A neural network model self-organizing a cognitive map using theta phase precession

The cognitive map is thought to consist of place cells in the hippocampus. We hypothesize that “theta phase precession”, a characteristic relation between the firing of the place cell and local field theta rhythm, contributes to the generation of the cognitive map as a network of place cells. We propose a network model of the hippocampus based on the hypothesis. In our model, the entorhinal cortex (EC) receives sensory-inputs of local views. The dynamic pattern of theta phase precession is generated by mutual synchronization of neural oscillators in the EC. In CA3, neural units receive inputs from the EC and the cognitive map is self-organized in the recurrent neural network. In the computer experiments, inheritance of theta phase precession in CA3 results in a kind of phase wave propagating along a 2-D plane. The 2-D plane is define as the desired array of place cells in accordance with their place fields. The wave propagation deriving from theta phase precession globally controls the synaptic modification among place cells so that their network is self-organized as the cognitive map