On the organization of grid and place cells: Neural de-noising via subspace learning
share
Place cells in the hippocampus are active when an animal visits a certain locations (referred to as place fields) within an environment and remain silent otherwise. Grid cells in the medial entorhinal cortex (MEC) respond at multiple locations, with firing fields that exhibit a hexagonally symmetric periodic pattern. The joint activity of grid and place cell populations, as a function of location, forms a neural code for space. An ensemble of codes, for a given set of parameters, is generated by selecting grid and place cell population and tuning curve parameters. For each ensemble, codewords are generated by stimulating a network with a discrete set of locations. In this manuscript, we develop an understanding of the relationships between coding theoretic properties of these combined populations and code construction parameters. These observations are revisited by measuring the performances of biologically realizable algorithms (e.g. neural bit-flipping) implemented by a network of place and grid cell populations, as well as interneurons, which perform de-noising operations. Simulations demonstrate that de-noising mechanisms analyzed here can significantly improve fidelity of this neural representation of space. Further, patterns observed in connectivity of each population of simulated cells suggest the existence of heretofore unobserved neurobiological phenomena.
READ FULL TEXT