Sleep Apnea Detection by Means of Analyzing Electrocardiographic Signal

Abstract

Obstructive sleep apnea (OSA) is a condition of cyclic, periodic obstruction (stenosis) of the upper respiratory tract. OSA could be associated with serious cardiovascular problems, such as hypertension, arrhythmias, hearth failure or peripheral vascular disease. Understanding the way of connection between OSA and cardiovascular diseases is important to choose proper treatment strategy. In this paper, we present a method for integrated measurements of biosignals for automatic OSA detection. The proposed method was implemented using a portable device with the application of the Support Vector Machine (SVM) classifier. The specific objective of this work is to analyze the minimum set of features for the ECG signal that could produce acceptable classification results. Those features can be further expanded using other biosignals, measured by the portable SleAp device. Additionally, the influence of the body movements and positions on measurement results with SleAp system are presented. The proposed system could help to determine the influence of OSA on the state of the cardiovascular system. © Springer International Publishing Switzerland 2014.