Research On The Design Of Hippocampal Functional Neural Network And FPGA Implementation

The Hippocampus is an important part of the brain.It is closely related to learning,memory and other functions.The establishment of functional network is of great significance to the in-depth study of its working mechanism.The simulation of its learning and memory function plays an important part in realizing brain-like computing.Utilizing Field Programmable Gate Array(FPGA)to realize hippocampal functional neural network is one of the research directions and difficulties of brain-like computing.Therefore,in this paper the learning and memory functions of hippocampal neurons were simulated by constructing a hippocampal functional neural network and by changing the structure of the network the simulation accuracy is improved and a good learning effect is obtained.Finally,the hippocampal functional network is realized by FPGA.Based on the understanding of hippocampal structure and function,this paper takes an experimental scenario as a prototype and uses the LIF neuron model as the single neuron model,combined with synaptic plasticity and the “winner-take-all” rule to construct a three-layer feed-forward network.It realizes the learning and memory function of hippocampal neurons.Secondly,in order to further analyze the characteristics of the constructed hippocampal functional network,this paper changes the network structure by increasing the number of layers of the feed-forward network and the number of neurons in the middle layer of the network,compares the learning effects of the networks before and after changing the structure and analyzes the differences in learning outcomes from multiple aspects.Finally,according to the constructed hippocampal functional network,the FPGA implementation of the network is completed.The hardware design of each component in the model are explained in detail,and the numerical operations and nonlinear function operations involved in the design process are solved.The non-multiplier hardware implementation method greatly saves the consumption of hardware resources.The hippocampal functional network in this paper can provide a scheme for the realization of the learning and memory function of brain-like computing.