Combining static and dynamic brain network analysis with machine learning for enhanced diagnosis of major depressive disorder.

Major Depressive Disorder (MDD) is a common mental disorder that severely impacts patients' quality of life and social functioning. Early diagnosis is crucial for improving treatment outcomes, so rapid and accurate diagnosis of MDD is of significant importance. This paper combines static and dynamic functional connectivity analyses based on fMRI data and proposes a novel feature selection method to identify and classify brain network abnormalities in MDD patients. The advantage of this method lies in enhancing classification accuracy through feature fusion while simultaneously reducing feature dimensionality. First, whole-brain correlation analysis is performed based on fMRI functional connectivity, followed by Rich Club analysis and sliding window methods to investigate the topological properties of the intrinsic functional brain network in MDD patients. Finally, the abnormal brain network is used as a feature to classify and diagnose MDD patients and healthy controls, achieving a classification accuracy of 90.28 %. This result validates that the identified abnormal brain networks in this study have clinical significance for assisting in the diagnosis of MDD.