Related Experiment Videos
Methods to reduce the variability of EMG power spectrum estimates
R V Baratta1, M Solomonow, B H Zhou
1Louisiana State University Medical Center, Department of Orthopaedic Surgery, New Orleans 70112, USA. rbarat@lsumc.edu
Summary
This study introduces three methods to reduce variability in electromyography (EMG) power density spectrum (PDS) by removing noise. These techniques improve median frequency (MF) estimation, especially during low-level contractions.
Area of Science:
- Biomedical Engineering
- Neuroscience
- Physiology
Background:
- Electromyography (EMG) power density spectrum (PDS) analysis is crucial for understanding muscle activity.
- Variability in PDS can arise from artifactual noise, complicating accurate analysis, particularly at low contraction levels.
Purpose of the Study:
- To describe and validate three methods for reducing EMG PDS variability by eliminating artifactual components.
- To enhance the accuracy of median frequency (MF) estimation from EMG signals.
Main Methods:
- Subtraction of power line noise in the time domain.
- Subtraction of system noise in the frequency domain.
- Employing techniques to train subjects for true maximal voluntary contraction (MVC) production.
Main Results:
- The described methods effectively reduce artifactual noise (power line and system noise) from EMG recordings across the full force range (0-100% MVC).
- Accurate estimation of median frequency (MF) was improved, especially during low-level contractions (0-25% MVC) with poor signal-to-noise ratios.
- Subject training for true MVC resulted in up to 30% higher force/torque output, significantly impacting PDS variable interpretation.
Conclusions:
- Noise reduction techniques preserve native EMG power while improving signal fidelity.
- Accurate MVC determination is critical for reliable interpretation of EMG PDS variables related to muscle contraction processes.