Dual credit assignment processes underlie dopamine signals in a complex spatial environment

• DOI: 10.1016/j.neuron.2023.07.017
• Publication Date: 2023-08-22T14:32:59Z
• Authors: Timothy A. Krausz; Alison E. Comrie; Ari E. Kahn; Loren M. Frank; Nathaniel D. Daw; Joshua D. Berke
• Keywords: reinforcement learning; Reward; Dopamine; Decision Making; 150; Animals; Brain; decision-making; dopamine; credit assignment; navigation; maze; Rats

Animals frequently make decisions based on expectations of future reward ("values"). Values are updated by ongoing experience: places and choices that result in assigned greater value. Yet, the specific algorithms used brain for such credit assignment remain unclear. We monitored accumbens dopamine as rats foraged rewards a complex, changing environment. observed brief pulses both at receipt (scaling with prediction error) novel path opportunities. Dopamine also ramped up ran toward ports, proportion to value each location. By examining evolution these place-value signals, we found evidence two distinct update processes: progressive propagation along taken paths, temporal difference learning, inference throughout maze, using internal models. Our results demonstrate within rich, naturalistic environments conveys place values via multiple, complementary learning algorithms.