decision neuroscience: an integrative...
TRANSCRIPT
DECISION NEUROSCIENCE
This page intentionally left blank
DECISIONNEUROSCIENCE
AN INTEGRATIVE PERSPECTIVE
Edited by
JEAN-CLAUDE DREHER
LEON TREMBLAY
Institute of Cognitive Science (CNRS), Lyon, France
AMSTERDAM • BOSTON • HEIDELBERG • LONDON
NEW YORK • OXFORD • PARIS • SAN DIEGO
SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO
Academic Press is an imprint of Elsevier
Academic Press is an imprint of Elsevier125 London Wall, London EC2Y 5AS, United Kingdom525 B Street, Suite 1800, San Diego, CA 92101-4495, United States50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United StatesThe Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom
Copyright © 2017 Elsevier Inc. All rights reserved.
No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical,including photocopying, recording, or any information storage and retrieval system, without permission in writing fromthe publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and ourarrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can befound at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by the Publisher (other than asmay be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and experience broaden ourunderstanding, changes in research methods, professional practices, or medical treatment may become necessary.
Practitioners and researchers may always rely on their own experience and knowledge in evaluating and using anyinformation, methods, compounds, or experiments described herein. In using such information or methods they should bemindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for anyinjury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use oroperation of any methods, products, instructions, or ideas contained in the material herein.
Library of Congress Cataloging-in-Publication DataA catalog record for this book is available from the Library of Congress
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
ISBN: 978-0-12-805308-9
For information on all Academic Press publicationsvisit our website at https://www.elsevier.com/
Publisher: Mara ConnerAcquisition Editor: April FarrEditorial Project Manager: Timothy BennettProduction Project Manager: Edward TaylorDesigner: Matthew Limbert
Typeset by TNQ Books and Journals
Contents
List of Contributors xiPreface xiii
IANIMAL STUDIES ON REWARDS,PUNISHMENTS, AND DECISION-
MAKING
1. Anatomy and Connectivity of theReward CircuitS.N. HABER
Introduction 3
Prefrontal Cortex 4
The Ventral Striatum 5
Ventral Pallidum 10
The Midbrain Dopamine Neurons 11
Completing the CorticaleBasal Ganglia RewardCircuit 14
References 15
2. Electrophysiological Correlates of RewardProcessing in Dopamine NeuronsW. SCHULTZ
Introduction 21
Basics of Dopamine Reward-Prediction ErrorCoding 21
Subjective Value and Formal Economic UtilityCoding 25
Dopamine and Decision-Making 27
Acknowledgments 29
References 29
3. Appetitive and Aversive Systems in theAmygdalaS. BERNARDI AND D. SALZMAN
Introduction 33
Conclusion 41
References 42
4. Ventral Striatopallidal PathwaysInvolved in Appetitive and AversiveMotivational ProcessesY. SAGA AND L. TREMBLAY
Introduction 47
The Corticobasal Ganglia Functional Circuits: TheRelation Between the Cortex and the BasalGanglia 49
The Direct and Indirect Pathways: From Inhibitionof Competitive Movement to AversiveBehaviors 49
Single-Unit Recording in Awake Animals toInvestigate the Neural Bases 50
Tasks to Investigate Appetitive Approach Behaviorand Positive Motivation 51
Ventral Striatum and Ventral PallidumAre Also Involved in Aversive Behaviors:First Evidence of Local Inhibitory Dysfunction 51
The Ventral Striatum and the Ventral PallidumEncode Aversive Future Events for Preparationof Avoidance and for Controlling Anxiety Level 53
Abnormal Aversive Processing Could BiasDecision-Making Toward PathologicalBehaviors 55
Conclusion 55
Acknowledgments 55
References 55
5. Reward and Decision Encodingin Basal Ganglia: Insights FromOptogenetics and Viral Tracing Studiesin RodentsJ. TIAN, N. UCHIDA AND N. ESHEL
Dopamine 59
Striatum 64
Conclusions 67
References 67
v
6. The Learning and Motivational ProcessesControlling Goal-Directed Action and TheirNeural BasesL.A. BRADFIELD AND B.W. BALLEINE
Learning “What Leads to What”: The Neural Basesof ActioneOutcome Learning 72
How Rewards Are Formed: Neural Bases ofIncentive Learning 74
Deciding What to Do: The Neural Bases ofOutcome Retrieval and Choice 76
How These Circuits Interact: Convergence on theDorsomedial Striatum 77
Conclusion 79
Acknowledgments 79
References 79
7. Impulsivity, Risky Choice, and ImpulseControl Disorders: Animal ModelsT.W. ROBBINS AND J.W. DALLEY
Delayed and Probabilistic Discounting: “ImpulsiveChoice” in Decision-Making 82
Premature Responding in the 5CSRTT 83
Stop-Signal Reaction Time 83
Neural Basis of Impulsivity 84
Neural Substrates of Waiting Impulsivity 84
Neural Substrates of Probability Discounting ofReward: Risky Choice 89
Motor Impulsivity: Stop-Signal Inhibition 90
Conclusion 90
Acknowledgments 91
References 91
8. Prefrontal Cortex in Decision-Making:The PerceptioneAction CycleJ.M. FUSTER
The PerceptioneAction Cycle 96
Inputs to the PerceptioneAction Cycle inDecision-Making 97
Prediction and Preparation Toward Decision 99
Execution of Decision 101
Feedback From Decision: Closure of thePerceptioneAction Cycle 102
References 103
IIHUMAN STUDIES ON MOTIVATION,PERCEPTUAL, AND VALUE-BASED
DECISION-MAKING
9. Reward, Value, and SalienceT. KAHNT AND P.N. TOBLER
Introduction 109
Value 110
Salience 113
Conclusions 118
References 118
10. Computational Principles of ValueCoding in the BrainK. LOUIE AND P.W. GLIMCHER
Introduction 121
Value and Choice Behavior 121
Value Coding in Decision Circuits 123
Context Dependence in Brain and Behavior 125
Neural Computations Underlying ValueRepresentation 127
Temporal Dynamics and CircuitMechanisms 130
Conclusions 133
References 133
11. Spatiotemporal Characteristics andModulators of Perceptual Decision-Makingin the Human BrainM.G. PHILIASTIDES, J.A. DIAZ AND S. GHERMAN
Introduction 137
Factors Affecting Perceptual Decision-Making 139
Conclusion 145
References 145
12. Perceptual Decision-Making: What DoWe Know, and What Do We Not Know?C. SUMMERFIELD AND A. BLANGERO
Introduction 149
What Are Perceptual Decisions? 149
Aims and Scope of This Chapter 150
Decision Optimality 150
Q1: How Is Information Integrated DuringPerceptual Decision-Making? 150
Q2: What Computations Do Cortical NeuronsPerform During Perceptual Decisions? 153
Q3: How Can We Study PerceptualDecision-Making in Humans? 155
Q4: How Do Observers Decide Whento Decide? 157
Q5: How Are Perceptual Decisions Biasedby Prior Beliefs? 158
Conclusions and Future Directions 160
Acknowledgments 160
References 160
13. Neural Circuit Mechanismsof Value-Based Decision-Making andReinforcement LearningA. SOLTANI, W. CHAISANGMONGKON AND X.-J. WANG
Introduction 163
Representations of Reward Value 164
Learning Reward Values 165
CONTENTSvi
Stochastic Dopamine-Dependent Plasticity forLearning Reward Values 167
Foraging With Plastic Synapses 167
Random Choice and Competitive Games 170
Probabilistic Inference With Stochastic Synapses 171
Concluding Remarks 173
Acknowledgments 174
References 174
IIISOCIAL DECISION NEUROSCIENCE
14. Social Decision-Making in NonhumanPrimatesM. JAZAYARI, S. BALLESTA AND J.-R. DUHAMEL
Introduction 179
Behavioral Studies on Social Decision-Making 180
Neuronal Correlates of Decision-Making in aSocial Context 182
Conclusions and Perspectives 185
References 185
15. Organization of the Social Brain inMacaques and HumansM.P. NOONAN, R.B. MARS, F.X. NEUBERT, B. AHMED,
J. SMITH, K. KRUG AND J. SALLET
Introduction 189
Medial Prefrontal Cortex 190
Superior Temporal Sulcus and Temporal ParietalJunction 193
A Social Brain Network 194
Summary and Perspectives 195
References 196
16. The Neural Bases of Social Influence onValuation and BehaviorK. IZUMA
Introduction 199
The Effect of Mere Presence of Others on ProsocialBehavior 199
The Effect of Others’ Opinions on Valuation 203
Concluding Remarks 206
References 206
17. Social Dominance Representations inthe Human BrainR. LIGNEUL AND J.-C. DREHER
Introduction 211
Learning Social Dominance Hierarchies 212
Interindividual Differences and SocialDominance 216
Neurochemical Approaches to SocialDominance and Subordination 218
Conclusion 222
Acknowledgments 222
References 222
18. Reinforcement Learning and StrategicReasoning During Social Decision-MakingH. SEO AND D. LEE
Introduction 225
Model-Free Versus Model-Based ReinforcementLearning 226
Neural Correlates of Model-Free ReinforcementLearning During Social Decision-Making 227
Neural Correlates of Hybrid ReinforcementLearning During Social Decision-Making 227
Arbitration and Switching Between LearningAlgorithms 228
Conclusions 230
References 230
19. Neural Control of Social Decisions:Causal Evidence From Brain StimulationStudiesG. UGAZIO AND C.C. RUFF
Introduction 233
Noninvasive Brain Stimulation Methods Used forStudying Social Decisions 234
Brain Stimulation Studies of Social Emotions 235
Brain Stimulation Studies of Social Cognition 236
Brain Stimulation Studies of Social BehavioralControl 238
Brain Stimulation Evidence for Social-SpecificNeural Activity 240
Conclusions 241
References 242
20. The Neuroscience of Compassion andEmpathy and Their Link to ProsocialMotivation and BehaviorG. CHIERCHIA AND T. SINGER
Introduction 247
The “Toolkit” of Social Cognition 248
The Neural Substrates of Empathy 249
The Psychological and Neural Bases of Compassion 250
Conclusion 255
References 255
CONTENTS vii
IVHUMAN CLINICAL STUDIES
INVOLVING DYSFUNCTIONS OFREWARD AND DECISION-MAKING
PROCESSES
21. Can Models of Reinforcement LearningHelp Us to Understand Symptoms ofSchizophrenia?G.K. MURRAY, C. TUDOR-SFETEA AND P.C. FLETCHER
Introduction 261
Reward Processing in Schizophrenia: A HistoricalPerspective 262
Dopamine, Schizophrenia, and ReinforcementLearning 263
The Possible Importance of Glutamate 263
Studies of Reward Processing/ReinforcementLearning in Psychosis: Behavioral Studies 264
Studies of Reward Processing/ReinforcementLearning in Psychosis: Neuroimaging Studies 265
Can an Understanding of Reward and DopamineHelp Us to Understand Symptoms ofSchizophrenia? 269
Summary 272
Acknowledgments 273
References 273
22. The Neuropsychology of Decision-Making: A View From the Frontal LobesA.R. VAIDYA AND L.K. FELLOWS
Introduction 277
Lesion Evidence in Humans 278
Decision-Making and the Frontal Lobes 279
Component Processes of Decision-Making 279
Summary 286
Acknowledgments 286
References 286
23. Opponent Brain Systems for Reward andPunishment Learning: Causal Evidence FromDrug and Lesion Studies in HumansS. PALMINTERI AND M. PESSIGLIONE
The Neural Candidates for Reward andPunishment Learning Systems 294
Evidence From Drug and Lesion Studies 296
Conclusions, Limitations, and Perspectives 300
References 300
24. Decision-Making and Impulse ControlDisorders in Parkinson’s DiseaseV. VOON
Introduction 305
The Role of Dopaminergic Medications andIndividual Vulnerability in Parkinson’s Disease 305
Reinforcing Effects and Associative Learning 307
Learning From Feedback 308
Risk and Uncertainty 309
Impulsivity 310
Summary 311
References 312
25. The Subthalamic Nucleus in ImpulsivityK. WITT
Introduction 315
