Dynamics of neural assemblies involved in memory-trace replay.
1. NSMA
2. Psychol, Univ. Arizona, Tucson, AZ
The replay of behaviorally induced multi-neuronal activity patterns during subsequent sleep is considered to play an important role in the consolidation process of certain types of memory. Studies using the explained variance method have shown that memory-traces are reactivated during slow-wave sleep in the hippocampus, in the neocortex and in the ventral striatum. The reactivation, however, often decays to undetectable levels within 1 hr, except for the ventral striatum where it shows little decline for up to 40 min. The presently available methods typically use all recorded neurons in their analysis, because it is difficult to detect neuronal sub-populations that specifically contribute to memory. In this study, we applied an analysis framework using spike train clustering (STC) and information geometry (IG) to long-term recordings from medial prefrontal cortex and hippocampus. The recordings span a 12-hour pre-task period, a 1-hour task period where the rat freely explored novel objects, and a 12-hour post-task period. An initial clustering on the basis of parwise correlation coefficients identifies groups of neurons of potential interest. Simulations results using biophysical model neurons show that IG analysis could reveal changes in correlations that were specifically due to effective synaptic strength modulations or due to modulations in external inputs. We show that such specific changes can be detected in these long term recordings and that different groups of neurons become significantly and selectively correlated at different times. Some sub-groups show correlations primarily during sleep or during awake periods but not both. Other sub-groups may be correlated with parameters such as animal position and different EEG states which remain to be determined.
These results indicate that the method using STC and IG are powerful tools for the detection and analysis of multi-neuronal spike patterns.
Grant/Other Support: MH046823
Keyword (Complete): reactivation; neural coding; timing