Camerer-神經(jīng)經(jīng)濟學(xué)presentation課件

Neuroeconomicsof

GamesandDecisions*

ColinCamerer,Caltech

Neuroeconomics:Groundingmicro-economics

indetailsofneuralactivityPartofbehavioraleconomics(usingpsychologytoinformtheoriesofrationalitylimits)Partofexperimentaleconomics(newtechniques)Partofneuroscience(higherordercognition)*”Neuroeconomics”Camerer,Loewenstein,PrelecJEconLit(85pp),ScanJEcon(25pp),“Whyeconomicsneedsbrains”NemmersPrizetalkMay7,2005CollaboratorsCaltech:MeghanaBhatt,MingHsu,RalphAdolphs,CedricAnen,SteveQuartzIowa:DanTranel

Baylor:BrooksKing-Casas,DamonTomlin,ReadMontagueNemmersPrizetalkMay7,2005ThreedirectionsinneuroeconomicsI.Supportforrational-choicemodels“Belief”neuronsExpected-valueneurons“Monkeyshopping”satisfiesGARPII:SupportforbehavioralalternativesLoss-aversioninmonkeyshoppingLearningintrustgamesAmbiguityvs.risk(Knight,Ellsberg)III:NewconceptsEquilibriumasa“stateofmind”Neuralcorrelatesof“strategicIQ”BiologicalbasisofdemandNemmersPrizetalkMay7,2005Economically-importantregionsofthehumanbrainNemmersPrizetalkMay7,2005Cingulate(yellow),orbitofrontal(pink),

amygdala(orange), somatosensory(green), insula(purple)NemmersPrizetalkMay7,2005ImportantfactsaboutthebrainFunctional“modularity”……but“plastic”esp.inchildhoodBehaviordependsoncircuits

Humanbrainisprimatebrain+neocortexLanguage,socialorganization(institutions)Infants,fraternitypartiesshowsimilarityManybiologicalfunctionsareautomated;consciousattentionisscarce(flickerparadigm)NemmersPrizetalkMay7,2005I:Rationalchoiceinthebrain

Midbrainneuronsanticipatereward(L),encodevaluefunctionV(.)learning(R)(Schultz,Dayan,MontagueSci97)NemmersPrizetalkMay7,2005Neuronfiringrates(yaxis)encodeexpectedvalue(x-axis)(Glimcher)NemmersPrizetalkMay7,2005Monkeysplaymixedequilibrium

ashumansdo(Dorris-GlimcherNeuron04)NemmersPrizetalkMay7,2005Capuchinmonkeysrespondtoprices(KeithChenetal05)NemmersPrizetalkMay7,2005II:BehavioraleconomicsinthebrainMonkeychoicesaresensitivetoreferencepointsReferencepoint(initialfoodrewardendowment)

1 2

1 2Outcome11 (1,2)(1,2)Choice%79%21% 71%29%NemmersPrizetalkMay7,2005Designgoal:Linkstimuliwithunobservedparametricprocesses/variableswithcircuitry0-stepthinking1-stepthinkingEquilibriumC=br(B)w(red)-P(red)NemmersPrizetalkMay7,2005OverviewoffMRINemmersPrizetalkMay7,2005DatatransformationsRealignmentSmoothingNormalisationGenerallinearmodelStatisticalparametricmap(SPM)Imagetime-seriesParameterestimatesDesignmatrixTemplateKernelp<0.05StatisticalinferenceNemmersPrizetalkMay7,2005AmbiguityAversion(withMingHsuetal)Thismaterialisinreviewandcannotbepubliclycirculatedatthistime.NemmersPrizetalkMay7,2005III:NewideasLimitedplanninginbargaining

limitedstepsofthinkingEquilibriumasa“stateofmind”BiologicalbasesofdemandNemmersPrizetalkMay7,2005Rubinstein-Stahlalternatingoffer

shrinking-piebargaining1offersdivisionof$5------------accept↓2offersdivisionof2.50-----------accept ↓1offersdivisionof1.25------------accept ↓(0,0)NemmersPrizetalkMay7,2005Limitedplanninginbargaining(Science,03)NemmersPrizetalkMay7,2005Cognitivehierarchythinkingingames

(Camerer,Ho,Chong,QJE04)Step0playerschooserandomlyStepkplayershavebeliefsgk(h)Stepkplayerschooses*i(k)=argmax

sΣhgk(h)πi(s,s*(h))One-step-belowgk(k-1)=1Nagel(1995),Stahl-Wilson(1995),Costa-gomes-Crawford-Broseta(2001)Nornalizedoverconfidencegk(h)=gk(h)/Σhk-1

gk(h)gk(h)=0forh>kLinktohierarchicalQRE(Palfrey-Rogers-Camerer,onthiscomputer)NemmersPrizetalkMay7,2005Limitedequilibration

BeautycontestgameNplayerschoosenumbersxiin[0,100]Computetarget(2/3)*(

xi/N)Closesttotargetwins$20NemmersPrizetalkMay7,2005NemmersPrizetalkMay7,2005Neuralcorrelatesofiteratedbelief

(Bhatt-CamererGEBinpress)8dominance-solvablegames.C,B,2BinrandomorderforeachgamePaidforchoice(x$.30)oraccuracyB,2B($15)againstliveopponentoutsidethescanner.(Enablesmeasureofscanneronbehavior.)N=16CaltechcommunitystudentsNemmersPrizetalkMay7,2005Example:1-step(easy)game;BdominatesA.

L-Rpayoffseparationallows“eyetracking”NemmersPrizetalkMay7,2005Conformitytoequilibrium:

Therearemanynonequilibriumtrials

Note:Cmatches2BmoreoftenthanBmatches2BNemmersPrizetalkMay7,2005

Equilibriumisastateofmind:

Expectedrewardtheoryofmind+

(inequilibrium↓)(↓out-of-equilibrium)NemmersPrizetalkMay7,2005Subjectcomplainingafteranexperiment(Zamir,2000)NemmersPrizetalkMay7,2005Ultimatumgames:Thisisyourbrainonunfairness

(Sanfey,Rillingetal,Sci13March’03)NemmersPrizetalkMay7,2005Consistent2nd-orderbeliefs(c=2b)vsinconsistent(“false”)(c≠2b)2nd-orderbeliefconsistencydifferentialactivatesdorsolateralprefrontalcortex(DLPFC)DLPFCalsoseeninultimatumgamesafterlowoffers(“intentionsmatter”)DLPFCispartof2nd-orderbeliefcircuitry?NemmersPrizetalkMay7,2005DeactivationininsulaandhighstrategicIQStrategicIQ(x-axis):Howmuchyouearnfromchoices&beliefsCorrelated(-)withactivityinLinsulainchoicetaskAreoverlyself-focussedpeoplepoorstrategicthinkers?NemmersPrizetalkMay7,2005InsulaandlowstrategicIQStrategicIQ(x-axis):Howmuchyouearnfromchoices&beliefsCorrelated(-)withactivityinLinsulainchoicetaskAreoverlyself-focussedpeoplepoorstrategicthinkers?NemmersPrizetalkMay7,2005CorrelatesofhigherstrategicIQNemmersPrizetalkMay7,2005Newideas:BiologicalbasisfordemandEconomicstakesdemandasgiven.But…InfluenceofadvertisingFamiliarityandhabitformation(“tightplaylist”radiostations)Imitationofmoviestars/TVshows“LALaw”boominlawschoolapplicationsSense-makingdrivedemandfor“closure”lawsuitsMedia:“Ifitbleeds,itleads”,NASCARracesDoestheamygdalacontroltheTVremote?Addiction:Isgolforshoe-