Characterizing context-dependent neural activity in the rat hippocampus and entorhinal cortex

The rat hippocampus and entorhinal cortex possess neurons with spatial receptive fields that modulate their firing properties under different behavioral contexts. Such context-dependent changes in neural activity are commonly studied through spiking data recorded during electrophysiological experiments in which a rat performs a continuous spatial alternation task on a 1-maze. Analysis of context-dependent activity relies on an accurate understanding of spiking data on several levels. Described within this work are principled methodologies for characterizing neural activity for a single trial, within a group of related experimental trials, and between two or more sets of related experimental trials. To capture the temporal variability of the spiking activity for a single spike train, a general cross-validation framework for firing rate estimation was developed. To quantitatively describe spiking over a set of related trials, visualization methods based on empirical distributions of firing rate trajectories were developed. Then, bootstrap methods were created to quantitatively characterize group structure and between-group differences for firing rates from sets of related experimental trials. Using these methodologies, the definition of differential firing was broadened to include context-dependent changes in the stochastic structure of the trial-to-trial firing rate variability. This analysis contrasts with the prevailing mean-based methods, which were designed to capture differential activity from cells that fire almost exclusively during a single context. This work enhances the understanding of directionally selective neurons in the rat CA1 and dorsocaudal medial entorhinal cortex (dcMEC) during continuous T- maze alternation. From these analyses, cells were identified from both CA1 and dcMEC with spiking activity that transiently encoded directional context. These neurons are termed intermittent context-dependent (ICD) cells. This work shows that, beyond splitter cells, there exist individual cells that contribute contextual information for small subsets of trials, and that groups of individually uninformative ICD cells robustly encode turn direction together as a population. Decoding algorithms were developed to predict a rat's future turn direction from spiking data for single cells and for populations. The results confirmed that ICD information is combined across cells within CA1 and dcMEC ensembles to encode context.