Learning in visual regions as support for the bias in future value-driven choice

Abstract

10Reinforcement learning can bias decision-making towards the option with the highest expected 11 outcome. Cognitive learning theories associate this bias with the constant tracking of stimulus values and 12 the evaluation of choice outcomes in the striatum and prefrontal cortex. Decisions however first require 13 processing of sensory input, and to-date, we know far less about the interplay between learning and 14 perception. This fMRI study (N=43), relates visual BOLD responses to value-beliefs during choice, and, 15 signed prediction errors after outcomes. To understand these relationships, which co-occurred in the 16 striatum, we sought relevance by evaluating the prediction of future value-based decisions in a separate 17 transfer phase where learning was already established. We decoded choice outcomes with a 70% accuracy 18 with a supervised machine learning algorithm that was given trial-by-trial BOLD from visual regions 19 alongside more traditional motor, prefrontal, and striatal regions. Importantly, this decoding of future 20 value-driven choice outcomes again highligted an important role for visual activity. These results raise 21 the intriguing possibility that the tracking of value in visual cortex is supportive for the striatal bias 22 towards the more valued option in future choice. 23 24