Lucheng Fang , Jie Cai , Zilin Huang , Aikebaier Tuohuti , Xiong Chen
{"title":"利用人工智能评估模拟鼾声,诊断阻塞性睡眠呼吸暂停。","authors":"Lucheng Fang , Jie Cai , Zilin Huang , Aikebaier Tuohuti , Xiong Chen","doi":"10.1016/j.sleep.2024.11.018","DOIUrl":null,"url":null,"abstract":"<div><h3>Background</h3><div>Performing simulated snoring (SS) is a commonly used method to evaluate the source of snoring in obstructive sleep apnea (OSA). SS sounds is considered as a potential biomarker for OSA. SS sounds can be easily recorded, which is a cost-effective method for prescreening purposes.</div></div><div><h3>Objective</h3><div>This study aimed to validate the performance of artificial intelligence (AI) models using SS sounds for OSA diagnosis.</div></div><div><h3>Methods</h3><div>All participants underwent full-night polysomnography (PSG) monitoring at the sleep center. SS sounds of the participants were recorded during the laryngoscopy procedure. The audio data were processed via Python, and relevant features were extracted. OSA diagnostic models were developed using three machine learning (ML) algorithms and one deep learning (DL) algorithm. The diagnostic performance was evaluated by multiple indicators.</div></div><div><h3>Results</h3><div>A total of 465 participants were included. For the support vector machine algorithm, the accuracy values at apnea-hypopnea index (AHI) levels of 5, 15, and 30 per hour were 0.914, 0.887, and 0.807, respectively. For the K-nearest neighbor algorithm, the accuracy values at AHI levels of 5, 15, and 30 per hour were 0.896, 0.872, and 0.756, respectively. For the random forest algorithm, the accuracy values at AHI levels of 5, 15, and 30 per hour were 0.905, 0.881, and 0.804, respectively. For the audio spectrogram transformer algorithm, the accuracy values at AHI levels of 5, 15, and 30 per hour were 0.926, 0.887, and 0.830, respectively.</div></div><div><h3>Conclusions</h3><div>Our study demonstrates that DL models can effectively screen and identify OSA with commendable performance. In addition, the identification ability of the DL models was better than that of any of the ML models.</div></div>","PeriodicalId":21874,"journal":{"name":"Sleep medicine","volume":"125 ","pages":"Pages 100-107"},"PeriodicalIF":3.8000,"publicationDate":"2024-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Assessment of simulated snoring sounds with artificial intelligence for the diagnosis of obstructive sleep apnea\",\"authors\":\"Lucheng Fang , Jie Cai , Zilin Huang , Aikebaier Tuohuti , Xiong Chen\",\"doi\":\"10.1016/j.sleep.2024.11.018\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><h3>Background</h3><div>Performing simulated snoring (SS) is a commonly used method to evaluate the source of snoring in obstructive sleep apnea (OSA). SS sounds is considered as a potential biomarker for OSA. SS sounds can be easily recorded, which is a cost-effective method for prescreening purposes.</div></div><div><h3>Objective</h3><div>This study aimed to validate the performance of artificial intelligence (AI) models using SS sounds for OSA diagnosis.</div></div><div><h3>Methods</h3><div>All participants underwent full-night polysomnography (PSG) monitoring at the sleep center. SS sounds of the participants were recorded during the laryngoscopy procedure. The audio data were processed via Python, and relevant features were extracted. OSA diagnostic models were developed using three machine learning (ML) algorithms and one deep learning (DL) algorithm. The diagnostic performance was evaluated by multiple indicators.</div></div><div><h3>Results</h3><div>A total of 465 participants were included. For the support vector machine algorithm, the accuracy values at apnea-hypopnea index (AHI) levels of 5, 15, and 30 per hour were 0.914, 0.887, and 0.807, respectively. For the K-nearest neighbor algorithm, the accuracy values at AHI levels of 5, 15, and 30 per hour were 0.896, 0.872, and 0.756, respectively. For the random forest algorithm, the accuracy values at AHI levels of 5, 15, and 30 per hour were 0.905, 0.881, and 0.804, respectively. For the audio spectrogram transformer algorithm, the accuracy values at AHI levels of 5, 15, and 30 per hour were 0.926, 0.887, and 0.830, respectively.