Anatomy, Physiology, and Function of CorticobasalGanglia Circuits 315
The Subthalamic Nucleus and Decision-Making:Evidence From Animal Studies 317
The Subthalamic Nucleus and Impulsivity:Evidence From Behavioral Observations 317
The Subthalamic Nucleus and Decision-Making:Evidence From Neuropsychological Studies 319
A Model of the Impact of the Subthalamic Nucleuson Decision-Making 321
References 322
26. Decision-Making in Anxiety and itsDisordersD.W. GRUPE
Introduction 327
Summary and Conclusions 335
References 336
27. Decision-Making in Gambling Disorder:Understanding Behavioral AddictionsL. CLARK
Introduction: Gambling and Disordered Gambling 339
Loss Aversion 340
Probability Weighting 341
Perceptions of Randomness 342
Illusory Control 344
Conclusion 345
Disclosures 345
References 345
CONTENTSviii
VGENETIC AND HORMONAL
INFLUENCES ON MOTIVATION ANDSOCIAL BEHAVIOR
28. Decision-Making in Fish: Genetics andSocial BehaviorR.D. FERNALD
Social System of the African Cichlid Fish,Astatotilapia burtoni 351
Domains of Astatotilapia burtoni Social Decisions 353
Summary 358
References 358
29. Imaging Genetics in Humans: MajorDepressive Disorder and Decision-MakingU. RABL, N. ORTNER AND L. PEZAWAS
Introduction 361
Major Depressive Disorder as a Disorderof Decision-Making 361
Imaging Genetics of Major Depressive Disorder 364
Conclusion 365
References 367
30. Time-Dependent Shifts in NeuralSystems Supporting Decision-Making UnderStressE.J. HERMANS, M.J.A.G. HENCKENS, M. JOELS AND
G. FERNANDEZ
Introduction 371
Large-Scale Neurocognitive Systems and Shiftsin Resource Allocation 372
Salience Network and Acute Stress 373
Executive Control Network and Acute Stress 376
Salience Network and Executive Control Networkand Recovery From Stress 377
Summary and Conclusion 378
References 380
31. Oxytocin’s Influence on SocialDecision-MakingA. LEFEVRE AND A. SIRIGU
Introduction 387
Oxytocin and Perception of Social Stimuli 388
Oxytocin and Social Decisions 389
Oxytocin and Social Reward 390
Oxytocin, Learning, and Memory 392
Perspectives 392
Acknowledgments 393
References 393
32. Appetite as Motivated Choice:Hormonal and Environmental InfluencesA. DAGHER, S. NESELILER AND J.-E. HAN
Introduction 397
Appetitive Brain Systems Promote Food Intake 398
Self-Control and Lateral Prefrontal Cortex: Role inAppetite Regulation and Obesity 401
Interaction Between Energy Balance Signals andDecision-Making 402
Conclusion 404
References 404
33. PerspectivesJ.-C. DREHER, L. TREMBLAY AND W. SCHULTZ
Identifying Fundamental ComputationalPrinciples: Produce Conceptual Foundations forUnderstanding the Biological Basis of MentalProcesses Through Development of NewTheoretical and Data Analysis Tools 411
Understanding the Functional Organization of thePrefrontal Cortex and the Nature of theComputations Performed in Various Subregions:Value-Coding Computations 412
Demonstrating Causality: Linking Brain Activity toBehavior by Developing and Applying PreciseInterventional Tools That Change NeuralCircuit Dynamics 413
Maps at Multiple Scales: Generate CircuitDiagrams That Vary in Resolution FromSynapses to the Whole Brain 414
The Brain in Action: Produce a Dynamic Picture ofthe Functioning Brain by Developing andApplying Improved Methods for Large-ScaleMonitoring of Neural Activity 415
The Analysis of Circuits of Interacting Neurons 415
Develop Innovative Technologies andSimultaneous Measures to Understand How theBrain Makes Decisions 416
Advancing Human Decision Neuroscience:Understanding Neurological/PsychiatricDisorders and Treating Brain Diseases 417
Conclusions 418
Reference 418
Index 419
CONTENTS ix
This page intentionally left blank
List of Contributors
B. Ahmed University of Oxford, Oxford, United Kingdom
B.W. Balleine University of Sydney, Camperdown, NSW,Australia
S. Ballesta Centre National de la Recherche Scientifique,Bron, France; Universite Lyon 1, Villeurbanne, France; TheUniversity of Arizona, Tucson, AZ, United States
S. Bernardi Columbia University, New York, New York,United States
A. Blangero University of Oxford, Oxford, United Kingdom
L.A. Bradfield University of Sydney, Camperdown, NSW,Australia
W. Chaisangmongkon King Mongkut’s University ofTechnology Thonburi, Bangkok, Thailand; New YorkUniversity, New York, NY, United States
G. Chierchia Max Planck Institute for Human Cognitive andBrain Sciences, Leipzig, Germany
L. Clark University of British Columbia, Vancouver, BC,Canada
A. Dagher McGill University, Montreal, QC, Canada
J.W. Dalley University of Cambridge, Cambridge, UnitedKingdom
J.A. Diaz University of Glasgow, Glasgow, United Kingdom
J.-C. Dreher Institute of Cognitive Science (CNRS), Lyon,France
J.-R. Duhamel Centre National de la Recherche Scientifique,Bron, France; Universite Lyon 1, Villeurbanne, France
N. Eshel Harvard University, Cambridge, MA, United States
L.K. Fellows McGill University, Montreal, QC, Canada
R.D. Fernald Stanford University, Stanford, CA, UnitedStates
G. Fernandez Radboud University Medical Centre,Nijmegen, The Netherlands
P.C. Fletcher University of Cambridge, Cambridge, UnitedKingdom
J.M. Fuster University of California Los Angeles, LosAngeles, CA, United States
S. Gherman University of Glasgow, Glasgow, UnitedKingdom
P.W. Glimcher New York University, New York, NY, UnitedStates
D.W. Grupe University of WisconsineMadison, Madison,WI, United States
S.N. Haber University of Rochester School of Medicine,Rochester, NY, United States
J.-E. Han McGill University, Montreal, QC, Canada
M.J.A.G. Henckens Radboud University Medical Centre,Nijmegen, The Netherlands
E.J. Hermans Radboud University Medical Centre,Nijmegen, The Netherlands
K. Izuma University of York, York, United Kingdom
M. Jazayari Centre National de la Recherche Scientifique,Bron, France; Universite Lyon 1, Villeurbanne, France
M. Joels University Medical Center Utrecht, Utrecht,The Netherlands
T. Kahnt Northwestern University Feinberg School ofMedicine, Chicago, IL, United States
K. Krug University of Oxford, Oxford, United Kingdom
D. Lee Yale University, New Haven, CT, United States
A. Lefevre Institut des Sciences Cognitives Marc Jeannerod,UMR 5229, CNRS, Bron, France; Universite Claude BernardLyon 1, Lyon, France
R. Ligneul Institute of Cognitive Science (CNRS), Lyon,France
K. Louie New York University, New York, NY, United States
R.B. Mars University of Oxford, Oxford, United Kingdom
G.K. Murray University of Cambridge, Cambridge, UnitedKingdom
S. Neseliler McGill University, Montreal, QC, Canada
F.X. Neubert University of Oxford, Oxford, United Kingdom
M.P. Noonan University of Oxford, Oxford, UnitedKingdom
N. Ortner Medical University of Vienna, Vienna, Austria
S. Palminteri University College London, London, UnitedKingdom; Ecole Normale Superieure, Paris, France
M. Pessiglione Institut du Cerveau et de la Moelle (ICM),Inserm U1127, Paris, France; Universite Pierre et MarieCurie (UPMC-Paris 6), Paris, France
L. Pezawas Medical University of Vienna, Vienna, Austria
M.G. Philiastides University of Glasgow, Glasgow, UnitedKingdom
U. Rabl Medical University of Vienna, Vienna, Austria
T.W. Robbins University of Cambridge, Cambridge, UnitedKingdom
C.C. Ruff University of Zurich, Zurich, Switzerland
Y. Saga Institute of Cognitive Sciences (CNRS), Lyon, France
J. Sallet University of Oxford, Oxford, United Kingdom
xi
D. Salzman Columbia University, New York, New York,United States; New York State Psychiatric Institute, NewYork, New York, United States
W. Schultz University of Cambridge, Cambridge, UnitedKingdom
H. Seo Yale University, New Haven, CT, United States
T. Singer Max Planck Institute for Human Cognitive andBrain Sciences, Leipzig, Germany
A. Sirigu Institut des Sciences Cognitives Marc Jeannerod,UMR 5229, CNRS, Bron, France; Universite Claude BernardLyon 1, Lyon, France
J. Smith University of Oxford, Oxford, United Kingdom
A. Soltani Dartmouth College, Hanover, NH, United States
C. Summerfield University of Oxford, Oxford, UnitedKingdom
J. Tian Harvard University, Cambridge, MA, United States
P.N. Tobler University of Zurich, Zurich, Switzerland
L. Tremblay Institute of Cognitive Science (CNRS), Lyon,France
C. Tudor-Sfetea University of Cambridge, Cambridge,United Kingdom
N. Uchida Harvard University, Cambridge, MA, UnitedStates
G. Ugazio University of Zurich, Zurich, Switzerland
A.R. Vaidya McGill University, Montreal, QC, Canada
V. Voon University of Cambridge, Cambridge, UnitedKingdom; Cambridgeshire and Peterborough NHSFoundation Trust, Cambridge, United Kingdom
X.-J. Wang New York University, New York, NY, UnitedStates; NYU Shanghai, Shanghai, China
K. Witt Christian Albrecht University, Kiel, Germany
LIST OF CONTRIBUTORSxii
Preface
Decision Neuroscience: an Integrative Perspective ad-dresses fundamental questions about how the brainmakes perceptual, value-based, and more complex de-cisions innonsocial andsocial contexts. This bookpresentsrecent and compelling neuroimaging, electrophysiolog-ical, lesional, and neurocomputational studies, in combi-nationwith hormonal and genetic studies, that have led toa clearer understanding of the neural mechanisms behindhow the brain makes decisions. The neural mechanismsunderlying decision-making processes are of criticalinterest to scientists because of the fundamental rolethat reward plays in a number of cognitive processes(such as motivation, action selection, and learning) andbecause they have theoretical and clinical implications forunderstanding dysfunctions of major neurological andpsychiatric disorders.
The idea for this book grew up from our edition of theHandbook of Reward and Decision Making (Academic Press,2009). We originally thought to revise and reedit thisbook, addressing one fundamental question about thenature of behavior: how does the brain process rewardand makes decisions when facing multiple options?However, given the developments in this active area ofresearch, we decided to feature an entirely different bookwith new contents, covering results on the neural sub-strates of rewards and punishments; perceptual, value-based, and social decision-making; clinical aspects suchas behavioral addictions; and the roles of genes andhormones in these various aspects. For example, anexciting topic from the field of social neuroscience is toknow whether the neural structures engaged withvarious forms of social interactions are cause or conse-quence of these interactions (Fernald, Chapter 28).
A mechanistic understanding of the neural encodingunderlying decision-making processes is of great interestto a broad readership because of their theoretical andclinical implications. Findings in this research field arealso important to basic neuroscientists interested in howthe brain reaches decisions, cognitive psychologistsworking on decision-making, as well as computationalneuroscientists studying probabilistic models of brainfunctions. Decision-making covers a wide range of topicsand levels of analysis, from molecular mechanisms toneural systems dynamics, neurocomputational models,and social system levels. The contributions to this book
are forward-looking assessments of the current andfuture issues faced by researchers. We were fortunate toassemble an outstanding collection of experts whoaddressed various aspects of decision-making processes.The book is divided into five parts that address distinctbut interrelated topics.
