EEG burst dynamics as an indicator of a progressive hypoxic state

Hypoxia disrupts perceptual and cognitive processes, posing serious risks for aircraft pilots. To investigate the neural mechanisms underlying these effects, we analyzed EEG subband power and burst dynamics in 27 participants reporting symptoms and performing cognitive tasks under progressive hypoxia. We hypothesized that specific burst features reflect declining oxygen availability and correlate with physiological, task performance, and symptom measures. Measurements were obtained during two conditions: normoxia, with O₂ maintained at 21%, and hypoxia, where O₂ was progressively decreased in four exposures (14.3%, 11.8%, 9.7%, and 8.1% O₂) of 5 min each after a 10-min normoxia baseline. EEG burst features-rate, duration, and cross-channel synchrony-were evaluated across overlapping subbands (1-7 Hz, 5-15 Hz, and 13-24 Hz) to identify neural signatures of hypoxic burden. Significant increases in burst features were observed in the 5-15 Hz and 13-24 Hz subbands starting at the second hypoxia exposure. In the 1-7-Hz subband, bursts emerged at the third exposure, coinciding with participant dropout. The strongest linear correlations of burst features with physiological, task performance, and symptom measures were found in the 5-15-Hz band: [Formula: see text] (-0.953), heart rate (0.963), task accuracy (-0.736), task completion time (0.653), and symptom rate (0.758). This study identifies novel EEG-based signatures of hypoxia, showing burst features increase with hypoxic exposure in a frequency- and time-dependent manner. These features strongly correlate with physiological decline and impaired cognitive performance. The findings may support real-time detection of hypoxia in aviation and other operational settings.<b>NEW & NOTEWORTHY</b> This study reveals how EEG burst activity changes with progressive hypoxia, offering insights into its neural correlates. Increased burst features in specific frequency bands correlate strongly with physiological, such as decreasing peripheral oxygen, task performance, such as declining accuracy, and symptom measures, such as increasing report rate. Our findings characterize a neural signature with potential use for real-time EEG monitoring to enhance hypoxia detection in aviation, potentially improving flight safety.