Some Computational Predictions on the Possibilities of Three-Dimensional Properties of Grid Cells in Entorhinal Cortex

Abstract

The discovery of grid cells in the entorhinal cortex (EC) of the rat (Hafting et al. 2005) has provided many hints of the mechanism of spatial computation in brain during animal movement. Since then, various experiments as well as computational modeling studies of grid cells have answered some of the key questions related to the properties of these cells. However, almost all of these studies are conducted on the rats and mice during their movement in horizontal space, and it is not clear whether the grid cells possess a three-dimensional firing field during movement in space that is either tilted or curved. In this paper, we make some predictions on the possibilities of three-dimensional shapes of grid fields by hypothesizing that they indeed possess such properties, and produce such three-dimensional fields during movement in tilted space. We show several polyhedral shapes that can be generated by our computational neural network model, and in case of movement in horizontal plane, our three-dimensional grid cell model is reduced to a two-dimensional model to generate grid fields similar to experimental findings.