FIRING RATE MODULATIONS: A SIMPLE STATISTICAL FRAMEWORK TO STUDY NEURONAL ENSEMBLE REACTIVATION

F.P. Battaglia¹*; B.L. McNaughton²; K.D. Harris³

1. LPPA, College de France, Paris, France

2. NSMA, Univ Arizona, Tucson, AZ, USA

3. CMBN, Rutgers Univ, Newark, NJ, USA

Memory consolidation is believed to involve the reactivation of cell assemblies activated by salient events, during subsequent sleep. Previous work has demonstrated such reactivation by relating pairwise correlations and temporal patterns during waking and sleep sessions. Under some conditions, however, a simpler consequence of assembly reactivation would be increased firing rate of those neurons taking part in reactivated assemblies, as observed for individual neurons by Pavlides & Winson (1989). During behavior in an environment a fraction, dependent on the size of the environment, of the hippocampal pyramidal cells is active and spatially selective. It is shown here that those cells show increased firing rate in subsequent sleep. The ratio of a cell's firing rate in sleep after behavior to sleep before behavior is related to the ratio of its rate during experience and initial sleep, by a power law over several orders of magnitude. This relationship was strongest for the first 10 minutes of sleep imme diately following behavior, and was stronger during hippocampal sharp waves than during non-sharp wave epochs. Across experiments, variations in firing rate reactivation mirrored variations in pairwise reactivation, suggesting the two measures reflect similar biological processes. Reactivation was stronger when rats were run on a cue-rich track, as compared to a familiar plain track. We suggest that the increased reactivation in this case reflects higher mnemonic saliency of the cue-rich track. While reactivation of rate modulations has so far received less attention than pairwise correlations, the comparative simplicity and statistical robustness of this approach suggests that firing rates analyses may prove valuable in the study of the factors affecting reactivation.

Support Contributed By: MH46823