Evaluation of an Adaptive Neurotechnology Based on Motor Imagery in Patients Who Suffered a Left Hemisphere Stroke

This randomized clinical trial will evaluate the feasibility and performance of a co-adaptive motor imagery brain-computer interface (BCI) for upper-limb rehabilitation in adults with subacute left-hemisphere stroke. The study will investigate whether continuous neurofeedback improves BCI control and learning compared with binary feedback based on correct or incorrect task performance. Participants will be randomly assigned to a neurofeedback group or a group receiving binary feedback. The intervention consists of BCI calibration followed by 10 real-time BCI training sessions. EEG and surface EMG signals will be recorded throughout the BCI sessions. BCI performance will be assessed using decoding accuracy and the algorithmic support provided by the Supportive Backward Adaptation (SBA) method, together with motor and neurophysiological measures.