Night-time brain inter-hemispheric asynchrony in sleep apnea patients carry information on neuropsychological impairment

Obstructive sleep apnea (OSA) is a serious sleep disorder with diurnal symptoms including neuropsychological impairments such as excessive daytime sleepiness and loss of attention. There are no efficient tools to measure these impairments in current clinical practice. In this paper, we explore the feasibility of measuring neuropsychological impairments using electroencephalography (EEG) data acquired during the standard clinical sleep diagnostic test known as polysomnography (PSG). We hypothesized that left-right hemispheric EEG asynchrony could quantitatively characterize neuropsychological impairment in OSA in a population of sleep laboratory patients. We acquired EEG data from 50 subjects undergoing routine PSG, using symmetric electrode derivations of C4-A2 and C3-A1. Their neuropsychological performance was assessed via a psychomotor vigilance task (PVT). We computed the left-right EEG asynchrony and developed a logistic regression model (LRM) to classify patients according to their PVT performance. Leave-one-out cross-validation studies on a LRM model with two-class PVT performance achieved a sensitivity of 83% (95% CI: 66-100%) and a specificity of 78% (95% CI: 64-92%). These results indicate that EEG asynchrony during sleep carries information on daytime neuropsychological impairments in OSA subjects.