Simplified Optimal Estimation of Time-Varying Electromyogram Standard Deviation (EMGσ): Evaluation on Two Datasets.

Area of Science:

• Biomedical Engineering
• Neuroscience
• Kinesiology

Background:

• Electromyography (EMG) signal whitening is crucial for accurate EMG-force relationship analysis.
• Understanding the impact of various whitening techniques, sampling rates, and noise correction is essential for broader EMG applications.

Purpose of the Study:

• To evaluate four EMG signal whitening procedures.
• To assess the influence of sampling rate and noise correction on EMG-force accuracy.
• To identify optimal methods for EMG signal processing in force-varying contractions.

Main Methods:

• Studied four whitening procedures, noise correction (root difference of squares - RDS), and sampling rates (1024, 2048, 4096 Hz).
• Analyzed EMG-force relationships in 64 subjects during elbow contractions (0-50% MVC) with force-varying and constant-force tasks.
• Compared EMG-force error across different whitening methods and sampling rates.

Main Results:

• Noise correction (RDS) reduced EMG noise by 5-10x.
• Higher sampling rates (4096 Hz) showed small but significant improvements in accuracy (~3-4%) over lower rates.
• Whitening significantly reduced EMG-force error (from 5.55% to ~4.7-4.9% MVC), with subject-specific filters performing slightly better than universal or simpler methods.

Conclusions:

• EMG signal whitening significantly improves EMG-force relationship accuracy.
• Noise correction and appropriate sampling rates are vital for reliable EMG data.
• The first difference whitening method offers a practical, calibration-free option for force-varying contractions.