Integrating EEG Source Localization and Resnet CNN for Advanced Stroke Rehabilitation

Upper limb motor impairments resulting from stroke greatly limit daily activities and diminish quality of life, making effective rehabilitation essential. This study addresses this issue by classifying motor tasks using EEG data from acute stroke patients, focusing on left-hand motor imagery, righthand motor imagery, and rest states. By using advanced source localization techniques combined with customized ResNetCNN architecture, we achieve superior spatial pattern recognition in EEG data. Our methodology results in classification accuracies of 91.03%, 89.07%, and 87.17, compared to 55.57% % to 72.2% achieved by traditional sensor domain methods. These findings demonstrate the potential of our approach to enhance braincomputer interfaces (BCIs) for more effective and personalized neurorehabilitation, ultimately improving recovery outcomes and quality of life for stroke patients. The results underscore the importance of advanced EEG classification techniques in providing clinicians with precise tools for developing individualized therapy plans, potentially leading to significant improvements in motor function recovery and patient outcomes.