Risky Decisions and Their Consequences: Neural Processing by Boys with Antisocial Substance Disorder ThomasJ.Crowley1*,ManishS.Dalwani1,SusanK.Mikulich-Gilbertson1,YipingP.Du2,CarlW.Lejuez3, Kristen M. Raymond1, Marie T. Banich2,4 1SubstanceDependenceDivision,PsychiatryDepartment,UniversityofColoradoDenver,Denver,Colorado,UnitedStatesofAmerica,2BrainImagingCenter,University ofColoradoDenver,Denver,Colorado,UnitedStatesofAmerica,3PsychologyDepartment,UniversityofMaryland,CollegePark,Maryland,UnitedStatesofAmerica, 4InstituteofCognitiveScience,DepartmentsofPsychologyandNeuroscience,UniversityofColoradoBoulder,Boulder,Colorado,UnitedStatesofAmerica Abstract Background:Adolescentswithconductandsubstanceproblems(‘‘AntisocialSubstanceDisorder’’(ASD))repeatedlyengage in risky antisocial and drug-using behaviors. We hypothesized that, during processing of risky decisions and resulting rewards and punishments,brain activationwould differbetweenabstinent ASD boysand comparison boys. Methodology/PrincipalFindings:We compared 20 abstinent adolescent male patients in treatment for ASD with 20 community controls, examining rapid event-related blood-oxygen-level-dependent (BOLD) responses during functional magneticresonance imaging.In90decisiontrialsparticipantschose tomakeeitheracautious responsethatearnedone cent,orariskyresponsethatwouldeithergain5centsorlose10 cents;oddsoflosingincreasedasthegameprogressed. Wealsoexaminedthosetimeswhensubjectsexperiencedwins,orseparatelylosses,fromtheirriskychoices.Wecontrasted decision trials against very similar comparison trials requiring no decisions, using whole-brain BOLD-response analyses of group differences, corrected for multiple comparisons. During decision-making ASD boys showed hypoactivation in numerousbrainregionsrobustlyactivatedbycontrols,includingorbitofrontalanddorsolateralprefrontalcortices,anterior cingulate, basal ganglia, insula, amygdala, hippocampus, and cerebellum. While experiencing wins, ASD boys had significantlylessactivitythancontrolsinanteriorcingulate,temporalregions,andcerebellum,withmoreactivitynowhere. DuringlossesASDboyshadsignificantlymoreactivitythancontrolsinorbitofrontalcortex,dorsolateralprefrontalcortex, brain stem,and cerebellum, with lessactivity nowhere. Conclusions/Significance:Adolescent boys with ASD had extensive neural hypoactivity during risky decision-making, coupledwithdecreasedactivity duringrewardandincreasedactivityduringloss.Theseneuralpatternsmayunderliethe dangerous,excessive,sustainedrisk-takingofsuchboys.Thefindingssuggestthatthedysphoria,rewardinsensitivity,and suppressed neural activity observed among older addicted persons also characterize youths early in the development of substanceuse disorders. Citation:CrowleyTJ,DalwaniMS,Mikulich-GilbertsonSK,DuYP,LejuezCW,etal.(2010)RiskyDecisionsandTheirConsequences:NeuralProcessingbyBoyswith AntisocialSubstanceDisorder.PLoSONE5(9):e12835.doi:10.1371/journal.pone.0012835 Editor:WangZhan,UniversityofCalifornia,SanFrancisco,UnitedStatesofAmerica ReceivedJuly15,2010;AcceptedAugust10,2010;PublishedSeptember22,2010 Copyright:(cid:2)2010Crowleyetal.Thisisanopen-accessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense,whichpermits unrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalauthorandsourcearecredited. Funding:NationalInstituteonDrugAbuse(www.nida.nih.gov)GrantsDA009842,011015;KaneFamilyFoundation(kanefamilyfoundation.org).Thefundershad noroleinstudydesign,datacollectionandanalysis,decisiontopublish,orpreparationofthemanuscript. CompetingInterests:Dr.CrowleyreceivestravelsupportfromtheAmericanPsychiatricAssociationtoparticipateinrevisingtheDiagnosticandStatistical ManualofMentalDisorders,andfromtheNationalInstituteonDrugAbuseforservingonitsNationalAdvisoryCouncil.Hisformerserviceonanadvisoryboardof CRSAssociateswassupportedbyReckittBenckiser.Theseinterestsdonotaltertheauthors’adherencetoallPLoSONEpoliciesondatasharingandmaterials.The otherauthorsreportnoconflictofinterest. *E-mail:[email protected] Introduction to take more risks than adults, but in laboratories and in real life ASD youths, even when abstinent, take more risks than other Some 200,000 adolescent admissions annually occur in adolescents [14,15]. Indeed, ASD’s symptoms of SUD and CD American substance-treatment programs [1]. Adolescent sub- (e.g., fire-setting, break-ins, and continued substance use despite stance use disorders (SUD) are so strongly comorbid with problems[16])epitomizeextremeriskybehaviors.Ofnote,ASD’s antisocial conduct disorder (CD) [2–4] that the combination risky behaviors are not necessarily ‘‘impulsive’’, i.e., done quickly may be termed ‘‘antisocial substance disorder’’ (ASD). Both without considering possible consequences. Indeed, they often antecedent genetic influences [5–8] and toxic effects of drugs require sustained preparation, such as ‘‘casing’’ a building before [9–12]maycontributetothesebehavioral problems,whichoften breaking in, or obtainingfalseidentification tobuyalcohol. persistfordecades[13].ASD’sgreatcosts,bothtothosewiththe TheexcessiveriskybehaviorsofASDyouthsmightresult,first, disorder and to society, make it important to understand this from aberrant neural processing of behavior-motivating rewards; condition’setiology. e.g., among normal adolescents a risk-taking propensity does ‘‘Risky behaviors’’ are behaviors that may result unpredictably correlate with more reward-related activation of nucleus accum- inrewardingand/oradverseoutcomes.Adolescentsgenerallytend bens (NAc) [17] (also see [18]). Second, aberrant processing of PLoSONE | www.plosone.org 1 September2010 | Volume 5 | Issue 9 | e12835 AntisocialBrains,Decisions behavior-inhibiting punishments could result in risky behaviors; e.g., after punished responses in reversal learning, children with psychopathictraitsshowabnormallyincreasedneuralactivationin ventromedialprefrontalcortex(vmPFC)andcaudate[19](alsosee [20]). Third, apart from initial processing of rewards or punishments, impaired integration of reward-punishment infor- mation in regions that decide on future behaviors could cause excessive risky behavior; e.g., under risky conditions substance- dependent adults under-recruit specialized conflict-monitoring