-wise connectivity (SR) was defined ashe sum of each region’s connectivity with all other regions, asalculated using absolute correlation values in each individual’snthresholded functional connectivity matrix. The measure re-ected how strongly connected each region was with the rest of theetwork. Edge-wise connectivity (SE) simply reflected the correla-

ion value for each pair of regional beta series (i.e., each element inach {78 � 78} functional connectivity matrix; see Figure 1).

etwork Topology AnalysisBefore generating graph-based representations of brain net-

trol (n � 25) Schizophrenia (n � 23) �2/t, p

13/12 14/9 .38, 542.12 (3.92) 20.45 (4.36) 1.40, .173.43 (9.25) 100.91 (12.07) 3.92, �.001

4.94 (3.19) 12.09 (2.59) 3.33, .0024.68 (2.31) 15.02 (2.02) –.53, .60

7.59 (3.96)7.05 (3.40)

44.36 (10.54)10.33 (2.57)51.14 (10.30)13 atypical

1 typical1 antidepressant1 benzodiazepine7 unmedicated

ients, it was 14 to 30 years.

Con

211

11

ork topology, we thresholded each individual’s {78 � 78} ab-

c

ei

ocmtrocctsee(wvt

R

B

f[iprorf

s bee

A. Fornito et al. BIOL PSYCHIATRY 2011;xx:xxx 3

solute functional connectivity matrix to remove spurious edges.Because this threshold is arbitrary, we computed topologicalmeasures for networks thresholded between 5% and 45% of allpossible edges. At each threshold, network topology was repre-sented as a binary, undirected graph from which five commonlystudied topological measures were computed: normalized pathlength (�), normalized clustering (�), small-worldness (�), globalcommunication efficiency (EG), and local communication effi-

iency (EL). The measures were averaged across thresholds be-fore statistical analysis (41). (Further details provided in SectionsS.2 and S.3 in Supplement 1.)

Statistical AnalysesBehavioral data were analyzed in PASW Statistics v. 18 (IBM,

Somers, New York) using multivariate analysis of variance with di-agnosis and trial type (i.e., AX, AY, BX, BY) as factors. For the neuro-imaging data, for each network property analyzed (SR, SE, �, �, and�), main effects of cue (A or B cue), diagnosis, and their interaction,were modeled using analysis of variance. Empiric p values wereestimated using permutation methods suitable for factorial designs(42,43). We performed 5000 permutations for each analysis.

Regional connectivity effects were corrected for multiplecomparisons using a false-positive adjustment, 1/N � .013,where N � the number of comparisons, which implies that, onaverage, there was � 1 false positive per analysis (19,44). Edge-wise connectivity effects were characterized using the network-based statistic (45). To compute the network-based statistic, ananalysis of variance was fitted to each of the (N2-N)/2 � 3003

dges (correlation values) in the {78 � 78} functional connectiv-

Figure 1. Schematic overview of the analysis strategy. The analysis couldevery event, and event-specific evoked responses were estimated usingrepresenting the degree to which each voxel’s activity was modulated btask conditions (cue A shown in blue, cue B show in red) and concatenata series) at each voxel. Third, an anatomic template was used to parceand correlations between every possible pair of regional beta series weparticipant and each task condition. Fourth, network connectivity anstructure. Network connectivity was analyzed using the values in the ccorrelation for each value in the functional connectivity matrix. For examof interest (ROI) 1 and ROI 2 would be .33. Region-wise connectivity (SR) wfor ROI 1 is computed as the sum of its correlation values with all other regof network structure as detailed in the main text and Section S.2 in Suppleas a blue circle plotted according to the stereotactic coordinates of its cea suprathreshold correlation between regional beta series. The graph harespect to approximate anatomical location.

ty matrix, yielding a p value matrix representing the probability a

f rejecting the null hypothesis for each effect at each edge. Aomponent-forming threshold, K, was applied to each p valueatrix, and the size of each connected component in these

hresholded matrices was computed. A connected componentepresents a subnetwork of edges that can be linked to eachther via suprathreshold connections. The size of the observedomponents was compared with a null distribution of maximalomponent sizes obtained through permutation testing to ob-ain component-wise p values corrected for multiple compari-ons (46). The method identifies connected subnetworks ofdges showing a particular effect of a size larger than would bexpected by chance (see Zalesky et al. for details and validation)

45). Here, we report results for k � .005 (i.e., retaining only edgesith p � .005). Results obtained at higher (k � .01) and lower (k � .001)

alues are provided in Figure S2 in Supplement 1 to illustrate sensitivityo parameter settings.

esults

ehavioral AnalysisAnalysis of performance accuracy revealed significant main ef-

ects of diagnosis [F (1,46) � 8220.41, p � .001] and trial typeF (3,44) � 19.30, p � .001], as well as a diagnosis � trial typenteraction [F (3,44) � 2.41, p � .01; Figure 2]. Bonferroni-adjustedost hoc testing revealed that patients were significantly less accu-

ate for BX trials [t (28.44) � 3.60, p � .004, corrected] but not forther conditions (p � .10), suggesting a specific deficit for trials

equiring cognitive control. For averaged reaction time (RT), weound significant effects of diagnosis [F (1,46) � 612.55, p � .001]

ivided into four broad steps. First, a unique regressor was generated foreral linear model. The result was a map of coefficients for each event,t event. Second, these event-specific maps were sorted into differentgenerate a condition-specific pseudo–time series of coefficients (i.e.,he brain in 78 regions. The mean series of each region was extracted,