</div></div><div><h3>Conclusions</h3><div>Our study demonstrates that DL models can effectively screen and identify OSA with commendable performance. In addition, the identification ability of the DL models was better than that of any of the ML models.</div></div>\",\"PeriodicalId\":21874,\"journal\":{\"name\":\"Sleep medicine\",\"volume\":\"125 \",\"pages\":\"Pages 100-107\"},\"PeriodicalIF\":3.8000,\"publicationDate\":\"2024-11-16\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Sleep medicine\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S1389945724005215\",\"RegionNum\":2,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"CLINICAL NEUROLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Sleep medicine","FirstCategoryId":"3","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1389945724005215","RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"CLINICAL NEUROLOGY","Score":null,"Total":0}
引用次数: 0
摘要
背景:模拟打鼾(SS)是评估阻塞性睡眠呼吸暂停(OSA)患者打鼾来源的常用方法。模拟鼾声被认为是 OSA 的潜在生物标志物。SS声易于记录,是一种经济有效的预筛方法:本研究旨在验证使用 SS 声诊断 OSA 的人工智能(AI)模型的性能:所有参与者都在睡眠中心接受了整夜多导睡眠图(PSG)监测。在喉镜检查过程中记录了参试者的SS声。通过 Python 处理音频数据并提取相关特征。使用三种机器学习(ML)算法和一种深度学习(DL)算法开发了 OSA 诊断模型。诊断性能通过多个指标进行评估:结果:共纳入 465 名参与者。支持向量机算法在呼吸暂停-低通气指数(AHI)水平为每小时 5、15 和 30 时的准确度值分别为 0.914、0.887 和 0.807。对于 K 近邻算法,每小时 AHI 水平为 5、15 和 30 时的准确度值分别为 0.896、0.872 和 0.756。随机森林算法在每小时 AHI 为 5、15 和 30 时的准确度值分别为 0.905、0.881 和 0.804。音频谱图转换算法在每小时 AHI 水平为 5、15 和 30 时的准确度值分别为 0.926、0.887 和 0.830:我们的研究表明,DL 模型能有效筛查和识别 OSA,其性能值得称赞。此外,DL 模型的识别能力优于任何 ML 模型。
Assessment of simulated snoring sounds with artificial intelligence for the diagnosis of obstructive sleep apnea
Background
Performing simulated snoring (SS) is a commonly used method to evaluate the source of snoring in obstructive sleep apnea (OSA). SS sounds is considered as a potential biomarker for OSA. SS sounds can be easily recorded, which is a cost-effective method for prescreening purposes.
Objective
This study aimed to validate the performance of artificial intelligence (AI) models using SS sounds for OSA diagnosis.
Methods
All participants underwent full-night polysomnography (PSG) monitoring at the sleep center. SS sounds of the participants were recorded during the laryngoscopy procedure. The audio data were processed via Python, and relevant features were extracted. OSA diagnostic models were developed using three machine learning (ML) algorithms and one deep learning (DL) algorithm. The diagnostic performance was evaluated by multiple indicators.
Results
A total of 465 participants were included. For the support vector machine algorithm, the accuracy values at apnea-hypopnea index (AHI) levels of 5, 15, and 30 per hour were 0.914, 0.887, and 0.807, respectively. For the K-nearest neighbor algorithm, the accuracy values at AHI levels of 5, 15, and 30 per hour were 0.896, 0.872, and 0.756, respectively. For the random forest algorithm, the accuracy values at AHI levels of 5, 15, and 30 per hour were 0.905, 0.881, and 0.804, respectively. For the audio spectrogram transformer algorithm, the accuracy values at AHI levels of 5, 15, and 30 per hour were 0.926, 0.887, and 0.830, respectively.
Conclusions
Our study demonstrates that DL models can effectively screen and identify OSA with commendable performance. In addition, the identification ability of the DL models was better than that of any of the ML models.
期刊介绍:
Sleep Medicine aims to be a journal no one involved in clinical sleep medicine can do without.
A journal primarily focussing on the human aspects of sleep, integrating the various disciplines that are involved in sleep medicine: neurology, clinical neurophysiology, internal medicine (particularly pulmonology and cardiology), psychology, psychiatry, sleep technology, pediatrics, neurosurgery, otorhinolaryngology, and dentistry.
The journal publishes the following types of articles: Reviews (also intended as a way to bridge the gap between basic sleep research and clinical relevance); Original Research Articles; Full-length articles; Brief communications; Controversies; Case reports; Letters to the Editor; Journal search and commentaries; Book reviews; Meeting announcements; Listing of relevant organisations plus web sites.