STRUCTURE OF THE BOOK
Adecision neuroscience perspective requiresmultiplelevels of analyses spanning neuroimaging, electrophysi-ological, behavioral, and pharmacological techniques, incombination with molecular and genetic tools. Theseapproaches have begun to build a mechanistic under-standing of individual and social decision-making. Thisbook highlights some of these advancements that haveled to the current understanding of the neuronal mech-anisms underlying motivational and decision-makingprocesses.
Part I is devoted to animal studies (anatomical,neurophysiological, pharmacological, and optogenetics)on rewards/punishments and decision-making. In theirnatural environment, animals face a multitude of stimuli,very few of which are likely to be useful as predictors ofreward or punishment. It is thus crucial that the brainlearns topredict rewards,providinga critical evolutionaryadvantage for survival. This first part of the book offers acomprehensive view of the specific contributions ofvarious brain structures as the dopaminergic midbrainneurons, the amygdala, the ventral striatum, and theprefrontal cortex, including the lateral prefrontal cortexand the orbitofrontal cortex, to the component processesunderlying reinforcement-guided decision-making, suchas the representation of instructions, expectations, andoutcomes; the updating of action values; and the evalua-tion process guiding choices between prospective re-wards. Special emphasis is made on the neuroanatomy ofthe reward system and the fundamental roles of dopa-minergic neurons and the basal ganglia in learningstimulusereward associations.
Chapter 1 (Haber SN) describes the anatomy andconnectivity of the reward circuit in nonhuman primates.It describes how corticalebasal ganglia loops are
xiii
topographically organized and the key areas of conver-gence between functional regions.
Chapter 2 describes three novel electrophysiologicalproperties of the classical dopamine reward-predictionerror (RPE) signal (Schultz W). Studies have identifiedthree novel properties of the dopamine RPE signal. Inparticular, concerning its roles in making choices, thedopamine RPE signal may not only reflect subjectivereward value and formal economic utility but could alsofit into formal competitive decision models. The RPEsignal may code the chosen value suitable for updatingor immediately influencing object and action values.Thus, the dopamine utility prediction error signalbridges the gap between animal learning theory andeconomic decision theory.
Chapter 3 focuses on the electrophysiological prop-erties of another important component of the rewardsystem in primates, namely the amygdala (Bernardi Sand Salzman D). The amygdala contains distinct appe-titive and aversive networks of neurons. Processing inthese two amygdalar networks can both regulate and beregulated by diverse cognitive operations.
Chapter 4 extends the concept of appetitive andaversive motivational processes to the striatum (Saga Yand Tremblay L). This chapter describes how the ventralstriatum and the ventral pallidum, two parts of thelimbic circuit in the basal ganglia, are involved not onlyin appetitive rewarding behavior, as classically believed,but also in negative motivational behavior. These resultscan be linked with the control of approach/avoidancebehavior in a normal context and with the expression ofanxiety-related disorders. The disturbance of thispathwaymay induce not only psychiatric symptoms, butalso abnormal value-based decision-making.
Chapter 5 (Tian J, Uchida N, and Eshel N) highlightsnew advances in the physiology, function, and circuitmechanism of decision-making, focusing especially onthe involvement of dopamine and striatal neurons. Usingoptogenetics in rodents, molecular techniques, andgenetic techniques, this chapter shows how these toolshave been used to dissect the circuits underlyingdecision-making. It describes exciting new avenues tounderstand a circuit, by recording from neurons withknowledge of their cell type and patterns of connectivity.Furthermore, the ability to manipulate the activity ofspecific neural types provides an important means to testhypotheses of circuit function.
Chapter 6 (Bradfield L and Balleine B) describes theneural bases of the learning and motivational processescontrolling goal-directed action. By definition, the per-formance of such action respects both the current valueof its outcome and the extant contingency between thataction and its outcome. This chapter identifies the neuralcircuits mediating distinct processes, including theacquisition of action-outcome contingencies, the
encoding and retrieval or incentive value, the matchingof that value to specific outcome representations, andfinally the integration of this information for action se-lection. It also shows how each of these individual pro-cesses are integrated within the striatum for successfulgoal-directed action selection.
Chapter 7 (Robbins TW and Dalley JW) describes an-imal models (mostly in rodents) of impulsivity and riskychoices. It reviews the neural and neurochemical basis ofvarious forms of impulsive behavior by distinguishingthree main forms of impulsivity: waiting impulsivity,risky choice impulsivity, and stopping impulsivity. Itshows that dopamine- and serotonin-dependent func-tions of the nucleus accumbens are implicated in waitingimpulsivity and risky choice impulsivity, as well ascortical structures projecting to the nucleus accumbens.For stopping impulsivity, dopamine-dependent functionsof the dorsal striatum are implicated, as well as circuitryincluding the orbitofrontal cortex and dorsal prelimbiccortex. Differences and commonalities between the formsof impulsive responding are highlighted. Importantly,various applications to human neuropsychiatric disor-ders such as drug addiction and attention deficit hyper-activity disorder are also discussed.
Chapter 8 (Fuster JM) proposes that the neuralmechanisms of decision-making are understandableonly in the structural and dynamic context of theperceptioneaction cycle, defined as the biocyberneticprocessing of information that adapts the organism to itsenvironment. It presents a general view of the role of theprefrontal cortex in decision-making, in the generalframework of the perceptioneaction cycle, includingprediction, preparation toward decision, execution, andfeedback from decision.
Part II covers the topic of the neural representation ofmotivation, perceptual decision-making, and value-based decision-making in humans, mostly combiningneurocomputational models and brain imaging studies.
Chapter 9 (Tobler P and Kahnt T) reviews severaldefinitions of value and salience, and describes humanneuroimaging studies that dissociate these variables.Value increases with the magnitude and probability ofreward but decreases with the magnitude and proba-bility of punishment, whereas salience increases with themagnitude and probability of both reward and punish-ment. At the neural level, value signals arise in striatum,orbitofrontal and ventromedial prefrontal cortex, andsuperior parietal areas, whereas magnitude-basedsalience signals arise in the anterior cingulate cortexand the inferior parietal cortex. By contrast, probability-based salience signals have been found in the ventro-medial prefrontal cortex.
Chapter 10 (Louie K and Glimcher PW) reviews anapproach centered on basic computations underlying
PREFACExiv
neural value coding. It proposes that neural informationprocessing in valuation and choice relies on computa-tional principles such as contextual modulation anddivisive normalization. Divisive normalization is anonlinear gain control algorithm widely observed inmultiple sensory modalities and brain regions. Identifi-cation of these computations sheds light on how theunderlying neural circuits are organized, and neuralactivity dynamics provides a link between biologicalmechanism and computations.
Chapter 11 (Philiastides M, Diaz J, and Gherman S)introduces the general principles guiding perceptualdecision-making. Perceptual decisions occur whenperceptual inputs are integrated and converted to form acategorical choice. It reviews the influence of a number offactors that interact and contribute to the decision pro-cess, such as prestimulus state, reward and punishment,speedeaccuracy trade-off, learning and training, confi-dence, and neuromodulation. It shows how these deci-sion modulators can exert their influence at variousstages of processing, in line with predictions derivedfrom sequential-sampling models of decision-making.
Chapter 12 (Summerfield C) reviews the neural andcomputational mechanisms of perceptual decisions. Itaddresses current controversial questions, such as howwe decide when to draw our decisions to a conclusion,and how perceptual decisions are biased by priorinformation.
Chapter 13 (Soltani A, Chaisangmongkon W, andWang XJ) presents possible biophysical and circuitmechanisms of valuation and reward-dependent plas-ticity underlying adaptive choice behavior. It reviewsmathematical models of reward-dependent adaptivechoice behavior, and proposes a biologically plausible,reward-modulated Hebbian synaptic plasticity rule. Itshows that a decision-making neural circuit endowedwith this learning rule is capable of accounting forbehavioral and neurophysiological observations in avariety of decision-making tasks.
Part III of the book focuses on the rapidly developingfield of social neuroscience, integrating neurosciencedata from both nonhuman primates and humans. Pri-mates are fundamentally social animals, and they mayshare common neural mechanisms in diverse forms ofsocial behavior. Examples of such behavior includetracking intentions and beliefs from others, beingobserved by others during prosocial decisions, orlearning the social hierarchy in a group of individuals. Itis also likely that at the macroscopic level, importantdifferences exist concerning social brain structures andconnectivity, and there is a need to directly comparebetween species to answer this fundamental question.Indeed, studies in both humans and monkeys report notonly an increase in gray matter density of specific brain
structures relative to the size of our social network, butalso species differences in prefrontaletemporal brainconnectivity. Furthermore, this part of the book presentsneurocomputational approaches starting to provide amechanistic understanding of social decisions. Forexample, reinforcement learning models and strategicreasoning models can be used when learning social hi-erarchies or during social interactions.
A social neuroscience understanding requires multi-ple approaches, such as electrophysiology and neuro-imaging in both monkeys (Chapters 14, 15, 19) andhumans (Chapters 16, 18, 20), as well as causal (Chapter21), neurocomputational (Chapters 17e19), endocrino-logical, genetics, and clinical approaches (Part V).
Chapter 14 (Duhamel JR and colleagues) presentsmonkey electrophysiological data revealing that theorbitofrontal cortex is tuned to social information. Forexample, in one experiment, macaque monkeys workedto collect rewards for themselves and two monkeypartners. Single neurons encoded the meaning of visualcues that predicted the magnitude of future rewards, themotivational value of rewards obtained in a socialcontext, and the tracking of social preferences and part-ner’s identity and social rank. The orbitofrontal cortexthus contains key neuronal mechanisms for the evalua-tion of social information. Moreover, macaque monkeystake into account the welfare of their peers when makingbehavioral choices bringing about pleasant or unpleasantoutcomes to a monkey partner. Thus, this chapter revealsthat prosocial decision-making is sustained by anintrinsic motivation for social affiliation and controlledthrough positive and negative vicarious reinforcements.
Chapter 15 (Sallet J and colleagues) reviews the sim-ilarities between monkeys and humans in the organiza-tion of the social brain. Using MRI-based connectivitymethods, they compare human and macaque socialareas, such as the organization of the medial prefrontalcortex. They revealed that the connectivity fingerprint ofmacaque area 10 best matched that of the human frontalpole, suggesting that even high-level areas share featuresbetween species. They also showed that animals housedin large social groups had more gray matter volume inbilateral mid-superior temporal sulcus and rostral pre-frontal cortex. Beyond species similarities, there are alsodistinct differences between human and macaqueprefrontaletemporal brain connectivity. For example,functional connections between the temporal cortex andthe lateral prefrontal cortex are stronger in humanscompared to connections with the medial prefrontalcortex in humans, but the opposite pattern is observed inmacaques.
Chapter 16 (Izuma K) focuses on two forms of socialinfluence, the audience effect, which is an increasedprosocial tendency in front of other people, and socialconformity, which consists in adjusting one’s attitude or
PREFACE xv
behavior to those of a group. This chapter discusses fMRIfindings in healthy humans in these two types of socialinfluence and also shows how reputation processing isimpaired in individuals with autism. It also links socialconformity and reward-based learning (reinforcementlearning).
Chapter 17 (Ligneul R and Dreher JC) examines howthe brain learns social dominance hierarchies. Socialdominance refers to relationships wherein the goals of oneindividual prevail over the goals of another individual in asystematic manner. Dominance hierarchies have emergedas a major evolutionary force to drive dyadic asymmetriesin a social group. This chapter proposes that the emer-gence of dominance relationships are learned incremen-tally, by accumulating positive and negative competitivefeedbacks associated with specific individuals and othermembers of the social group. It considers such emergenceof social dominance as a reinforcement learning probleminspired by neurocomputational approaches traditionallyapplied to nonsocial cognition. This chapter also reportshow dominance hierarchies induce changes in specificbrain systems, and it reviews the literature on interindi-vidual differences in the appraisal of social hierarchies, aswell as the underlying modulations of cortisol, testos-terone, and serotonin/dopamine systems, which mediatethese phenomena.
