Gender dependant snore sound based multi feature obstructive sleep apnea screening method.

Area of Science:

• Sleep Medicine
• Biomedical Engineering
• Signal Processing

Background:

• Obstructive Sleep Apnea (OSA) is a prevalent sleep disorder often undiagnosed due to a lack of convenient screening methods.
• Snoring is a primary symptom of OSA, caused by upper airway vibrations, and may exhibit gender-specific characteristics.
• Current diagnostic methods like polysomnography are resource-intensive, necessitating simpler screening tools.

Purpose of the Study:

• To develop and validate a gender-specific, snore-based screening technique for Obstructive Sleep Apnea.
• To investigate if incorporating gender differences in snore sound analysis improves OSA detection performance.
• To establish the feasibility of a low-cost, non-contact, and unattended OSA screening method.

Main Methods:

• A multi-parametric Obstructive Sleep Apnea screening technique analyzing snore sounds was developed, incorporating gender-specific features.
• Logistic regression models were employed to classify subjects into Obstructive Sleep Apnea/non-Obstructive Sleep Apnea categories.
• K-fold cross-validation was performed on gender-specific datasets (51 males, 36 females) recorded over 6-8 hours, with results compared against polysomnography.

Main Results:

• The gender-specific snore analysis achieved high diagnostic performance: 93±9% sensitivity and 89±7% specificity for females.
• For males, the technique demonstrated 91±8% sensitivity and 89±12% specificity.
• These results indicate a significant improvement by accounting for gender-specific variations in snore signals for OSA detection.

Conclusions:

• A gender-specific, snore-based approach is effective for screening Obstructive Sleep Apnea.
• The proposed technique offers a promising foundation for developing affordable, user-friendly, non-contact OSA screening devices.
• This method could significantly improve OSA diagnosis rates by enabling widespread community screening.