Spatial rate/phase codes provide landmark-based error correction in a temporal model of theta cells

Location-Controlled Oscillator Model

This poster was presented at the Society for Neuroscience 2014 meeting:

J. D. Monaco, K. Zhang, and H. T. Blair. (2014) Spatial rate/phase codes provide landmark-based error correction in a temporal model of theta cells. Society for Neuroscience 2014, Washington, D.C. 752.07.

Abstract

The spatial firing patterns of place cells in hippocampus and grid cells in entorhinal cortex form a spatial representation that is stable during active navigation but also able to encode changes in external landmarks or environmental structure. One class of model that has been investigated as a possible mechanism for generating these spatial patterns relies on temporal synchronization between theta cells, which fire strongly with the septohippocampal theta rhythm (6–10 Hz) and are found throughout the hippocampal formation, that act as velocity-controlled oscillators. However, a critical problem for these models is that the oscillatory interference patterns that they generate become unstable in the presence of phase noise and errors in self-motion signals. Previous studies have proposed hybridizing temporal models with attractor network models or integrating environmental feedback from sensory cues. Preliminary data from subcortical regions in rats suggest that some theta cells exhibit spatially selective firing similar to hippocampal place fields or entorhinal/subicular boundary fields. These cells also demonstrate a consistent phase relationship across space, relative to ongoing hippocampal theta and to other simultaneously recorded cells, that is correlated with the firing rate at a given location. Inspired by this data, we present a novel synchronization model in which place cells or boundary-vector cells provide a stable, landmark-based excitatory input that drives a rate-to-phase mechanism to generate a population of cells that act as location-controlled oscillators. These cells fire preferentially at theta phases that are specific to a given location, determined by the presence of external landmarks. We show that these location-controlled oscillators provide a stable spatial reference that corrects phase errors in velocity-controlled oscillators, thus preventing the encoded position from drifting with respect to the environment and the actual position of the animal. Thus landmark-based rate/phase correlations in extrahippocampal areas may provide the sensory feedback required by temporal models of neural representations of space.