Chapter 18 (Seo H and Lee D) describes reinforcementlearning models and strategic reasoning during socialdecision-making. It shows that dynamic changes inchoices and decision-making strategies can be accountedfor by reinforcement learning in a variety of contexts. Thisframework has also been successfully adopted in a largenumber of neurobiological studies to characterize thefunctions of multiple cortical areas and basal ganglia. Forcomplex decision-making, including social interactions,this chapter shows that multiple learning algorithmsmayoperate in parallel.
Chapter 19 (Ugazio G and Ruff C) reports brainstimulation studies on social decision-making, which testthe causal relationship between neural activity anddifferent types of processes underlying these decisions,including social emotions, social cognition, and socialbehavioral control.
Chapter 20 (Chierchia G and Singer T) shows that twoimportant social emotions, empathy and compassion,engage distinct neurobiological mechanisms, as well asdifferent affective and motivational states. Empathy forpain engages a network including the anterior insula andanterior midcingulate cortex, areas associated withnegative affect; compassionate states engage the medialorbitofrontal cortex and ventral striatum and are associ-atedwith feelings ofwarmth, concern, and positive affect.
Part IV of the book focuses on clinical aspectsinvolving disorders of decision-making and of the
reward system that link together basic research areas,including systems, cognitive, and clinical neuroscience.Dysfunction of the reward system and decision-makingis present in a number of neurological and psychiatricdisorders, such as Parkinson’s disease, schizophrenia,drug addiction, and focal brain lesions. The study ofpathological gambling, for example, and other motivatedstates associated with, and leading to, compulsivebehavior provides an opportunity to learn about thedysfunctions of reward system activity, independent ofdirect pharmacological activation of brain reward cir-cuits. On the other hand, because drugs of abuse directlyactivate brain systems, they provide a unique challengein understanding how pharmacological activation in-fluences reward mechanisms leading to persistentcompulsive behavior.
Chapter 21 (Murray GK, Tudor-Sfetea C, and FletcherPC) shows that principles of reinforcement learning areuseful to understand the neural mechanisms underlyingimpaired learning, reward, and motivational processesin schizophrenia. Two symptoms characteristic of thisdisease is considered in this framework, namely de-lusions and anhedonia.
Chapter 22 (Vaidya AR and Fellows LK) takes aneuropsychological approach to review focal frontal lobedamage effects on value-based decisions. It reveals thenecessary contributions of specific subregions (ventro-medial, lateral, and dorsomedial prefrontal cortex) todecision-making, and provides evidence as to the disso-ciability of component processes. It argues that theventromedial frontal lobe is required for optimal learningfrom reward under dynamic conditions and contributesto specific aspects of value-based decision-making. It alsoshows a necessary contribution of the dorsomedialfrontal lobe in representing action-value expectations.
Chapter 23 (Palminteri S and Pessiglione M) reviewsreinforcement learning models applied to reward andpunishment learning. These studies include fMRI andneural perturbation following drug administration and/or pathological conditions. They propose that distinctbrain systems are engaged, one in reward learning(midbrain dopaminergic nuclei and ventral prefrontos-triatal circuits) and another in punishment learning,revolving around the anterior insula.
Chapter 24 (Voon V) discusses decision-making im-pairments and impulse control disorders in Parkinson’sdisease. The author reports enhancement of the gainassociated with levodopa, reinforcing properties ofdopaminergic medications, and enhancement of delaydiscounting in these patients. Lower striatal dopaminetransporter levels preceding medication exposure, anddecreased midbrain D2 autoreceptor sensitivity, mayunderlie enhanced ventral striatal dopamine release andactivity in response to salient reward cues, anticipatedand unexpected rewards, and gambling tasks.
PREFACExvi
Impairments in decisional impulsivity (delay discount-ing, reflection impulsivity, and risk taking) implicate theventral striatum, orbitofrontal cortex, anterior insula,and dorsal cingulate. These findings provide insight intothe role of dopamine in decision-making processes inaddiction and suggest potential therapeutic targets.
Chapter 25 (Witt K) reports that motor control is theresult of a balance between activation and inhibition ofmovement patterns. It points to a central role of thesubthalamic nucleus within the indirect basal gangliapathway, acting as a brake on the motor system. Thissubthalamic nucleus function occurs when an automaticresponse must be suppressed to have more time tochoose between alternative responses.
Chapter 26 (Grupe DW) discusses value-baseddecision-making as one of a key behavioral symptomspresent in anxiety disorders. This chapter highlights al-terations to specific processes: decision representation,valuation, action selection, outcome evaluation, andlearning. Distinct anxious phenotypes may be charac-terized by differential alterations to these processes andtheir associated neurobiological mechanisms.
Chapter 27 (Clark L) presents a conceptualization ofdisordered gambling as a behavioral addiction driven byan exaggeration of multiple psychological distortionsthat are characteristic of human decision-making, andunderpinned by neural circuitry subserving appetitivebehavior, reinforcement learning, and choice selection.The chapter discusses the neurobiological basis of path-ological gambling behavior in loss aversion, probabilityweighting, perceptions of randomness, and the illusionof control.
Part V focuses on the roles of hormones and genesinvolved in motivation and social decision-making pro-cesses. The combination of molecular genetic, endocri-nology, and neuroimaging has provided a considerableamount of data that help in the understanding of thebiological mechanisms influencing decision processes.These studies have demonstrated that genetic and hor-monal variations have an impact on the physiologicalresponse of the decision-making system. These varia-tions may account for some of the inter- and intra-individual behavioral differences observed in socialcognition.
Chapter 28 (Fernald RD) presents an originalapproach for cognitive neuroscientists by focusing on thedifficult question of how an animal’s behavior orperception of its social and physical surroundings shapesits brain. Using a fish model system that depends oncomplex social interactions, this chapter reports how thesocial context influences the brain and, in turn, alters thebehavior and neural circuitry of animals as they interact.Gathering of social information vicariously producesrapid changes in gene expression in key brain nuclei and
these genomic responses prepare the individual tomodify its behavior to move into a different social niche.Both social success and failure produce changes inneuronal cell size and connectivity in key brain nuclei.This approach bridges the gap between social informa-tion gathering from the environment and the levels ofcellular and molecular responses.
Chapter 29 (Rabl U, Ortner N, and Pezawas L) ex-amines the use of imaging genetics to explore the re-lationships between major depressive disorder anddecision-making.
Chapters 30e32 report neuroendocrinological find-ings in social decision-making, likening variations in thelevels of different types of hormones (cortisol, oxytocin,ghrelin/leptin) to brain systems engaged in social de-cisions and food choices. Chapter 30 (Hermans EJ andcolleagues) integrates knowledge of the effects of stressat the neuroendocrine, cellular, brain systems, andbehavioral levels to quantify how stress-related neuro-modulators trigger time-dependent shifts in the balancebetween two brain systems: a “salience” network, whichsupports rapid but rigid decisions, and an “executivecontrol” network, which supports flexible, elaborate de-cisions. This simple model elucidates paradoxical find-ings reported in human studies on stress and cognition.
Chapter 31 (Lefevre A and Sirigu A) reviews evidencefor a role for oxytocin in individual and social decision-making. It discusses animal and human studies to linkthe behavioral effects of oxytocin to its underlyingneurophysiological mechanisms.
Chapter 32 (Dagher A, Neseliler S, and Han JE) ex-amines the neurobehavioral factors that determine foodchoices and food intake. It reviews findings on the in-teractions between brain systems that mediate feedingbehavior and the gut and adipose peptides that signalthe current state of energy balance.
Chapter 33 (Dreher, Tremblay, and Schultz) concludesthis decision neuroscience book by integrating perspec-tives from all contributors.
We anticipate that while some readers may read thevolume from the first to the last chapter, other readersmay read only one or more chapters at a time, and notnecessarily in the order presented in the book. This iswhy we encouraged an organization of this volumewhereby each chapter can stand alone, while makingreferences to others andminimizing redundancies acrossthe volume. Given the consistent acceleration of ad-vances in the various approaches described in this bookon decision neuroscience, you are about to be dazzled bya first look at the new stages of an exciting era in brainresearch. Enjoy!
Jean-Claude DreherLeon Tremblay
PREFACE xvii
This page intentionally left blank
C H A P T E R
33
PerspectivesJ.-C. Dreher1, L. Tremblay1, W. Schultz2
1Institute of Cognitive Science (CNRS), Lyon, France; 2University of Cambridge, Cambridge, United Kingdom
AbstractDiscovering how the brain makes decisions is one of themost exciting challenges of neurosciences that has emergedin recent years. The evolution of the field of decisionneuroscience has benefited from the advance of novel tech-nological capabilities in neurosciences, and the pace at whichthese capabilities have been developed has accelerateddramatically in the past decade.
Discovering how the brain makes decisions is one ofthe most exciting challenges of neurosciences that hasemerged in recent years. The evolution of the field of de-cision neuroscience has benefited from the advance ofnovel technological capabilities in neurosciences, andthe pace at which these capabilities have been devel-oped has accelerated dramatically since 2005.
It is certainly difficult to predict what will be the mostexciting developments in decision neuroscience inthe future and somewhat arbitrary to organize potentialperspectives along a coherent line. We have asked thecontributors to this book to give us their respectiveperspectives for developments in their research do-mains. We have taken the liberty to build these perspec-tives based on these views and along lines outlined in aneuroscience report from the Brain Research throughAdvancing Innovative Neurotechnologies (BRAIN)Initiative [1]. This BRAIN Initiative should help usdevelop and apply new tools and technologies to under-stand the brain at multiple levels. In parallel, theHuman Brain Project from the EU, the Brain/MINDSProject from Japan (BrainMapping by Integrated Neuro-technologies for Disease Studies), CanadaBrain, and anational brain project under way in China should alsofoster technological innovation for great discoveriesthat should lead to a revolution in our understandingof how the brain makes decisions, from a multilevelperspective.
IDENTIFYING FUNDAMENTALCOMPUTATIONAL PRINCIPLES:
PRODUCE CONCEPTUAL FOUNDATIONSFOR UNDERSTANDING THE
BIOLOGICAL BASIS OF MENTALPROCESSES THROUGH DEVELOPMENTOF NEW THEORETICAL AND DATA
ANALYSIS TOOLS
Theory and mathematical modeling are advancingour understanding of complex, nonlinear brainfunctions where human intuition fails. New kinds ofdata are accruing at increasing rates, mandatingnew methods of data analysis and interpretation. Toenable progress in theory and data analysis, decisionneuroscience will need to foster collaborationsbetween experimentalists and researchers from physics,mathematics, engineering, and computer science.
One general approach widely used in the field offMRI research and more recently applied to monkeyelectrophysiology is to use a so-called model-basedapproach, allowing us to identify the computationsperformed by a given brain region. This approach se-lects the best model fitting behavior among a set ofmodels and allows us to regress brain activity withoutput parameters from these models. One classicalmodel-based fMRI approach concerning learning ofbasic stimulusereinforcer associations used predictionerrors as regressors. Similar model-based fMRI ap-proaches have been used to study social learning,such as learning social hierarchies based on victoriesand defeats in a competitive game, or modeling of stra-tegic reasoning (see Ligneul and Dreher, Chapter 17;Palminteri and Pessiglione, Chapter 23; and Lee, Chap-ter 18).