puted to generate a {78 � 78} functional connectivity matrix for eachology were analyzed using graph-based representations of network

ation matrix. Edge-wise connectivity (SE) reflected the strength of thein the connectivity matrix shown, SE for the connection between region

mputed as the sum of each region’s correlation values. For example, SR

Topologic measures were computed using graph-based representations1. In the graph-based representation shown, each region is represented

d. Each connection, or edge, is represented as a red line and representsn overlaid on a cortical surface rendering to provide an orientation with

be da geny thaed tollate tre comd toporrelple,as co

ions.mentntroi

nd trial type [F (3,44) � 108.75, p � .001]. The diagnosis � trial type

www.sobp.org/journal

c

Stiies

c4titn

tlttpe

rol sub

4 BIOL PSYCHIATRY 2011;xx:xxx A. Fornito et al.

w

interaction trended towards significance [F (1,44) � 2.75, p � .054;Figure 2].

Functional Connectivity AnalysisRegion-Wise Connectivity. Regional main effects of cue sur-

viving false-positive correction were found in 14 regions, localizedto bilateral hippocampi, subgenual cingulate cortices, lingual andangular gyri; temporal poles; left amygdala, angular and posteriorcingulate gyri; and right middle frontal gyrus (Figure 3, top row;Table S1 in Supplement 1). At uncorrected levels, there was moreextensive involvement of left posteromedial and midcingulate cor-tices, right lateral parietal cortex, and bilateral orbitofrontal regions.For all regions, functional connectivity was increased during B cuetrials. These findings reflected significant effects of cognitive con-trol on regional functional connectivity that were common to pa-tients and control subjects.

Regional main effects of diagnosis were widespread and in eachcase reflected reduced functional connectivity in patients (Figure 3,bottom row). At uncorrected levels, the reductions were significantin 50% of regions (n � 39). These effects survived false-positive

orrection in 20% of regions (n � 16; Table S1 in Supplement 1).

Figure 2. Mean accuracy (left) and reaction time (right) of patients and cont.05, corrected.

Figure 3. Brain regions showing significant main effects of cue (top row)

connectivity in the more difficult B cue condition; main effects of diagnosis reflsubjects. Effects surviving false-positive correction for multiple comparisons are s

ww.sobp.org/journal

ignificant effects were distributed bilaterally, primarily localizingo frontal, temporal, and parietal areas, with an additional reductionn left visual cortex. These findings reflected generalized, context-ndependent connectivity reductions in patients that were appar-nt in both task conditions. No cue-by-diagnosis interaction effectsurvived false-positive correction.

Edge-Wise Functional Connectivity. For the main effect ofue, one connected component, comprising 56 edges connecting2 regions, was statistically significant (p � .001, corrected; Figure 4,op left). For all but one edge in this subnetwork, connectivityncreased in the more difficult B cue condition. Functional connec-ivity of right medial temporal regions showed particularly pro-ounced effects (29 hippocampal and 14 amygdala connections).

To better understand the regional distribution of these connec-ions, we classified each region as belonging to one of five broadobar subgroupings: frontal, temporal, parietal, occipital, or subcor-ical. We then categorized each edge in the affected subnetwork onhe basis of the lobes they connected (e.g., frontotemporal, tem-oroparietal, etc.) and counted the proportion of edges falling intoach category. Most connections showing a significant effect of cue

jects in each task condition. Error bars represent standard deviations. *p �

iagnosis (bottom row). Main effects of cue reflected increased functional

and d ected reduced functional connectivity in patients compared with controlhown in yellow; those significant at p � .05, uncorrected are shown in red.

n

td

C

pscrrtcf

wfonmBamh

ptcroimCp

A. Fornito et al. BIOL PSYCHIATRY 2011;xx:xxx 5

were frontotemporal (36%) or temporosubcortical (25%; Figure 4,bottom left). Similar findings were obtained across different valuesof K (Figure S2 in Supplement 1).

For the main effect of diagnosis, one component comprising200 edges connecting 54 regions was significant (p � .001, cor-rected; Figure 4, top middle). For all edges, functional connectivitywas reduced in schizophrenia patients. The most affected regionsincluded bilateral hippocampi, middle temporal, and lateral pre-frontal cortex. Most of these connections linked frontal cortex toposterior regions: 34% were frontotemporal, and 20% were fronto-parietal (Figure 4, bottom middle). Most of the remaining connec-tions were either temporoparietal (18%) or occipitotemporal (10%).Similar findings were obtained across different values of K (FigureS2 in Supplement 1).

One connected component, comprising seven edges connect-ing eight regions, demonstrated a significant cue-by-diagnosis in-teraction (p � .042, corrected; Figure 4, top right). Approximatelytwo thirds of these involved frontal regions: 29% were frontofrontaland 29% were frontoparietal. For each edge, control subjectsshowed a B cue � A cue task effect on connectivity, whereas pa-tients showed the reverse pattern (Figure S3 in Supplement 1).These findings were not significant for k � .01 or k � .001.

