CaltechTHESIS
  A Caltech Library Service

Multiple forms of valuation in the human brain

Citation

Wunderlich, Klaus (2010) Multiple forms of valuation in the human brain. Dissertation (Ph.D.), California Institute of Technology. http://resolver.caltech.edu/CaltechTHESIS:11112009-132836900

Abstract

Our lives are defined by the decisions we make, often involving choices between different actions or goods. An important open problem in decision neuroscience is, what value signals are used in guiding the different types of choices, where are they stored in the brain, and how does the brain compare them to make a choice. We used fMRI in human subjects to address these questions in a variety of different choice settings: decisions between actions, economic choices, and more complex hierarchical decisions. We found evidence for a separate representation of two main forms of value signals in the human brain: precursors of choice, such as signals relating to the value of each available action or stimulus, and signals reflecting the consequence of the decision process by encoding the expected value of the option that is subsequently chosen. On the precursor side, we found action value signals in supplementary motor cortex and stimulus value signals in medial prefrontal cortex. Separate brain regions, most prominently ventromedial prefrontal cortex, were involved in encoding the value of the chosen action or stimulus. Importantly, we found value chosen signals in stimulus decisions even when no actions were associated with choosing the stimuli, providing evidence for the hypothesis that the brain doesn’t need the motor system to make such decisions but is capable of making economic choices completely within an abstract representation of goods. Furthermore, in action decisions, we found that activity in dorsomedial frontal cortex resembles the output of a decision comparator, implicating this region in the computation of the decision itself. In a real world setting where multiple stimuli could potentially influence outcomes, an individual may consider a number of theories about which features are relevant for giving reward. We found that decision variables based on simultaneous integration of all evidences were better able to explain subjects’ behavior and activity in prefrontal cortex than those generated by an attention-gated approach, i.e. by first picking the theory that is most likely correct and then choosing accordingly. These results demonstrate that the human brain is capable of optimally integrating information, similar to an ideal Bayesian observer.

Item Type:Thesis (Dissertation (Ph.D.))
Subject Keywords:decision making, neuroeconomics, reward learning, fmri, vmpfc, sma
Degree Grantor:California Institute of Technology
Division:Engineering and Applied Science
Major Option:Computation and Neural Systems
Thesis Availability:Public (worldwide access)
Research Advisor(s):
  • O'Doherty, John P.
Thesis Committee:
  • Andersen, Richard A. (chair)
  • Bossaerts, Peter L.
  • Adolphs, Ralph
  • Shimojo, Shinsuke
Defense Date:9 November 2009
Author Email:kwunder (AT) caltech.edu
Funders:
Funding AgencyGrant Number
Gordon and Betty Moore FoundationUNSPECIFIED
Record Number:CaltechTHESIS:11112009-132836900
Persistent URL:http://resolver.caltech.edu/CaltechTHESIS:11112009-132836900
Default Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:5364
Collection:CaltechTHESIS
Deposited By: Klaus Wunderlich
Deposited On:02 Apr 2012 19:26
Last Modified:26 Dec 2012 03:18

Thesis Files

[img]
Preview
PDF - Final Version
See Usage Policy.

3894Kb

Repository Staff Only: item control page