A Comprehensive Review: Computational Models for Obstructive Sleep Apnea Detection in Biomedical Applications

Area of Science:

• Sleep Medicine
• Biomedical Engineering
• Computer-Aided Diagnosis

Background:

• Obstructive sleep apnea (OSA) is a sleep disorder with serious health consequences, including hypertension and heart failure.
• Traditional diagnosis via polysomnography (PSG) is cumbersome and uncomfortable for patients.
• Computer-aided diagnosis (CAD) offers a promising avenue for efficient and automated OSA detection.

Purpose of the Study:

• To provide a comprehensive overview of computer-aided diagnosis approaches for obstructive sleep apnea.
• To survey recent advancements in screening and detection methods for OSA events.
• To identify current research challenges and gaps in OSA diagnosis.

Main Methods:

• Literature review of sleep apnea research from the past decade.
• Analysis of various screening approaches for OSA identification.
• Examination of preprocessing, feature extraction, selection, and classification techniques in CAD for OSA.

Main Results:

• The study surveys diverse methods for identifying OSA events using physiological signals.
• It details the software-based knowledge contributing to OSA detection.
• Key techniques in signal processing and machine learning for OSA classification are presented.

Conclusions:

• Computer-aided diagnosis holds significant potential for improving OSA detection and influencing treatment decisions.
• Further research is needed to address identified challenges and gaps in current diagnostic methodologies.
• Automated OSA detection systems can enhance diagnostic efficiency and patient comfort.