Single Neuron Correlates of Learning, Value, and Decision in the Human Brain

Author: Gallo Aquino, Tomas

Year: 2022

Degree: Dissertation (Ph.D.)

Advisor: O'Doherty, John P.

Committee Members: Adolphs, Ralph; O'Doherty, John P.; Rutishauser, Ueli; Allman, John Morgan

Option: Computation and Neural Systems

DOI: 10.7907/has2-gk35

Abstract

In this thesis, I present several new results on how the human brain performs value-based learning and decision-making, leveraging rare single neuron recordings from epilepsy patients in vmPFC, preSMA, dACC, amygdala, and hippocampus, as well as reinforcement learning models of behavior. With a probabilistic gambling task we determined that human preSMA neurons integrate computational components of stimulus value such as expected values, uncertainty, and novelty, to encode an utility value and, subsequently, decisions themselves. Additionally, we found that post-decision related encoding of variables for the chosen option was more widely distributed and especially prominent in vmPFC. Additionally, with a Pavlovian conditioning task we found evidence of stimulus-stimulus associations in vmPFC, while both vmPFC and amygdala performed predictive value coding, establishing direct evidence for model-based Pavlovian conditioning in human vmPFC neurons. Finally, in a Pavlovian observational learning paradigm, we found a significant proportion of amygdala neurons whose activity correlated with both expected rewards for oneself and others, and in tracking outcome values received by oneself or other agents, further establishing amygdala as an important center in social cognition. Taken together, our findings expand our understanding of the role of several human cortical brain regions in creating and updating value representations which are leveraged during decision-making.

Files