Associative memory and segmentation in a network composed of Izhikevich neurons

Zhang Wei,Qiao Qingli,Zheng Xuyuan

DOI: https://doi.org/10.1109/ICNC.2008.538

2008-01-01

Abstract:Associative memory is one of the brain's main function. This paper presents a new artificial neural network composed of Izhikevich neuron models to simulate the associative memory and segmentation of human brain. The stored memory patterns are coded with the connection weight. The memory is represented in the spatio-temporal firing pattern of the neurons. The stored memory patterns can be retrieved and segmented through the adjusting of connection weight when the network is presented with corrupted input patterns. The simulation results prove that connection weight plays an important role in the associative memory and segmentation of human brain, by changing the connection weight the neural network can implement the associative memory and segmentation functions of human brain. © 2008 IEEE.