Measuring Sensorimotor Rhythms During Active and Resistive Upper-Limb Movement Execution

Sensorimotor rhythms (SMR) are widely employed in brain–machine interface applications, although their reliable detection during overt upper-limb movements remains challenging because of motion artifacts and concurrent peripheral activity. This study investigated whether SMR modulation can be decoded during elbow flexion and extension performed under active and resistive conditions resembling robot-assisted rehabilitation, with and without voluntary modulation. Nine healthy participants executed four experimental conditions while electroencephalography (EEG), electromyography (EMG), and inertial measurement unit (IMU) signals were synchronously acquired. EEG data were analyzed using a filter bank common spatial pattern (FBCSP)-based classification pipeline, whereas EMG signals were used to estimate exerted force. Subjective workload was evaluated through the NASA-TLX questionnaire. Classification accuracies generally exceeded the theoretical chance level, reaching average values above 73% ± 14% and above 80% ± 12% when discriminating voluntary SMR modulation independently of movement type. Only a few individual cases were at or below the theoretical chance level. EMG analyses confirmed higher force production during resistive tasks, while event-related desynchronization/synchronization analyses revealed distinct cortical modulations in the $\mu$ and $\beta$ bands, indicating the presence of task-related cortical modulation underlying the decoded features. These findings demonstrate the feasibility of detecting SMR modulation during overt upper-limb movements and support the potential of the proposed framework for future adaptive robot-assisted rehabilitation systems, where movement resistance could be modulated according to real-time neural activity.