The High Order Correlation Maximum Entropy Model Of Neuronal Populations Encoding And Direction Distinction

Neuronal population encoding and direction distinction is the basic issue of neuroscience research.However,the high-order correlation between neurons has a great influence on the neuronal population encoding and the direction of distinction,which can not be neglected.The maximum entropy model can effectively describe the existing data,find out the probability distribution,and quantify the effect of higher-order correlation.In this paper,the maximum entropy model is used to explore the direction distinction in brain science.At the neuron level,the higher-order correlations between external stimuli and population response(spike activity)affect neuronal population activity theoretical and practical analysis.The first chapter of this article mainly introduces the research background,the principle of maximum entropy,the knowledge of neuron clusters and related models.In the second chapter,we design the related visual judgment test to study the judgment of whether a subject is parallel to a group of lines in different environmental modes(line number,color,width and length).According to the principle of maximum entropy,the Generalized Iterative Scaling algorithm(GIS)is used to establish the maximum entropy model,and the corresponding probability distribution is obtained.At the same time,the performance of the model is quantified by the change of entropy.The correlation between environmental conditions was found to have a greater impact on human brain.In the third chapter,we consider the influence of time and use the maximum entropy model which contains the spatio-temporal relationship.We use the Galuber model to generate the simulation data,compare the model probability with the data.We use J-S divergence and entropy ratio to quantify the extent of its impact on visual judgment,indicating that time correlation has a certain impact,can not be ignored.In the fourth chapter,we study how to distinguish some similar external stimuli from the neuron level.Based on the Exponential integrate-and-fire model,a neuronal population activity(firing)is generated using a noisy current as the stimulus,and the maximum entropy model is used to characterize and quantify the distinguishing effect of different stimuli.In the fifth chapter,we summarize the whole thesis.Furthermore,we put forward the questions that need to be improved and the related research prospects.