Aviation fatigue detection based on cardiorespiratory coupling: A study under normoxic and mild hypoxic conditions

• Cardiorespiratory Coupling (CRC) as a Novel Biomarker for Aviation Fatigue • Investigating dynamic variation patterns of CRC characteristics during the progression of aviation fatigue • Constructing a feature-fused fatigue detection model to enhance detection accuracy • Model validation under aviation hypoxia conditions With rising long-duration flights, pilot fatigue increasingly threatens flight safety and efficiency. While the electrocardiogram (ECG) provides a stable and portable solution for aviation fatigue monitoring, reliable biomarkers beyond heart rate variability (HRV) remain scarce. This study investigates the dynamic changes of cardiorespiratory coupling (CRC) during fatigue progression and evaluates its potential for fatigue detection, including its performance under mild hypoxic conditions. We recruited eight healthy male participants (age: 20.6 ± 1.8 years) for two separate 6-hour simulated flight experiments under normoxic and mild hypoxic (equivalent to 3000 m altitude) conditions, with continuous ECG recorded for CRC analysis. Results demonstrated that the CRC power was significantly higher in fatigued states compared to alert states. As fatigue deepened, notable changes were observed in CRC power, power distribution across frequency bands, spectral peak locations, and spectral centroids, with CRC high-frequency components exhibiting the strongest correlation with fatigue. Moreover, the Granger causal coupling of the respiration sequence on the RR interval sequence decreased significantly with increasing fatigue. A fatigue detection model integrating CRC with HRV features outperformed the model using HRV features alone, improving the random forest classifier's training accuracy by 8.4% and testing accuracy by 4.4% under normoxic condition. This study introduces CRC to aviation fatigue research, reveals its dynamic changes with fatigue, and provides a novel technical pathway for fatigue monitoring in aviation and related fields.