Tracking motor improvement at the subtask level during robot-aided neurorehabilitation of stroke patients

Robot-aided neurorehabilitation can provide intensive, repetitious training to improve upper-limb function after stroke. To be more effective, motor therapy ought to be progressive and continuously challenge the patient's ability. Current robotic systems have limited customization capability and require a physiotherapist to assess progress and adapt therapy accordingly.

The authors aimed to track motor improvement during robot-assistive training and test a tool to more automatically adjust training.

A total of 18 participants with chronic stroke were trained using a multicomponent reaching task assisted by a shoulder-elbow robotic assist. The time course of motor gains was assessed for each subtask of the practiced exercise. A statistical algorithm was then tested on simulated data to validate its ability to track improvement and subsequently applied to the recorded data to determine its performance compared with a therapist.

Patients' recovery of motor function exhibited a time course dependent on the particular component of the executed task, suggesting that differential training on a subtask level is needed to continuously challenge the neuromuscular system and boost recovery. The proposed algorithm was tested on simulated data and was proven to track overall patient's progress during rehabilitation.

Tuning of the training program at the subtask level may accelerate the process of motor relearning. The algorithm proposed to adjust task difficulty opens new possibilities to automatically customize robotic-assistive training.