To examine medication effects, we compared SE values betweenpatients on or off antipsychotics for each of the 200 edges impli-cated in the subnetwork showing a diagnosis main effect. Only 26edges (13%) showed significant effects at p � .05 uncorrected, and

Figure 4. Subnetworks of functional connections showing a significant maidiagnosis interaction (right column). The top row presents graph-based rep

lotted according to the stereotactic coordinates of its centroid, and eachhe main effect of cue represents a single edge where connectivity was highue condition. For the main effect of diagnosis, all suprathreshold edges refleenderings to provide an orientation with respect to approximate anatomicaf reference. Bottom row presents the proportion of each type of conne

nterconnects. Amyg, amygdala; Angular, angular gyrus; Fron Inf Oper, inferioiddle frontal gyrus; Fron Sup, superior frontal gyrus; Fron Sup Med, superio

ing, midcingulate cortex; Occ Sup, superior occipital gyrus; Par Inf, inferiorrecuneus; R, right.

o differences survived multiple comparison correction. None of r

hese edges were part of the subnetwork showing a significantiagnosis � cue interaction.

orrelations Between Brain Connectivity and Behavior

To assess the relationship between functional connectivity andatients’ cognitive control performance deficits, we focused on theubnetwork showing a significant cue-by-diagnosis interaction. Be-ause these seven edges specifically showed a cognitive-control-elated functional connectivity deficit in schizophrenia patients, weeasoned that they were the most relevant for understanding pa-ients’ cognitive control impairments. For each edge, we defined aognitive-control-related connectivity index as the normalized dif-erence between B cue and A cue connectivity:

[SE-B(i) - SE-A(i)] ⁄ [SE-B(i) + SE-A(i)],

here SE-A (i) and SE-B (i) refer to the edge-wise connectivity valuesor edge i during A cue and B cue trials, respectively. Positive valuesn this scale reflect greater connectivity during trials requiring cog-itive control (i.e., B � A cue). Similarly, we derived a behavioraleasure of cognitive control as the normalized difference between

X and AX trial accuracy and RT [i.e., (BX � AX)/(BX � AX)]. For theccuracy scale, higher values reflect better cognitive control perfor-ance (i.e., relatively more accurate responses); for the RT scale,

igher values reflect poorer cognitive control performance (i.e.,

ct of cue (left column), main effect of diagnosis (middle column) and cue �tations of these subnetworks, with each region represented as a blue circlethreshold edge represented as a red line. The green edge in the graph forA cue than B cue trials. For all other edges, connectivity was higher in the Breduced connectivity in patients. The graphs are overlaid on cortical surfacetion. The names of key regions in each subnetwork have been noted for ease

in each subnetwork, as categorized according to the lobes each edgetal operculum; Fron Inf Tri, inferior frontal gyrus, pars Triangularis; Fron Mid,tal gyrus, medial aspect; Hipp, hippocampus; L, left; Ling, lingual gyrus; Midal lobe; Postcentral, postcentral gyrus; Precentral, precentral gyrus; Precun,

n efferesensupraer inctedl locactionr fronr fron

pariet

elatively slower responding).

www.sobp.org/journal

r

sc

G

wcritdctfseps2

apiwtcecdrho

timesubje

���EE

6 BIOL PSYCHIATRY 2011;xx:xxx A. Fornito et al.

w

The RT cognitive control measure was significantly and specifi-cally correlated with cognitive-control-related functional connec-tivity in one edge, linking right midcingulate and superior frontalregions (i.e., R Mid Cing and R Fron Sup, respectively; Figure 4, topright) in control subjects ( � .45, p � .026) but not patients ( �–.09, p � .689; Figure 5). This effect did not survive Bonferronicorrection for seven comparisons (� � .05/7 � .007). There were nosignificant correlations between functional connectivity and cogni-tive control accuracy. There were no significant associations be-tween connectivity measures and symptom ratings or IQ.

Network Topology AnalysisGroup means and key statistical findings for each topological

measure studied are presented in Table 2. The only significanteffects on network topology were main effects of cue for � and EL,,eflecting a reduction in both during B cue trials.

Discussion

Schizophrenia is often characterized as a disorder of aberrantbrain connectivity, although mapping putative connectivity distur-bances at the level of whole-brain circuits, and their relation tocognitive impairments, has proved challenging. We analyzedwhole-brain maps of task-related functional connectivity measuredduring cognitive control performance and found evidence for awidespread disturbance of functional connectivity in people withfirst-episode schizophrenia, which occurred regardless of task con-text and which was most pronounced in frontoposterior systems. Inaddition, we identified a more specific deficit in functional connec-tivity of frontoparietal regions that was associated with the imple-mentation of cognitive control. Together these findings suggestthat schizophrenia is associated with a generalized and widespreadfunctional connectivity deficit upon which are superimposed more

Figure 5. Scatterplot of the association between cognitive control reactionand superior frontal gyri (R Mid Cing and R Fron Sup in Figure 4) for control

Table 2. Summary Statistics for Network Topological Properties as a Functi

Control Subjects Patients

A Cue B Cue A Cue

1.154 (.047) 1.138 (.051) 1.165 (.052)1.773 (.338) 1.730 (.417) 1.883 (.395)1.488 (.299) 1.480 (.344) 1.559 (.361)

G .513 (.032) .511 (.036) .519 (.032)

L .739 (.034) .721 (.035) .743 (.030)

Data represent means. Standard deviations are presented in parentheses.

ww.sobp.org/journal

pecific, context-dependent alterations of interregional functionaloupling.

eneralized Functional Connectivity Deficits in SchizophreniaOur analysis of region-wise functional connectivity revealed

idespread and generalized (context-independent) patient defi-its localized predominantly to frontotemporal and medial poste-ior cortices. In addition, our analysis of edge-wise connectivitydentified a subnetwork of 200 connections linking nearly 70% ofhe regions studied in which patients showed a context-indepen-ent functional connectivity reduction. More than half of theseonnections linked frontal cortex with posterior regions. Togetherhese results point to a generalized and pervasive reduction inunctional connectivity between frontal and other brain regions inchizophrenia. Our findings are consistent with a wide range ofvidence implicating frontal dysfunction as a core feature of schizo-hrenia pathophysiology (5,10,47–51), and other connectivitytudies implicating frontotemporal disturbances in particular (21–3,26,52–56).

