Myoelectric Control Based on A Generic Musculoskeletal Model: Towards A Multi-User Neural-Machine Interface

Area of Science:

• Biomedical Engineering
• Neuroscience
• Rehabilitation Engineering

Background:

• Electromyography (EMG)-based neural-machine interfaces (NMIs) are crucial for prosthetic control.
• Existing NMIs often require extensive user-specific calibration and customization.
• Predicting coordinated metacarpophalangeal (MCP) and wrist motion is essential for naturalistic hand function.

Purpose of the Study:

• To develop a user-generic EMG-based NMI for predicting coordinated MCP and wrist flexion/extension.
• To minimize calibration requirements for NMIs.
• To evaluate the performance of a generic musculoskeletal model against user-specific models.

Main Methods:

• Developed a user-generic musculoskeletal model using data from 6 able-bodied subjects and 9 upper limb postures.
• Implemented an EMG-based NMI centered on this generic model.
• Conducted on-line evaluations with able-bodied subjects and a transradial amputee performing virtual hand/wrist posture matching tasks.
• Compared the generic model's performance against a user-specific model.

Main Results:

• All subjects successfully used the user-generic NMI to perform virtual tasks.
• The generic model demonstrated comparable completion times to the specific model.
• The generic model resulted in fewer overshoots and improved path efficiency compared to the specific model.
• Able-bodied subjects outperformed the amputee subject in task performance.

Conclusions:

• A user-generic EMG-based NMI using a musculoskeletal model can predict coordinated MCP and wrist motion for multiple users, including amputees.
• This approach addresses the limitations of current NMIs that require lengthy customization.
• The developed NMI shows potential for future applications in powered prosthetic arms.