Gray Matter Network Anomalies In Schizophrenia Based On MRI Data

Schizophrenia is a kind of basic personality change,the existence of thinking,emotion,behavior split disease,its etiology and pathogenesis,treatment and prevention,has been the central topic of psychiatry research.According to previous studies,schizophrenia is always associated with abnormal brain connections,so schizophrenia is generally considered to be a mental illness caused by abnormal brain connections.In recent years,the development of complex network theory based on graph theory has provided necessary tools and analytical methods for the study of human brain connectome.A large number of studies have shown that both structural and functional brain networks exhibit some characteristics of complex networks,such as modularity,network efficiency,centrality,and small-world attributes.In addition,the analysis of anatomical connectivity of the brain by clustering method shows that some areas of the cerebral cortex form modules with specific functions,which are closely connected within the modules and separated between the modules.In schizophrenia,changes in the structure and function of some specific modules will affect the patients.The development of high-resolution brain imaging technology has made it possible to study human brain activities and diseases.With the update and development of medical imaging technology,structural magnetic resonance imaging has become popular due to its obvious advantages such as easy access,high signal-to-noise ratio and relatively insensitive artifacts.Existing research at the group level to construct the schizophrenic brain network structure,and focus on a single area of the brain research schizophrenia brain network anomalies,but these methods may ignore the influence of individual differences,limited the normal individual variation in study,especially to identify single-subject brain structure changes or abnormal morphology related applications.That's why it's important to build a single-subject gray matter network in order to observe subtle changes in the brain.Therefore,using graph theory to explore the changes of network topology and specific modules of single-subject gray matter network in schizophrenic population,and applying the differences extracted from the whole brain network and modules to the diagnosis of schizophrenic population,will provide a new idea for the research of auxiliary diagnosis of schizophrenia.However,the changes of single-subject gray matter networks with schizophrenia are still unknown,and the differences between voxel-based and region-based brain structural connections need to be further explored.By constructing and analyzing the single-subject gray matter network of schizophrenia,this paper will discuss the abnormal condition of the structural network of schizophrenia patients,and provide a new thinking direction for the auxiliary diagnosis of schizophrenia.The main research contents and results are as follows,(1)Schizophrenia based on single-subject gray matter networks in brain regions.The brain was divided into 90 nodes by AAL template,and the probability density function of gray matter volume of each node was calculated by kernel density estimation,and the single-subject gray matter network was constructed by the KL dispersion value of each distribution function.Applying it to the study of the structural brain networks in schizophrenia,we found that the abnormal topology of the single-subject gray matter networks in the global topological properties,modules and nodes of schizophrenia patients showed a decrease in the global efficiency of the network,as well as a decrease in the connectivity of the frontal parietal region and between the frontal parietal network and the visual network,and the attributes were related to the severity of symptoms.(2)Study on single-subject gray matter networks based on voxels in schizophrenia.Through the brain is divided into 6mm*6mm*6mm cubes,calculate the Pearson correlation of cubes to build single-subject gray network,using the GPU parallel computing network properties,found that the grey matter of schizophrenia brain networks in the global,modules,and abnormal node topology,performance in the global efficiency increase,enhanced module connection,brain region showed that schizophrenia covariant consistency,brain interval increase influence each other.(3)Voxel-based improvement of single-subject gray matter networks in schizophrenia.The above voxel-based construction method was standardized to AAL template and nodes were unified to analyze the individual gray matter network of schizophrenia.The results shortened the running time and improved the computing efficiency,further proving that the individual gray matter network of schizophrenia patients was abnormal.(4)Combining single-subject gray matter network with machine learning to build a diagnostic model for schizophrenia.In order to better serve the results of network analysis for the early auxiliary diagnosis of schizophrenia,the network attributes with significant differences between groups were extracted,the feature space was constructed,the classifier was used to train the classification model,and the diagnostic ability of single-subject gray matter network was tested.The results showed that the single-subject gray matter network method had a high accuracy of 85.6% in disease classification,which provided a certain methodological basis for early diagnosis of schizophrenia.