A novel channel selection method for multiple motion classification using high-density electromyography.

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Signal Processing

Background:

• Effective myoelectric control requires optimal surface electromyography (EMG) channel selection and electrode placement.
• Existing channel selection methods (SFS, FMS) are dependent on specific EMG features or classification algorithms, leading to variability.
• A novel method is needed for robust EMG channel selection in pattern recognition-based myoelectric control systems.

Purpose of the Study:

• To propose and evaluate a new Multi-Class Common Spatial Pattern (MCCSP) method for surface EMG channel selection.
• To assess the MCCSP method's performance independent of specific EMG features and classification algorithms.
• To compare the MCCSP method against existing channel selection techniques (SFS, FMS).

Main Methods:

• High-density EMG recordings were obtained from twelve mildly-impaired traumatic brain injury (TBI) patients.
• The proposed MCCSP method was used to select a subset of optimal EMG channels.
• The selected channels were utilized for motion classification using pattern recognition, with variations in electrode configurations, feature sets, and classifiers.

Main Results:

• The MCCSP method demonstrated superior motion classification performance compared to SFS and FMS.
• A consistent combination of selected EMG channels was achieved using the MCCSP method.
• The study validated MCCSP's effectiveness across different electrode configurations, features, and classifiers.

Conclusions:

• The MCCSP method presents a practical approach for surface EMG channel selection in myoelectric control.
• This method facilitates the development of more effective myoelectric control systems for rehabilitation applications.
• MCCSP can enhance active rehabilitation for TBI patients and improve prosthetic limb control for amputees.