Longitudinal Case Study of Regression-Based Hand Prosthesis Control in Daily Life

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Neuroprosthetics

Background:

• Conventional prosthetic hands often use limited electromyographic (EMG) signals for single-function control.
• Current advanced controllers offer more functions but still in a sequential manner.
• Regression-based mapping of EMG signals offers potential for simultaneous, independent control of prosthetic functions.

Purpose of the Study:

• To evaluate the real-world, daily-life performance of a regression-based EMG control system for hand prostheses.
• To compare the novel control method against a participant's existing conventional prosthesis over a two-month period.

Main Methods:

• A two-month case study involving a participant with a transradial amputation.
• Functional tests and questionnaires were administered at the beginning and end of the study.
• Performance was compared between the regression-based control and the participant's conventional prosthesis.

Main Results:

• The participant using regression-based control outperformed their conventional prosthesis in two of three functional metrics after just one day of training.
• No model retraining was necessary throughout the study.
• Daily use led to further performance improvements, surpassing conventional control in all functional metrics.

Conclusions:

• Regression-based prosthesis control demonstrates high fidelity and is transferable from laboratory settings to daily life.
• This advanced control method offers a significant improvement over conventional sequential control for prosthetic hands.
• The findings support the clinical viability of regression-based EMG control for enhanced prosthetic function.