Alterations in functional connectivity between frontal cortexnd other brain regions has been demonstrated in schizophreniaatients using a variety of techniques, some of which specifically

solated task-dependent connectivity changes (57–59), and othershich did not (22,54,56,60). Of these, the only studies that at-

empted to distinguish between generalized and task-dependentonnectivity changes have been those using dynamic causal mod-ling, which allows assessment of group differences in task-relatedausal interactions (effective connectivity), as well as so-called en-ogenous, or context-independent, connectivity, between specific

egions of interest (61,62) (Section S1 in Supplement 1). This workas identified altered effective connectivity between frontal andther brain regions at different illness stages of schizophrenia and

(RT) performance and functional connectivity between right mid-cingulatects (left, blue) and patients (right, red).

Cue and Diagnosis

F, p

ue Cue Diagnosis Cue � Diagnosis

(.076) 4.06, .042 .55, .470 �.001, .953(.322) 1.22, .263 .96, .335 .13, .711(.312) .10, .752 .62, .443 .03, .867(.034) .19, .664 1.27, .259 .69, .405(.027) 12.57, �.001 .41, .527 �.001, .999

on of

B C

1.1481.7981.534

.525

.726

ggbattq(

T

iidn(gtmfdS

snmtmfii

L

nltfi8nttm

aitncvasbfir

C

scss

A. Fornito et al. BIOL PSYCHIATRY 2011;xx:xxx 7

across a range of paradigms, including those assessing perceptualdiscrimination (23), memory (25), and cognitive control (26,63), butthe most replicated differences have been for the context-indepen-dent endogenous connectivity parameters. These results concurwith our finding of a widespread and generalized connectivity def-icit in schizophrenia, which particularly affects functional integra-tion between frontal cortex and other regions. Measuring connec-tivity differences across a range of tasks in the same sample will helpto clarify precisely how generalized these effects are. Another im-portant question concerns whether these generalized connectivitydeficits reflect alterations in underlying anatomical connectivity(64 – 66). Studies combining fMRI with diffusion-based imaging arenecessary to address this question.

Specific, Cognitive-Control-Related Functional ConnectivityDeficits in Schizophrenia

Superimposed against a pervasive and generalized functionalconnectivity deficit, patients showed a specific cognitive-controlrelated connectivity impairment in a subnetwork of seven connec-tions linking eight regions located in frontal and parietal cortex. Foreach connection in this network, control subjects generally showedincreased connectivity in the B cue condition, whereas patientsshowed the reverse pattern (i.e., A � B cue connectivity). Thesefindings complement and extend prior activation- and connectivi-ty-based studies of the AX Continuous Performance Task in schizo-phrenia (27,29,67,68) by characterizing network abnormalities at awhole-brain level and mapping the specific dysfunctional subnet-work associated with cognitive control performance in patients.The importance of frontoparietal dysfunction for understandingcognitive control deficits in schizophrenia is underscored by theconvergence of findings between this past work and our own.

Patients’ apparent failure to increase frontoparietal connectivityappropriately in response to the B cue may reflect difficulty recruit-ing key elements of the neural network required to implementcognitive control. This is consistent with the observed correlationsbetween connectivity and task performance. In control subjects,greater connectivity between dorsal cingulate and prefrontal cor-tex in the B cue condition was associated with slower B cue RTs. It isthought that dorsal cingulate activity signals the need for engage-ment of cognitive control processes mediated by lateral prefrontalregions (69 –71). This signal typically initiates when task perfor-mance elicits conflict. Higher conflict is associated with greater taskdifficulty and a greater need for strategic control (72,73). One pre-diction of this view is that cingulofrontal connectivity should behighest when task difficulty is high, indexed by increased conflictand an increased need for cognitive control. Our results accord withthis prediction, because healthy individuals who found B trials moredifficult, as reflected by relatively slower RTs, showed greater cingu-lofrontal connectivity during these trials. This association betweenconnectivity and behavior was not present in the patient group,suggesting a breakdown of conflict monitoring and a relative fail-ure to implement cognitive control, evidenced by patients’ signifi-cantly lower accuracy rates for BX trials. Together these data sug-gest that patients did not appropriately alter frontoparietalfunctional connectivity in response to increased cognitive controldemands, resulting in a decoupling of network dynamics from taskperformance and a relative failure to implement control processes.

We found a significant main effect of task on both region- andedge-wise connectivity in a distributed network of regions incorpo-rating primarily frontal and temporal cortices. These findings areconsistent with activation-based studies showing greater prefron-tal activation for B relative to A cue trials (27,29,67). In addition, we

found strong task effects on connectivity of medial temporal re- a

ions, areas that have not been identified as key task-related re-ions in prior work. This likely reflects the differing sensitivities ofeta series connectivity- and activation-based methods. The formerre sensitive to task effects on interregional correlations in trial-to-rial variations of task-evoked responses, whereas the latter charac-erize mean differences in the strength of these responses. Conse-uently, the two do not always yield convergent findings (28)

Section S1 in Supplement 1).

opologic PropertiesBoth patients and control subjects showed increased global

ntegration (i.e., lower normalized path length, �) and reduced localnformation processing (i.e., lower local efficiency, EL) in the moreifficult B cue condition, suggesting that global properties of brainetwork organization are responsive to changing task contexts

74). These findings are consistent with evidence that increasedlobal integration of functional brain networks is a marker of adap-

ive behavior (75,76). The findings were not uniform across othereasures of similar topological properties (e.g., we found no effects

or EG or �), however, and should be interpreted with respect toifferences in how these measures are computed (Section S2 inupplement 1).

