Research On Neuron Model Based On Ion Channel And Its Application In Spiking Neural Network

Different ion channels lead to different voltages.The diversity of cellular functions may be due to their morphology and also the nature and distribution of ion channels on the cell membrane.The computational neuroscience analysis method in the past achievements is not very mature in terms of theoretical basis and experimental conditions,and can't satisfy the study of complex phenomena of biological nervous system.Therefore,by virtue of the mode of information processing between neurons,this paper studies the nonlinear dynamic characteristics of the neural network by simulating the information processing mechanism of neurons and the behavioral characteristics of the neural network.It is very important to reveal the information transmission and processing function by adjusting the interconnection relationship between neuron nodes and then carrying out information processing.It's useful for constructing the relevant artificial neural network and studying its information transmission and processing function.With the development of neuron and artificial neural network,it is more urgent than any other period in the history to deeply understand the essence of neuron information processing and apply it reasonably.The proposal of this topic is also based on the in-depth analysis of the neuron information processing and its related applications.The main contributions of this paper are as follows:First,this paper discussed the common ion channels neuron model,based on the experimental data and the simulation of models,pointed out that the inherent nonlinear dynamics of neuron model.Based on the existing ion channel neuron models,this paper studies the randomness and certainty of the discharge rhythm under noise-induced conditions.The randomness and certainty of the discharge rhythm in the case of noise induction,and new discharge characteristics are proposed to enrich the discharge rhythm.In addition,based on the STDP learning rules,this paper considers the context of neuron firing pulses,and combines the Winner-Take-All competition rules to propose a new STDP learning rule to analyze different neuron information processing capabilities.The variation and robustness of the membrane voltage before and after training were studied under the premise of the new learning rules.Since the same algorithm is different in efficiency on different models,this part also combines different neuron models to study the influence of neuron models on spikes firing and information transmission.Finally,using a neuron model with time characteristics added to the ion channel neuron model,this paper proposes a novel neural network structure based on traditional spiking neural network,including the input layer,learning layer,output layer,and internal coding methods of the entire network,connection structures,etc.,and then applying them in different data sets to verify their validity.