Novel Methods for Surface EMG Analysis and Exploration Based on Multi-Modal Gaussian Mixture Models.
Anna Magdalena Vögele1, Rebeka R Zsoldos2, Björn Krüger3
1Multimedia, Simulation and Virtual Reality Group, Institute of Computer Science II, University of Bonn, Bonn, Germany.
Plos One
|July 1, 2016
Summary
This study presents a new Gaussian mixture model (GMM) method for analyzing animal muscle activation during locomotion using surface EMG (sEMG) data. The approach isolates muscle activation patterns and enables hierarchical clustering for deeper insights into locomotion biomechanics.
Area of Science:
- Biomechanics
- Animal Locomotion Analysis
- Data Science in Biology
Background:
- Analyzing muscle activation during locomotion is crucial for understanding animal movement.
- Surface electromyography (sEMG) provides valuable data but requires sophisticated analysis techniques.
- Existing methods may not fully capture the complexity of muscle activation patterns throughout locomotion cycles.
Purpose of the Study:
- Introduce a novel data analysis method for animal muscle activation during locomotion.
- Develop a technique to isolate and analyze specific muscle activation components within sEMG data.
- Explore the application of Gaussian mixture models (GMMs) for dissecting complex sEMG signals.
Main Methods:
- Fitting Gaussian mixture models (GMMs) to surface EMG (sEMG) data.
- Utilizing the resulting Gaussian modes as building blocks for hierarchical clustering.
- Applying the method to sEMG data from 14 horses during walk and trot.
Main Results:
- Successfully identified composite peak models representing general muscle activation patterns per sensor location.
- Demonstrated the method's applicability in isolating distinct muscle activation phases during locomotion.
- Validated the approach for analyzing muscle activation in horses across different gaits.
Conclusions:
- The GMM-based method offers a powerful tool for detailed analysis of muscle activation during locomotion.
- This approach enhances the exploration of sEMG data by providing interpretable components.
- The identified composite peak models offer insights into the biomechanics of equine locomotion.
