The Study Of Neural Network Information Transmission And Functional Connection

The brain is a nonlinear system with extremely complex structures and functions.The normal work of the brain based on conduction of neural information.Neural network is an important carrier of brain information transmission,and plays an important role in cognitive behaviors.The detection and transmission of information in neuronal networks is an important issue of neuroscience.The transmission of neural information is influenced by the connection of the brain,so understanding the brain connection is a physiological basis for a deep understanding of neural network information transmission.Therefore,this paper constructs a network model under the external stimulus,reveals how the inherent parameters of the neural network and external stimuli affect the detection and transmission;and based on the cortical neural network discharge time series data,analyzes the layered conduction characteristics of the cortical neural network,and characterizes the coding and transmission mechanisms of information in the cortical network.First of all,the detection and conduction laws of information in a single-layer neural network model are studied.Construct a single-layer neural network model,apply sinusoidal periodic stimulation of high and low frequencies to the network,and use resonance theory to characterize the detection and conduction of information.The vibration resonance of the network is sensitive to the period of the high frequency signal.The existence of an optimal high frequency signal makes the resonance phenomenon most significant.By analyzing the network parameters,it is found that the synaptic strength,especially the inhibitory synaptic coupling affects the detection and conduction of weak signals in the network.Second of all,the law of the intrinsic characteristics of the feedforward network is studied.A three-layer feedforward network is constructed using the Izhukevich model.A sinusoidal periodic input is applied to the network,and studies the effects of intrinsic state and external input current on information transmission.Studies have found that the ratio of inhibitory synaptic coupling and the proportion of inhibitory cluster discharge neurons can significantly promote layer-by-layer conduction of information in a multi-layer network.Finally,according to the experimental data,the cortical neural network connection is depicted to explore the conduction law of information.Granger causality methods were used to estimate functional connections between neurons at different time scales.According to the layered characteristics of the cortical network,the functional connection characteristics between the different layer are analyzed.The connections between neurons in different layers dominate.For different time scales,the directionality of the connections changes,and there is a significant connection reversal phenomenon.It indicates that the time scale will affect the information transmission in the cortical neural network to some extent.This paper studies the detection and transmission of information by neural networks.The influence of high frequency input and network structure on information transmission is analyzed and combined with the functional connection of cortical neural network.The conclusions obtained have some inspiration for understanding the coding and transmission of cortical network information.