An Amplitude Of Low-Frequency Fluctuation And Support Vector Machine Analysis Potential Neuroimaging Biomarker Of Schizophrenia

Objective: Schizophrenia is a group of chronic and highly relapsing heavy psychiatric disorders whose etiology is currently unknown.Currently,clinicians’ subjective judgments of specific clinical symptoms are the main basis for distinguishing schizophrenia in patients with schizophrenia from healthy populations with schizophrenia.The lack of corresponding objective biomarkers for the diagnosis of schizophrenia to distinguish schizophrenic patients from healthy populations often results in misdiagnosis and underdiagnosis of patients.We aimed to explore the potential value of support vector machine methods combined with resting-state brain function amplitude of low-frequency fluctuation as a diagnosis of schizophrenia.Methods: One hundred and thirty-one patients with schizophrenia who were hospitalized at the Wuhan Mental Health Center from December 2020 to December 2021 were recruited,and 128 sex-and age-matched healthy controls were recruited openly at universities and communities in Hubei during the same period.Two trained psychiatric residents collected general information from participants,scored by the Positive vs.Negative Symptom Scale(PANSS scale)and the Repetitive Set Neuropsychological Status Rating Scale(RBANS scale),and completed resting-state functional magnetic resonance imaging(rs-f MRI)scans,and analyzed the results using amplitude of lowfrequency fluctuation and support vector machine methods.Differences in years of education,PANSS scale and RBANS scale results between patients with schizophrenia and healthy controls were compared using a two-sample t-test,while differences in gender distribution between these groups were tested by chi-square test using SPSS 23.0.p<0.05 was considered statistically significant.Differences between the two groups were determined by performing a voxel-by-voxel analysis of covariance on individual wholebrain amplitude of low-frequency fluctuation maps for these two groups.Results were thresholded at P<0.01.Results: Six brain regions were found to have significantly lower amplitude of lowfrequency fluctuation values in patients with schizophrenia compared to healthy controls:the left angular gyrus,the syrinx,the anterior cingulate cortex,the right cerebellum,the bilateral middle temporal gyrus,and the precuneus.No significantly increased values of amplitude of low-frequency fluctuation were found in brain regions of patients with schizophrenia compared to healthy controls.The results of the subsequent support vector machine analysis showed that the decrease in amplitude of low-frequency fluctuation in the precuneus bilaterally could effectively distinguish schizophrenic patients from healthy controls with 73.36% precision,91.60% sensitivity and 54.69% specificity,respectively.Conclusion: Alterations in the functional activity of local brain regions are present in the resting state of schizophrenic patients,and reduced amplitude of low-frequency fluctuation values in the precuneus bilaterally could potentially be a candidate biomarker for the diagnosis of schizophrenia.