One fundamental theoretical view about the brain,put forward by leading researchers such as Karl Friston
411Decision Neuroscience
Copyright © 2017 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/B978-0-12-805308-9.00033-6
and Rajesh Rao, is that the brain performs Bayesiancomputations in general, and particularly when makingdecisions. According to this view, decision-making andaction selection are treated as an inference problemsolving the problem of selecting behavioral sequencesor policies. Choices are based upon beliefs about alter-native policies, whereby the most likely policy mini-mizes the difference between attainable and desiredoutcomes. Policies are then selected under the priorbelief that they minimize the difference (relative en-tropy) between a probability distribution over statesthat can be reached and states that agents believethey should occupy. Future developments in the fieldof decision neuroscience will be to test this Bayesianview of the brain in various contexts, not only forperceptual and value-based decisions but also for socialdecision-making. Perspectives for research include thearticulation of neurocomputational definitions of valuecoding with a general Bayesian brain perspective. Thepioneering work of Karl Friston has been developedalong this line but will need to be extended to find pre-dictable experimental validations.
Yet another perspective for future research is toextend classical reinforcement learning approachesto social decision-making (see Ligneul and Dreher,Chapter 17) and strategic reasoning (see Lee, Chapter18). Social interactions are often repeated in a particularsetting, making it possible for decision-makers toimprove their strategies through experience. Therefore,the exact nature of learning algorithms utilized duringiterative social decision-making and the correspondingneural substrates are important topics for psychologicaland neurobiological research. Previous research hasshown that humans and nonhuman primates rely ona dynamic mixture of multiple learning algorithms forboth social and nonsocial decision-making. For simple,model-free reinforcement learning, strategies arerevised exclusively based on the observed outcomesfrom previously chosen actions, whereas for model-based reinforcement learning and belief learning,observed behaviors of other decision-makers and infer-ences about them also influence future choices. Thesedifferent types of learning algorithms might be imple-mented in different regions of the association cortexand basal ganglia, but how the neuronal activity ineach of these brain areas contributes to the specifictypes of computations for learning remains poorly un-derstood. For example, how the brain can update thevalues for multiple actions through mental simulationis difficult to study, because the activity related tosuch simulation may not be tightly linked to anyobservable sensory and motor events. In addition,although the brain must continuously support multiplelearning algorithms, how the outputs of various
learning algorithms are combined and how potentialconflicts between them get resolved need to be investi-gated further.
UNDERSTANDING THE FUNCTIONALORGANIZATION OF THE PREFRONTALCORTEX AND THE NATURE OF THECOMPUTATIONS PERFORMED IN
VARIOUS SUBREGIONS: VALUE-CODINGCOMPUTATIONS
The subdivisions of the prefrontal cortex and thecomputations performed by these subregions will bekey to providing a mechanistic understanding ofdecision-making. For example, the roles of subdivi-sions of the medial and lateral orbitofrontal cortexwill need to be specified, together with their participa-tion in multiple modulatory loops with other impor-tant structures such as the nucleus accumbens,ventral pallidum, amygdala, and hypothalamus, aswell as modulation with autonomic input from thegut (see Dagher, Chapter 32).
Similarly, functional divisions of the dorsolateral pre-frontal cortex will need to be further characterized. Thisbrain region, considered the highest level of the execu-tive hierarchy, temporally coordinates the perceptioneaction cycle by means of its cognitive executive func-tions upon the posterior cortex (see Fuster, Chapter 8).In addition, the dorsolateral prefrontal cortex has thecapability of anticipating (predicting) perception, action,and outcome; this confers to that cortex the functions ofplanning and preadapting that are critical for effectivedecision-making. The ventromedial prefrontal cortexcloses the perceptioneaction cycle by collecting neuralfeedback from reward, monitoring of outcome, andrisk assessment; it also has predictive capability of antic-ipated reward. All our decisions are to some degreeBayesian, based on the updating of prior hypotheses ofperception, action, or outcome, whether their “data-base” is conscious, unconscious, or intuitive; therefore,any reasonable computational neuroscience ofdecision-making should include probability as an essen-tial variable.
One key organizing concept in the field of decisionneuroscience is the concept of value. Understandingthe computational principles of value coding in thebrain has received considerable attention from re-searchers to understand the neurobiological basis ofdecision-making. This progress has illuminated bothwhere decision processing occurs in the brain andwhat information is represented in relevant neural ac-tivity. For example, neurophysiological and neuroi-maging studies have identified specific brain areas
33. PERSPECTIVES412
V. GENETIC AND HORMONAL INFLUENCES ON MOTIVATION AND SOCIAL BEHAVIOR
involved in option valuation and selection, includingthe frontal and parietal cortices, amygdala, and basalganglia including midbrain dopamine neurons, ventralstriatum, and pallidum. Neural activity in these areashas been shown to correlate with diverse decision vari-ables relevant to choice behavior, such as rewardmagnitude, risk, ambiguity, and delay to reinforce-ment. A central principle derived from this researchis that information about the idiosyncratic subjectivevalue of choice options is represented in the neuralactivity of decision-related brain areas and that it isthis idiosyncratic representation that appears to driveactual choice behavior. However, a critical aspect ofdecision-making processes (and one oddly relevant toeconomics and psychology) remains largely unex-plored: how neural circuits represent value informa-tion. In an information-processing system, the form ofinformation representation is a key intermediate levelmediating the link between low-level implementationand high-level goals. Information coding is a particu-larly significant issue for biological systems, whichface inherent constraints such as energetic costs andbiophysical limitations. Because such constraints limitthe information-coding capacity of neural systems,they require a transformation between the input(the variable to be encoded) and the output (the neuralactivity representing that variable) of a neural circuit.For example, the representation of the vast range of po-tential rewarding outcomes with the finite dynamicrange of neural activity necessitates a compressiveinputeoutput computation that can havesignificant implications for what we choose and whenwe choose it.
Thus, to understand decision-related inputeoutputfunctions (i.e., value coding), it will be critical for studiesnot simply to demonstrate correlation between neuralactivity and value but to quantify the precise relation-ships between the two. Experimental studies havebegun to quantify these neural value-coding computa-tions in brain regions such as the orbitofrontal and pos-terior parietal cortices. A notable finding of this initialwork is that these value inputeoutput functions are flex-ible and dynamic, changing in very specific ways inresponse to contextual influences such as the architec-ture of the choice set a decision-maker faces and the his-tory of past rewards encountered by that decision-maker. Importantly, this contextual value coding is, atleast in part, mediated by well-described computationssuch as divisive normalization that are prominent insensory processing, an observation arguing for a generalmechanism for information coding in the brain.
In addition to identifying, quantifying, and modelingthese value-coding computations, two specific direc-tions are important targets for future research. First,
how are value-coding computations related to thestructure and connectivity of the underlying biologicalcircuits? One important approach to answering thisquestion will be the examination of various circuitcomponents, including cells with different functionalroles (i.e., excitation versus inhibition), laminar loca-tions, and connectivity patterns. New devices andtechniques, such as large electrode arrays and optoge-netics, will be crucial to this process. Another promisingapproach in this direction is a dynamical analysis ofneural activity, focusing on fast-timescale (i.e., milli-second level) changes in firing rates rather than activityaveraged over long windows; such dynamics can revealkey details about the functional connectivity of neuralcircuits and the resulting patterns of information flow.Second, how do value-coding computations affectchoice behavior? Given the inherent constraints ofinformation processing in neural circuits, biologicaldecision-making can never reach the optimality pre-dicted by normative models that have no real biologicalconstraints. Quantifying the relationship between value-coding computations and choice behavior will illumi-nate both the constraints faced by biological choicesystems and how neural computational algorithmscompensate for those constraints.
DEMONSTRATING CAUSALITY:LINKING BRAIN ACTIVITY TO
BEHAVIOR BY DEVELOPING ANDAPPLYING PRECISE INTERVENTIONAL
TOOLS THAT CHANGE NEURALCIRCUIT DYNAMICS
To enable the immense potential of circuit manipula-tion, a new generation of tools for optogenetics, chemo-genetics, and biochemical and electromagneticmodulation should be developed for use in animalsand eventually in human patients.
Since the pioneering work by Wolfram Schultz andcolleagues, we have learned a great deal about thenature of dopamine responses during learning. Inparticular, we know that dopamine neurons signalreward-prediction error, or the difference between thereward that an animal expects and the reward it actu-ally receives. This signal is thought to reinforcerewarding actions and suppress alternative actions,potentially through corticobasal ganglia loops definedby expression of different dopamine receptors. Howev-er, we are only at the beginning stages of understand-ing how dopamine neurons calculate these responses.Given the number of different possible sources ofinput, how do dopamine neurons converge on suchsimilar prediction error responses? In what ways are
V. GENETIC AND HORMONAL INFLUENCES ON MOTIVATION AND SOCIAL BEHAVIOR
DEMONSTRATING CAUSALITY: LINKING BRAIN ACTIVITY TO BEHAVIOR BY DEVELOPING 413
dopamine neurons homogeneous versus heteroge-neous? Are they involved in learning from punish-ments as well as rewards? Furthermore, there aremany unanswered questions about how dopaminerelease affects downstream circuits in vivo. What arethe differential roles of phasic versus tonic dopaminefiring in motivating learning and behavior? What isthe effect of dopamine release on striatal and corticalneurons in vivo? How do striatal D1 and D2 neuronsinteract during behavior? What types of learningrequire dopamine, and what types are dopamine inde-pendent? To address these fundamental questions,newly developed molecular, genetic, and recordingtechniques will be critical.
By directly activating and inhibiting populations ofneurons in a behavioral context, neuroscience is pro-gressing from correlative measures to understandingof causal brain regions [transcranial magnetic stimula-tion (TMS), neuropsychology]. Methods such as TMSor transcranial direct current stimulation (tDCS) arethus likely (see Ruff, Chapter 19) to establish causalmechanisms for a given brain region, complementingclassical neuropsychological approaches in patientswith focal brain lesions (see Fellows, Chapter 22). Acentral challenge for a neuropsychological perspectiveon the role of the prefrontal cortex in value-baseddecision-making is to continue to dissect decisionprocesses at the level of brain mechanisms. We havegeneral guides to this now: clearly there are specificregions within the frontal lobes, for example, contrib-uting in specific ways to value-based choice. However,the mechanisms that are engaged remain unclear, inpart because the component processes of decision-making remain ill-defined, with likely multiple routesto decision-making in any given situation. We need totake advantage of converging methods to providerobust tests of well-specified, mechanistic models ofdecision-making. This of course is true for cognitiveneuroscience in general, but it seems particularly truefor decision neuroscience, in which, for the most part,models remain very general. Progress in the neuropsy-chological study of decision-making requires goodbehavioral measures of the constructs of interest.Although the past several years of work now betterequip us in this regard, there is still much to be done.Creative approaches that go beyond button-presschoices and reaction times, such as eye tracking, auto-nomic measures, and assessments of physical andcognitive effort, hold promise for uncovering the“microbehaviors” underlying value assessment andchoice. It is also increasingly clear that we cannottake an isolationist perspective on decision-making.Decision behaviors do not emerge fully formed fromsome specialized “economic” module of the brain,but rather are interlinked with attention, memory,
socialeemotional, and action-selection processes. Abetter understanding of these interactions will accel-erate advances, particularly as many of these relatedprocesses are much more thoroughly studied. Finally,decision neuroscience must aim to understand value-based choice broadly construed: in economic, but alsopolitical, social, and esthetic contexts. Testing the gen-erality of explanatory models across the whole gamutof motivated behavior will, in the end, yield the mostpowerful insights.
Finally, causality can also be assessed using compu-tational models, which allow researchers to assessprobabilistic causality in humans. Building on theoriesof nonlinear dynamical systems, whole-brain computa-tional models have been used to efficiently characterizenetwork-level communication across distributed sets ofbrain areas (i.e., functional connectivity) to investigatethe spatiotemporal dynamics of brain organizationand complex cognitive architectures [2]. This dynamiccharacterization can incorporate time-dependent activ-ity operating on varying timescales, which may capturea more complete picture of the spatiotemporal proper-ties inherent to decision-making.
