Analysis of Bipolar Disorder Using fMRI

Abstract In this work, we have analysed bipolar disorder using fMRI based on brain regional activity measurements. In this research work, we have located the eights regions as per the literature review which are strongly affected brain regions because of BPD. For those regions, the below operations are carried out. Initially functional points are identified using independent component analysis and their connectivity has been established using correlation coefficients. Then located the activated points using Hierarchical Modular Analysis based on the strength of the interregional connectivity. Followed by constructing network between the activated points of each brain regions. Property of network parameters like centrality page rank and centrality degree, centrality closeness, assortativity and clustering coefficients are extracted. Finally, adaboost classifier consist of three weak classifiers, KNN, GDA and naïve bayes are experimented. It was found that this work had given 94.2% accuracy comparatively better than earlier research works.