circuitry in posterior mesofrontal cortex [21]; also see [22,23]. Toaddressthesethreepossibilities,weaskedwhetherASDyouths underconditionsofriskprocessdecisions,rewards,orpunishments differently fromcommunity-comparison youths. Only a few studies have compared brain activation in ASD youths and controls. ASD youths did show greater activation in amygdalaandregionsofthedefaultnetworkwhileperformingthe Stroop task [24]. In a go/no-go task marijuana-using youths (without CD) had more activation frontally (and elsewhere) than controls [25]. Conversely, youths with familial risk for ASD had lessfrontalactivationthancontrolsduringamotorinhibitiontask [26], perhaps like substance-involved adults who, when consider- Figure 1. Schematic illustration of dysphoria induced by ing risky decisions, showed hypoactivity in brain regions repeated intoxication-withdrawal cycles. Each intoxication leads processing potential losses andresponse conflicts [21]. to a subjective ‘‘high’’, with enhanced response to reward due to Structural alterations of brain have been associated with the reduced reward thresholds in medial forebrain bundle. Each acute withdrawaleventresultsinsubjectivedysphoriawithreducedresponse risk-taking of ASD youngsters, even among those merely torewardduetoelevatedrewardthresholds.Frequentcyclerepetitions vulnerable to ASD through family history. Youngsters with CD graduallysuppresssubjective‘‘highs’’,deepeningdysphoriabyfurther reportedlyhavereducedvolumeininsulaandamygdala[27],and raising reward thresholds. Increasingly, the drug is used to escape in temporal lobes, hippocampus, and vmPFC [28]. Compared dysphoria and achieve normal mood. During abstinence, mood withcontrols,alcohol-na¨ıvesonsofalcoholicmenreportedlyhave recoversveryslowly.(BasedonKoobandVolkow[33]). widespread gray-matter volume reductions, the severity of which doi:10.1371/journal.pone.0012835.g001 correlateswiththeseverityofinattention,impulsivity,hyperactiv- ity, and conduct problems [29]. Aggression and defiance brieflywouldrelievethatdysphoria,negativelyreinforcingfurther negatively correlate with right ACC gray-matter volume among drug use (Fig.1 and[34]). community boys not selected for ASD [30], while impulsivity Withsuchconflictingsuggestionsintheliterature,wecouldnot negatively correlates withvmPFC volume [31]. make a directional hypothesis for this study. Thus, we simply BecauseASDyouthscombineantisocialconductproblemswith hypothesized that, as adolescent boys repeatedly decide between SUD, recent publications suggest partially conflicting possibilities doingariskyoracautiousbehavior,andastheyexperiencewins for the neural underpinnings of their problems. First, like adults or losses from their risky choices, functional magnetic resonance with antisocial or psychopathic traits (but substance-free) [32], imaging (fMRI) will show that youths with ASD have different ASD youths’ repeated risk-taking might occur because they brain activation patterns than community-control boys. Unlike experience increased dopaminergic response to reward anticipation. some previous adolescent studies, our z-shim procedure [35] AmongantisocialadultsimpairedamygdalaandvmPFCfunction allowed good visualization of orbitofrontal regions that are also are thought [23] to reduce responses to punishment or loss. importantinprocessingrewardandpunishment[18].Ourresults Increased response to reward and decreased response to strongly supported our hypothesis. punishmentcouldcauseexcessivepursuitofexcitingrewardswith failure toinhibit behaviors that maybe punished. Methods Alternatively, reviewing human and animal studies, Koob and Participants and Assessments Volkow [33] suggest that repeated intoxication-withdrawal cycles from addictive drugs are associated with decreased dopaminergic EthicsStatement:Writteninformedconsent(adults)andassent response to reward, due to increased stimulation thresholds in (minors) was obtained from all research subjects. The Colorado compromised reward circuits (see also [18]). These processes Multiple Institutional ReviewBoard approved allprocedures. would produce ‘‘reward insensitivity’’, reducing motivation for Patients and controls were males, ages 14–18 years (inclusive) non-drug stimuli. Koob and Volkow [33] also indicate that withIQ$80,withoutknownMRIcontraindications(claustropho- chronic drug use disrupts frontal activity in ACC, OFC, and bia,orthodonticbraces,colorblindness,ferricmetalinthebody), DLPFC, a disruption continuing well into protracted abstinence. and without history of unconsciousness .15minutes, serious Becausethoseareascontributetodecision-makingandbehavioral neurologicalillness,orneurosurgery.Theyandtheirparentsspoke inhibition, such disruption would facilitate recurring risk-taking sufficient English for consenting. After explanation of procedures and relapses. These authors further propose that repeated 18-year-old subjects provided written informed consent for intoxication-withdrawal cycles activate a brain stress system participation; those ,18 years old provided written assent and mediated by corticotropin releasing factor (CRF) and other parentsprovidedconsent.Subjectswerepaid$50,wonameanof neurotransmitters [33]. They suggest that in human addicts $6.25 more in the behavioral task, and earned $3 more if head hypodopaminergic reward insensitivity and stress activation movement was,2mmduring theMRI. presentassubjectivedysphoria,a‘‘negativeemotionalstate’’that Patients’ inclusion criteria were: in treatment in our programs continueslongintoprotractedabstinence(Fig.1).Relapsesatleast for youths (most referred by criminal-justice or social-service PLoSONE | www.plosone.org 2 September2010 | Volume 5 | Issue 9 | e12835 AntisocialBrains,Decisions agencies and on probation); serious antisocial problems including Ofthefourtobacco-experiencedcontrolsubjects,onereported DSM-IV [16] CD symptoms; DSM-IV [16] substance abuse or using tobacco regularly. All 20 patients reported smoking in the dependence on a non-nicotine substance; and multisubstance last6months,but14werenowinaresidentialtreatmentprogram urine and saliva tests drug-free $7 days before assessment. that vigorously suppressed smoking. Thus, we estimated