Related Experiment Video
Updated: Jan 11, 2026
07:40
Multi-Modal Home Sleep Monitoring in Older Adults
Published on: January 26, 2019
8.1K
Multichannel machine learning for polysomnographic diagnosis of obstructive sleep apnea: a Bayesian meta-analysis
Shahana Rani1, Esther Yanxin Gao1,2,3, Joel Zuo Er Ong1
1Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore.
Summary
Artificial intelligence (AI) models show high accuracy in diagnosing obstructive sleep apnea (OSA), with neural networks performing best. Further validation is needed for integrating AI into sleep medicine to improve patient access to diagnosis.
Area of Science:
- Sleep Medicine
- Medical Diagnostics
- Artificial Intelligence
Background:
- Obstructive sleep apnea (OSA) affects over 1 billion people globally, with more than 80% undiagnosed.
- Traditional polysomnography is accurate but labor-intensive, causing diagnostic delays and increasing healthcare costs.
- Artificial intelligence (AI) presents a potential solution for efficient and accurate OSA diagnosis.
Purpose of the Study:
- To evaluate the diagnostic accuracy of AI models for obstructive sleep apnea (OSA).
- To compare the performance of different AI subtypes, including neural networks, against the apnea-hypopnea index (AHI).
Main Methods:
- A systematic literature search was conducted across major databases (PubMed, Embase, Scopus, Web of Science, IEEE Xplore).
- Bayesian bivariate meta-analysis and meta-regression were employed on data from 7 studies involving 19 AI models.
- Risk of bias and evidence quality were assessed using QUADAS-2 and GRADE criteria.
Main Results:
- AI models demonstrated a pooled sensitivity of 89.2% and specificity of 87.1% for OSA diagnosis.
- Neural networks (NNs) exhibited superior performance with a sensitivity of 92.8% and specificity of 87.8%.
- No significant impact of age or sex was observed, and no publication bias was detected.
Conclusions:
- AI models, particularly neural networks trained on polysomnography data, show excellent diagnostic accuracy for OSA.
- External validation and further research are crucial for integrating AI into routine sleep medicine practice.
- AI has the potential to enhance access to efficient and accurate OSA diagnosis.
