EEG-based brain network analysis in stroke patients during a motor execution task

Abstract

Post-stroke survivors often suffer motor function disorders, which are usually associated with anatomical and functional alterations of brain network. Previous EEG-based brain network analyses mainly focused on stroke-linked brain network in resting state and single aspect (globally or regionally), leaving the pattern of functional connectivity (FC) in stroke patients during specific motion task uncovered yet. In this study, we investigated stroke specific FC patterns in patients who suffered unilateral hemispheric stroke during a motor execution task. Partial correlation coefficients between multiple electroencephalography (EEG) channels were computed to construct the functional networks for healthy controls and stroke patients. The graph-based analysis was then performed to characterize specific FC patterns in stroke patients. Results suggested that brain networks were characterized in stroke patients by lower global efficiency and clustering coefficient in alpha and beta band, compared to healthy controls. Regionally, stroke patients exhibited weaker local connection in motor area of affected hemisphere during motor execution, which may explain their motor deficits. The findings of our study may offer new insight to study the neural plasticity and brain reorganization after stroke.