Mental Fatigue Estimation Using EEG in a Vigilance Task and Resting States.

Area of Science:

• Neuroscience and Brain-Computer Interfaces (BCI)
• Cognitive Psychology and Psychophysiology

Background:

• Prolonged mental exertion leads to fatigue, impairing performance and increasing accident risks.
• Electroencephalogram (EEG) monitoring is emerging as a key BCI tool for assessing mental fatigue.

Purpose of the Study:

• To utilize EEG signals for quantifying mental fatigue levels by predicting reaction time (RT) in a psychomotor vigilance task (PVT).
• To evaluate the feasibility and robustness of EEG-based fatigue detection.

Main Methods:

• Conducted a 36-hour sleep deprivation experiment with 18 subjects, recording EEG data every four hours.
• Collected EEG data during a 6-minute PVT and two 3-minute resting states (eyes closed/open).
• Employed multiple linear regression (MLR) with selected EEG features to estimate RT, validated using leave-one-subject-out (LOSO).

Main Results:

• Mean RT increased significantly over the 36-hour period, correlating with rising fatigue.
• EEG data from the PVT task achieved a high mean correlation coefficient of 0.81 ± 0.16 for RT prediction.
• Resting-state EEG showed lower correlations (eyes-closed: 0.65 ± 0.20; eyes-open: 0.50 ± 0.30).

Conclusions:

• EEG-based monitoring of mental fatigue is feasible and robust, particularly during active tasks like PVT.
• This method holds significant potential for real-time fatigue assessment in operational environments.
• Further research may explore optimizing resting-state EEG analysis for fatigue detection.