that 6 Patients’ exclusion criteria were: psychosis; current high risk of non-residential patients and one control had used tobacco in the suicide,violence,orfiresetting;orintreatmentandabstinent$30 fewdaysbeforeimaging.Nosubjectssmokedduringthe1hrpre- days (to minimize treatment effects on risk-taking). We obtained MRI training. assent/consent on 28 patients, excluding 1 because of past embedded metal, 2 for not meeting substance diagnostic criteria, Behavioral Tasks and Analyses 4formotionduringimaging,and1forbrainabnormalitiesnoted In a mock scanner subjects practiced our Colorado Balloon during scanning.Twenty others completed allprocedures. Game (CBG; Fig. 2), which is conceptually different from the Tomaximizesimilaritywithpatientswerecruitedcontrolsinzip- Balloon Analogue Risk Task that we previously employed with code areas from which previous patients had come. One was similarpatients[15].Wethenconductedrapidevent-relatedfMRI referred by a previous control. All others were contacted by a ofneuralprocessing(a)assubjectsdecidedbetweendoingarisky telemarketing company that phoned, described the project, and oracautiousbehavior,and(b)astheyexperiencedwinsorlosses invited families with possibly-qualifying children to accept a call fromriskybehaviors.‘‘DecisionBalloons’’(DecBa)weretesttrials fromtheresearchers,whothenmetwiththeyouthandaparentor thatforcedachoicebetweendoingariskyoracautiousbehavior guardiantoexplaintheproject,invitingwrittenparentalconsent, and then provided relatively large monetary wins or losses after and youth assent or consent, to participate. Regarding age, risky behaviors. ‘‘Directed Balloons’’ (DirBa) were ‘‘baseline gender,English-languageskills,andIQ,inclusioncriteriawerethe comparison’’ trials that required no decisions and provided only same as patients’. Exclusion criteria were: court convictions a small monetary reward for following a direction. DecBa and (exceptminortrafficorcurfewoffenses);substance-relatedarrests, DirBasharedidenticalmotorresponsesandalmostidenticalvisual treatment, school-expulsions; obvious psychosis; physical illness; andauditorystimuli(exceptfortheinitialfull,vs.half,yellowlight urine or breath tests containing non-prescribed substances a few (Fig. 2B)), but only DecBa forced decisions and gave larger days, or immediately, before scans; meeting criteria for DSM-IV rewards or losses for risky decisions. Thus, we reasoned that CD in the last year; or non-tobacco substance dependence. As subtractingbaselineDirBabrainactivationfromDecBaactivation samplesaccumulated,weskewedcontrolrecruitment(e.g.,seeking should remove visual-, auditory-, and motor-related activation, olderboys)tomaintainpatient-controlcomparability.Twenty-five while highlighting decision-related, and win-or-loss-related, acti- controlcandidatesprovidedassent/consent,butweexcluded1for vation. a substance-positive test, 2 for MRI-incompatible metal, 1 for motionduringimaging,and1forsignallossfromalargesinus;20 Imaging Neural Processing others completed all procedures. Psychosocial assessments were completed several days before Goggles and earphones delivered CBG’s stimuli (Fig. 2A,B). fMRI’s. Senior staff trained Bachelor-level interviewers and TheCBGpresented90pairsofballoons,eachpairincludingone examined all records for accuracy. Typical interview time was DecBa(‘‘Youdecidewhichbuttontopress’’)andoneDirBa(‘‘The 2 hrs for controls and 3hrs for patients (who reported more computerwillplaythegame.Youdon’tneedtodecideanything’’). symptoms). Assessments were: Child Behavior Checklist (CBCL) Subjectsrespondedwithrightandleftindexfingersonfiber-optic and Youth Self Report (YSR) [36,37] for symptom severity of button response pads. Balloons within a pair usually were attention-deficit/hyperactivity disorder (ADHD), anxiety, and separated by balloons from other pairs (average 2.9, range 1–5, depression; Diagnostic Interview Schedule for Children (DISC- balloons,programmedwith‘‘optseq2’’[48].Eachsubject’ssession IV) [16,37,38] for CD symptoms and diagnoses; Composite was divided into 3 runs, each presenting 30 identically-ordered International Diagnostic Interview-Substance Abuse Module DecBa-and-DirBaballoonpairs.EachDecBaorDirBatrialended (CIDI-SAM) [37,39–41] for DSM-IV abuse or dependence for with a fixation screen (Fig. 2B), usually 2 sec, but in each of the 11 substance categories; Peak Aggression Rating Scale [37]; subject’s 3 runs four trials were ‘‘jittered’’ to 4 sec. Subjects Carroll Self-Rating Scale for depression severity [37,42,43]; received the amount on the counter at game’s end; that amount Synergy Interview [37] for education, legal issues, and medical/ could not fallbelow $3.00. psychologicalhistory;ModifiedHollingshead-RedlichSocialClass Across the 90 DecBa trials the reward schedule changed. To Rating [44]; Wechsler Abbreviated Scale of Intelligence (WASI) model real-life shifting of reward contingencies, risky right [45]VocabularyandMatrixReasoningforIQestimates;Eysenck responses were very likely to be rewarded early, and punished JuniorImpulsivenessScale[46];andhandednesspreference[47]. later,inthegame(Fig.2C).Moreover,risky-responsepunishments Treating therapists tested patients’ urine about weekly for (10 cents)werelargerthanrewards(5cents)tofurtherencourage substances. Researchers tested patients and controls with urine gradual shifting from risky-right to cautious-left responding. (AccuTestTM) and saliva (AlcoScreenTM) dipsticks about 1 week, Subjects only were advised, ‘‘Try to guess whether [the balloon] and immediately, before scanning. willpopfromwhatthelastfewwhole-yellowlightballoonsdid.If thelastfewpopped,maybethisonewillpop.Ifthelastfewdidn’t Estimating Abstinence Duration pop,maybe thisonewon’tpop.’’ At treatment admission 14 patients, most referred from strictly Imaging Decision-Making. During DecBa’s 4-sec yellow controlledenvironments,producedanadmissionurinesamplefreeof light(Fig.2B),subjectsdecidedwhethertheywouldmakealeftor unprescribeddrugs;12ofthosealsodeniedanysubstanceuseinthe a rightresponsewhenthegreenlightcameon. Sincetheyellow- previous 30 days and continued producing substance-free urine light preceded responding, these 2 TRs reflected processing of samples.Forthose12weestimatedabstinencedurationas:(30days)+ decision,notresponse.DecBawasthetesttrialandDirBawasthe (number of days between admission and imaging). For all other ‘‘baseline comparison’’ trial; DirBa, unlike DecBa, required only patients abstinence duration was the length in days of a continuous compliance with a simple direction and no risky-vs.