Muscle Force Estimation Based on Neural Drive Information From Individual Motor Units

Area of Science:

• Biomedical Engineering
• Neuroscience
• Kinesiology

Background:

• Muscle contraction force estimation commonly uses macroscopic surface electromyography (SEMG) features.
• Microscopic neural drive information from individual motor unit (MU) activities remains underexplored for force estimation.

Purpose of the Study:

• To propose and validate a novel method for muscle force estimation using microscopic neural drive information.
• To leverage individual MU action potential waveforms and firing sequences for enhanced force prediction accuracy.

Main Methods:

• Decomposition of high-density SEMG (HD-SEMG) to isolate individual MU activities.
• Supervised machine learning to determine MU twitch force from action potential waveforms.
• Physiologically meaningful muscle force model integrating individual MU contributions.

Main Results:

• The proposed method achieved a root mean square difference (RMSD) of 8.3% ± 2.8% in force estimation.
• Significantly outperformed four common force estimation methods (RMSD: 11.7%–20%).
• Demonstrated effectiveness in estimating muscle force from HD-SEMG data in healthy subjects.

Conclusions:

• The novel method effectively utilizes neural drive information for precise muscle force estimation.
• Offers a valuable tool for applications in precise motor control, sports science, and rehabilitation medicine.
• Highlights the potential of microscopic neural analysis for advancing biomechanical research.