1 comment

[ 3.2 ms ] story [ 27.2 ms ] thread
Grounded in reinforcement learning. General idea is hippocampal replay is valuable, and a rational agent should arbitrate between work and replay in order to maximize future reward