HD-EMG Electrode Count and Feature Selection Influence on Pattern-based Movement Classification Accuracy.

Area of Science:

• Biomedical Engineering
• Neuroscience
• Robotics

Background:

• Electromyography (EMG) pattern classification is crucial for prosthetic and robotic control.
• Higher electrode counts in EMG systems generally correlate with improved accuracy.

Purpose of the Study:

• To determine the optimal number of high-density EMG (HD-EMG) channels for accurate myoelectric control.
• To evaluate the impact of electrode count and feature selection on classifying hand and forearm movements.

Main Methods:

• Recorded HD-EMG signals from volunteers performing various movements.
• Extracted time-domain features from selected EMG channels.
• Trained Support Vector Machine (SVM) classifiers with varying numbers of input channels.

Main Results:

• Classification accuracy significantly improved with increased electrode counts.
• Diminishing returns observed, with marginal accuracy gains beyond 100 electrodes.
• Feature selection and electrode count impact overall classification performance for 17 distinct movements.

Conclusions:

• HD-EMG offers enhanced myoelectric control accuracy.
• An optimal range of approximately 100 electrodes exists for balancing accuracy and system complexity.
• This research informs the design of more efficient and effective prosthetic and robotic control systems.