Brain-Inspired Spatial Representation Based on Grid Cells with Finite-Spacing Firing Field

Grid cell is a kind of important neuron cell related to spatial cognition and navigation in animal brain. It has a hexagonal firing field extending to the whole space. Grid cells exist in the form of modules, whose number is finite, and the grid cells between adjacent modules have discrete firing- field spacing with constant ratio. The existing models can simulate the constant ratio discrete of firing-field spacing between modules, but the finite characteristic of firing-field spacing cannot be simulated. We propose a recurrent attractor network model of grid cell, which can generate grid cells with finite firing-field spacing by analyzing the connections between grid cells in a single network. Then, the grid cells in different modules could possess discrete firing-field spacing with constant ratio through the competition among the recurrent networks. The recurrent attractor network model can provide reference for the construction of brain-inspired navigation system of unmanned platform.