Research On The Evolution Method Of Brain Map Network In Depressed Patients

With the development of science and technology,the era of the Internet has risen rapidly,and more and more affairs are connected with the establishment of the network.For example,the research results of brain diseases have changed from the original imaging only theory to the joint analysis of imaging reports and brain network evaluation indicators.With the acceleration of urban life rhythm,the incidence of depression is increasing year by year,and the incidence of the disease is spreading to young people on campus,and the pathogenesis is not clear.Therefore,this paper builds an evolutionary model based on brain network data of depressed patients,studies the pathogenesis of depressed patients and simulates the process of brain network pathology.Create a memory graph convolution evolution network(MGCEN)based on graph convolutional neural network(GCN)and Long Short-Term Memory(LSTM)to predict the future development trend of the network.In view of the fact that the onset of depression is mainly caused by changes in brain structure over time,this paper builds a brain network model based on structural magnetic resonance images and diffusion tensor images.First,this paper divides the brain MRI image into 116 brain regions according to the AAL template,and takes 90 brain regions as the research object,and analyzes the GA value of the gray matter volume of each brain region.It was found that the difference between the brain regions of patients with depression and normal people is mainly reflected in the frontal and temporal regions.Then use the brain area as a node to build a brain network with the correlation between each brain area as an edge,and analyze the differences in network topology such as clustering coefficient,degree distribution,and local efficiency in the brain network of depressed patients and healthy people.Afterwards,in order to study the network evolution process,this paper builds a depression brain network evolution model.The model uses the healthy human brain network as the initial network,regards the brain network of depressed patients as the target network,and combines the QUATRE(QUasi-Affifine TRansformation Evolutionary)algorithm to obtain the optimal hyperparameter solution vector of the network evolution formula.Through computer simulation experiments,it is found that the evolutionary network in the steady state of the network has the same topological properties as the brain network of the real depressed patients.The change of network topology in the depressive brain network evolution model is determined by the evolutionary random formula in the model.Artificially established evolution formulas have deviations in the real evolution direction of the network.In order to seek the evolution characteristics of the network,this paper creates a direction graph convolution model based on GCN,introduces the evolution direction vector into the model,and builds a two-layer structure to obtain the second-order neighbor information of the nodes in the graph network.The generated evolution network is close to the real brain network structure of depressed patients in network structure.In order to further predict the future development trend of network evolution,this article has established an MGCEN model based on LSTM and GCN.The model has the characteristics of extracting time series information and graph structure information in the network at the same time.Compared with traditional network prediction models,more attention is paid to the combination of network timing information and structural information,which can accurately predict the network state at the next moment.