MyoGestic: EMG interfacing framework for decoding multiple spared motor dimensions in individuals with neural lesions

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Technology

Background:

• Restoring motor function after neurological damage (spinal cord injuries, strokes) or limb loss (amputations) remains a significant clinical challenge.
• Surface electromyography (EMG) can detect voluntary motor commands from spared motor neurons, even without visible muscle activation.
• Existing EMG-based control systems often require extensive calibration and lack adaptability.

Purpose of the Study:

• To develop and validate a novel wireless, high-density EMG system (MyoGestic) for intuitive myoelectric control.
• To enable rapid, real-time decoding of motor intent from individuals with diverse neurological conditions and limb loss.
• To facilitate a participant-centered approach for developing adaptive myocontrol algorithms.

Main Methods:

• Development of a wireless, high-density EMG bracelet and a user-friendly software framework (MyoGestic).
• Implementation of machine learning algorithms for rapid adaptation and real-time decoding of motor intent.
• Real-time decoding of motor intent from participants with traumatic spinal cord injury (SCI), spinal stroke, and amputations.

Main Results:

• Successful real-time decoding of multiple motor dimensions from all participants within minutes.
• Demonstrated control of a digital hand, orthosis, prosthesis, and 2D cursor using decoded EMG signals.
• Achieved rapid adaptation of machine learning models tailored to individual user needs.

Conclusions:

• MyoGestic provides an intuitive and adaptive platform for EMG-based control, significantly reducing setup time.
• The system effectively translates neural signals into functional control for assistive devices, applicable to SCI, stroke, and amputation.
• A participant-centered design approach enhances the development of effective and personalized myocontrol solutions.