Repeated automatic sleep scoring based on ear-EEG is a valuable alternative to manually scored polysomnography

Area of Science:

• Sleep Science
• Biomedical Engineering
• Neuroscience

Background:

• Polysomnography (PSG) is the gold standard for sleep quantification but is obtrusive and resource-intensive.
• Emerging technologies offer less invasive sleep monitoring alternatives, yet require rigorous clinical validation.
• Ear-EEG presents a promising, unobtrusive method for sleep stage assessment.

Purpose of the Study:

• To clinically validate the ear-EEG method against concurrent polysomnography (PSG) for sleep staging.
• To assess the accuracy and precision of automatic ear-EEG scoring for various sleep metrics.
• To compare the reliability of ear-EEG derived sleep metrics with PSG over multiple nights.

Main Methods:

• Twenty healthy subjects underwent four nights of simultaneous ear-EEG and PSG recordings.
• Two independent technicians manually scored PSG data, while an automated algorithm scored ear-EEG data.
• Analysis focused on sleep stages and eight key sleep metrics, including Total Sleep Time (TST), Sleep Efficiency, and REM latency.

Main Results:

• Automatic ear-EEG scoring demonstrated high accuracy and precision for Total Sleep Time, Sleep Onset Latency, Sleep Efficiency, and Wake After Sleep Onset compared to manual PSG scoring.
• REM latency and REM fraction showed high accuracy but low precision.
• The automated system systematically overestimated N2 sleep fraction and underestimated N3 sleep fraction.

Conclusions:

• Ear-EEG provides a clinically validated, less obtrusive alternative to PSG for sleep staging.
• Automatic ear-EEG scoring offers reliable sleep metric estimation, particularly advantageous for multi-night recordings.
• Ear-EEG is a viable and potentially superior alternative to PSG for both single and extended sleep monitoring.