Motor improvement estimation and task adaptation for personalized robot-aided therapy: a feasibility study.

Area of Science:

• Robotics in Medicine
• Rehabilitation Engineering
• Neurorehabilitation

Background:

• Robotic systems are increasingly used in upper limb rehabilitation.
• Current clinical studies have not confirmed superior efficacy of robotic therapy over conventional methods.
• Personalizing robot-aided therapy to individual motor deficits is key to improving outcomes.

Purpose of the Study:

• To present a model-based approach for personalizing robot-aided rehabilitation therapy within training sessions.
• To automatically adapt rehabilitation training based on continuous estimation of motor improvement.

Main Methods:

• Combined robot-recorded motor performance measures to estimate patient motor improvement.
• Developed a personalization routine for adaptive rehabilitation training.
• Tested the approach using an upper-limb exoskeleton with healthy subjects and subacute stroke patients.

Main Results:

• The model accurately estimated motor improvement in fast and slow adapting healthy subjects.
• Individualized training protocols were proposed based on adaptation speed.
• Stroke patients retained motor improvements after personalized training sessions.

Conclusions:

• The model-based approach is feasible for personalizing robot-aided rehabilitation therapy.
• Pilot tests with stroke patients show encouraging results for clinical applicability.
• Automated training adaptation is timely and tailored to individual patient abilities.