Brain Network Feature Extraction And Analysis Of First Episode Schizophrenic Patients Based On Microstate

Brain science is an important research direction in today’s world,and countries around the world have launched brain science research programs.Research on neurological and psychiatric diseases in brain science mainly includes Huntington’s disease,Parkinson’s disease,depression and schizophrenia.Brain networks are a research hotspot in this field,and with the development of graph theory,brain network research provides a basis for understanding the state of the brain.However,at present,the brain network has problems such as weak spatial explanatory and neglect of the relationship between channels in schizophrenia.In view of the above problems,this paper studies the brain function network of EEG signaling in schizophrenia based on microstate method and Douglas peucker algorithm.Specific contents include:(1)Based on graph theory,the EEG brain network and feature analysis of first-time fine patients were constructedEEG data from 103 first episode of schizophrenia patients and 92 healthy controls were preprocessed using Douglas peucker’s algorithm,and brain networks were constructed by selecting scalp electrodes as nodes,Phase Locking Value(PLV)as a functional connectivity index and selecting appropriate thresholds.The global and local attribute values of the two networks were statistically analysed to select significant features for de-classification and to explore the difference frequency bands,attributes and brain regions.We found better results for local attribute classification.The local efficiency of frequency bands,node degree and node median,frequency band node degree and clustering coefficient,frequency band clustering coefficient and local efficiency were all differential attributes.The area under the curve(AUC)of clustering coefficients was lower in both patient groups than in the control group.Patients had weaker information processing capacity in the frontal and temporal lobes.(2)Analysis of brain network characteristics changes in patients with first fine separation after eight weeks of treatmentA brain network was constructed for 23 patients who had been treated for eight weeks and their characteristics were analyzed,and the differential attributes were analyzed in relation to clinical information.The results showed that the local efficiency of the patients in the frequency band was improved,and the feature path length and clustering coefficient were reduced.There was no significant improvement in the left frontal and occipital lobes,and the local efficiency and clustering coefficients of the parietal and temporal lobes were improved.The difference attribute was significantly correlated with scale scores,but not with drug dose.(3)Based on the microstate analysis method,the characteristics of the first fine patients were analyzedGlobal field power was calculated for both EEG data sets,and the topology at the peak of the global field power sequence was obtained using a modified K-means clustering algorithm to obtain microstate categories,which were then fitted to the EEG data.Microstate parameters were calculated and statistically tested to obtain difference characteristics.The results showed significant differences between the topographies of the two groups of microstates B and D.The duration and temporal coverage of microstates in category A was significantly lower in the patient group compared to the control group,and the duration and temporal coverage of category C was significantly higher compared to the control group.There was a significant decrease in the number of occurrences of category A and B in patients.Finally,comparing the classification results of the complex network approach and the microstate approach,we found that the average classification effects of the network features,microstate features and combined features were 77.5%,80.5% and 87.2%,respectively.