Thursday Computational Psychiatry seminar series - Vasilisa Skvortsova

Title: Computational noise in reward-guided learning drives behavioral variability in volatile environments

Speaker: Vasilisa Skvortsova (UCL)

Abstract:

When learning the value of actions in volatile environments, humans often make seemingly irrational decisions which fail to maximize expected value. Prominent theories describe these ‘non-greedy’ decisions as the result of a compromise between choosing a currently well-valued action vs. exploring more uncertain, possibly better-valued actions – known as the ‘exploration-exploitation’ trade-off. However, we have recently shown that the accuracy of human decisions based on multiple ambiguous cues is bounded not by variability in the choice process, but rather by computational noise arising from the underlying inference process. I will present the results of two behavioral and one fMRI experiment (a total of 90 participants) aimed to investigate the role of learning noise in driving behavioral variability in canonical restless bandit tasks. At the behavioral and computational levels, I will argue that more than half of non-greedy decisions observed in this kind of tasks are triggered by learning noise rather than by choice stochasticity. At the neurophysiological level, I will show that the trial-to-trial variability of sequential learning steps and their impact on behavior could be predicted both by BOLD responses in the dorsal anterior cingulate cortex (dACC) and by phasic pupillary dilation – suggestive of neuromodulatory fluctuations driven by the locus coeruleus-norepinephrine (LC-NE) system. Together, these findings indicate that most of the behavioral variability observed in volatile environments is due to the limited computational precision of reward guided learning.

Thursday 13 February 14:00

Max Planck UCL Centre for Computational Psychiatry and Ageing Research

Russell Square House, 10-12 Russell Square, second floor