MAPS AT MULTIPLE SCALES: GENERATECIRCUIT DIAGRAMS THAT VARY IN
RESOLUTION FROM SYNAPSES TO THEWHOLE BRAIN
It is increasingly possible to map connected neuronsin local circuits and distributed brain systems, enablingan understanding of the relationship between neuronalstructure and function. It is now possible to envisionimproved technologiesdfaster, less expensive,scalabledfor anatomic reconstruction of neural circuitsat all scales, from noninvasive whole human brain imag-ing to dense reconstruction of synaptic inputs and out-puts at the subcellular level.
For example, understanding of the circuit diagramsthat underlie impulsivity, risky choice, and impulse con-trol disorders is now possible to attain based on animalmodels. Impulsivity has emerged as a major dimen-sional construct in psychiatry with relevance to a rangeof disorders from addiction to attention deficit hyperac-tivity disorder (ADHD) and from Parkinson’s diseaseto depression, mania, and dementia. As a heritable,disorder-associated trait, impulsivity is broadlyacknowledged to affect the quality of decision-makingthrough effects on risk sensitivity, subjective value-based judgments (e.g., temporal discounting of delayedrewards), and cognitive control mechanisms responsiblefor the inhibition of ongoing behavior. Several decadesof research in humans and experimental animals haverevealed divergent but often interacting neural circuitry
V. GENETIC AND HORMONAL INFLUENCES ON MOTIVATION AND SOCIAL BEHAVIOR
33. PERSPECTIVES414
that underlies various impulsivity phenotypes,including the inability to await rewards, inability toterminate initiated behavior, preference for risky choice,or tendency to incompletely process information prior todecision-making. Yet formidable challenges lie ahead.For example, at present we lack a detailed understand-ing of the biological origins and neural circuitry of traitimpulsivity, including environmental interactions,and how these collectively contribute to poor impulsecontrol. Addressing this shortfall requires preclinicalscientists to study predictive biomarkers and neurode-velopmental trajectories for impulsivity in muchyounger animals. By continuing to explore the behav-ioral diversity of impulsivity and adopting translationalneural imaging, genomic, and objective behavioral ap-proaches, we expect to see further advances in our un-derstanding of trait impulsivity. This work requires adetailed dimensional analysis of impulsivity, character-ized in aggregate by variation in genes, molecules, andcircuits, in addition to a therapeutic focus away frombrain monoaminergic systems (e.g., in the form of medi-cation with Ritalin for ADHD) toward novel brain mech-anisms and hence new neuropharmacological targets.
THE BRAIN IN ACTION: PRODUCE ADYNAMIC PICTURE OF THE
FUNCTIONING BRAIN BY DEVELOPINGAND APPLYING IMPROVED METHODSFOR LARGE-SCALE MONITORING OF
NEURAL ACTIVITY
One important challenge in the future will be to re-cord dynamic neuronal activity from densely sampleddand in some test cases completedneural networks, overlong periods of time, in all areas of the brain, in bothmammalian systems and diverse model organisms,while making various types of decisions. There arepromising opportunities both for improving existingtechnologies and for developing entirely new technolo-gies for neuronal recording, including methods basedon electrodes, optics, molecular genetics, and nano-science and encompassing various facets of brainactivity.
The combination of existing techniques using multi-modal neuroimaging approaches in both nonhumanand human primates is also likely to bring insightsinto how the brain makes decisions. For example, thecombination of intracranial EEG (iEEG) recordings inpatients with epilepsy (whether with single cells or mac-roelectrodes) and fMRI, or single/multiple-cell record-ings combined simultaneously with fMRI in monkeys,should bring a better understanding of the precisetemporal dynamics at the systems level. Similarly, thenew PETefMRI scanners, which allow us to map
simultaneously both radiotracers and to acquire bloodoxygen level-dependent (BOLD) responses during deci-sion-making tasks, should bring exciting new findingsto the community. Converging approaches using thesame paradigms with different imaging modalities(e.g., EEG or MEG) and fMRI, together with physiolog-ical measures (e.g., pupil dilation, heart beat, etc.)should allow us to specify the dynamics of decisionstogether with a broader view at the neurophysiologicallevel.
Another interesting perspective from our field comesfrom the observation that the social environment shapesneural structures and processes, and vice versa. InChapter 28, Fernald gave a few examples of these inter-relationships using genetics and social behavior in ani-mals. Social animals interact with others and theirenvironment to survive and reproduce if possible. Todo this, animals acquire, evaluate, and translate infor-mation about their social and physical situation into de-cisions about what to do next. The information gatheredand the resulting decisions can profoundly alter both thebehavior and the physiology of an animal. These choicesin the brain are both produced by and result in a diversearray of cellular and molecular actions. The challenge isto discover where decisions are made and, in particular,what information is used to guide specific choices. Withnew genetic techniques, animal studies directed at un-derstanding how the brain decides are not restricted toa limited number of “model organisms” but any animalwith an interesting decision-making behavior.
THE ANALYSIS OF CIRCUITS OFINTERACTING NEURONS
The circuits of interacting neurons are particularlyrich in research opportunities, with potential for revolu-tionary advances. This area of research represents a realknowledge gap. We can now study the brain at veryhigh resolution by examining individual genes, mole-cules, synapses and neurons, or we can study large brainareas at low resolution with whole-brain imaging. Thechallenge remaining is what lies in betweendthe thou-sands and millions of neurons that constitute functionalcircuits.
One example is to understand the essential circuitrythat mediates the neural bases of goal-directed action.Bradfield and Balleine point that current research inneuroscience is predominantly technique driven and,as a consequence, it can be a challenge to maintainthe balance between doing what is expedient andasking questions that are worth answering. Not allrecently developed techniques are equally useful instudying complex psychological capacities, somethingthat is particularly true of studies investigating goal-
V. GENETIC AND HORMONAL INFLUENCES ON MOTIVATION AND SOCIAL BEHAVIOR
THE ANALYSIS OF CIRCUITS OF INTERACTING NEURONS 415
directed action in animals. In such experiments theevents to which the nervous system is exposed are pre-dominantly under the animal’s control rather than theexperimenters’, meaning, therefore, that, because theinitiating and terminating conditions for actions arefluid, the dynamics of the neural processes thatmediate both acquisition and subsequent performancecan be very complex. The challenge for the future isto bring this complexity under control. To the extentthat is achieved it may become possible to addressone of the most important open questions: it is stillnot known with any precision what learning rulesmediate the acquisition of goal-directed actions. Estab-lishing the essential circuitry supporting this learningprocess should help in that regard but there are impor-tant behavioral constraints to bear in mind. Forexample, different learning processes appear to beengaged at different rates by different schedules ofreward: ratio schedules generate more consistentgoal-directed learning and higher rates of performancethan interval schedules even when parameters areselected that match rates of reward delivery or interres-ponse times. Whether such distinctions can be capturedin associative or computational terms is still an openquestion. A number of researchers have recentlyclaimed that goal-directed learning is best captured,computationally, by model-based reinforcementlearning, using which a model of the environment isconstructed to ensure that action selection maximizeslong-run future reward. However, the performance ofgoal-directed actions respects the causal value of an ac-tion with respect to its specific outcome, and causalvalue does not necessarily coincide with reward maxi-mization. Indeed, considerable evidence suggests thatanimals prefer causal actions to both equally rewardingnoncausal actions and to performing no actions at all.Establishing the essential circuitry that mediates goal-directed action and the computational processes imple-mented in that circuit that make such actions possibleis one of the most important research problems andmost difficult challenges for future research.
DEVELOP INNOVATIVE TECHNOLOGIESAND SIMULTANEOUS MEASURES TO
UNDERSTAND HOW THE BRAIN MAKESDECISIONS
Consenting humans who are undergoing diagnosticbrain monitoring or receiving neurotechnology for clin-ical applications provide an extraordinary opportunityfor scientific research. This setting enables research onhuman brain function, the mechanisms of humanbrain disorders, the effect of therapy, and the value of di-agnostics. Seizing this opportunity requires closely
integrated research teams performing according to thehighest ethical standards of clinical care and research.New mechanisms are needed to maximize the collectionof this priceless information and ensure that it benefitsboth patients and science.
Examples include linking hormones and BOLDresponse during behavioral tasks (see Hermans andFernandez, Chapter 30, and Lefbvre and Sirigu, Chap-ter 31). Another related example concerns the effectsof acute stress on decision-making, which are just begin-ning to be understood. Such stress-induced shift from“reflective” to “reflexive” behavior may map twodistinct large-scale neural systems. This mapping isbased on a vast body of animal findings of effects ofstress-related neuromodulators within individual brainregions. It is essential that this cross-species inference iscorroborated in humans. There is, however, a paucity ofhuman pharmacological work detailing region-specificeffects and time-dependent effects of catecholaminessuch as dopamine and norepinephrine. In particular,we highlight the lack of human work on stress-induced dopamine release, which to our knowledge islimited to one seminal paper showing increased dopa-mine release using PET. Understanding the specificroles of dopamine and norepinephrine in the centralresponse to stressors will be critical to developing anunderstanding not only of immediate effects ondecision-making processes, but also of the specific vul-nerabilities that occur in response to acute stress in therealm of psychopathology. Regarding corticosteroids, afruitful road for further exploration will be to specifythe role of corticosteroids in limiting or terminatingthe acute response to stressors and promoting “reflec-tive” types of decision-making to enhance long-termadaptation. In particular, this role of corticosteroidshas not been explored fully in relation to stress-relatedpsychopathology. In investigating this, it will be impor-tant to distinguish the roles of baseline shifts and phasicresponses to stressors. One particularly promisingavenue is to further explore the potential of corticoste-roids in enhancing various forms of extinction-basedtherapy. Another large gap in our knowledge is howrapid and comprehensive shifts in neural activity aregenerated across large-scale neural systems. We high-lighted the potential contribution of stress-related neu-romodulators to this process, but these probably havedownstream effects on the balance between excitatoryand inhibitory neurotransmitters, which remain poorlyunderstood. Finally, the combination of basic neurosci-ence work with network-level analyses using functionalneuroimaging in humans has yielded important newinsights about the architecture of human cognition andits regulation at various levels of stress and arousal.One important future challenge is to translate thesenetwork-level findings back to basic neuroscience, in
V. GENETIC AND HORMONAL INFLUENCES ON MOTIVATION AND SOCIAL BEHAVIOR
33. PERSPECTIVES416
which these network-level effects can be studied inmuch more spatiotemporal detail using, for instance,in vivo electrophysiological recordings and optogeneticmanipulations.
ADVANCING HUMAN DECISIONNEUROSCIENCE: UNDERSTANDING
NEUROLOGICAL/PSYCHIATRICDISORDERS AND TREATING BRAIN
DISEASES
Clinical developments coming from the field ofdecision neuroscience and reward processing are vastand likely to bring new promises. For example, inParkinson’s disease, the main current treatment is thedopamine precursor drug, L-dopa, but its efficacydecreases over time while severe side effects increase.Understanding the brain’s motor circuits and decisionalsystem with deep brain stimulation, which can restoremotor circuit function in patients with Parkinson’s dis-ease for up to several years, may also help to understandhow we form a decision. Which factors specificallyinvolve the inhibitory cortical network interacting withsubthalamic nucleus (STN) function in the decision-making process? Is it the decision conflict per se or otherfactors such as choice difficulty, appetitive/aversivevalence of the choices, or information integration that in-fluence STN activity and adjustment of response thresh-olds? Changing dynamically the response thresholdmight be a universal function in decision conflict ormight be task specific. Therefore it has to be shown ifdifferent neuronal circuits/mechanisms are involved,for example, adopting risk-taking strategies or actingunder time pressure along the line of an accuracyespeed trade-off. In a clinical perspective the exact elec-trode position in relation to changes in inhibitory controlshould give us further insights into the exact fiber tractsthat are involved in the adjustment of responsethreshold. High-frequency stimulation has a negativeimpact on decision threshold. In analogy it should beclarified if low-frequency stimulation improves thedecision-making process, reflecting the other side.Similar research concerning deep brain stimulation ofvarious areas into brain circuits for mood and emotionhave the potential to advance psychiatry in similarways.
