The Modeling And Simulation Of Izhikevich Neural Network

Objective: In this paper, an artificial neural network is built to study the role of connection weight in the synchronization of the neurons in the neuronal network. The role of connection weight in the associative memory and pattern segmentation is analyzed.Methods: An artificial neural network with Izhikevich neuron model is used to simulate the associative memory and pattern segmentation when there is Gaussian white noise in the network; the role of the connection weight in the associative memory and pattern segmentation is studied.In the paper the stored memory patterns in the neural network are stored by the connection weight. The neural network is presented with Gaussian white noise. When the network are presented individually with perfect and polluted input patterns, the firing of neurons become synchronous gradually with the increment of the connection weight.Result: The network is presented with both perfect and corrupted input pattern when there is only one stored memory pattern in the neural network The neurons in the neural network fire randomly when the connection weight is small. With the increment of the connection weight, the firing of the neurons which belong to the same memory pattern in the network becomes synchronous, then, the memory of stored memory pattern is achieved. The network is presented with both perfect and corrupted input patterns when there are two stored memory patterns in the neural network. The neurons in the neural network fire randomly when the connection weight is small. With the increment of the connection weight, the firing of the neurons which belong to the same pattern in the network becomes synchronous. The memory and segmentation of stored memory patterns are achieved.Conclusion: In this paper an artificial neural network is built. The neurons in the neural network can fire synchronously with the increment of the connection weight.. The neuron model in the neural network is Izhikevich neuron model. The results show that the Izhikevich model is suitable for the simulation of large scale neural network. The connection weight plays an important role in the memory and segmentation of stored memory patterns.