Task discrimination for non-weight-bearing movements using muscle synergies

Area of Science:

• Biomedical Engineering
• Rehabilitation Robotics
• Neuroprosthetics

Background:

• Effective myoelectric control of lower limb prostheses relies on accurately distinguishing muscle activation patterns.
• Current methods often struggle with the nuanced muscle synergies underlying diverse movements.

Purpose of the Study:

• To present a novel method utilizing muscle synergies for discriminating non-weight-bearing lower limb movements.
• To evaluate the performance of this method against established machine learning algorithms.

Main Methods:

• Non-negative matrix factorization (NMF) was employed to identify muscle synergies from electromyography (EMG) data recorded from eight thigh muscle sites.
• Non-negative least squares (NNLS) was used for estimating EMG patterns.
• Classification accuracy was assessed for various movements including knee and ankle joint actions.

Main Results:

• The muscle synergy-based method demonstrated effective discrimination of non-weight-bearing lower limb movements.
• Classification accuracy was higher for knee joint movements compared to ankle joint movements.
• The proposed algorithm achieved performance comparable to Linear Discriminant Analysis (LDA) in offline analysis.

Conclusions:

• Muscle synergies offer a viable approach for decoding complex muscle activation patterns in lower limb prostheses.
• The developed method shows promise for improving the intuitive control of prosthetic devices.
• Further research may explore real-time implementation and integration into advanced prosthetic systems.