Study Of The Significance Difference Of Brain Network In Patients With Depression Based On ICA

With the development of the times,people are under increasing pressure,and depression has become a hot issue in society.Depression has seriously affected people's lives,work,and study,and caused great distress to people.There are many technologies currently used in the auxiliary diagnosis of depression.EEG is one of the important technologies for studying depression.It has the advantages of high time resolution,low cost,simple operation and no invasive damage to the body of the subject.Loved by the majority of researchers.The combination of EEG data and graph theory to construct a complex network of the brain can avoid the influence of human subjective factors to a certain extent,and can maximize the use of the rich information provided by EEG data.Most of the traditional brain network construction methods use electrodes as nodes,and the volume conductor effect causes the correlation between the signals to deviate,which affects the analysis of the brain network topology.ICA can remove the mutual interference between EEG signals and make the processed signal components as independent as possible.Therefore,this study builds a complex brain network of patients with depression and healthy controls based on ICA,explores the significant differences between the two groups of ICA brain networks,and finds the significant difference nodes of the two groups of ICA brain networks,which is depression The study of the disease provides new ideas.In this study,we analyzed the ICA brain network of depression patients and healthy controls in response to the significant difference between patients with depression.First,pre-process the two groups of subjects,extract the independent components of the signal through ICA,and build the ICA brain network with the independent component as the node and the correlation between the component power spectral density as the connection strength;secondly,the seven types of the binary network Attribute analysis.Aiming at the problem that the seven attributes of the binary network are not significantly different,a new network attribute,information dimension,is proposed to analyze the significant difference between the depressed patients and the healthy control group;finally use A new fusion attribute finds the significant difference nodes in the ICA brain network,and initially locates the different brain regions of depression.The main research contents of this study are as follows:(1)Pre-process the original EEG data of patients with depression and healthy control group,use ICA to extract the independent components of the signal,use the micro-state method to segment the independent components,define the independent components as the nodes of the brain network,and the power spectral density of the components The correlation coefficient is used as the connection strength between the nodes,and the binary network and the weighted network of the two groups of subjects are constructed.(2)To analyze the 7 attributes of the binary network of the depressed patients and the healthy control group,in order to solve the problem that the 7attributes of the two groups of subjects were not significantly different under the binary network,this study proposed a new brain network attribute——Information dimension,analyze the ICA brain network from the perspective of fractal,calculate the significance of the difference between the network information dimension of the depressive patients and the healthy control group under the binary network and the weighted network,and calculate the classification accuracy rate.(3)Aiming at the problem of inconsistency of the significant difference nodes obtained by multiple single attributes,this study proposes a multi-attribute fusion method,which uses the analytic hierarchy process to fuse the selected four single attributes to obtain new fusion attributes.According to the fusion attributes,the ICA brain network Analyze multiple nodes of ICA to find the significant difference nodes of the ICA brain network,verify the significant difference nodes by calculating the invulnerability of the network,and use s LORETA to trace the ICA components corresponding to the significant difference nodes for preliminary positioning Get differences in depression brain regions.