Machine Learning Assisted Wearable Wireless Device for Sleep Apnea Syndrome Diagnosis

Abstract

Sleep apnea syndrome (SAS) is a common but underdiagnosed health problem related to impaired quality of life and increased cardiovascular risk. In order to solve the problem of complicated and expensive operation procedures for clinical diagnosis of sleep apnea, here we propose a small and low-cost wearable apnea diagnostic system. The system uses a photoplethysmography (PPG) optical sensor to collect human pulse wave signals and blood oxygen saturation synchronously. Then multiscale entropy and random forest algorithms are used to process the PPG signal for analysis and diagnosis of sleep apnea. The SAS determination is based on the comprehensive diagnosis of the PPG signal and blood oxygen saturation signal, and the blood oxygen is used to exclude the error induced by non-pathological factors. The performance of the system is compared with the Compumedics Grael PSG (Polysomnography) sleep monitoring system. This simple diagnostic system provides a feasible technical solution for portable and low-cost screening and diagnosis of SAS patients with a high accuracy of over 85%.

Developing Probabilistic Ensemble Machine Learning Models for Home-Based Sleep Apnea Screening using Overnight SpO2 Data at Varying Data Granularity

“…Numerous studies have leveraged a subset of signals derived from PSG, such as electroencephalogram (EEG), electrocardiogram (ECG), airflow, and blood oxygen saturation levels (SpO2), for automated sleep apnea screening. Performance outcomes vary based on signal modality and computational models utilized [4][5][6][7][8][9][10][11][12][13]. Yet, most sensing modalities, including EEG, ECG, and airflow, are not readily available for home use, hindering widespread adoption for at-home apnea screening.…”

Patent and Bibliometric Analysis of the Scientific Landscape of the Use of Pulse Oximeters and Their Prospects in the Field of Digital Medicine

“…Wang et al developed an apnea diagnostic system using a photopletismography sensor to synchronously collect human pulse wave signals and oxygenate the blood. Machine learning was used to process data [114]. The high accuracy of the system was revealed-more than 85%.…”

“…The integration of digital technologies into pulse oximetry significantly enhances healthcare delivery by streamlining the flow of patient data, improving patient safety, enabling timely medical care, and augmenting the objectivity of clinical results and the accuracy of clinical outcomes, while reducing both the time and use of material resources. Generalized approaches to risk reduction when using pulse oximetry based on the example of a number of patents and scientific articles [104,113,114,120,130,[132][133][134][135][136]. Generalized approaches to risk reduction when using pulse oximetry based on the example of a number of patents and scientific articles [104,113,114,120,130,[132][133][134][135][136].…”

“…Generalized approaches to risk reduction when using pulse oximetry based on the example of a number of patents and scientific articles [104,113,114,120,130,[132][133][134][135][136]. Generalized approaches to risk reduction when using pulse oximetry based on the example of a number of patents and scientific articles [104,113,114,120,130,[132][133][134][135][136].…”

“…Figure 9. Generalized approaches to risk reduction when using pulse oximetry based on the example of a number of patents and scientific articles[104,113,114,120,130,[132][133][134][135][136].…”