Classification of schizophrenia patients vs healthy controls with dynamic functional network connectivity

Introduction: Assessing and understanding the functional differences in healthy vs. diseased brain and successful classification of d ifferent types of brains based on these functional differences are among the main goals of functional brain imaging. Different functional imaging outcomes have been previously used as input to classification algorithms in this manner. In this work, we use dynamic functional network connectivity (DFNC), time -windowed correlations between time-courses of different brain networks (components) estimated via spatial independent component analysis (sICA) from resting state fMRI data. The new approach identified meaningful inter-component linkages and allowed us to classify healthy and schizophrenic patients with reasonable success rates (70 -73% average).