-cautious seriesofduring-treatmentnegativeurinesamplesbeforeimaging. decision-making. PLoSONE | www.plosone.org 3 September2010 | Volume 5 | Issue 9 | e12835 AntisocialBrains,Decisions Figure2.ColoradoBalloonGame.A.Decision-Balloonscreen,yellowlightilluminated.Counterinitially$5.B.Eventsduringpresentationof90 pairedtrials,eachDecisionBalloon(DecBa)followedbyaDirectedBalloon(DirBa).Top:timing(seconds).Coloredcirclesrepresentstoplightlights. DecBabegins(B,upper):yellowlightilluminated4sec,subjectdecidestopressleft(L)orright(R)button.Greenlight(0.5sec),subjectexecutes response. Red light, consequence appears (3.5sec). Risky right press consequence, either: (a) ‘‘smiley face’’, expanding balloon, puffing sounds, counteradds5cents,or(b)‘‘pop’’sound,shrinkingballoon,‘‘frowneyface’’,counterloses10cents.Cautiousleftpressconsequence:+1centon counter,dull‘‘thud’’sound,unchangedballoon.Then,‘‘jittered’’fixation.DirBa’s(B,lower)areidenticaltotheirpairedDecBa’sexcept:onlyhalfof initialyellowlightilluminates,signaling(i)startofaDirBaand(ii)buttontopressduringgreenlight(e.g.,rightillumination–pressright)–thesame buttonchosenduringprecedingpairedDecBa.Green-lightpressondirectedbutton:+2centsoncounter.Thenballoonrepeatstheconsequence (puffup,pop,ornochange)ofpreviouspairedDecBa;subjectwastoldthatDirBaconsequenceswouldnotaffectearnings.Finally,jitteredfixation screen.C.DuringDecBa,decliningproportionofrightpressesprogrammedtowinasgameprogresses.Mostlypressingleftlateringamesaves earnings.D.VisualAnalogScales(VAS).Aftersessionssubjectsratedtheiropinionsaboutthestatedquestionson100mmlines.Markedpositions representall-subjectmeans;groupsdidnotdiffersignificantly.UpperVAS:subjects’understandingofdecision-makingsourceforDecBavs.DirBa. LowerVAS:Differentemotionalresponsestopuff-upsorpopsofDecBa,vs.DirBa.E.Meannumbers,DecBaselections. doi:10.1371/journal.pone.0012835.g002 PLoSONE | www.plosone.org 4 September2010 | Volume 5 | Issue 9 | e12835 AntisocialBrains,Decisions ImagingReward-PunishmentProcessing. After the4-sec The50-minsessionendedwithT1FLAIRimages(T1-weighted yellow light (Fig.2B) subjects responded during the0.5-sec green spin-echo data set: 31 slices of part head, matrix=2566192, light, and then during the 3.5-sec red light they observed the NEX=2, TE/TR/TI=7.3ms/2000ms/860ms; imaging time= consequences (risky right-response win: smiley face, puff-up sound, 4 min, 25sec). Additionally, we acquired one IR-EPI (TR= balloon enlarges, counter increases 5 cents; risky right-response loss: 2000ms, TE=26ms) volume (with excellent contrast between frowney face, pop sound, balloon shrinks, counter decreases gray and white matter) to improve coregistration between EPIs 10cents;cautiousleftresponse:noface,dullthudsound,nochangein andtheIR-SPGR. balloon, counter increases 1 cent). Hence, combining the 2 TR’s Our fast z-shimmed image acquisition was designed to reduce thatspannedthegreen-andred-lightperiods(4sectotal;Fig.2B) inferior frontal susceptibility artifact [35]. Compensation was permitted us to assess the processing of reward or punishment appliedonlytoafewslicescoveringtheinferiorfrontalregionto (across-subject mean: 29 wins, and separately, 23 losses from 52 improve temporal resolution in a whole brain scan. Slice- rightpresses (Fig.2E)). acquisition order assured effective, constant repetition time in In the mock-scanner practice session subjects learned that both the z-shim slices and other slices. Moreover, we applied z- during DirBa’s red-light periods the counter increased 2 cents if shim compensation to 5 of the 31 slice locations in the OFC the subject responded on the signaled side, regardless of region. To optimize the amplitude of z-shim compensation subsequent audio-visual consequences (Fig. 2B); the latter were gradient, G, we ran on each subject a trial scan with 3 different c identicalineachDecBa-DirBapair.DirBa’salways-predictable2- G values (i.e., 0.55, 0.70, and 0.85G ), where G is an c null null cent reward for compliance was risk-free, certain, and consider- amplitude that nulls the MRI signal in regions without ably smaller than the 5-cent ‘‘win’’ reward, or the 10-cent ‘‘loss’’ susceptibility effect. We determined that a G of 0.70 G gave c null punishment, that followed DecBa’s risky choices. Accordingly, to optimalsignalrecoveryintheventral-medialOFC.ThisG value c assesswin-or-lossrelatedactivationwesubtractedDirBaactivation produced robust OFCactivation. fromDecBaactivationduringthe4-secgreen-and-redlightperiod (Fig. 2B). In these analyses the first 0.5sec included green-light Image Analysis motorresponding,butineachpairoftrialstheDecBaandDirBa Data preprocessing included motion correction, coregistration green-lightstimuliandresponseswereidentical(Fig.2B),aswere to structural images, normalization to standard Montreal Neuro- the red-light audio-visual stimuli. The only DecBa-DirBa differ- logical Institute (MNI) space, and smoothing. For within-subject encewasthemeaningofthosered-lightstimuli(DirBa,2-centgain. fMRIanalyseswefittedpreprocesseddatawiththegenerallinear DecBa risky response: 5 cent win or 10cent loss; DecBa cautious model (GLM) of Statistical Parametric Mapping [49] software, response: 1 cent gain). Thus, subtracting DirBa activation from filtering low frequency noise, correcting for temporal autocorre- DecBa activation was designed to cancel out green-light-related lation,andconvolvingwithasinglecanonicalHRFsignal.A128-s activation, while highlighting activation associated with experi- highpassfilterremovedsignaldriftandlow-frequencyfluctuation. encinga win ora loss. The GLM model included these trial periods: decision, outcome Other Data. We recorded occurrence of left or right (win or loss), and fixation. We generated single-subject contrast responses and reaction times (from green-light onset to maps with SPM-2, analyzing brain-function