Quantitative evaluation of motor function recovery process in chronic stroke patients during myoelectric controlling robot-assisted elbow training

Objective: To investigate the changes in electrophysiological, kinetic and kinematical parameters during myoelectric controlling robot-assisted elbow training in stroke patients, and to provide a more comprehensive and quantitative evaluation method. Method: Eight subjects with chronic upper extremity paresis after stroke attended a 20-session elbow training using a myoelectric controlling robot. EMG of biceps and triceps, elbow torques and angle signals were recorded synchronously during the experiment. Result: After the 20-session training, there were statistically significant improvements in Ashworth scale and Fugl-Meyer scale for elbow (P<0.05). After training in maximal voluntary isometric contraction (MVC) experiment, elbow flexion, and extension MVC torques increased significantly (P<0.01). The moment torque-EMG ratio of triceps increased significantly (P<0.05). Root mean square error (RMSE) between target angle and motion angle also decreased significantly (P<0.05). Conclusion: The myoelectric controlling robot-assisted elbow training could improve joint moment torques, muscle efficiency, and motion accuracy for stroke patients. These parameters could quantitatively reflect motor function of stroke patients from different aspects, and possesed the potential value in applied in clinical evaluation of motor function.