A Deep Learning Approach for Mental Fatigue State Assessment.

Area of Science:

• Sports Science
• Biomedical Engineering
• Artificial Intelligence

Background:

• Traditional methods for detecting mental fatigue in sports often rely on heart rate variability (HRV) analysis.
• Existing deep learning models like CNNs and LSTMs have shown potential but can be improved for complex physiological signal analysis.

Purpose of the Study:

• To develop and validate a hybrid deep neural network model for accurate mental fatigue detection in sports.
• To compare the proposed model's performance against traditional machine learning and other deep learning techniques.

Main Methods:

• A hybrid deep neural network integrating Residual Networks (ResNet) and Bidirectional Long Short-Term Memory (Bi-LSTM) was employed for feature extraction.
• A transformer architecture was utilized for feature fusion.
• The model was trained and tested using original ECG data, 2D spectral characteristics, and physiological information.

Main Results:

• The proposed hybrid deep learning model achieved a high accuracy of 95.29% in identifying mental fatigue.
• The model significantly outperformed conventional methods like Support Vector Machines (SVMs) and Random Forests (RFs).
• Experimental results also showed superiority over other deep learning models such as Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM).

Conclusions:

• The study presents a highly accurate and effective deep learning-based method for recognizing mental fatigue from physiological signals.
• This approach offers a promising tool for enhancing sports performance and physical fitness training.
• The findings suggest a new direction for fatigue monitoring beyond traditional HRV analysis.