EEG Correlates of Upper Limb Function During the Box and Block Test After Stroke

Stroke frequently results in upper limb motor impairment, making accurate and objective motor assessment essential for guiding rehabilitation. Traditional clinical tools such as the Box and Block Test (BBT) provide global performance scores but offer limited insight into the underlying motor and motor planning processes, thereby limiting their integration with modern neurorehabilitation technologies that increasingly rely on biosignals for real-time and quantitative feedback. Fifteen healthy individuals and eleven stroke survivors performed an EEG-adapted version of the BBT while electroencephalographic, kinematic, and behavioral data were recorded. Event-related desynchronization/synchronization (ERD/ERS) was computed in alpha, beta, and gamma bands during the three BBT phases: reaching, block handling, and coming back. Grand-average spectral analysis revealed a significant reduction of medial-frontal desynchronization during the block handling phase in stroke patients compared to healthy controls. Moreover, beta-band ERD over sensorimotor areas correlated positively with Fugl-Meyer Assessment scores. A phase-classification analysis based on ERD/ERS features achieved high discrimination performance in both groups, with macro-average AUC values up to 0.76 in the beta band and 0.86 in the low-gamma band, while classification performance was reduced when stroke patients performed the task with the affected limb. These results demonstrate the feasibility of phase-resolved EEG analysis during a complex functional task and support the use of EEG-derived markers for a more detailed characterization of post-stroke upper limb function.