We found no group differences in network topology. In contrast,tudies of chronic patients have reported differences in a variety ofetwork parameters, during both rest and working memory perfor-ance (18,19,77,78). One hypothesis explaining this discrepancy is

hat, despite showing widespread functional connectivity impair-ent, patients’ global network topology is relatively intact during a

rst psychotic episode but subsequently deteriorates with ongoingllness. Longitudinal studies are required to test this postulate.

imitations

We used an a priori anatomic template to define the networkodes used in our analyses. This approach is widely used in the

iterature (19,38,39), but the node definitions remain arbitrary. Fur-her work examining the effects of template selection on reportedndings will be important to determine their generalizability (79-1). In addition, our analyses suggested that antipsychotics couldot account for our case– control differences, although more sys-

ematic investigation of medication effects on functional connec-ivity is required given emerging evidence that connectivity can be

odulated by antipsychotic use (82,83).The main effects on functional connectivity observed in our

nalyses were robust to different analysis parameter settings. Thenteraction effect was more sensitive to such variations, suggestinghat it may have been present across a more restricted range ofetwork component sizes. In addition, the correlation betweenonnectivity and task performance in control subjects did not sur-ive multiple comparison correction. Replication of these effects inlarger sample is warranted. However, the consistency of our re-

ults with past findings, as well as the strength of the connectivity-ehavior association (20% shared variance), suggest that ourndings are indeed robust and relevant for understanding the neu-al basis of cognitive control deficits in schizophrenia.

onclusions

Cognitive control impairments are regarded as a core feature ofchizophrenia, but relatively little is known about the large-scaleonnectivity abnormalities that give rise to such deficits. Our resultsuggest that the first episode of schizophrenia is associated withpecific cognitive-control-related functional connectivity deficits in

circumscribed network of frontoparietal regions, which occur

www.sobp.org/journal

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

8 BIOL PSYCHIATRY 2011;xx:xxx A. Fornito et al.

w

against a background of a more pervasive and generalized impair-ment affecting the connectivity of frontal regions with the rest ofthe brain. These connectivity impairments parallel, and may there-fore underlie, the profile of general and specific cognitive deficitsknown to characterize the disorder.

AF was supported by a National Health and Medical ResearchCouncil CJ Martin Fellowship (Grant No. 454797). JHY was supportedby the National Institutes of Health. AZ was supported by the Austra-lian Research Council (Grant No. DP0986320).

ETB was employed half-time by GlaxoSmithKline. All other authorsreported no biomedical financial interests or potential conflicts of in-terest.

Supplementary material cited in this article is available online.

1. Barch DM, Keefe RS (2010): Anticipating DSM-V: Opportunities and chal-lenges for cognition and psychosis. Schizophr Bull 36:43– 47.

2. Bora E, Yucel M, Pantelis C (2010): Cognitive impairment in schizophre-nia and affective psychoses: Implications for DSM-V criteria and beyond.Schizophr Bull 36:36 – 42.

3. Keefe RS, Fenton WS (2007): How should DSM-V criteria for schizophre-nia include cognitive impairment? Schizophr Bull 33:912–920.

4. Heinrichs RW, Zakzanis KK (1998): Neurocognitive deficit in schizophre-nia: A quantitative review of the evidence. Neuropsychology 12:426 –445.

5. Lesh TA, Niendam TA, Minzenberg MJ, Carter CS. (2011): Cognitive con-trol Deficits in schizophrenia: Mechanisms and meaning. Neuropsycho-pharmacology 36:316 –338.

6. Burdick KE, Goldberg JF, Harrow M, Faull RN, Malhotra AK (2006): Neu-rocognition as a stable endophenotype in bipolar disorder and schizo-phrenia. J Nerv Ment Dis 194:255–260.

7. Tyson PJ, Laws KR, Roberts KH, Mortimer AM (2004): Stability of set-shifting and planning abilities in patients with schizophrenia. PsychiatryRes 129:229 –239.

8. Wood SJ, Pantelis C, Proffitt T, Phillips LJ, Stuart GW, Buchanan J-A, et al.(2003): Spatial working memory ability is a marker for risk-of-psychosis.Psychol Med 33:1239 –1247.

9. Snitz BE, Macdonald AW 3rd, Carter CS (2006): Cognitive deficits inunaffected first-degree relatives of schizophrenia patients: A meta-an-alytic review of putative endophenotypes. Schizophr Bull 32:179 –194.

10. Minzenberg MJ, Laird AR, Thelen S, Carter CS, Glahn DC (2009): Meta-analysis of 41 functional neuroimaging studies of executive function inschizophrenia. Arch Gen Psychiatry 66:811– 822.

11. Ragland JD, Yoon J, Minzenberg MJ, Carter CS (2007): Neuroimaging ofcognitive disability in schizophrenia: Search for a pathophysiologicalmechanism. Int Rev Psychiatry 19:417– 427.

12. MacDonald AW 3rd, Thermenos HW, Barch DM, Seidman LJ (2009):Imaging genetic liability to schizophrenia: Systematic review of FMRIstudies of patients’ nonpsychotic relatives. Schizophr Bull 35:1142–1162.

