Comparison of sEMG processing methods during whole-body vibration exercise

Area of Science:

• Biomechanics
• Exercise Physiology
• Signal Processing

Background:

• Quantifying muscle activity using surface electromyography (sEMG) during whole-body vibration (WBV) exercises is challenging due to artifacts.
• Artifacts manifest as spikes in the sEMG signal at the vibration frequency and its harmonics, potentially compromising data accuracy.
• Standard sEMG processing methods may not adequately address these vibration-induced artifacts.

Purpose of the Study:

• To evaluate the impact of different sEMG processing methods on muscle activity quantification during WBV exercises.
• To compare the effectiveness of band-stop filters, band-pass filters, and spectral linear interpolation in artifact removal.
• To determine the optimal processing strategy for accurate sEMG analysis during WBV.

Main Methods:

• sEMG data were collected from participants performing squats with and without WBV.
• Artifacts were removed using three methods: band-stop filter, band-pass filter, and spectral linear interpolation.
• Intraclass correlation coefficients (ICC) and bias were analyzed to compare methods against unfiltered sEMG during no-vibration trials.

Main Results:

• Spectral linear interpolation demonstrated the highest ICC values, indicating excellent agreement with the reference measure.
• A band-stop filter also showed good agreement, though it introduced a greater bias at higher mean sEMG values.
• A band-pass filter performed poorly unless sEMG(RMS) was corrected for bias, after which its agreement improved significantly.

Conclusions:

• Spectral linear interpolation or a band-stop filter are recommended for artifact removal in sEMG during WBV.
• If using a band-stop filter, correcting sEMG(RMS) for systematic bias is advised to enhance performance.
• Accurate muscle activity quantification during WBV requires careful selection and application of sEMG processing techniques.