RNN-based Mental Fatigue Prediction using EEG Data in High-Stress Environments

Mental fatigue significantly affects cognitive function and safety in high-stress settings, including aviation, healthcare, and industrial activities. Traditional fatigue detection approaches, including self-reports and performance assessments, often lack objectivity and are inadequate for real-time monitoring. This research introduces an innovative method for forecasting mental fatigue using Electroencephalogram (EEG) data and Recurrent Neural Networks (RNNs). EEG signals provide significant insights into cerebral function, allowing continuous and objective monitoring of mental states. RNNs, recognized for their ability to simulate sequential dependencies, are used to analyze temporal patterns in EEG data and predict fatigue levels. The proposed model analyzes diverse EEG properties, such as alpha, beta, and theta waves, to elucidate the dynamics of mental fatigue over time. The method is assessed using authentic EEG recordings from individuals under high-stress conditions, demonstrating its efficacy in accurately predicting mental fatigue. The findings suggest that RNN-based models can serve as a reliable tool for real-time fatigue diagnosis, thereby enhancing safety and performance in critical settings. This methodology facilitates the development of advanced fatigue management systems in critical occupations.