Application of Graph Theory Features for the Objective Diagnosis of Depressive Patients with or without Anxiety: an Rs-fMRI Study

Purposes: To probe abnormality that may lead to anxiety in depressive patients. Procedures: This study investigated the graph theory features ahead of machine learning feature selection procedure. Classification methods were applied afterwards. Methods: Graph theory, statistical analysis and forward sequential feature se lection were combined to find features. SVM classifier was also involved. Results: 1 global and 22 local features were found correlated with clinical anxiety factor. Conclusions: Anxiety is correlated with emotion and cognitive loop and other regions.