13. Friston KJ (1998): The disconnection hypothesis. Schizophr Res 30:115–125.

14. Bullmore ET, Frangou S, Murray RM (1997): The dysplastic net hypothe-sis: An integration of developmental and dysconnectivity theories ofschizophrenia. Schizophr Res 28:143–156.

15. Andreasen NC, Paradiso S, O’Leary DS (1998): “Cognitive dysmetria” asan integrative theory of schizophrenia: A dysfunction in cortical-subcor-tical-cerebellar circuitry? Schizophr Bull 24:203–218.

16. Stephan KE, Baldeweg T, Friston KJ (2006): Synaptic plasticity and dys-connection in schizophrenia. Biol Psychiatry 59:929 –939.

17. Friston KJ (1994): Functional and effective connectivity in neuroimag-ing: A synthesis. Hum Brain Mapp 2:56 –78.

18. Liu Y, Liang M, Zhou Y, He Y, Hao Y, Song M, et al. (2008): Disruptedsmall-world networks in schizophrenia. Brain 131:945–961.

19. Lynall ME, Bassett DS, Kerwin R, McKenna PJ, Kitzbichler M, Muller U,Bullmore E (2010): Functional connectivity and brain networks in schizo-phrenia. J Neurosci 30:9477–9487.

20. Garrity AG, Pearlson GD, McKiernan K, Lloyd D, Kiehl KA, Calhoun VD

(2007): Aberrant “default mode” functional connectivity in schizophre-nia. Am J Psychiatry 164:450 – 457.

ww.sobp.org/journal

1. Meyer-Lindenberg A, Poline J-B, Kohn PD, Holt JL, Egan MF, WeinbergerDR, Berman KF (2001): Evidence for abnormal cortical functional con-nectivity during working memory in schizophrenia. Am J Psychiatry 158:1809 –1817.

2. Meyer-Lindenberg AS, Olsen RK, Kohn PD, Brown T, Egan MF, Wein-berger DR, Berman KF (2005): Regionally specific disturbance of dorso-lateral prefrontal-hippocampal functional connectivity in schizophre-nia. Arch Gen Psychiatry 62:379 –386.

3. Mechelli A, Allen P, Amaro E Jr, Fu CH, Williams SC, Brammer MJ, et al.(2007): Misattribution of speech and impaired connectivity in patientswith auditory verbal hallucinations. Hum Brain Mapp 28:1213–1222.

4. Fornito A, Bullmore ET (2010): What can spontaneous fluctuations of theblood oxygenation-level-dependent signal tell us about psychiatric dis-orders? Curr Opin Psychiatry 23:239 –249.

5. Benetti S, Mechelli A, Picchioni M, Broome M, Williams S, McGuire P(2009): Functional integration between the posterior hippocampus andprefrontal cortex is impaired in both first episode schizophrenia and theat risk mental state. Brain 132:2426 –2436.

6. Crossley NA, Mechelli A, Fusar-Poli P, Broome MR, Matthiasson P, JohnsLC, et al. (2009): Superior temporal lobe dysfunction and frontotempo-ral dysconnectivity in subjects at risk of psychosis and in first-episodepsychosis. Hum Brain Mapp 30:4129 – 4137.

7. Yoon JH, Minzenberg MJ, Ursu S, Walter R, Wendelken C, Ragland JD, etal. (2008): Association of dorsolateral prefrontal cortex dysfunction withdisrupted coordinated brain activity in schizophrenia: Relationship withimpaired cognition, behavioral disorganization, and global function.Am J Psychiatry 165:1006 –1014.

8. Rissman J, Gazzaley A, D’Esposito M (2004): Measuring functional con-nectivity during distinct stages of a cognitive task. Neuroimage 23:752–763.

9. Barch DM, Carter CS, Braver TS, Sabb FW, MacDonald AM, Noll D, CohenJD (2001): Selective deficits in prefrontal cortex function in medication-naive patients with schizophrenia. Arch Gen Psychiatry 58:280 –288.

0. Barch DM, Carter CS, MacDonald AM, Braver TS, Cohen JD (2003): Con-text-processing deficits in schizophrenia: Diagnostic specificity, 4-weekcourse, and relationships to clinical symptoms. J Abnorm Psychol 112:132–143.

1. MacDonald AM, Pogue-Geilie MF, Johnson MK, Carter CS (2003): A spe-cific deficit in context processing in the unaffected siblings of patientswith schizophrenia. Arch Gen Psychiatry 60:57– 65.

2. Bullmore E, Barnes A, Bassett DS, Fornito A, Kitzbichler M, Meunier D,Suckling J (2009): Generic aspects of complexity in brain imaging dataand other biological systems. Neuroimage 47:1125–1134.

3. Bullmore E, Sporns O (2009): Complex brain networks: Graph theoreticalanalysis of structural and functional systems. Nat Rev Neurosci 10:186 –198.

4. Rubinov M, Sporns O (2009): Complex network measures of brain con-nectivity: Uses and interpretations. Neuroimage 52:1059 –1069.

5. Barch DM, Carter CS, Braver TS, Sabb FW, MacDonald A 3rd, Noll DC,Cohen JD (2001): Selective deficits in prefrontal cortex function in med-ication-naive patients with schizophrenia. Arch Gen Psychiatry 58:280 –288.

6. MacDonald AW, 3rd, Carter CS (2003): Event-related FMRI study of con-text processing in dorsolateral prefrontal cortex of patients with schizo-phrenia. J Abnorm Psychol 112:689 – 697.

