“Comparison of machine learning and deep learning classifier to detect sleep apnea using single-channel ECG and HRV: A Systematic Literature Review”

Abstract

sleep apnea (SA) detection has popularly been performed using the gold standard polysomnography (PSG). This method is complex, time-consuming, and essentially requires a skilled person to record multi-channel sleep data. PSG data acquisition to detect OSA requires a dedicated sleep laboratory and hence it is expensive. The main aim of this paper is to select the most suitable modality single-channel electrocardiogram (ECG) and heart rate variability (HRV) to detect OSA using machine learning and deep learning classifiers. This paper includes selected articles written in the English language and published during the last decade. A total of 273 papers were found and 25 were selected using the preferred reporting items for systematic literature reviews and meta-analyses (PRISMA) approach. It includes articles compiled by searching in google scholar, PubMed, and Web of Science databases. The main contribution of this systematic literature review (SLR) is to identify the research gap and to Adhoc novel, methodologies to fill the gap for sleep apnea detection using ECG and HRV in machine learning (ML) and deep learning (DL) classifiers. The study concludes that DL algorithms for SA using ECG and HRV modalities are easy and robust. It was reported in this review that during the year 2021, most of the research work was carried out in DL for SA detection which establishes its significance. This SLR also provides important information to choose an efficient model for SA detection and characterization using a single-channel modality.