High-Efficiency Graph Measures in Discriminating Schizophrenia Patients from Healthy Controls Using Structural and Functional Connectivity

The human brain is a complex network characterized by structural and functional connections within and between regions. It is proven that in some brain diseases (such as schizophrenia), these connections are disrupted. Graph theory is a suitable tool to analyze brain network disruptions. The data analyzed in this study include the functional and structural connections of 27 schizophrenic patients and 27 healthy controls. By means of Independent Component Analysis (ICA) and joint ICA (jICA) algorithms, the brain network is divided into 10 sub-networks, in which structural connections play an important role. We look for graph parameters that distinguish patients from healthy people. We observed that the local parameters of joint subnetwork #3 such as clustering efficiency, strength, and local efficiency can be used to distinguish patients from controls. The global parameters in joint subnetwork #1, such as characteristic path length, radius, diameter, and max modularity have the same effect. To investigate nervous system disorders such as schizophrenia it is better to analyze the joint sub-networks of the brain than the whole brain.