differences in response), as well as the resulting consequences (i.e., counter contrasts of interest (e.g. DecBa vs.DirBa)as fixedeffects. changes and the balloon’s puff, pop, or no-change). Post-session For between-subject whole-brain analyses, we compared debriefings asked about in-magnet experiences, game strategies, groups’ single-subject contrast maps generated by SPM2 using etc. On Visual Analogue Scales (VAS; Fig. 2D) subjects rated (a) SPM5’s random effects models. We used SPM5’s ANCOVA to the extent to which they or the computer made the left-right adjustallfMRIanalysesforageandIQ(IQmean:patients97.1; response decision for DecBa and for DirBa, and (b) their happy- controls 104.9, t (38)=2.67; p=0.011) before producing final sad reactions to balloon puff-ups or pops. T-tests and chi-square statistical maps. testscomparedthegroupsondemographicandclinicalvariables, Agreeing that ‘‘research needs to move beyond the simple and on debriefing responses regarding DecBa and DirBa. Mixed identificationofsinglestructures’’[50],weconductedwhole-brain, modelsexaminedgroupandrundifferences(seebelow)inCBG’s rather than region-of-interest, analyses, considering all structures Total riskyrightpresses, andlast-session risky rightpresses. exceedingcluster-definedthresholds.Weexpectedtohavegreatest power in analyses that examined DecBa-DirBa differences in a Image Acquisition single group of 20 subjects, considering all 90 trials together. For Ina3TGeneralElectricMRIscanner,withstimulisynchronized those analyses we controlled for false positive results with voxel- to trigger pulses, subjects first observed a video during a 3D T1 levelfamily-wise error (FWE)correction (p ,0.05). corr. anatomicalscan(IR-SPGR,TR=9ms,TE=1.9ms,TI=500ms In analyses expected to have less power (comparing 2 groups, flipangle=10u,matrix=2566256,FOV=220mm2,1241.7mm and/or examining fewer trials), we used the cluster- level FWE thickcoronalslices;9min12sec). correction(AFNI’sAlphaSimprogram[51]),asusedpreviouslyby Sessions then presented 90 paired DecBa-and-DirBa trials, us [24] and by others publishing in PLoS One or other excellent divided into 3 runs. Each echo-planar (EPI) run (TR=2000ms, journals (e.g., [52–57]). In comparison with voxel-level FWE, TE=26ms, flip angle=70u, FOV=220mm2, 642 matrix, 36 cluster-level FWE controls for false-positive results and achieves slices, 4 mm thick, no gap, angled parallel to the planum p ,0.05bysimultaneouslyrequiringalesssignificantdifference corr sphenoidale) lasted 10min, 23sec, and had 30 paired DecBa- in activation at each individual voxel (p ,0.005), but also uncorr. and-DirBapresentations.One-minuterestimages(abstractnature requiring a simulation-determined minimum number of contigu- drawing) separated the3 runs. ous activated voxels in each cluster. In other words, voxel-level Individualtrialswerediscountedifthesubjectfailedtorespond FWE can identify an intensely activated, single-voxel ‘‘hotspot’’; behaviorally during the 0.5sec green light. Data from individual cluster-level FWE identifies multivoxel ‘‘warmspots’’ that, al- trialswithspike-likemovementofthehead.2 mmwerereplaced though less intensively activated, must be larger and so (like the with dummy fixation data. The subject was excluded from voxel-levelfocus)wouldonlyoccurbychanceinwholebrainata analyses if10ormoretrials ina30-trial runfailedthose criteria. multiple-comparisons-corrected probability of p ,0.05. Con- corr PLoSONE | www.plosone.org 5 September2010 | Volume 5 | Issue 9 | e12835 AntisocialBrains,Decisions sidering our 6 mm full-width-half-maximum smoothing, 1000 prevalence of CD, anxiety and depression (dysphoria) scores, Monte Carlo simulations estimated the overall significance level attention problems, estimated mean IQ, number of SUD (probabilityofafalsedetection)forthresholdingthe3Dfunctional symptoms, and prevalence of SUD. Patients had many more z-mapimageovertheentirebrainvolume,regardlessofactivation legal problems. withinthatmap.Thesesimulationsindicatedthatrequiringcluster Although symptom minimization is not uncommon in our size #97 voxels, and each voxel with an activation difference at patients’ self-reports, 19 reported symptoms meeting criteria for p =0.005, provided a whole-brain family-wise corrected DSM-IVconductdisorder(Table1).Fourteenreportedsymptoms (uncorr) false positiverate p =.05. qualifyingforDSM-IVsubstancedependenceonatleastonedrug (FWE-corr) We examined potential confounds in our data with SPM’s other than nicotine, and the other six reported symptoms ‘‘glass-brain’’ images that show all beyond-threshold areas of qualifying for non-nicotine substance abuse without dependence. activation. Those images were statistically adjusted with two We conclude that these patientshad ASD. continuous measures (ADHD severity (from CBCL) and depres- Nosubjects’urineorsalivacontainedalcoholornon-prescribed sion severity (Carroll rating)). Three other potential confounds drugsjustbeforescans.Weestimatedthatpatientswereabstinent werecategorical.Forthemwere-analyzedthedataafterexcluding a meanof38.6(range9–59)daysbefore imaging.Inthe30days 3 left-handed subjects,andseparately 7 currenttobacco smokers, before imagingonecontrol reportedusingalcoholon2daysand and separately after excluding subjects reporting prescribed another used cannabis on one day. We estimated that 6 patients medicationusearoundthetimeofscanning(6medicatedpatients andonecontrolhadusedtobaccointhefewdaysbeforeimaging; (A used amphetamine-dextroamphetamine and risperidone; B, theycouldusetobaccoadlibitumbeforecomingtothelaboratory, fexofenadine; C, fluoxetine, quetiapine; D, unidentified ‘‘ulcer but abstained foronehour before scans. drug’’;E,methylphenidate;F,unidentified‘‘asthmainhaler’’)and 4 medicated controls (A and B, amphetamine-dextroamphet- Behavior amine; C,albuterol; D,topiramate)). A post-fMRI VAS (Fig. 2D, upper line) showed that subjects Names for clusters’ regions of maximum activation follow the understoodthatdecisionsweremadebytheminDecBaandnotin NRW Research atlas [58]. Our ‘‘broad regions’’ include parts or DirBa(0mm=‘‘Itoldmyself’’;100mm=‘‘thecomputertoldme’’; allofBrodmannareas(BA)10–14and47inOFC[18];portionsof