Assessment of Cognitive Engagement in Stroke Patients From Single-Trial EEG During Motor Rehabilitation

We propose a novel method for monitoring cognitive engagement in stroke patients during motor rehabilitation. Active engagement reflects implicit motivation and can enhance motor recovery. In this study, we used electroencephalography (EEG) to assess cognitive engagement in 11 chronic stroke patients while they executed active and passive motor tasks involving grasping and supination hand movements. We observed that the active motor task induced larger event-related desynchronization (ERD) than the passive task in the bilateral motor cortex and supplementary motor area (SMA). ERD differences between tasks were observed during both initial and post-movement periods (p <; 0.01). Additionally, differences in beta band activity were larger than differences in mu band activity (p <; 0.01). EEG data was used to help classify each trial as involving the active or passive motor task. Average classification accuracy was 80.7±0.1% for grasping movement and 82.8±0.1% for supination movement. Classification accuracy using a combination of movement and post-movement periods was higher than in other cases (p <; 0.05). Our results support using EEG to assess cognitive engagement in stroke patients during motor rehabilitation.