A Caltech Library Service

The Neurocomputational Basis of Self-Control Success and Failure


Sullivan, Nicolette Jean (2015) The Neurocomputational Basis of Self-Control Success and Failure. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/Z90P0X8V.


Individuals often have difficulty delaying gratification – that is, forging smaller sooner rewards in favor of larger rewards delivered at a delay. Common examples of this deficit in self-control are difficulties in saving for retirement, going to the gym, or eating healthy foods. Despite an extensive literature on the neural substrates of decision-making, relatively little is still known about the sources of underlying individual variation in the ability to successfully execute self-control. This manuscript presents three studies examining the sources of individual differences in delay of gratification, with the goal of obtaining a more complete understanding of the neural mechanisms underlying choice. The main question this thesis addresses is: what features of the brain’s decision process allow individuals to down-regulate the appeal of smaller sooner rewards, in order to forgo them in favor of greater future reward?

In the first study, I present a novel method of measuring decision process dynamics, in which we harness the power of fine temporal resolution in recording computer mouse movements in dietary choices. We find that up to 39% of individual variation in dietary self-control can be explained by differences in the speed with which the decision-making circuitry processes basic attributes, such as tastiness, versus more complex, abstract, attributes, such as healthfulness.

In the second study, we extend this novel approach to a classic experimental economics paradigm, intertemporal choice. We found large individual variance in the speeds with which immediate and delayed reward values were processed. We found that about 25% of the individual differences can be explained by differences in the speed at which delayed rewards are processed. We also found that the relative speed at which immediate and delayed rewards are processed fluctuated across trials: subjects processed delayed rewards faster than immediate rewards when they made patient choices, but the order of processing speeds was reversed during impulsive choices. Together, these results show that a sizable fraction of variation in the ability to postpone gratification might be attributable to variables that affect the speed at which different types of rewards are processed, and not to differences on deep preference parameters like the temporal discount rate used by the brain’s valuation systems.

Across the lifespan, self-control improves in many choice domains. The third and final study capitalizes on this phenomenon of behavioral change with age to investigate the neural mechanisms underlying improvements in self-control. I use functional magnetic resonance imaging (fMRI) to examine the neural correlates of changes in discounting of future monetary rewards across the lifespan from adolescents (13 years old) to seniors (70 years old). We find that neural response to value in reward-related striatal brain regions dramatically decrease with age. In contrast, we find that the left dorsolateral prefrontal cortex, often found to be related to successful self-control, increases its functional connectivity to key valuation, reward, and future-thinking brain regions with age during very tempting trials. These results suggest a mechanism through which increased self-control is improved.

Taken together, these studies argue that individual features of the decision process have a large influence on the overall ability to exert self-control in both dietary and monetary choice domains. Specifically, we find that the speed with which abstract future attributes such as health information, relative to more concrete attributes such as taste, are processed have a large influence on individual self-control ability. We also find that decreased reward sensitivity, paired with increased effective connectivity between control and valuation regions specifically when control is required most, allow for increased ability to delay gratification with age.

Item Type:Thesis (Dissertation (Ph.D.))
Subject Keywords:self-control, decision making, economics, neuroeconomics, temporal discounting, intertemporal choice
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):
  • Rangel, Antonio
Thesis Committee:
  • Adolphs, Ralph (chair)
  • Quartz, Steven R.
  • O'Doherty, John P.
  • Camerer, Colin F.
Defense Date:5 May 2015
Non-Caltech Author Email:nikkisjunkmail (AT)
Record Number:CaltechTHESIS:05292015-162052753
Persistent URL:
Default Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:8932
Deposited By: Nicolette Sullivan
Deposited On:07 Mar 2017 19:48
Last Modified:04 Oct 2019 00:08

Thesis Files

PDF - Final Version
See Usage Policy.


Repository Staff Only: item control page