mean(6SD)scores(mm):DecBa,patients6.5(3.8),controls12.1 BA6,8,9,and46inDLPFC[58];inferiorBA8,andBA44and (3.8); DirBa, patients 95.2 (5.0), controls 88.0 (5.0); DecBa or 45inventrolateralPFC(VLPFC)[59];andportionsofBA6,8,9, DirBa trial type F(1,38)=333.8, p,0.0001; group NS). Another and10,andBA24,25,and32inmedialprefrontalcortex (Med VAS(Fig2D,lowerline)showedthat,asintended,thepuff-upsor PFC) [60,61]; BA 10 and 11, and the inferior and subgenual pops of DecBa produced stronger emotions than those of DirBa regionsofACC(BA24and32)invmPFC[23].Theatlasdoesnot (0mm=‘‘really, really happy’’; 100mm=‘‘really, really sad’’. identify NAc; we considered it bounded by MNI coordinates DecBa: DirBa x Puff: Pop interaction, F=47.6 (df=1,3); x=(6) 4 to 15; y=0 to 22; z=2 to 210 [62]. Our procedures p,0.0001; patient-control main effect, NS). Hence, the data cannot resolve ventral tegmental area (VTA) and adjacent indicated that subjects clearly understood the different expecta- substantia nigra (SN) from surrounding structures, so we labeled tions of DecBa and DirBa trials, and that the rewards and the region within coordinates x=614, y=14 to 28, z=24 to punishments elicited theexpectedemotional responses. 216as‘‘Midbrain (SN/VTA)’’ [63];our ‘‘y’’polarity isreversed Patients and controls did not differ in mean reaction times fromthisreference.Rarely,SPM5placedaclusteroutsideofgray (Table 2). However, patients failed to respond within the 0.5sec matter (e.g., in white matter), perhaps because of registration green-light limit on DecBa slightly but significantly more often errors.Iftheclusterwas#3mmfromgraymatter,welabeleditin than controls (patient mean 2.5 (6SD1.8) trials, controls 1.4 thenearest gray matter;ifit was.3mm,wedonot report it. (61.4), t(38)=2.03,p=0.049). Some of our data are expressed in standard SPM activation Considering all subjects together, the number of risky right units. The mean activation of all brain voxels (white and gray pressesdecreasedsignificantlyacrossthethree30trialruns.Mixed matter)duringtheentiresessionisnormalizedat100percent,and modelanalysisofriskyrightpressesevaluatedpotentialgroup,run, mean activation in each region during DecBa, and separately and group x run effects. Only the run effect was significant: F during DirBa, is scaled proportionately in percentages. For (1,40)=49.0, p,0.0001. The estimated mean decrease was 7.3 example, if in some cortical region, patients’ mean activation (1.0)presses. duringDecBais180percentoftheirmeanactivationinallbrain We thought that patients, compared to controls, might make voxels,andiftheirmeanactivationinthatregionduringDirBais more right presses overall, and especially on the last 30 DecBa 179 percent of mean activation in all brain voxels, then patients’ trials.However,thegroupsdidnotdifferinoverallrightresponses mean DecBa-minus-DirBa activation difference in that region is (patients’mean(6SD)52.2(2.3);controls’52.4(1.9),t(38)=0.07; 180-179=1 SPM activationunit. NS), nor in right responses in the final 30-trial run (patients 15.1 (4.8); controls 15.0(3.7), t (38)=20.07;NS). Results With no differences in right pressing, the groups did not differ significantly in wins (patients 28.3 (5.2); controls 29.9 (4.6); t Demographics (38)=21.0; p=0.32) or losses (patients 23.8 (6.0); controls 22.5 All patients had been referred to our program for youths with (4.8);t(38)=0.76;p=0.45).Thissimilarwin-lossexperiencehelps serious antisocial and substance problems. Fourteen were in in evaluating neural activation differences, since neither group residentialtreatment,4inday-treatment,and2wereoutpatients. experienced morefrustration-inducing losses. Patients and controls did not differ significantly in age or racial distribution (Table 1). However, patients’ mean socioeconomic Brain Activation During Decision-Making status score (equating to Social Class IV, lower middle class) was In DecBa trials subjects decided about their next response significantlylowerthancontrols’score(III,uppermiddleclass).As duringyellow-lightperiods,andinDirBatrialstheyweredirected expectedpatientshadsignificantlyworse(Table1)aggressionand on how to respond during yellow-light periods. In many frontal impulsivenessscores,conductproblems,numberofCDsymptoms, and subcortical regions deciding recruited significantly more PLoSONE | www.plosone.org 6 September2010 | Volume 5 | Issue 9 | e12835 AntisocialBrains,Decisions Table1. Subjectcharacteristics. Patient Control Test p-value N=20 N=20 DemographicandPsychiatric MeanAge(SD) 16.5(1.0) 16.5(1.6) t-test NS Caucasian(n) 12 15 chi-square NS Non-Caucasian(n) 8 5 SESScore:Mean(SD) 47(18) 35(16) t-test p,.03 SocialClass III IV AggressionScore:Mean(SD) 5.7(3.2) 0.5(1.1) M-WU p,.0005 EysenckImpulsivenessScore:Mean(SD) 11.9(6.0) 6.7(4.5) t-test p,.005 YouthSelfReport:CPMean(SD) 69.0(7.7) 53.8(4.6) t-test p,.0005 CDLifetimeSymptom:Mean(SD) 6.8(2.3) 0.5(0.6) M-WU p,.0005 CDLifetimeDiagnosis(n1) 19 1 chi-square p,.0005 CBCL,YSR2,Anx-Dept-score:Mean(SD) 57.4(8.6) 52.0(7.3) M-WU p,0.007 CBCL,YSR2,Att-ProbScalet-score:Mean(SD) 58.3(8.2) 53.5(4.5) t-test p=0.029 CarrollDepressionRatingScore:Mean(SD) 8.5(6.7) 4.1(3.8) M-WU p,.02 IQfull-scalet-score:Mean(SD) 97.1(9.3) 104.9(9.0) t-test p,.02 SubDepSymptoms,AcrossDrugs:Mean(SD) 12.4(7.2) 0.2(.67) M-WU p,.0005 SubstanceUseDisorders3,4 TobaccoDependence 13 1 chi-square p,.0005 AlcoholAbuse 8 0 FisherExact p,.004 AlcoholDependence 8 0 FisherExact p,.004 CannabisAbuse 7 0 FisherExact p,.009 CannabisDependence 10 0 chi-square p,.0005 CocaineAbuse 2 0 FisherExact NS CocaineDependence 2 0 FisherExact NS ClubDrugsAbuse 3 0 FisherExact NS ClubDrugDependence 4 0 FisherExact NS HallucinogenAbuse 2 0 FisherExact NS AmphetaminesDependence 2 0 FisherExact NS HallucinogenDependence 1 0 FisherExact NS LegalProblems5 LifetimeCourtAppearances:Mean(SD) 11(11) 0 LifetimeAdmissionstoDetentionorJail:Mean(SD) 3(4) 0 DaysonProbation,Last6Months:Mean(SD) 139(70) 0 Abbreviations Att-Prob,AttentionProblemsScale.Anx-Dep,Anxious-DepressedScale.CBCL,ChildBehaviorChecklist.CD,ConductDisorder.CP,ConductProblemst-score.M-WU, Mann-WhitneyUtest.SES,SocioeconomicStatus.SubDep,SubstanceDependenceSymptoms.YSR,YouthSelf-Report. Footnotes 1NocontrolsmetDSM-IV’spast-yearCDdiagnosticcriteria. 