Validation of a New 2-Channel Portable Sleep Staging Device: Quantitative Analysis of Electroencephalography, Electromyography, and Electrooculography

Abstract

Abstract Sleep scores obtained using a 2-channel portable telemetry polysomnographic (PSG) system, named SleepGraph, demonstrated an almost perfect correlation when compared to standard laboratory PSG. The components of sleep-stage dependent electroencephalographic (EEG), electromyographic (EMG), and electrooculographic (EOG) results were strikingly consistent between the SleepGraph and the standard PSG. The distinct EEG components were elucidated using the EEG power spectrum. What sets the SleepGraph apart is its innovative approach to EMG and EOG derivation from the lower eyelid to the opposite chin. This unique configuration not only reduces the number of required facial electrodes but also enhances the sleep stage specificity of both EMG and EOG measurements. By using digital band-pass filters, the EMG and EOG records were distinctly separated. The specificity of the EEG power in the δ, α, spindle, and β ranges, along with the EMG integral and the EOG variance, proves to be particularly strong, suggesting their potential as reliable parameters for the development of artificial intelligence-based sleep diagnostic algorithms.