Clinical-level screening of sleep apnea syndrome with single-lead ECG alone is achievable using machine learning

Aug 5, 2026

Upload by Riley

Purpose: establish a simple, noninvasive sleep apnea (SA) screening approach that reduces burden on potential patients. Objective: verify whether past and future single-lead electrocardiogram (ECG) data near SA occurrence sites improve machine-learning estimation accuracy for SA and sleep apnea syndrome (SAS). Methods: Apnea-ECG dataset of 70 ECG recordings; time-window size optimized by detection accuracy; compare SA detection and SAS diagnosis performance.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Abstract

Purpose

Methods

Results

Conclusions

Introduction

Sleep apnea syndrome and clinical impact

Diagnosis with polysomnography and limitations

Prior ECG-based detection methods and research gap

Sleep and Breathing (2025) 29:156 [https://doi.org/10.1007/s1](https://doi.org/10.1007/s1) 1325-025-03316-0 SLEEP BREATHING PHYSIOLOGY AND DISORDERS • ORIGINAL ARTICLE Clinical-level screening of sleep apnea syndrome with single-lead ECG alone is achievable using machine learning with appropriate time windows Takahiro Yamane1 · Masanori Fujii2,3 · Mizuki Morita1,4 Received: 2 September 2024 / Revised: 3 February 2025 / Accepted: 26 March 2025 © The Author(s) 2025 Abstract Purpose To establish a simple and noninvasive screening test for sleep apnea (SA) that imposes less burden on potential patients. The specific objective of this study was to verify the effectiveness of past and future single-lead electrocardiogram (ECG) data from SA occurrence sites in improving the estimation accuracy of SA and sleep apnea syndrome (SAS) using machine learning. Methods The Apnea-ECG dataset comprising 70 ECG recordings was used to construct various machine-learning models. The time window size was adjusted based on the accuracy of SA detection, and the performance of SA detection and SAS diagnosis (apnea‒hypopnea index ≥ 5 was considered SAS) was compared. Results Using ECG data from a few minutes before and after the occurrence of SAs improved the estimation accuracy of SA and SAS in all machine learning models. The optimal range of the time window and achieved accuracy for SAS varied by model; however, the sensitivity ranged from 95.7 to 100%, and the specificity ranged from 91.7 to 100% . Conclusions ECG data from a few minutes before and after SA occurrence were effective in SA detection and SAS diagnosis, confirming that SA is a continuous phenomenon and that SA affects heart function over a few minutes before and after SA occurrence. Screening tests for SAS, using data obtained from single-lead ECGs with appropriate past and future time windows, should be performed with clinical-level accuracy. Keywords Disease screening · Sleep apnea syndrome (SAS) · Single-lead ECG · Artificial intelligence · Machine learning 􀀍 Mizuki Morita [mizuki@okayama-u.ac.jp](mizuki@okayama-u.ac.jp) 1 Department of Biomedical Informatics, Graduate School of Interdisciplinary Science and Engineering in Health Systems, Okayama University, Okayama, Japan 2 Department of Geriatric Medicine, Faculty of Medicine, Dentistry and Pharmaceutical Sciences, Okayama University, Okayama, Japan 3 Department of Allergy and Respiratory Medicine, Okayama University Hospital, Okayama, Japan 4 Faculty of Health Sciences, Okayama University Medical School, Okayama, Japan Introduction Sleep apnea syndrome (SAS) is a disorder in which breathing frequently ceases or decreases during sleep. Symptoms include snoring during sleep, extremely heavy daytime sleepiness, fatigue, and difficulty concentrating. SAS significantly impacts both the cardiovascular and endocrine systems and is closely linked to several lifestyle-related diseases, such as hypertension, heart failure, diabetes, and cerebrovascular disease [1] . SAS can be classified into three types: central sleep apnea (CSA), obstructive sleep apnea (OSA), and a combination of these two types. In CSA, breathing during sleep is disturbed by an abnormality in the respiratory center located in the medulla oblongata of the brain. In OSA, the pharynx is obstructed, and air cannot pass through [2] . The number of potential patients with OSA has been estimated to be more than that of patients 1 3 treated with continuous positive airway pressure (CPAP), the standard treatment for moderate to severe OSA, regardless of the country. In Japan, for example, the former is estimated to be 22 million [3], while the latter is > 400,000 [4] . In France, the former is approximately 24 million [3], and the latter is approximately 830,000 [5] . These data indicate that the number of patients with OSA who are not properly diagnosed and benefit from treatment is very small. SAS is diagnosed using polysomnography (

FAQs

What was the purpose of the study on sleep apnea screening?

To create a simple, noninvasive screening test for sleep apnea that is less burdensome for potential patients.

How were machine-learning models built and evaluated?

Using the Apnea-ECG dataset of 70 ECG recordings, the study adjusted the ECG time-window size based on SA detection accuracy and compared SA detection and SAS diagnosis performance.

What timing of single-lead ECG data improved estimation accuracy?

ECG data from a few minutes before and after SA occurrence improved estimation accuracy for SA and SAS across all models.

Aug 5, 2026

Upload by Riley