Objective:To investigate the feasibility of flight fatigue being detected via photoplethysmography (PPG) and regional cerebral oxygen saturation (rScO 2) in order to address the challenges posed by flight fatigue during prolonged or multiple consecutive flights. Methods:A total of 16 healthy male volunteers were enrolled. A wireless cerebral oximetry monitor headband was employed to collect PPG and rScO 2 data from the forehead while a multi-lead physiological data acquisition system was used concurrently to record three-lead electrocardiograms (ECGs). After 18 h of sleep deprivation, each volunteer performed a flight-simulating task, which was divided into 4 stages: the baseline period (T1), relaxation period (T2), early fatigue period (T3) and severe fatigue period (T4). Five-minute data was collected from each stage for analysis using AcqKnowledge 6.0. Heart rate (HR) and 3 HR variability (HRV) metrics, namely standard deviation of NN intervals (SDNN), root mean square of successive differences (RMSSD), and low frequency to high frequency power ratio (LF/HF), were computed independently from both ECG and PPG traces. The mean rScO 2 value for each stage was used to represent the cerebral oxygen saturation during that stage. The intra-class correlation coefficient (ICC) was employed to assess the consistency of the measurements, and the differences in HR and HRV indicators of the volunteers in the 4 stages of the experiment were analyzed. Results:The HR measured by ECG and PPG was highly consistent across the 4 stages among the 14 volunteers ( ICC=0.951, 0.963, 0.962, 0.963, P=0.013, 0.011, 0.021, 0.015), so were SDNN, RMSSD and LF/HF values ( ICC=0.935-0.983, all P<0.05). HR values calculated with either method showed significant differences across the 4 stages in the 14 volunteers ( F=21.63, 20.52, P=0.007, 0.008). HR gradually declined from T1 to T4, and was significantly lower at T4 than at T1 ( P=0.011, 0.009). There were significant differences in SDNN ( F=22.31, 24.26, P=0.006, 0.003), RMSSD ( F=22.30, 22.26, P=0.006, 0.006), and LF/HF ( F=20.37, 25.13, P=0.009, 0.002) across the 4 stages among the 14 volunteers. SDNN and RMSSD kept increasing as fatigue was intensified, while LF/HF decreased correspondingly. Statistically significant differences were found in SDNN, RMSSD and LF/HF values between T4 and T1 (all P<0.01). rScO 2 measured during the flight-simulating trial also differed significantly across the 4 stages ( F=21.39, P=0.007). rScO? at both T3 and T4 was significantly lower than at T1 ( P=0.009, 0.007). Conclusions:PPG can replace ECG for monitoring HR and HRV indicators under flight fatigue. Furthermore, the combination of PPG with rScO 2 monitoring allows for earlier detection of flight fatigue. This study is expected to offer a user-friendly and non-invasive approach to management of pilot fatigue.

Objective:To investigate the feasibility of flight fatigue being detected via photoplethysmography (PPG) and regional cerebral oxygen saturation (rScO 2) in order to address the challenges posed by flight fatigue during prolonged or multiple consecutive flights. Methods:A total of 16 healthy male volunteers were enrolled. A wireless cerebral oximetry monitor headband was employed to collect PPG and rScO 2 data from the forehead while a multi-lead physiological data acquisition system was used concurrently to record three-lead electrocardiograms (ECGs). After 18 h of sleep deprivation, each volunteer performed a flight-simulating task, which was divided into 4 stages: the baseline period (T1), relaxation period (T2), early fatigue period (T3) and severe fatigue period (T4). Five-minute data was collected from each stage for analysis using AcqKnowledge 6.0. Heart rate (HR) and 3 HR variability (HRV) metrics, namely standard deviation of NN intervals (SDNN), root mean square of successive differences (RMSSD), and low frequency to high frequency power ratio (LF/HF), were computed independently from both ECG and PPG traces. The mean rScO 2 value for each stage was used to represent the cerebral oxygen saturation during that stage. The intra-class correlation coefficient (ICC) was employed to assess the consistency of the measurements, and the differences in HR and HRV indicators of the volunteers in the 4 stages of the experiment were analyzed. Results:The HR measured by ECG and PPG was highly consistent across the 4 stages among the 14 volunteers ( ICC=0.951, 0.963, 0.962, 0.963, P=0.013, 0.011, 0.021, 0.015), so were SDNN, RMSSD and LF/HF values ( ICC=0.935-0.983, all P<0.05). HR values calculated with either method showed significant differences across the 4 stages in the 14 volunteers ( F=21.63, 20.52, P=0.007, 0.008). HR gradually declined from T1 to T4, and was significantly lower at T4 than at T1 ( P=0.011, 0.009). There were significant differences in SDNN ( F=22.31, 24.26, P=0.006, 0.003), RMSSD ( F=22.30, 22.26, P=0.006, 0.006), and LF/HF ( F=20.37, 25.13, P=0.009, 0.002) across the 4 stages among the 14 volunteers. SDNN and RMSSD kept increasing as fatigue was intensified, while LF/HF decreased correspondingly. Statistically significant differences were found in SDNN, RMSSD and LF/HF values between T4 and T1 (all P<0.01). rScO 2 measured during the flight-simulating trial also differed significantly across the 4 stages ( F=21.39, P=0.007). rScO? at both T3 and T4 was significantly lower than at T1 ( P=0.009, 0.007). Conclusions:PPG can replace ECG for monitoring HR and HRV indicators under flight fatigue. Furthermore, the combination of PPG with rScO 2 monitoring allows for earlier detection of flight fatigue. This study is expected to offer a user-friendly and non-invasive approach to management of pilot fatigue.