As noted previously, reinforcement learning com-bined with model-based fMRI has proven a valuabletool to reveal the brain regions computing predictionerrors during learning stimulusereward/punishmentassociations. It is now possible to use such tool to under-stand various neurological and psychiatric diseases,such as schizophrenia. Critically, this perspective linksclinical observations to a vibrant and rapidly developing
cognitive neuroscience field. More complex andsophisticated models of reinforcement learning arebeginning to demonstrate the importance of adaptationsin key parameters such as prediction error and learningrate. By explicitly studying this adaptivity and how itmay be perturbed in mental illness, we are likely todevelop an ever-richer explanatory link between keysymptoms of mental illness and alterations in brain,behavior, and cognition. Progress in refining our under-standing in this regard could ultimately pave the wayfor the introduction of precision medicine (scientificallybased, individually tailored treatment) interventions inpsychiatry.
Similarly, understanding the neuronal bases ofnegative motivational behavior including avoidancewill be crucial points to elucidate aversive behaviorrelated to psychiatric disorders such as anxiety.Nonhuman primate models would be essential for pre-clinical study. It would be required to find a neuronalcircuit for aversive behavior and observation of itsabnormal state. These processes would pave the wayto understanding psychiatric disorders and developingtreatments.
One example comes from the field of anxiety disor-ders. Research on psychiatric disorders has increasinglyfocused on broad biological and psychological mecha-nisms that can confer risk for psychopathology gener-ally speaking, with specific manifestations of disordersinfluenced by environmental factors experienced atdifferent developmental time points. A huge challengecurrently faced by the field is delineating what thesekey domains of functioning are that may confer suchbroad risk when disrupted, and how these disruptionsare neurobiologically characterized, all to better under-stand who may develop these conditions and treat orideally prevent clinical anxiety. In the search for thesebroad underlying mechanisms of anxiety disorders,the research domain of decision-making has beenlargely ignored, with most human neuroimagingstudies focusing instead on the passive elicitationof fear or anxiety. While phenomenologically valid,this approach falls short in demonstrating the adaptiveor maladaptive behavioral consequences of anxiety,including the choices one makes between potentiallyrewarding and punishing outcomes. Along withemerging investigations of value-based decision-making in anxiety and its disorders, extant data thatdo not explicitly probe decision-making processes pro-vide evidence for disruptions to neurobiological mecha-nisms throughout the decision-making process. Futureresearch that systematically explores alterations to spe-cific aspects of the decision-making process and associ-ated changes in brain function or structure, and linksthese changes with symptoms of anxiety and associatedpsychopathology, has the potential to advance our
V. GENETIC AND HORMONAL INFLUENCES ON MOTIVATION AND SOCIAL BEHAVIOR
ADVANCING HUMAN DECISION NEUROSCIENCE: UNDERSTANDING NEUROLOGICAL/PSYCHIATRIC DISORDERS 417
ability to diagnose, treat, and prevent the emergence ofanxiety disorders.
Our understanding of brain mechanisms underlyingdecision-making is also likely to bring new knowledgeto the understanding of drug and behavioral addictions.For example, gambling serves as a real-world exampleof risky decision-making and an activity that becomesexcessive for some people. Chapter 27 by Clark exploreswhat we currently know about decision-making and itsunderlying brain basis in gambling, with a focus ongambling disorder, the first recognized behavioraladdiction in the Diagnostic and Statistical Manual ofMental Disorders, fifth edition. Despite long-standingdiscussion in behavioral economics as to why peopleplay such games, given their negative expected value,it is only recently that researchers have begun to inves-tigate phenomena like loss aversion and the illusion ofcontrol in groups of participants separated in terms ofgambling involvement.
CONCLUSIONS
Collectively, the chapters from this book, DecisionNeuroscience, illustrate that: (1) theories and experimentsin neuroscience are helping to illuminate the mecha-nisms underlying decisions; (2) much remains to bedone regarding complex decisions; (3) social decisionneuroscience offers a special challenge of addressingmore complex problems that depend on predicting theintentions of others; (4) the social environment shapesneural structures and processes, and vice versa; and(5) new experimental methods (optogenetics) or nonin-vasive causal methods (e.g., TMS, tDCS) will help re-searchers to decipher the necessary brain regionsengaged in specific processes when making differenttypes of decisions.
To conclude, this book opens up three mainperspectives:
1. Pursue human studies and nonhuman models in parallel.The goal is to understand the human brain, but manymethods and ideas are developed first in animalmodels, both vertebrate and invertebrate.Experiments should take advantage of the uniquestrengths of diverse species and experimentalsystems. The research on animals has been and willremain crucial to determining the neural basis of theunderlying mechanisms of decision-making.
2. Cross boundaries in interdisciplinary collaborations. Nosingle researcher or discovery will solve the brain’smysteries. The most exciting approaches will bridgefields, linking experiments to theories, biology toengineering, tool development to experimentalapplication, human neuroscience to nonhumanmodels in innovative ways.
3. Integrate spatial and temporal scales. A unified view ofthe brain will cross spatial and temporal levels,recognizing that the nervous system consists ofinteracting molecules, cells, and circuits across theentire body, and important functions can occur inmilliseconds or minutes, or take a lifetime.
The most important perspective of the field of deci-sion neuroscience will be a comprehensive, mechanisticunderstanding of how the brain makes decisions thatemerges from synergistic applications of new technolo-gies and conceptual structures.
Reference
[1] Jorgenson LA, et al. The BRAIN Initiative: developing technologyto catalyse neuroscience discovery. Philos Trans R Soc 2015;370(1668).
[2] Deco G, Tononi G, Boly M, Kringelbach ML. Rethinking segrega-tion and integration: contributions of whole-brain modelling. NatRev Neurosci July 2015;16(7):430e9.
V. GENETIC AND HORMONAL INFLUENCES ON MOTIVATION AND SOCIAL BEHAVIOR
33. PERSPECTIVES418
Index
‘Note: Page numbers followed by “f” indicate figures, “t” indicate tables and “b” indicate boxes.’
AAcute stressbiphasic-reciprocal model, 379e380,
380fexecutive control network,
376e377behavioral effects, 376brain systems level effects,376e377
recovery, 377e378, 377finduction/hydrocortisone
administration, 379, 379flarge-scale neurocognitive systems,
372e373, 372fneural resources reallocation, 379e380,
380foverview, 371e372resource allocation shifts, 372e373salience network, 373e376behavioral effects, 374e375brain systems-level effects, 375e376cellular effects, 373e374neuroendocrine, 373e374recovery, 377e378, 377f
Amygdala, 179e180, 184, 184f, 193, 299,330, 340, 362e365, 374
overview, 33e41valence, 34e36cognitive processes, 39e41, 41fcognitive regulation, 36e39, 38fextinction, 36e39, 38flearning, 36e39, 38f
Anxiety disordersaction selection, 332e333learning, 334e335neural and behavioralmechanisms,
328e330, 329foutcome evaluation, 333e334overview, 327e335, 328frepresentation, 329f, 330e331valuation, 329f, 331e332
Appetitive brain systemsamygdala, 398e399decisionmaking, 402e404energy balance signals, 402e404food intake, 398e401, 399fghrelin, 403e404glucagon-like peptide 1, 404hippocampus (HC), 398e399insula, 399insulin, 402e403leptin, 403orbitofrontal cortex (OFC), 400
overview, 397e398, 398fpeptide YY, 404prefrontal regions, 400e401, 401fsatiety hormones, 404self-control and lateral prefrontal cortex,
401e402striatum, 400
Astatotilapia burtoni, 351e353, 352fconsequences, 353deceptive males, 353e354, 354ffemale mate choice consequences,
357e358, 358thierarchy, 354e355, 355fmale behavior, 355e356social system, 353transitive inference (male), 356e357,
356fAttention deficit hyperactivity disorder
(ADHD), 414e415Autism spectrum disorders (ASD), 201,
203Aversive behaviorsdirect and indirect pathways,
49e50pathological behaviors, 55ventral pallidum, 51e55ventral striatum, 51e55
BBalloon Analogue Risk Task (BART),
310, 331Basal gangliaappetitive approach behavior, 51aversive behaviors, 49e55cortex, 48f, 49overview, 47e49positive motivation, 51single-unit recording, 50ventral pallidum, 51e53avoidance preparation and anxietylevel controlling, 53e55
local inhibitory dysfunction,51e53
ventral striatumavoidance preparation and anxietylevel controlling, 53e55
local inhibitory dysfunction,51e53
Bayesian ideal observer, 150Biased competition, 99Blood oxygen level-dependent (BOLD),
189, 192, 306, 415e417
Brain Research through AdvancingInnovative Neurotechnologies(BRAIN), 411
Brain stimulation studiesoverview, 233e234social behavioral control, 238e240,
240fsocial cognition, 236e238, 238fsocial decisions, 234e235, 235fsocial emotions, 235e236, 237fsocial-specific neural activity, 240e241,
241f
CCatechol-O-methyltransferase (COMT),
144, 365Certainty equivalent (CE), 26Channelrhodopsin (ChR2), 35e36, 61Choice behavior, 121e123, 123fChosen value, 27e28Classical receptive field (CRF), 126Compassionempathy, 249prosocial motivation, 250
neural bases, 251e253, 253foverview, 247prosocial behavior, 253e255, 254fpsychological bases, 250e251, 251fsocial cognition, 248e249emotional contagion, 248e249mentalizing, 248mimicry, 248e249
Conditioned stimulus (CS), 34e35Corpus callosum (CC), 73e74Corticalebasal ganglialateral habenula, 14e15pedunculopontine tegmental nucleus,
15raphe serotonergic systems, 15subthalamic nucleus, 14thalamus, 14e15
DDBS. See Deep brain stimulation (DBS)Decision circuits, 123e125, 124fDeclarative, rational discounting
component, 88e89Deep brain stimulation (DBS), 315,
317e321Default mode network (DMN), 192Designer receptors exclusively activated
by designer drugs (DREADDs), 74
419
Diagnostic and Statistical Manual ofMental Disorders, fifth edition(DSM-5), 340, 361e362
Divisive normalization, 127e129DMF. See Dorsomedial frontal damage
(DMF)Dopamineblocking, 61fcircuitry, 62fe63fdecision-makingdecision variables updating, 27e28,28f
direct influences on choices, 28e29,29f
defined, 59e64function, 61e62neurons, 29overview, 21physiology, 59e61reinforcement learning, 263reward-prediction error, 21e23response components, 23reward learning, 23e25, 24f
schizophrenia, 263striatum, 64e67adaptive decision-making, 66circuits, 66e67habits, 64e66, 65finitiating movement, 64e66, 65fmotor skills, 64e66, 65fvalue representations, 66
subjective value, 25e26risk, 26utility, 26e27
Dorsal anterior cingulate cortex (DACC),4, 4f, 201, 204f, 332e333
