HIPPOCAMPAL REPRESENTATIONS OF SPATIAL AND ABSTRACT VARIABLES IN A DYNAMIC CUE-DIRECTED T-MAZE TASK

We seek to examine whether hippocampal neurons can encode both spatial and non-spatial abstract variables simultaneously, and how the generalized task structure is encoded and utilized by the hippocampal-prefrontal network. We designed a Dynamic Cue-Directed T-Maze (DCT) task where rats make a binary choice based on a visual cue received at the onset of each trial. While the spatial configurations change across trials, the task structure, i.e. the relationship between visual cues and correct decision, remains consistent. We run the DCT task in a custom-designed dynamic ‘Omniroute’ maze apparatus, a 3x3 grid of interconnected octagonal compartments with walls that can go up and down to produce dynamically changing configurations. We developed a training protocol for rats to learn the complex DCT task and successfully trained seven rats. We collected electrophysiological data from the Hippocampal CA1 region in four of these rats. Analysis of the collected data revealed cells that encoded different spatial and non-spatial task-relevant variables. We are quantifying how much information each cell encodes about task-relevant variables, and performing ensemble analyses to determine how the collective activity of these cells represent the task structure. We find that hippocampal activity seems to represent a combination of abstract and physical structures of the task. We further hypothesize that this representation will be complemented by a generalized task structure in the anterior cingulate cortex. We will perform simultaneous hippocampal/prefrontal recording to elucidate the relationship between the representations in both these structures.