A remote quantitative Fugl-Meyer assessment framework for stroke patients based on wearable sensor networks

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Wearable Sensor Networks

Background:

• Traditional Fugl-Meyer Assessment (FMA) is time-consuming and complex for remote monitoring.
• Stroke patients require continuous assessment and training in non-clinical settings.
• Wearable sensors offer a potential solution for remote patient data collection.

Purpose of the Study:

• To develop a novel remote quantitative Fugl-Meyer Assessment (FMA) framework for stroke patients.
• To enable accurate monitoring of upper limb, wrist, and finger movement function using wearable sensors.
• To facilitate remote rehabilitation training and evaluation in both clinical and home settings.

Main Methods:

• Utilized two accelerometers and seven flex sensors to capture upper limb movement data.
• Developed an extreme learning machine-based ensemble regression model to map sensor data to FMA scores.
• Applied the RRelief algorithm for optimal feature selection and designed seven training exercises to simplify FMA.

Main Results:

• The proposed quantitative FMA model accurately predicted FMA scores from wearable sensor data, achieving a coefficient of determination of 0.917.
• Experiments conducted in both clinical and home settings validated the model's precision.
• The framework demonstrated effectiveness in quantifying rehabilitation progress.

Conclusions:

• The developed wearable sensor framework provides a reliable method for remote quantitative FMA.
• This approach offers a potential solution for continuous, objective assessment and training of stroke survivors.
• The system supports extended use of wearable sensor networks for stroke patient management outside clinical environments.