2 Abstract

In this thesis, I combine insights from economics, psychology, and neuroscience to develop ‘Counterfactual Predicted Utility Theory’ (CPUT) as a neurobiologically-plausible theory of decision-making under risk. CPUT is inspired by the observation that sub-second fluctuations in the levels of the neurotransmitter dopamine seemingly reflect factual and counterfactual information. I propose that people incorporate anticipated counterfactual events when making risky decisions. This leads to behavior that is considered ‘irrational’ from a classical economic perspective as described by Expected Utility Theory (EUT). To assess counterfactual predicted utility theory’s validity as a theory of decision-making under risk, I compared variations of CPUT and EUT on human choice data from a sure-bet or gamble task using hierarchical Bayesian modeling techniques. I quantified model fit with multiple methods. This includes comparing marginal likelihood model evidence and leave-one-out cross validation predictive accuracy. Compared to EUT, I found that CPUT does not offer a better explanation for the data collected, nor does it generalize as well. However, the task design limits the available counterfactual information, and the data collected may not be sufficient to accurately assess CPUT. Further research into the neurobiological mechanisms for processing factual and counterfactual information and the downstream behavioral consequences is warranted.