Novel pseudo-wavelet function for MMG signal extraction during dynamic fatiguing contractions.

Area of Science:

• Biomechanics
• Signal Processing
• Sports Science

Background:

• Muscle fatigue assessment is crucial in sports performance and injury prevention.
• Accurate detection of muscle fatigue aids in optimizing training and preventing overexertion.
• Mechanomyography (MMG) offers a non-invasive method for monitoring muscle status.

Purpose of the Study:

• To develop and validate a genetic algorithm-based approach for classifying muscle fatigue.
• To optimize the detection of muscle fatigue using a novel pseudo-wavelet function.
• To enhance the accuracy of muscle fatigue classification in sports-related dynamic contractions.

Main Methods:

• Recorded Mechanomyography (MMG) signals from the biceps muscle during dynamic contractions to fatigue.
• Employed a genetic algorithm to evolve a pseudo-wavelet function for fatigue detection.
• Trained and tested the algorithm on labeled MMG data (Non-Fatigue and Fatigue) using a cross-validation approach.

Main Results:

• The evolved pseudo-wavelet function demonstrated superior performance in classifying muscle fatigue compared to standard wavelet functions.
• Achieved an average correct classification rate of 80.63% for muscle fatigue detection.
• Showed statistically significant improvements in classification rates, ranging from 4.70 to 16.61 percentage points (p < 0.05).

Conclusions:

• The developed genetic algorithm-evolved pseudo-wavelet function is effective for accurate muscle fatigue classification.
• This method offers a promising advancement for real-time monitoring of muscle fatigue in sports.
• The findings support the utility of advanced signal processing techniques in sports science applications.