EEG Microstate Analysis of Brain Dynamics During Motor Imagery in Stroke Patients

Stroke is a chronic neuropsychiatric disorder with complex neuropathological features that often leads to varying degrees of consciousness impairment and limb paralysis. Understanding the dynamic changes in brain activity during motor imagery (MI) of paralyzed limbs in stroke patients is essential for exploring the neurocognitive mechanisms underlying motor function impairment. This study employed Electroencephalography (EEG) microstate analysis to characterize dynamic brain activity in stroke patients during four upper limb MI tasks: forward arm raising, forward arm lowering, lateral arm raising, and lateral arm lowering. Unlike conventional EEG analysis, microstate analysis provides millisecond-level spatiotemporal insights into neural activity. A modified K-means clustering algorithm was used to identify four representative microstate classes (A, B, C, and D). Microstate features including occurrence rate, duration, coverage, and global explained variance (GEV) were calculated to assess rapid transitions in brain states during task execution. Differences between movement directions were more pronounced than those between raising and lowering, with lateral tasks exhibiting higher microstate duration and GEV compared to forward tasks, particularly in microstate A. Compared to lowering movements, raising movements yielded higher values in microstate duration and GEV, while occurrence rate and coverage showed variability across different microstate classes. This study demonstrates the potential of microstate features as biomarkers for understanding individual differences in motor impairment in stroke patients and may contribute to more effective, personalized rehabilitation strategies.