MD-LMS Algorithm Based Brain Functional Connectivity Analysis in Resting State fMRI

This paper proposes a new method to analyze resting state fMRI images for brain functional connectivity extraction. The proposed method regards the brain areas as nodes and the brain functional connectivity as a diffusion network, and then applies multitask diffusion LMS (MD-LMS) algorithm and hierarchical clustering to extract the valid networks. MD-LMS algorithm is a signal analysis method in a diffusion network. The proposed method was applied to three healthy subjects, and extracted multiple brain functional connectivity networks. The proposal method examined the brain hierarchical functional structure, and extracted temporal change of the brain functional connectivity. The novelty of the method is that the proposed method can extract transition of brain functional connectivity networks.