ABSTRACT
Sleep apnea (SLA) is one of the most common sleep-related disorders, and the effects of undiagnosed SLA can range from high blood pressure to cardiac arrest. However, the majority of people are also unaware of their disorder. Overnight polysomnography (PSG) in a sleep laboratory is the global standard method for the diagnosis and monitoring of SLA. Moreover, these tests are costly and inaccessible, and require an experienced specialist to score. Meanwhile, in comparison to the number of SLA patients, the proportion of available research labs and beds is limited. Multiple researchers have suggested and developed automated scoring processes based on fewer detectors and automated classification algorithms to resolve these problems. An automatic detection system will allow for a high-rate diagnosis and the analysis of even more patients. Deep learning is achieving high importance due to the availability of databases and recently developed methods. As a result, SLA study has been currently gaining a lot of attention in deep learning. In this study, we discuss several algorithms focused on cutting-edge deep-learning methods for automatic feature extraction and identifying SLA episodes in cardiopulmonary signals. The present work aims to review published research papers from the last 10 years in order to find the outcome of research objectives such as how to develop specific deep neural networks, what other kind of preprocessing or extraction of features is required, as well as the advantages and limitations of different types of networks. First, we will discuss the background and evaluation of the problem. Second, we present a detailed analysis of the current work and discuss a potential application. After that, we compare the SLA classification algorithm based on accuracy, specificity, sensitivity, and other parameters. Overall, our analysis provides comprehensive and detailed information for researchers looking to add to this field.
