Obstructive sleep apnea (OSA) affects an estimated 2-4% of middle–aged adults yet we are still exploring how best to delineate the neurophysiological deficits that accompany this disorder. Untreated OSA leads to an increased risk of motor vehicle accidents. Traditional polysomnographic (PSG) metrics do not consistently correlate with daytime functioning. There is a clinical need for simple biomarkers to identify individuals susceptible to OSA-related cognitive deficits.
There is a close relationship between EEG-based changes in brain activity and daytime function in healthy sleepers. No studies have explored quantitative EEG (qEEG) biomarkers during baseline sleep and resting wakefulness (baseline) as correlates of waking neurobehavioural performance during extended wakefulness in OSA.
The aims of the thesis were 1) to identify qEEG biomarkers of neurobehavioural dysfunction and sleepiness in OSA and controls during 40-hours (h) of extended wakefulness, and 2) to develop and validate automated EEG artefact processing methods for subsequent qEEG analysis of waking and sleep EEG. EEG biomarkers were derived using conventional power spectral analysis and a novel qEEG analysis technique, detrended fluctuation analysis (DFA).
This study showed that wake qEEG markers significantly correlated with impaired performance and increased sleepiness across 40-h of extended wakefulness in both groups. Baseline waking measures of the DFA scaling exponent, but not power spectra, were associated with impaired simulated driving after 24-h awake in OSA. Furthermore, OSA patients with greater EEG slowing during REM sleep showed a marked decline in performance after 24-h awake. These key findings highlight the potential utility of qEEG analysis to yield simple biomarkers of neurobehavioural impairment and sleepiness.
Automated EEG artefact processing methods for resting awake and PSG recordings were developed and validated against a reference-standard of manual artefact recognition as part of this study. These proven artefact processing methods will be pivotal for exploring qEEG biomarkers in future studies.