Resting State Functional Connectivity Analysis of Multiple Sclerosis and Neuromyelitis Optica Using Graph Theory

Abstract

The aim of this study is to investigate resting state functional connectivity in brain tissue using fMRI data in order to differentiating Neuromyelitis Optica (NMO) from Multiple Sclerosis (MS). In this method, normalized mutual information accompanied with graph theoretical analysis and community structure measures were used to differentiating. The fMRI time series were extracted for 264 nodes in which were selected based on neurological principals. Normalized mutual information (NMI) between each two time series was calculated and constructed NMI based connectivity matrix. The graph of pairwise mutual information is thresholded such that the top 3471 un-directed links are preserved. The graphs are analyzed graph theoretic measures and community structure procedure was applied in order to modularity detection. There are no significant differences between NMO and MS patients in terms of classical graph theoretic measures. However a significant difference was found in modularity statistics between NMO and MS patients.