1 comment

[ 5.0 ms ] story [ 9.7 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