Classification of muscle activity based on effort level during constant pace running

Area of Science:

• Biomechanics
• Exercise Physiology
• Signal Processing

Background:

• Running involves psychologic and physiologic changes affecting perceived effort, muscle properties, and movement strategies.
• Electromyographic (EMG) activity changes reflect alterations in muscle properties and movement strategies during running.
• Understanding these EMG changes can provide insights into running effort and fatigue.

Purpose of the Study:

• To test if EMG signals change systematically during a run.
• To determine if these changes correlate with the runner's effort level.
• To evaluate the effectiveness of wavelet-transformed EMG data for discriminating running effort phases.

Main Methods:

• Fifteen female recreational runners completed 1-hour treadmill runs at a constant speed.
• EMG signals were recorded from tibialis anterior, gastrocnemius medialis, vastus lateralis, and semitendinosus muscles.
• Wavelet transformed EMG data were analyzed using a support vector machine (SVM) to classify effort phases.

Main Results:

• Support vector machine (SVM) achieved >80% recognition rates for differentiating running effort stages across all muscles.
• Average recognition rates were: tibialis anterior (TA) - 89.2%, gastrocnemius medialis (GM) - 88.3%, vastus lateralis (VL) - 84.6%, and semitendinosus (ST) - 94.0%.
• The choice of SVM parameters (penalty C, cross-validation folds n) had minimal impact on the classification accuracy.

Conclusions:

• Lower extremity EMG signals, analyzed with wavelet methods, exhibit systematic differences between low and high effort stages of prolonged running.
• These findings suggest EMG signal analysis is a viable method for objectively assessing running effort.
• Wavelet-based EMG analysis combined with SVM offers a promising approach for non-invasive monitoring of running intensity and fatigue.