The Diagnostic Model of Obstructive Sleep Apnea Hypopnea Syndrome Based on Artificial Neural Networks

Abstract

Obstructive sleep apnea hypopnea syndrome (OSAHS) is a kind of breathing regulator disorder disease in sleep, which is related to many complex factors and not yet fully elucidated the pathological state. This pathological state is not only related to snoring, excessive daytime sleepiness (EDS), also due to hypopnea or apnea caused by repeated episodes of hypoxia and hypercapnia, which can lead to the complications of cardiopulmonary and other vital organs even to sudden death. So OSAHS is a potential fatal disease, which has been widely appreciated in clinic. The target of dealing with OSAHS was not only to decrease the mortality rate, but also to minimize the potential adverse effects. Therefore, it was important to detecting, preparation and treatment early in clinic, especially how to detect as soon as possible. In this study, the diseases diagnosis was realized by using multilayer, probabilistic, logic model, and generalized regression neural networks. The diseases dataset was prepared by patients' case reports from No.307 hospital of PLA's database and volunteers' experiments. By which, built a model to analysis a number of parameters associated to OSAHS.