Obstructive Sleep Apnea (OSA) remains underdiagnosed in 2022, as a result of the unawareness of its serious health-related consequences and the lack of diagnosis accessibility. Respiratory polygraphy (PV) is widely used as a screening tool and sometimes a diagnosis test, although polysomnography (PSG) remains the gold standard investigation as it provides complete information about sleep architecture and arousals. Thus, it has been shown that the Apnea Hypopnea Index (AHI) and Respiratory Disorder Index (RDI) are underestimated by PV vs PSG. Approaches to substitute PSG by simpler but equally efficient diagnosis tests have included devices aiming to record complementary signals and to analyze them with Artificial Intelligence. In this context, ASEEGA algorithm has demonstrated its performance for automatic sleep scoring in healthy individuals and patients with various sleep disorders, based on a single channel EEG analysis.

This study aims at comparing the real-life performance and feasibility of added single channel EEG automatic sleep scoring using ASEEGA to PV versus standard PV and PSG in adults referred to a regional sleep reference center for suspected OSA.

We hypothesize that this approach (1) is as accurate as PSG and more accurate that PV for AHI analysis, and (2) is less time-consuming than PSG.

Inclusion Criteria:

- Hospitalized at the Center for Sleep Medicine and Respiratory Diseases, Hôpital de la Croix-Rousse, Lyon Academic Hospital, Lyon between 09/2022 and 11/2022 for PSG following suspected OSA .

Exclusion Criteria:

• Unreliable sleep scoring due to artefacts (> 10% of the recording) or to epileptic abnormalities on the electroencephalography recordings
• Presence of other suspected sleep disorder: central hypersomnia, rapid-eye-movement (REM) or non-REM sleep parasomnia, sleep-related motor disorders...
• Ongoing psychotropic medication
• Opposition of the patient to the use of their data.

Patients Hospitalized at the Center for Sleep Medicine and Respiratory Diseases