Dorsolateral prefrontal cortex (DLPFC),138e139, 235e237, 239e241,362e364
Dorsomedial frontal damage (DMF), 279,279f
Dorsomedial striatum (DMS),72, 75f
Dynamic normalization model, 131,132fe133f
EElectrodermal activity (EDA), 341fElectroencephalography (EEG), 137e138,
199EsbereHaselgrove model, 113e114
FFeedback-related negativity (FRN), 205Five-choice serial reaction time task
(5CSRTT), 81e87, 90, 317, 318bFrontal lobesdecision-making, 279, 279fcomponent processes, 279e286, 280fdecisions over time, 285e286decision strategies, 284e285learning value, 282e283option identification, 280value and attention, 283e284, 283f
value construction, 280e284humans lesion evidence, 278designs, 278
overview, 277
GGambling disorderillusory control, 344e345loss aversion, 340e341, 341foverview, 339e340probability weighting, 341e342randomness perceptions, 342e343, 343f
Gambling rats, 90Gonadotropin-releasing hormone 1
(GnRH1), 351Good-based value, 165Gray matter (GM), 194, 195f
HHebbian rule, 165Hybrid reinforcement, 227e228, 229fHypothetical competitive decision
model, 29
IIllusory control, 344e345Immediate-early gene (IEG), 357Impulse control disorders (ICDs)associative learning, 307e308dopaminergic medications, 305e307,
306fimpulsivity, 310e311individual vulnerability, 305e307, 306foverview, 305parkinson disease, 305e307, 306f, 309freinforcement learning, 309freinforcing effects, 307e308risk, 309e310uncertainty, 309e310
Impulsivity, 310e311, 317e319defined, 81five-choice serial reaction time task
(5CSRTT), 81e82premature responding, 83
impulsive choice, 82e83motor impulsivity, 90neural basis, 84reward temporal discountingamygdala, 88e89corticolimbic influences, 88e89nucleus accumbens, 87
risky choice, 89e90stop-signal inhibition, 90stop-signal reaction time (SSRT), 83e84waiting impulsivity neural substrates5CSRTT premature responding,84e87, 85f
Intraparietal sulcus (IPS), 115f, 116Iowa Gambling Test (IGT), 264, 330
LLateral habenula (LHb), 63e64Lateral intraparietal area (LIP), 123e125,
124f, 127, 129e130, 130f
Lateral intraparietal region (LIP), 153Lateral prefrontal cortex (LPFC), 185Learning/motivational processesaction-outcome learning, neural bases
of, 72e74, 73fdorsomedial striatum convergence,
77e79incentive learning, 74e76, 75foutcome retrieval and choice, 76e77overview, 71e72prelimbic cortexemediodorsal
thalamus disconnection, 73fLinear decoder, 39
MMajor depressive disorder (MDD)decision-making, 361e364, 364fdefined, 361imaging genetics, 364e365
Medial frontal cortex (MPFC), 201Medial portion of the orbitofrontal cortex
(mOFC), 72, 76Mediodorsal thalamus (MD), 72Medium spiny neurons (MSNs), 64e65Midbrain dopamine neurons, 11e13, 11fafferent projections, 11f, 12efferent projections, 11f, 12striatonigrostriatal network, 13, 13f
Mixed selectivity, 99Model-free reinforcement, 227Monetary Incentive Delay Task (MIDT),
204e205Multi-Source Interference Task (MSIT),
204e205
NNash equilibrium, 225Neural computations, 127e130, 130fNeuromodulation, 144e145
OOne-stage theory, 270Optimal decision models, 151fOptogenetic stimulation, 66Optogenetic tagging, 60, 60fOrbitofrontal cortex (OFC), 37e38, 38f, 127Oxytocindefined, 387e388learning, 392memory, 392mode of release, 387e388, 388fnonsocial reward, 391e392, 392fperspectives, 392e393social brain, 387social decisions, 389e390, 389fcooperative behaviors, 389e390empathy, 390prosociality, 390social norms, 390trust games, 389e390
social reward, 390e392, 391fe392fsocial stimuliemotion perception, 388, 389fsensory perception, 388
INDEX420
PParkinson disease (PD), 59, 81, 297, 298f,
305e307, 306f, 309f, 315, 366e367,417
Perceptioneaction (PA) cycledecision execution, 101e102decision feedback, 102e103decision-making, 97e99, 98fdefined, 96e97, 96fe97fprediction and preparation, 99e101,
100fe101fPerceptual decision-makingaims and scope, 150beliefs, 158e160adaptive gain control, 159e160predictive coding, 159probability cues, 159
cortical neurons computations,153e154
mixed selectivity, 154posterior parietal cortex neurons,153e154
rodent, 153e154decision optimality, 150defined, 137e139, 149e150dorsolateral prefrontal cortex (DLPFC),
138e139electroencephalography (EEG),
137e138factors affecting, 139e145confidence, 143e144learning and training, 142e143neuromodulation, 144e145prestimulus state, 139e140reward and punishment, 140e141speed versus accuracy trade-off,141e142
human studies, 155e157BOLD signals, 155discrete computational stages,155e157, 156f
EEG signals, 155, 156finformation integration, 150e153lateral intraparietal region firing rates,153
observers sample informationsequentially, 150e152
sequential probability ratio test(SPRT), 150e152, 151f
observers, 157e158optimal bound collapses, 158speed and accuracy, 157e158, 157ftime-varying urgency signal, 157f, 158
overview, 149Posterior cingulate cortex (PCC), 192e193,
233, 235, 285e286, 321, 401fPosterior parietal cortex (PPC), 96e97, 112,
123e124, 153e154Posttraumatic stress disorder (PTSD), 53Prediction error, 99, 216Predictive machine, 102Prefrontal cortex, 39, 412e413decision execution, 101e102decision feedback, 102e103decision-making, 97e99, 98f
defined, 96e97, 96fe97fprediction and preparation, 99e101,
100fe101fPrimatessocial decision-makingamygdala, 184behavioral studies, 180e182, 181flateral prefrontal cortex (LPFC), 185neuronal correlates, 182e185, 183fOFC, 182e183, 183foverview, 179e180, 180f
Probability weighting, 341e342
RReinforcement learning (RL), 21e22,
33e34, 36, 121e122, 143, 163e176,203e204, 204fe205f, 212e216, 308,309f, 365e366, 412
computational models, 291be292b, 293fhybrid, 227e228model-free versus model-based, 226psychosisbehavioral studies, 264e265neuroimaging studies, 265e269,266fe267f
schizophrenia, 263social decision-making, 227
RescorlaeWagner models, 22, 27e28, 36,164, 215e216
Reward circuitcomplex network features, 15corticalebasal ganglialateral habenula, 14e15pedunculopontine tegmentalnucleus, 15
raphe serotonergic systems, 15subthalamic nucleus, 14thalamus, 14e15
midbrain dopamine neurons, 11e13, 11fafferent projections, 11f, 12efferent projections, 11f, 12striatonigrostriatal network, 13, 13f
overview, 3e4, 4fprefrontal cortex, 4e5ventral pallidum, 10e11, 10fventral striatum, 5e10amygdala/hippocampalprojections, 9
connections, 6e10, 8fcortical projections, 6e9efferent projections, 9e10features, 6, 7fthalamic projections, 9
Reward-prediction error (RPE) signal, 164,166e167, 166f, 291b, 294, 297
Reward processingpsychosisbehavioral studies, 264e265neuroimaging studies, 265e269,266fe267f
schizophrenia, 262Reward/punishment learninganterior insula and dorsal striatum,
299, 299fbehavioral tasks, 293b
dopaminergic opponent system, 295e296drug and lesion studies, 296e300, 298flimitations, 300neural candidates, 294e296opponent system, 295other opponent systems, 296overview, 291e292, 293f, 295fperspectives, 300serotoninergic opponent system, 296
Rewards valuecompetitive games, 170e171dopamine, 164flearning, 165e167overview, 163e164plastic synapses, foraging with, 166f,
167e170, 169frandom choice, 170e171representations, 164e165stochastic dopamine-dependent
plasticity, 166f, 167stochastic synapses, probabilistic
inference with, 171e173, 172fRight supramarginal gyrus (rSMG), 236,
237f
SSaliencebehavioral function, 111f, 113e114brain regions encoding, 115fcompound stimuli, 114fconfounds, 114e115defined, 113e114human brain signals, 116e117probability-based salience, 117fVMPFC, 117, 117f
Salvelinus fontinalis, 357Schizophreniaanhedonia, 271e272, 272fdelusions, 269e270dopamine, 263glutamate, 263e264overview, 261e262reinforcement learning psychosisbehavioral studies, 264e265neuroimaging studies, 265e269,266fe267f
reward processing psychosisbehavioral studies, 264e265neuroimaging studies, 265e269,266fe267f
Sensorimotor circuits, 123, 124f, 127Sequential probability ratio test (SPRT),
150e152, 151fSkill-chance confusion, 344Social brainhypothesis, 212medial prefrontal cortex (MPFC),
190e193, 190fe191fnetwork, 194e195, 195foverview, 189e190, 190fperspectives, 195e196superior temporal sulcus (STS),
193e194, 194ftemporal parietal junction (TPJ),
193e194, 194f
INDEX 421
Social conformity, 199, 203e206Social decision-makingreinforcement learningalgorithm, 228e229, 230fhybrid reinforcement, 227e228, 229fmodel-free reinforcement, 227model-free versus model-based, 226overview, 225e226
Social dominanceinterindividual differences, 216e222appraisal intercultural differences,216e218, 217f
social status and personality, 218neurochemical approaches, 219adrenocortical systems, 219e221, 220fdopamine, 220f, 221e222neural systems, 219e221, 220freproductive systems, 219e221, 220fserotonin, 220f, 221e222subordination, 219
reinforcement learning, 212e216social cognition, 212e215subordination, 214fe215f, 215e216
social hierarchieshealth and well-being, 211societies evolutionary pressure andfeature, 212, 213f
Social influenceoverview, 199prosocial behavior, 199e203autism observer effect, 201e203, 202fsocial reward processing, 199e201
valuation opinionsfuture directions, 206social conformity, 203e206, 204f
Social network size (SNS), 189Somatic marker hypothesis (SMH), 330SPRT. See Sequential probability ratio test
(SPRT)Standard economic theories, 226Subcallosal anterior cingulate cortex
(sACC), 362e364Subthalamic nucleus (STN), 417
corticobasal ganglia circuits, 316fanatomy, 315e317function, 315e317physiology, 315e317
decision-makinganimal studies, 317, 318bimpact model, 321e322, 322fneuropsychological studies, 319e321
direct pathway, 315e316impulsivity, 317e319indirect pathway, 315e316overview, 315
TTemporal difference (TD), 36, 163e165Temporal Experience of Pleasure Scale
(TEPS), 272, 272fTemporoparietal junction (TPJ), 115f,
116e117, 233, 238TEPS. See Temporal Experience of Pleasure
Scale (TEPS)Theory of mind/mentalizing, 199e200Three-circuit network model, 172e173Tonically active neurons (TANs), 66Transcranial direct current stimulation
(tDCS), 216, 234e235, 235f, 239e241,240fe241f, 414
Transcranial electrical stimulation (tES),234e235
Transcranial magnetic stimulation (TMS),199, 234e236, 235f, 414
UUnconditioned stimulus (US), 34e36,
38e39, 53e54, 54f
VValuebehavioral function, 110e111, 111fbrain regions encoding, 115fcompound stimuli, 114fdefined, 110e111
human brain signals, 111e113, 112foverview, 109e110ventral striatum, 112f
Value-based decision-making, 55, 140,163e176, 226, 277, 279, 327e328,376e377
Value codingbrain/behavior context dependence,
125e127, 126f, 128fe129fchoice behavior, 121e123, 123fcomputations, 412e413decision circuits, 123e125, 124fmechanism and behavior, 121, 122fneural computations, 127e130, 130foverview, 121temporal dynamics and circuit
mechanisms, 130e133, 132fe133fVentral striatum, 5e10amygdala/hippocampal projections, 9connections, 6e10, 8fcortical projections, 6e9efferent projections, 9e10features, 6, 7fthalamic projections, 9
Ventral tegmental area (VTA), 60e62, 306Ventrolateral prefrontal cortex (VLPFC),
193, 212Ventromedial frontal damage (VMF), 279,
279f, 283fVentromedial prefrontal cortex (VMPFC),
112e113, 117, 117f, 200e201, 330,332, 334
VMF. See Ventromedial frontal damage(VMF)
VMPFC. See Ventromedial prefrontalcortex (VMPFC)
Voxel-based lesion-symptom mapping,278, 283f
WWhite matter (WM), 194, 195f
INDEX422