7. Barch DM, Carter CS, MacDonald AW, 3rd, Braver TS, Cohen JD (2003):Context-processing deficits in schizophrenia: Diagnostic specificity,4-week course, and relationships to clinical symptoms. J Abnorm Psychol112:132–143.

8. Liu H, Liu Z, Liang M, Hao Y, Tan L, Kuang F, et al. (2006): Decreasedregional homogeneity in schizophrenia: A resting state functional mag-netic resonance imaging study. Neuroreport 17:19 –22.

9. Achard S, Salvador R, Whitcher B, Suckling J, Bullmore E (2006): A resil-ient, low-frequency, small-world human brain functional network withhighly connected association cortical hubs. J Neurosci 26:63–72.

0. Tzourio-Mazoyer N, Landeau B, Papathanassiou D, Crivello F, Etard O,Delcroix N, et al. (2002): Automated anatomical labeling of activations inSPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage 15:273–289.

1. Bassett DS, Bullmore E, Verchinski BA, Mattay VS, Weinberger DR,

Meyer-Lindenberg A (2008): Hierarchical organization of human corticalnetworks in health and schizophrenia. J Neurosci 28:9239 –9248.

5

5

5

5

5

5

5

6

6

6

6

6

6

6

6

6

6

7

7

7

7

7

7

7

7

7

7

8

8

8

8

A. Fornito et al. BIOL PSYCHIATRY 2011;xx:xxx 9

42. Suckling J, Bullmore E (2004): Permutation tests for factorially designedneuroimaging experiments. Hum Brain Mapp 22:193–205.

43. Anderson MJ, Ter Braak CJF (2003): Permutation tests for multi-factorialanalysis of variance. J Stat Computaion Simul 73:85–113.

44. Bullmore ET, Suckling J, Overmeyer S, Rabe-Hesketh S, Taylor E, Bram-mer MJ (1999): Global, voxel, and cluster tests, by theory and permuta-tion, for a difference between two groups of structural MR images of thebrain. IEEE Trans Med Imaging 18:32– 42.

45. Zalesky A, Fornito A, Bullmore ET (2010): Network-based statistic: Iden-tifying differences in brain networks. Neuroimage 53:1197–1207.

46. Nichols TE, Holmes AP (2002): Nonparametric permutation tests forfunctional neuroimaging: A primer with examples. Hum Brain Mapp15:1–25.

47. Ellison-Wright I, Glahn DC, Laird AR, Thelen SM, Bullmore E (2008): Theanatomy of first-episode and chronic schizophrenia: An anatomical like-lihood estimation meta-analysis. Am J Psychiatry 165:1015–1023.

48. Fornito A, Yucel M, Dean B, Wood SJ, Pantelis C (2009): Anatomicalabnormalities of the anterior cingulate cortex in schizophrenia: Bridg-ing the gap between neuroimaging and neuropathology. Schizophr Bull35:973–993.

49. Weinberger DR, Berman KF (1998): Prefrontal function in schizophrenia:Confounds and controversies. In: Roberts AC, Robbins TW, Weiskrantz L,editors. The Prefrontal Cortex: Executive and Cognitive Functions. NewYork: Oxford University Press, 165–179.

50. Lewis DA, Hashimoto T, Volk DW (2005): Cortical inhibitory neurons andschizophrenia. Nat Rev Neurosci 6:312–324.

51. Fornito A, Yucel M, Patti J, Wood SJ, Pantelis C (2009): Mapping greymatter reductions in schizophrenia: An anatomical likelihood estima-tion analysis of voxel-based morphometry studies. Schizophr Res 108:104 –113.

2. Fletcher P, McKenna PJ, Friston KJ, Frith CD, Dolan RJ (1999): Abnormalcingulate modulation of fronto-temporal connectivity in schizophrenia.Neuroimage 9:337–342.

3. Spoletini I, Cherubini A, Di Paola M, Banfi G, Rusch N, Martinotti G, et al.(2009): Reduced fronto-temporal connectivity is associated with frontalgray matter density reduction and neuropsychological deficit in schizo-phrenia. Schizophr Res 108:57– 68.

54. Wolf DH, Gur RC, Valdez JN, Loughead J, Elliott MA, Gur RE, Ragland JD(2007): Alterations of fronto-temporal connectivity during word encod-ing in schizophrenia. Psychiatry Res 154:221–232.

5. Ford JM, Mathalon DH, Whitfield S, Faustman WO, Roth WT (2002):Reduced communication between frontal and temporal lobes duringtalking in schizophrenia. Biol Psychiatry 51:485– 492.

6. Lawrie SM, Buechel C, Whalley HC, Frith CD, Friston KJ, Johnstone EC(2002): Reduced frontotemporal functional connectivity in schizophre-nia associated with auditory hallucinations. Biol Psychiatry 51:1008 –1011.

7. Henseler I, Falkai P, Gruber O. (2010): Disturbed functional connectivitywithin brain networks subserving domain-specific subcomponents ofworking memory in schizophrenia: Relation to performance and clinicalsymptoms. J Psychiatr Res 44:364 –372.

8. Woodward ND, Waldie B, Rogers B, Tibbo P, Seres P, Purdon SE (2009):Abnormal prefrontal cortical activity and connectivity during responseselection in first episode psychosis, chronic schizophrenia, and unaf-fected siblings of individuals with schizophrenia. Schizophr Res 109:182–190.

9. Honey GD, Pomarol-Clotet E, Corlett PR, Honey RA, McKenna PJ, Bull-more ET, Fletcher PC (2005): Functional dysconnectivity in schizophre-nia associated with attentional modulation of motor function. Brain128:2597–2611.

