A Per-sample Digitized Algorithm for Automatically Detecting Apnea and Hypopnea Events from Airflow and Oximetry

Area of Science:

• Sleep Medicine
• Biomedical Engineering
• Signal Processing

Background:

• Sleep apnea significantly impacts quality of life, necessitating accurate and early diagnosis.
• Manual diagnosis of sleep apnea is prone to variability in scoring criteria and inter-scorer differences.
• Automated detection methods can enhance diagnostic reliability and consistency.

Purpose of the Study:

• To develop and validate a novel automated algorithm for detecting apnea and hypopnea events.
• To assess the algorithm's performance against manual scoring based on established guidelines.
• To explore the applicability of the algorithm in portable sleep monitoring devices.

Main Methods:

• Utilized airflow and pulse oximetry signals from 30 Sleep Heart Health Study polysomnography records.
• Developed a per-sample digitized algorithm analyzing peak signal excursion from airflow for apnea/hypopnea detection.
• Filtered hypopneas using a ≥3% oxygen desaturation threshold from pulse oximetry data.

Main Results:

• The automated algorithm demonstrated high performance in detecting apneas and hypopneas.
• Achieved a sensitivity of 98.1% and a positive predictive value of 95.3% for event detection.
• The algorithm's accuracy was validated against manual scoring by a trained sleep physiologist.

Conclusions:

• The developed automated algorithm provides accurate detection of sleep apnea events.
• This novel method overcomes limitations associated with manual sleep apnea diagnosis.
• The algorithm is suitable for integration into portable sleep monitoring devices for widespread clinical use.