Search Papers | Poster Sessions | All Posters
Poster A65 in Poster Session A - Tuesday, August 6, 2024, 4:15 – 6:15 pm, Johnson Ice Rink
Drift-diffusion dynamics of the hippocampal replay
Zhongxuan Wu1, Xue-Xin Wei1 (); 1University of Texas at Austin
The hippocampus and associated brain areas exhibit striking replay activities. Replays are thought to be important in learning, memory and planning, and have important implications in developing learning algorithms in machine learning. Surprisingly, how to characterize the structure of replays remains controversial. Most existing methods rely on restrictive assumptions, by detecting replay activities based on the sequentiality of spike trains or the posterior probability decoded from a Bayesian framework. We develop a general and high-interpretable drift-diffusion framework to understand the structure of replays. The two key parameters (drift & diffusion parameters) in the model can be directly mapped onto the speed and quality of a replay event. Applications of this framework provide new opportunities to address important open questions in the study of replays, including: (I) whether replays follow random walk; (ii) whether many of the replay events are stationary; (iii) whether preplay exists. We expect that our approach will be broadly applicable in studying the structure and dynamics of replays.
Keywords: neural coding spatial learning replay generative models