0. Jeong B, Wible CG, Hashimoto R, Kubicki M (2009): Functional andanatomical connectivity abnormalities in left inferior frontal gyrus inschizophrenia. Hum Brain Mapp 30:4138 – 4151.

1. Friston K (2009): Dynamic causal modeling and Granger causality. Com-ments on: The identification of interacting networks in the brain usingfMRI: Model selection, causality and deconvolution [published online

ahead of print September 19]. Neuroimage.

2. Friston KJ, Harrison L, Penny W (2003): Dynamic causal modelling. Neu-roimage 19:1273–1302.

3. Allen P, Stephan KE, Mechelli A, Day F, Ward N, Dalton J, et al. 2010Cingulate activity and fronto-temporal connectivity in people with pro-dromal signs of psychosis. Neuroimage 49:947–955.

4. Zalesky A, Fornito A, Seal ML, Cocchi L, Westin CF, Bullmore ET, et al.(2011): Disrupted axonal fiber connectivity in schizophrenia. Biol Psychi-atry 69:80 – 89.

5. Skudlarski P, Jagannathan K, Anderson K, Stevens MC, Calhoun VD,Skudlarska BA, Pearlson G (2010): Brain connectivity is not only lowerbut different in schizophrenia: A combined anatomical and functionalapproach. Biol Psychiatry 68:61– 69.

6. Honey CJ, Sporns O, Cammoun L, Gigandet X, Thiran JP, Meuli R, Hag-mann P (2009): Predicting human resting-state functional connectivityfrom structural connectivity. Proc Natl Acad Sci U S A 106:2035–2040.

7. MacDonald AM, Carter CS (2004): Event-related fMRI study of context-processing in dorsolateral prefrontal cortex of patients with schizophre-nia. J Abnorm Psychol 112:689 – 697.

8. MacDonald AW 3rd, Carter CS, Kerns JG, Ursu S, Barch DM, Holmes AJ, etal. (2005): Specificity of prefrontal dysfunction and context processingdeficits to schizophrenia in never-medicated patients with first-episodepsychosis. Am J Psychiatry 162:475– 484.

9. Cohen JD, Botvinick M, Carter CS (2000): Anterior cingulate and prefron-tal cortex: Who’s in control? Nat Neurosci 3:421– 423.

0. Botvinick M, Braver TS, Barch DM, Carter CS, Cohen JD (2001): Conflictmonitoring and cognitive control. Psychol Review 108:624 – 652.

1. Botvinick MM, Cohen JD, Carter CS (2004): Conflict monitoring andanterior cingulate cortex: An update. Trends Cogn Sci 8:539 –546.

2. Kerns JG, Cohen JD, MacDonald AM, Cho RY, Stenger VA, Carter CS(2003): Anterior cingulate conflict monitoring and adjustments in con-trol. Science 303:1023–1026.

3. Milham MP, Banich MT, Claus ED, Cohen NJ (2003): Practice-relatedeffects demonstrate complementary roles of anterior cingulate andprefrontal cortices in attentional control. Neuroimage 18:483– 493.

4. Bassett DS, Meyer-Lindenberg A, Achard S, Duke T, Bullmore E (2006):Adaptive reconfiguration of fractal small-world human brain functionalnetworks. Proc Natl Acad Sci U S A 103:19518 –19523.

5. van den Heuvel MP, Stam CJ, Kahn RS, Hulshoff Pol HE (2009): Efficiencyof functional brain networks and intellectual performance. J Neurosci29:7619 –7624.

6. Bassett DS, Bullmore ET, Meyer-Lindenberg A, Apud JA, Weinberger DR,Coppola R (2009): Cognitive fitness of cost-efficient brain functionalnetworks. Proc Natl Acad Sci U S A 106:11747–11752.

7. Micheloyannis S, Pachou E, Stam CJ, Breakspear M, Bitsios P, Vourkas M,et al. (2006): Small-world networks and disturbed functional connectiv-ity in schizophrenia. Schizophr Res 87:60 – 66.

8. Rubinov M, Knock SA, Stam CJ, Micheloyannis S, Harris AW, Williams LM,Breakspear M (2009): Small-world properties of nonlinear brain activityin schizophrenia. Hum Brain Mapp 30:403– 416.

9. Zalesky A, Fornito A, Harding IH, Cocchi L, Yucel M, Pantelis C, BullmoreET (2010): Whole-brain anatomical networks: Does the choice of nodesmatter? Neuroimage 50:970 –983.

0. Fornito A, Zalesky A, Bullmore ET (2010): Network scaling effects ingraph analytic studies of human resting-state FMRI data. Front SystNeurosci 4:22.

1. Hayasaka S, Laurienti PJ (2010): Comparison of characteristics betweenregion- and voxel-based network analyses in resting-state fMRI data.Neuroimage 50:499 –508.

2. Tost H, Braus DF, Hakimi S, Ruf M, Vollmert C, Hohn F, Meyer-LindenbergA. (2010): Acute D2 receptor blockade induces rapid, reversible remod-eling in human cortical-striatal circuits. Nat Neurosci 13:920 –922.

3. Lui S, Li T, Deng W, Jiang L, Wu Q, Tang H, et al. (2010): Short-term effectsof antipsychotic treatment on cerebral function in drug-naive first-epi-sode schizophrenia revealed by “resting state” functional magnetic res-

onance imaging. Arch Gen Psychiatry 67:783–792.

www.sobp.org/journal