A one-parameter neural activation to muscle activation model: estimating isometric joint moments from electromyograms

Area of Science:

• Biomechanics
• Muscle Physiology
• Computational Modeling

Background:

• Nonlinearities in the electromyography (EMG)-force relationship are known but often omitted in Hill-type muscle models.
• Existing EMG-driven models typically do not incorporate these physiological nonlinearities, particularly at low force levels.

Purpose of the Study:

• To present a novel one-parameter transformation model (A-model) that captures physiological nonlinearities in the EMG-force relationship.
• To integrate this A-model into EMG-driven Hill-type muscle models for improved accuracy.

Main Methods:

• Developed a one-parameter A-model based on phenomenological data for the curvilinear EMG-force relationship.
• Implemented the A-model within an EMG-driven Hill-type muscle model framework.
• Utilized optimization techniques to determine muscle-specific curvature parameters.

Main Results:

• The A-model successfully accounts for physiological nonlinearities observed at low force levels.
• Integration of the A-model into EMG-driven models significantly improved the fit of measured joint moments.
• Optimization allowed for subject-specific tuning of the muscle force-EMG relationship.

Conclusions:

• The proposed one-parameter A-model is a simple yet effective method to incorporate nonlinearities into EMG-driven muscle models.
• This approach enhances the predictive accuracy of muscle force and joint moment estimations.
• The A-model offers a valuable tool for more realistic biomechanical simulations.