This event has already occurred.
Multiple Learning Strategies in Prefrontal-basal ganglia and cerebellar networks
Ph.D. candidate at the Okinawa Institute of Science and Technology and Nara Institute of Science and Technology
2 Riverside Circle, Roanoke, VA 24016
Abstract: We proposed a three-system parallel model for action selection and learning from reward and trial-and-error. When knowledge of the task dynamics is unavailable, a valued-based method is used for exploration and learning predictions of future rewards from sensory states and actions. If an internal model exists, a model-based method is implemented by mental simulation to predict future states reached by hypothetical actions and speed up learning. After repeated experiences, a memory-based method is used for fast selection of actions found successful for a given sequence of sensory states. I will discuss and present behavioral and fMRI data that support this hypothesis.
About the Speaker:
Alan is a Ph.D. candidate at the Okinawa Institute of Science and Technology and Nara Institute of Science and Technology.