Study On Characteristics Of EEG Microstates In First-episode Drug-naive Adolescents With Depression

BackgroundAs a common mental disease,depression seriously endangers people’s physical and mental health.At present,the diagnosis of depression mainly depends on the evaluation of doctor scale,and there is no quantitative standard.Therefore,finding objective quantitative criteria to help doctors evaluate depression is a research hotspot in academic circles.EEG microstate analysis is a powerful tool to study brain electrical activities.It can consider the electrical signals of the whole cerebral cortex at the same time.More and more studies show that there are significant abnormalities in EEG microstate parameters in patients with depression,but most of the subjects in these studies are patients with drugs,and the experimental results may be affected by drugs.In addition,there is no report on the abnormal changes of EEG microstate time series in patients with depression.ObjectiveBased on the microstate analysis algorithm of EEG,the temporal and spatial characteristics of EEG and abnormal changes of brain function in first-episode untreated adolescent depression were explored;Based on machine learning technology,the obtained EEG microstate features are tried to be used for early auxiliary diagnosis and classification of patients with depression.MethodsIn this study,34 first-episode untreated adolescent depression patients and 34 healthy controls were recruited to collect EEG data of all subjects in the resting state with their eyes closed.After preprocessing,the microstate analysis algorithm was used to study the EEG signals of the two groups of subjects,and four typical brain topographic topologies（A,B,C and D）were calculated.On this basis,three time parameters of the duration,occurrence and coverage of the four microstates and the nonlinear complexity of the microstate time series were extracted to study the temporal and spatial characteristics of EEG and abnormal changes of brain function in patients with depression.Finally,taking the obtained EEG microstate parameters as features,support vector machine（SVM）and k-nearest neighbor classifier（KNN）were used to classify patients with depression,so as to explore whether microstate features can be used as objective biomarkers for early diagnosis of depression.ResultsThe results showed that four typical EEG microstate topologies（A,B,C and D）were obtained in both groups,but compared with the healthy control group,the microstate C topology of patients with depression was significantly abnormal.In terms of EEG microstate parameters,the duration of microstate B,C and D in the patient group increased significantly,the occurrence and coverage of microstate B increased significantly,and the occurrence and coverage of microstate A and C decreased significantly in the depression group.In addition,the complexity analysis of the microstate time series of the two groups showed that the Samp En and LZC values of the patient group were significantly higher than those of the healthy group.Pearson correlation analysis showed that Samp En and LZC were significantly negatively correlated with HDRS score.When the EEG microstate features are combined with the complexity features of microstate sequence,the classification accuracy of 90.9% is obtained.ConclusionThe results show that patients with early depression have shown abnormal EEG microstate,which reflects the potential abnormality of the allocation of neural resources and the transformation between different brain network states.In addition,EEG microstate and the complexity of microstate sequence can be used as neurophysiological markers for early auxiliary diagnosis of depression.