Behavior of a Hippocampal Spiking Network and FPGA Implementation

The hippocampus in human brain is responsible for memory processing and tasks learning, which has long been one of the main interests of many researchers. In this paper, a spiking neural network of the hippocampus is realized based on a designed task. In the task, the model rat firstly gets familiar with the environment and finds item in one of the pots and finally gets a reward for making the correct response. Through the network, the feature of the hippocampal neurons can be simulated in software successfully. The three-layer network was built based on the spike-timing dependent synaptic plasticity and the synaptic weights will be modified between layers during task which can finally achieve the memory-related behavior of the model rat. Besides software realization, we also utilize Field Programmable Gate Array (FPGA) to reproduce the characteristics of the network in real time. The results show that the spiking neural network of the hippocampus can mimic the memory-related behavior of the model rat.