eva‎luation of functional networks in MS patients using fMRI data based on working memory

3/14/2024 12:00:00 AM

MS is an autoimmune disease in which the myelin cell sheath of the nerve cells is damaged in the brain and spinal cord. More than half of MS patients suffer from cognitive disorders such as working memory disorder. Working memory plays an important role in temporary storage and information processing and updating them. For eva‎luating working memory in MS patient we used n-back task. The human brain is a sophisticated network of interaction areas, one of the ways to eva‎luate this complex network is graph theory. The purpose of this study is to investigate the functional networks in MS patients using graph theory and to separate the MS and control groups. To construct functional brain networks, we employed an automated anatomical labelling (AAL) atlas with 116 regions and the connectivity matrix was obtained using Pearson correlation. After extracting local features including clustering coefficient, page rank centrality, degree, eigenvector centrality, k-coreness centrality and betweenness centrality, the Wilcoxon rank sum test was used to select discriminative and significant features, and eventually the data were classified using SVM, decision tree and KNN classifier. The results demonstrate that the features extracted from the 3-back network, which is a network with a high cognitive load, performed better than other networks. It seems that the areas related to the cerebellum in the 1-back network and the orbitofrontal region in the 3-back network were more significant. The features of degree, k-coreness centrality, eigenvector centrality extracted from 3-back network and betweenness centrality feature extracted from 1-back network were able to distinguish two groups with 95.5% accuracy. 3-back network has performed better than other networks. It seems that with the increase of cognitive load, a better separation between the two groups occurred, although the 1-back network performed better than the 2-back network. It is possible that the effectiveness of working memory capacity was different in different cognitive loads and MS patients probably had more disorders in 2-back connections.

حافظه كاري , اف ام آر آي , ام اس , تئوري گراف , ارتباط عملكردي

working memory , fMRI , Multiple sclerosis , Graph theory , functional connectivity

Sareh Yousefi

Mohammad Reza Daliri