Opponent intracerebral signals for reward and punishment prediction errors in humans

Abstract

Whether maximizing rewards and minimizing punishments rely on distinct brain systems remains debated, inconsistent results coming from human neuroimaging and animal electrophysiology studies. Bridging the gap across species and techniques, we recorded intracerebral activity from twenty patients with epilepsy while they performed an instrumental learning task. We found that both reward and punishment prediction errors (PE), estimated from computational modeling of choice behavior, correlated positively with broadband gamma activity (BGA) in several brain regions. In all cases, BGA increased with both outcome (reward or punishment versus nothing) and surprise (how unexpected the outcome is). However, some regions (such as the ventromedial prefrontal and lateral orbitofrontal cortex) were more sensitive to reward PE, whereas others (such as the anterior insula and dorsolateral prefrontal cortex) were more sensitive to punishment PE. Thus, opponent systems in the human brain might mediate the repetition of rewarded choices and the avoidance of punished choices.