Detecting Reach to Grasp Activities using Motion and Muscle Activation Data

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Human-Computer Interaction

Background:

• Activities of daily living (ADLs) recovery relies on motor function restoration post-stroke.
• Home-based rehabilitation necessitates effective methods for tracking limb and body motion.
• Quantifying upper extremity interaction with the environment, specifically reach-to-grasp (RTG) tasks, remains a challenge.

Purpose of the Study:

• To evaluate the Myo armband sensor as a noninvasive tool for monitoring RTG task performance.
• To assess the feasibility of using wearable sensor data for rehabilitation progress tracking.

Main Methods:

• Utilized a commercial Myo armband sensor to collect accelerometer and rate gyroscope data.
• Recorded muscle activation data during various RTG tasks.
• Developed a classifier integrating motion and muscle activation data to differentiate gestures.

Main Results:

• Significant variations in accelerometer and rate gyroscope data were observed across different RTG task types.
• The developed classifier achieved 93% accuracy in distinguishing between various gestures.
• Demonstrated the Myo armband's capability to capture relevant kinematic and muscle activation patterns.

Conclusions:

• The Myo armband sensor shows promise as a noninvasive method for monitoring upper extremity RTG performance.
• This technology can potentially enhance the assessment of motor recovery in home-based settings.
• Accurate gesture recognition has implications for personalized rehabilitation strategies.