Psychophysiology,]]](2011),1–8.WileyPeriodicals,Inc.PrintedintheUSA.Copyrightr2011SocietyforPsychophysiologicalResearchDOI:10.1111/j.1469-8986.2011.01205.x

Thisoughttobegood:Brainactivityaccompanyingpositiveandnegativeexpectationsandoutcomes
YULIAO,a,bKLAUSGRAMANN,cWENFENGFENG,a,bGEDEONO.DEA
K,dandHONGLIa,ba
b
KeyLaboratoryofCognitionandPersonality(SWU),MinistryofEducation,Chongqing,China
c
SchoolofPsychology,SouthwestUniversity,Chongqing,Chinad
SwartzCenterforComputationalNeuroscience,UniversityofCalifornia,SanDiego,LaJolla,California,USADepartmentofCognitiveScience,UniversityofCalifornia,SanDiego,LaJolla,California,USA
Abstract
Thecurrentstudyemployedamodi edgamblingtask,inwhichprobabilisticcueswereprovidedtoelicitpositiveornegativeexpectations.Event-relatedpotentials(ERPs)to‘‘ naloutcome’’and‘‘probabilisticcues’’wereanalyzed.Differencewavesbetweenthenegativeconditionandthecorrespondingpositiveconditionwereexamined.Theresultscon rmthatfeedbackrelatednegativity(FRN)amplitudeismodulatedbytheinteractionofoutcomevalenceandexpectancybyshowinglargerFRNdifferencewavesforunexpectedthanexpectedoutcomes.Moreinterestingly,thedifferencewavebetweenERPselicitedbypositiveandnegativeexpectationsshowedanegativede ection,withafrontalmidlinesourcedensityaround280msafteronsetofthepredictivecue.NegativeexpectationswereassociatedwithlargerFRNamplitudesthanpositiveexpectations.ThissuggeststhatFRNiselicitedbyprobabilisticcuestopendingoutcomes.Descriptors:ERN,Errorsignals,Feedback,FRN,Rewardexpectancy,Reinforcementlearning
Ourbrainscontinuouslypredictfutureoutcomesandcompareexpectedorworsethanexpected,arespectivepositiveornegativethesepredictionswithsubsequentoutcomes.ThediscrepancypredictionerrorsignalisissuedbythemidbraindopaminesystembetweenanoutcomeandapriorpredictionisoftencalledtoACC,wherethatactivitymanifestsatthescalpasanamplitudepredictionerror.PredictionerrorsarecriticalforlearningandmodulationoftheFRN(Holroyd&Coles,2002).LargerFRNdecisionmakingbecausetheyconstitutefeedbackthatcanmod-amplitudesareusuallyobservedtobeassociatedwithoutcomesifyfuturedecisionsandincreaseaccuracyinsimilareventsorthatareworsethanexpected(e.g.,Holroyd&Krigolson,2007;situations(Rescorla&Wagner,1972;Sutton&Barto,1998).Miltneretal.,1997;Ruchsow,Grothe,Spitzer,&Kiefer,2002;seeThus,predictionerrorfeedbackisnecessarytoimprovedecisionNieuwenhuisetal.,2004,forareview).However,becausethemakingandperformance.
FRNisanalyzedbyusingadifferencewavesapproach,thediffer-Reinforcementlearning-ERNtheory(RL-ERNtheory)pro-encebetweennegativefeedbacktrialsandpositivefeedbacktrialsposesthatanevent-relatedpotential(ERP)componentknownasmightalsobecausedbyalargerpositive-goingde ectioninthefeedback-relatednegativity(FRN,alsocalledfeedbackerror-timerangeofthefERNelicitedbypositivefeedback(Holroyd,relatednegativityorfERN)isarewardpredictionerrorsignalPakzad-Vaezi,&Krigolson,2008).
(Holroyd&Coles,2002).FRNisafrontocentralnegative-goingConsistentwiththeRL-ERNtheory,previousstudieshaveERPthatoriginatesintheanteriorcingulatecortex(ACC;Am-shownthattheamplitudeoftheFRNismodulatedbyanin-iez,Joseph,&Procyk,2005;Brown&Braver,2005;Niki&teractionoffeedbackvalenceandexpectancy:Speci cally,largerWatanabe,1979;Ridderinkhof,Ullsperger,Crone,&Nieuwen-FRNsareelicitedbyunexpectednegativefeedback(Holroyd&huis,2004;Shidara&Richmond,2002).Itpeaksaround200toColes,2002;Holroyd,Nieuwenhuis,Yeung,&Cohen,2003;300msfollowingfeedbackanddiffersforunfavorableversusNieuwenhuis,Nielen,Mol,Hajcak,&Veltman,2005).More-favorableoutcomes(Gehring&Willoughby,2002;Holroyd&over,themodulationoftheinteractionbetweenoutcomevalenceColes,2002;Miltner,Braun,&Coles,1997;Nieuwenhuis,Hol-andexpectancyonFRNamplitudehasalargereffectwhentheroyd,Mol,&Coles,2004).AccordingtoRL-ERNtheory,anprediction–outcomeassociationisstronger.LargerFRNsareevaluativesystemmonitorsongoingeventstopredictwhetherelicitedforoutcomesthatareatoddswithrelativelymoresalienttheseeventswillendwellorpoorly.Whenthesystemrevisesitssubjectiveoutcomepredictions(Hajcak,Moser,Holroyd,&predictions,suchthatcircumstancesareeitherbetterthan
Simons,2007;Moser&Simons,2009)oriftheoutcomewaslesspredictable(Sailer,Fischmeister,&Bauer,2010).Inotherwords,Addresscorrespondenceto:Dr.HongLi,KeyLaboratoryofCog-rewardpredictionerrorseemstodependnotonlyonvalence,butnitionandPersonalityofMinistryofEducation,SouthwestUniversity,alsoonthedegreeoftheprediction–outcomediscrepancy.
No.1TianShengRd.,Beibei,Chongqing,China400715.E-mail:Asthebrainconstantlyupdatesoutcomepredictionsbasedlihong1@swu.edu.cn
onpastexperiencesandcurrentstimuli,predictionerrorscanbe
1
热门文档