2ForonepatientwithnoChildBehaviorChecklist,YouthSelfReportscorewassubstituted. 3Fordrugsnotlisted,noknowncases. 4Multipledisordersinsomesubjects,sonumberssum.20. 5Nostatisticaltests,duetolackofvarianceincontrolsubjects. doi:10.1371/journal.pone.0012835.t001 activation than following a direction. This was true for controls Methods) to determine a cluster-size threshold ($97 contiguous (Table 3;inthisandeach subsequent table, afootnoteshows the voxels, each at p =0.005); such clusters were unlikely uncorr contrast analyzed) and for patients, although patients activated (p,0.05)tooccur by chanceinour whole-brain analyses. many fewer voxels and regions (Table 4). The high t-values in By that analysis, decision-making (compared to following a these single-group, all-trial, FWE analyses reflect their consider- direction)activatedalargesetofregionssignificantlymoreamong ablepower.Aformaltwo-groupcomparisonofregionsdifferently controls than among patients (Table 5; Fig. 3 (Decision)). activatedbycontrolsandpatientshadlesspowerthanone-group Conversely, in the reverse contrast no brain regions activated analyses, and the stringent FWE procedure found no group moreinpatients thanincontrols.Thediscrepancy (controls.pa- differences. Therefore, we used a simulation procedure (see tients,6233voxels(Table5);patients.controls,0voxels)strongly PLoSONE | www.plosone.org 7 September2010 | Volume 5 | Issue 9 | e12835 AntisocialBrains,Decisions Table2. Mean(SD) reaction times, msec.a difference was significantly greater among controls. Strikingly, however,insomeregionspatients’(butnotcontrols’)valueswere negative, indicating less activation during the decision-requiring DecBa than after thesimpledirectionsof DirBa. Control Patient t-value p To test whether patients’ negative DecBa-minus-DirBa differ- DecBab 257(30) 261(26) 20.47 NS ences in Fig. 4 were due to abnormally high DirBa values in DirBac 275(28) 273(31) 0.30 NS patients, we examined between-group differences in DirBa activation within the regions listed in Fig. 4. For each subject Footnotes aFromgreen-lightonsettoresponse. during DirBa, the mean activation within each region of interest bAllDecisionBalloonswithresponseduringgreen-lightperiod. was normalized (i.e., expressed in ‘‘SPM activation units’’ (see cAllDirectedBalloonswithresponseduringgreen-lightperiod. Methods)). We then conducted between-group t-tests on those doi:10.1371/journal.pone.0012835.t002 SPM activation values. In no brain regions did patients’ DirBa valuessignificantlyexceedthoseofcontrols(datanotshown).This supports the conclusion that patients had less activation than suggeststhat,incomparisontocontrols,patients’negativeDecBa- controls during riskydecision-making. minus-DirBavalueswereduetoreducedDecBaactivation,rather To illustrate sources of group differences, Fig. 4 shows each than toenhanced DirBa activation. group’s mean DecBa-minus-DirBa activation difference (not adjusted for age or IQ) for each cluster in Table 3. In some Brain Activation While Experiencing Wins or Losses regions both groups had more activation during decision-making In the 90 DecBa trials subjects averaged 52 risky right (DecBa) than while following a direction (DirBa), but that responses, producing a mean of about 29 5-cent wins and 23 Table3. Controls’lociofactivation duringdecision-making.A,B Structure BrodmannAreaorSideC ActivatedVoxels MaximumActivationD t x y z Sup&MidFrGy MainlyR9,10 293 36 54 10 19.1 SupFrGy MainlyR10 65 26 52 32 11.5 MidFrGy L10 16 230 44 30 9.3 MidFrGy R9 44 36 30 40 10.6 MedFrGy R9 15 22 36 28 7.7 ACCG MainlyR24,32 1776G 6 22 40 15.1 MedFrGytoSupFrGyG L,R6,8,9 4 34 36 18.0 SupFrGy(Pre-SMA)G MainlyR6 2 10 62 12.6 InfFrGyH R,L45,47 R3260H;L216H 36 24 210 12.9 InsulaH R,L13 40 18 2 11.7 NAcH R,L 12 14 22 11.7 CaudateH R,L 6 14 10 14.4 PutamenH R,L 214 6 24 12.1 Midbrain(SN,VTA)H R,L 10 212 214 11.0 ThalamusH R,L 34 6 212 14 13.7 PostCingGy R 51 2 226 30 9.4 CerebTonsil L 39 214 252 248 11.9 CerebAntLobeI R 43I 6 260 236 10.1 UvulaVermisI R 2 264 238 8.9 CerebTuber,Tonsil L 194 234 264 238 11.1 Clusters,10voxelsE - 54 - - - - TotalActivatedVoxels - 6100 - - - - Abbreviations ACC,anteriorcingulatecortex.Ant,anterior.Cereb,cerebellar.Cing,cingulate.Ctr,controls.DecBa,DecisionBalloons.DirBa,DirectedBalloons.Gy,gyrus.Inf,inferior.L, left.Med,medial.Mid,middle.NAc,nucleusaccumbens.Occ,occipital.Par,parietal.Post,posterior.Pt,patient.R,right.Sec,secondary.SMA,supplementarymotorarea. SN,substantianigra.Sup,superior.Temp,temporal.Uncorr,uncorrectedformultiplecomparisons.VTA,ventraltegmentalarea. Footnotes AProcedurefordeterminingsignificance:voxel-levelfamily-wiseerrorcorrection(p ,0.05). corr BContrastexamined:(DecBa)Ctr-(DirBa)Ctr. CIfbilateral,thelargestmaximumisshown. DMontrealNeurologicalInstitutecoordinates,mmfromanteriorcommissure. ECombinedvolumeofallclusterscomprising,10voxels. G-JRegionsbearingthesamesuperscriptcompriseoneactivatedcluster. doi:10.1371/journal.pone.0012835.t003 PLoSONE | www.plosone.org 8 September2010 | Volume 5 | Issue 9 | e12835 AntisocialBrains,Decisions Table4. Patients’ lociofactivationduring decision-making.A,B Structure BrodmannAreaorSideC ClusterSizeinVoxels MaximumActivationD t x y z Caudate,putamen,NAc R 169 14 22 210 11.1 Putamen,NAc L 58 216 18 210 7.9 ACCE MainlyR24,32 929E 6 28 34 15.3 MedFrGyE MainlyR6 0 16 50 8.9 SupFrGyE R8 2 14 56 9.3 InfFrGy R,L47 R184;L109 36 18 28 12.6 Midbrain(SN,VTA) R 18 4 214 210 8.5 Midbrain L 10 26 212 218 8.0 Midbrain R 11 4 226 26 8.5 Clusters,10voxelsF - 5 - - - - TotalActivatedVoxels - 1493 - - - - Abbreviations:AsinTable3. Footnotes AProcedurefordeterminingsignificance,asinTable3. BContrastexamined:(DecBa)Pt-(DirBa)Pt. CIfbilateral,thelargestmaximumisshown. DMontrealNeurologicalInstitutecoordinates,mmfromanteriorcommissure. ERegionsbearingthesamesuperscriptcompriseoneactivatedcluster. FCombinedvolumeofallclusterscomprising,10voxels. doi:10.1371/journal.pone.0012835.t004 10-cent losses (Fig. 2E). We contrasted DecBa’s 4-sec green-and- ratingsofdepressionandADHDseverity(Fig.5,Cells2A,2B);or redlightperiodswiththoseofDirBa,whichpaid2centsforeach afterweexcluded6patientsand1controlthoughttobecurrent, directedresponse.Weanalyzedwins,andseparatelylosses,finding