S01-195 - Predicting spatial behavior from complex hippocampal oscillatory codes

Session Name
1510 - Poster Session 01 - Section: Emergent Dynamics in Neural Networks (ID 501)
Date
10.07.2022
Session Time
09:30 AM - 01:00 PM

Abstract

Abstract Body

Theta and gamma oscillations are believed to organise hippocampal activity and function. The current dominant view posits the existence of two gamma frequency sub-bands, occurring at different theta phases in CA1, produced by different generators and sub-serving exclusive cognitive operations. However, this view relies on averaging over many oscillatory events, possibly overshadowing recently suggested manifold theta-nested gamma contents. We thus explored agnostically the diversity and informational capacity of individual theta-gamma bouts in LFPs recorded in the mouse dorsal hippocampus (n=5) during the learning of a spatial reference memory task. We found gamma bouts with nearly every combination of frequency, amplitude and theta-phase whichever the hippocampal layer. Importantly, although their average can resemble the two classic sub-bands pattern, they mainly reflect few strong power events. Furthermore, this diversity would not depend entirely on external generators: a spiking computational model of local recurrent circuitry involving one excitatory and one inhibitory (fast-spiking-like) populations can generate similarly complex oscillations fluctuating over the entire gamma spectrum without fine-tuning for most parameter combinations. Finally, gamma diversity is functional: a machine learning approach could accurately decode mice running speed and coarse maze position from individual theta-gamma oscillatory events, for all layers. The theta-gamma “code” nonetheless differed across layers and learning stages. Altogether, our findings suggest that hippocampal LFP oscillatory diversity is not mere noise but carries an actual encoding of context and behavior. These population codes are complex in nature and not reducible to simpler descriptions relying on few reference bands.

Hide