Using centrality measures to extract core pattern of brain dynamics during the resting state

Abstract

The patterns of brain dynamics were studied during resting state on a macroscopic scale for control subjects and multiple sclerosis patients. Macroscopic brain dynamics is defined after successive coarse-grainings and selection of significant patterns and transitions based on Markov representation of brain activity. The resulting networks show that control dynamics is merely organized according to a single principal pattern whereas patients dynamics depict more variable patterns. Centrality measures are used to extract core dynamical pattern in brain dynamics and classification technique allow to define MS dynamics with relevant error rate.

Keywords: Brain dynamics; Centrality measures; Coarse-graining; Discrete dynamics; Entropy rate; Macroscopic transition graph; Multiple sclerosis; Resting state; kNN classifier.