regular cigarette smokers (Cells 3A, 3B); or after exclusion of 6 very distinctpatterns in thepatients andcontrols. patients and 4 controls using prescription medications (Cells 4A, In the DecBa-minus-DirBa contrast, controls as a group 4B); orafter exclusionof 3 left-handed subjects (Cells 5A,5B). significantly activated many structures, involving over 13,000 Similarly, Fig. 5, Cell 1C, shows regions in which controls’ voxels,whilepatientsactivatedfewerstructuresandabouthalfas activationexceededpatients’duringwins(fromTable8),andCell manyvoxels(Tables6,7).Inaformalcomparisonseekingregions 1D shows the reverse (patient.control) contrast. Again, the moreactivatedbycontrolsthanpatients,severalregionsandmany complete absence of patient.control activity in Cell 1D carries voxels activated significantly (Table 8; Fig. 3 (Win)); the opposite down after adjustment for ADHD and depression (Cell 2D), or contrast (patients.controls) found no regions activating signifi- after we excluded smokers (Cell 3D), or medicated subjects (Cell cantly. These observations indicate that controls were more 4D), orleft-handers (Cell 5D). sensitivetowinsthan patients. Finally, the shadowgram of Fig. 5, Cell 1E, shows that during Incontrolslosses(Table9)activatedfewerstructuresandvoxels losses controls’ activation exceeded patients’ in no regions, than wins (Table 6). Moreover, unlike wins, losses actually whereas many areas were activated in the patient.control activated slightly fewer voxels in controls (Table 9) than in contrast (Cell 1F). These differences persisted when we adjusted for ADHD and depression (Cells 2E, 2F), or when we excluded patients (Table 10). Indeed, in formal comparisons of the two smokers(Cells3D,3F)orleft-handers(Cells5D,5F).Thepattern groups we found no voxels more activated in controls than in patients,whilethepatient.controlcontrastfoundmanyactivated was broken only after we excluded 10 medicated subjects; then, among the remaining 10 there was greater activation in right structures and voxels (Table 11; Fig. 3 (Loss)); the largest cluster superior frontal gyrus (BA6) among controls, compared with wasinprefrontalcortex.Thesefindingsindicatethatpatientswere patients(Cell4E).However,thatsinglefinding,relyingononly10 more sensitivetolosses than controls. subjectsandanaverageof23DecBa-DirBatrials,doesnotnegate our conclusion: it appears unlikely that these potential confounds Possible Confounds explainthelargepatient-controlneural-activationdifferencesthat Comparedwithcontrols,patients’neuralfunctionwasreduced wereport. duringdecision-makingandwins,andenhancedduringlosses,and we sought confounds that might explain these differences. ‘‘Glass Discussion brains’’(Fig.5),2-dimensionalshadowgramsofallactivatedareas, obscure details but visually summarize important large-scale In this fMRI study we investigated the hypothesis that, while patterns. The shadowgram in Fig. 5, Cell 1A (Row 1, Column deciding between doing a risky or a cautious behavior, or while A), presents the data of Table 5 (Decision period, all trials, experiencing wins or losses from risky choices, youths with ASD control.patient activation regions), showing numerous activated would have different brain activation patterns than community- regions.Cell1Bshowsthereverse(patient.control)contrast,and control boys. Important design features included an adolescent noregionsactivate.ThestarkdifferencesbetweenCells1Aand1B sample with very serious antisocial and substance problems, a z- persisted when we simultaneously adjusted brain activity for shimproceduretoenhanceorbitofrontalimaging,andanoveltask PLoSONE | www.plosone.org 9 September2010 | Volume 5 | Issue 9 | e12835 AntisocialBrains,Decisions Table5. Lociactivating significantly more incontrols than inpatients duringdecision-making.A,B Structure BrodmannAreaorSideC ClusterSizeinVoxels MaximumActivationD t x y z SupFrGyK R10 344K 21 54 6 3.8 MidFrGyK R10 32 50 6 4.1 MedFrGyL L10 1724L 220 42 24 4.1 MidFrGyL L11 224 38 26 2.8 ACCL L24,32 224 34 18 4.1 ACCL R,L32 0 36 20 3.5 InsulaL L13 246 22 14 3.4 ClaustrumL L 224 20 12 3.5 ACCM R24,32 444M 12 20 36 3.7 MedFrGyM R9 20 36 24 2.9 CaudateN R 137N 19 24 6 3.1 PutamenN R 20 10 22 3.2 Insula R13 245 32 26 22 2.8 MidFrGy R6 169 32 22 48 3.7 Amygdala R 118 22 28 212 3.2 MedFrGyO L6 278O 216 210 58 4.0 SupFrGy(Pre-SMA)O L6 218 26 68 3.8 Pre-CentralGyO L4 216 226 60 3.7 MedFrGyP R6 369P 18 216 60 3.5 SupFrGy(Pre-SMA)P R6 22 214 71 4.2 MidTempGy R21 121 50 224 214 3.8 Hippocampus L 157 234 232 212 3.6 Post-CentralGy L3 154 260 216 50 3.6 Post-CentralGyQ L3 387Q 232 236 52 3.4 InfParietalLobuleQ L40 232 252 58 3.2 SupTempGy L41 206 242 236 6 3.9 PrecuneusR L31 256R 220 246 30 4.0 CingGyR L31 220 246 28 3.0 SupramarginalGy R40 175 60 254 26 3.5 LingualGy R19 126 30 258 2 2.9 AntLobe L 154 28 246 232 3.4 UvulaVermisS R 327S 6 262 236 5.1 CulmenVermisS R 2 262 230 3.6 PostLobeCerebellarTonsilS R 16 260 248 4.2 PostLobePyramisVermis L 342 210 276 234 4.1 TotalActivatedVoxels - 6233 - - - - Abbreviations:AsinTable3. Footnotes AContrastexamined:(DecBa-DirBa)Ctr-(DecBa-DirBa)Pt. BProcedurefordeterminingsignificance:Forvoxel-wiseuncorrectedp,0.005MonteCarlosimulationsindicatethatwhole-brainclusterwisethresholdp,0.05 (correctedformultiplecomparisons)requires.96clusteredvoxels. CIfbilateral,thelargestmaximumisshown. DMontrealNeurologicalInstitutecorrdinates,mmfromanteriorcommissure. K-RRegionsbearingthesamesuperscriptcompriseoneactivatedcluster. doi:10.1371/journal.pone.0012835.t005 that compared risk-taking trials with almost-identical risk-free and Volkow [33] predicted that during protracted abstinence trials. Wediscuss three mainfindings. addicts would show ‘‘disrupted activity of frontal regions, includ- ing dorsolateral prefrontal regions, cingulate gyrus, and orbito- Patients’ Neural Hypoactivity During Decision-Making frontal cortex’’, a disruption ‘‘hypothesized to underlie their While deciding between doing a risky or a cautious behavior, impaired inhibitory control and impulsivity … [contributing] to patients’ brains showed extensive neural hypoactivity. Koob relapse’’. PLoSONE | www.plosone.org 10 September2010 | Volume 5 | Issue 9 | e12835