Quantifying Neural Processing Changes During An Auditory Oddball Task Under Hypoxia Using Electroencephalography

To investigate neural phenotypes specific to hypoxia, we quantify the capability of electroencephalography (EEG) dynamics to determine if an individual is in a normoxic or hypoxic state during a passive auditory task and identify EEG-based features that provide maximal discriminability. Two distinct modeling approaches are explored, each intended to capture contrasting aspects of the relationship between neural activity and the hypoxic physiological state. With an analysis that exploits Auditory Deviance Response landmarks to characterize cognitive-neurophysiological changes during hypoxia, we achieve a receiver operating characteristic area under the curve (AUC) of 0.76 using only two fronto-temporal channels, on par with larger EEG channel set benchmarks. With a novel EEG correlation structure approach that examines finer neural coordination patterns across a spatially diverse channel set, we achieve an AUC of 0.85. Our methods of EEG analysis may find applications in understanding and detecting environmental impact on neural health.