Detail

Neural Reinforcement Learning Lab (NeuRLab)

Introduction

Living in an uncertain environment, we desire to pursue good things and to avoid bad things. We are interested in how the brain recognizes different situations and learns to make better decisions. Related questions are: How does the brain represent reward or punishment? How does the brain remember something good and pursue it? How does the brain choose one action out of multiple options? What makes one animal more intelligent than another animal? What can we learn about how the brain works from artificial intelligence?

Reinforcement learning (RL) theory provides theoretical and computational frameworks to these problems. Interestingly, it has been shown that dopamine activity in the brain resembles the teaching signal in one of reinforcement learning theories, temporal difference (TD) learning. However, the detailed neural mechanisms of adaptive behaviors remain elusive. We perform experiments using animals and analyze data using computational models derived from artificial intelligence (AI) to understand the biological mechanisms of reinforcement learning.

Selected Recent Publications

1. Kim HR*, Malik AM*, Mikhael JG, Bech P, Tsutsui-Kimura I, Sun F, Zhang Y, Li Y, Watabe-Uchida M, Gershman SJ, Uchida N (2020) A unified framework for dopamine signals across timescales. Cell (lead author)

2. Kim HR, Angelaki DE, DeAngelis GC (2017) Gain Modulation as a Mechanism for Coding Depth from Motion Parallax in Macaque Area MT. Journal of Neuroscience 37 (34), 8180-8197

3. Kim HR, Angelaki DE, DeAngelis GC (2015) A novel role for visual perspective cues in the neural computation of depth. Nature Neuroscience 18(1), 129-137.