Non-linear and spectral EEG dynamics as stage-specific biomarkers of neuropsychological recovery post-stroke

Resting-state electroencephalography (EEG) offers promising window into the neuroplastic changes underlying recovery after stroke. This longitudinal study aimed to characterize the stage-specific dynamics of non-linear (fractal dimension, FD) and linear (power spectral density, PSD) EEG metrics and to evaluate their associations with recovery of neuropsychological functions.

We recruited 44 patients with left-hemisphere ischemic stroke at various recovery stages and 25 age-matched healthy controls. All participants underwent high-density EEG recording at two time points (~5-7 weeks apart). The EEG was analyzed for FD, PSD across standard frequency bands.

A rehabilitation-related increase in FD in the right hemisphere was observed in all patient groups except those in the early recovery stage, where a unique decrease in left-hemisphere FD occurred. This right-hemispheric FD increase correlated with improvements in motor control function. Spectral changes were also stage-specific: a decrease in frontal beta power (14-20 Hz) was linked to longer inter-session intervals, while a right posterior increase in slow-wave activity (2-6 Hz) correlated with gains in auditory-verbal memory. Conversely, a reduction in alpha power (7-9 Hz) was associated with improved visuospatial ability and motor functions, but only in early and intermediate recovery stages. Recovery after stroke is subserved by dynamic, phase-specific neurophysiological processes reflected in the FD and spectral composition of spontaneous brain activity. The right hemisphere appears to play a crucial compensatory role, with FD serving as a sensitive biomarker of motor planning recovery.

These findings underscore the necessity of stage-specific neurorehabilitation framework, informed by objective EEG biomarkers, to individualize and enhance therapeutic efficacy.