L. Paul Sands

The paper shows that extracellular dopamine in human striatum encodes both reward and punishment prediction errors through separate pathways during a learning task.

This paper introduces a valence-partitioned reinforcement learning model that predicts human choice behavior, subjective experience, and BOLD responses by processing reward and punishment independently.

The paper reviews the challenges in understanding conscious experience and proposes a Dynamic Affective Core Hypothesis integrating affective value with neural representations.