Construction Of Functional Network And Turning Behavior Decoding Of Pigeons Based On Spike Trains

The brain network is one of the most complex networks in nature.Hundreds of millions of neurons in the brain constitute a large and complex brain structure network.The connection among different neurons,neuron clusters or brain regions in the time scale on dynamic synchronization forms a function network.From the perspective of network to analyze the neural information processing mechanism is one of the hot spots on the neuroscience,control science and other interdisciplinary research.At present,the study of brain function network mainly analyzes the functional connection between different brain regions through large-scale information acquisition techniques such as electroencephalogram,magnetoencephalography and magnetic resonance imaging.While the research based on microelectrode array to study the neuron-based unit's valuable role in explorating the brain mechanism is relatively few,and many problems still need to be clarified.In this paper,the construction method of neural functional network is studied,and the topology of the neural network is analyzed.The neural network is used to decode the movement of the pigeon.It is proved that the neural network is efficient in the analysis of brain information processing mechanism.The main research contents and achievements are as follows:1)Using the normalized mutual information algorithm to measure the functional connection among the spike trains,and construct the neural functional network.The performance of the network construction algorithm is verified by introducing the Izhikevich neuron release model.The results show that the connection between the neural functional network based on the mutual information algorithm and the real network is highly consistent,and its performance is superior to the traditional correlation coefficient algorithm.2)The topology characteristics of three typical networks,such as random network,regular network and small world network,are analyzed.On the basis of this,we study the influence of the spike rate and neuron number on network topologicalcharacteristics.The results show that the overall trend of the network topology shows a parabolic shape with the increasing of the spike rate;With the increasing of the number of neurons,the network topology characteristics show a decreasing trend as a whole.3)Using the feature of neural functional network and the support vector machine decoding algorithm to decode the movement of pigeons in the goal-oriented task of cross maze,and compared with the feature decoding algorithm based on spike trains.The results show that the neural functional network has better decoding performance,whether it is global efficiency or clustering coefficient,the decoding rate is better than that of neuron cluster spike rate.