Automatic screening of Obstructive Sleep Apnea from single-lead Electrocardiogram

Obstructive Sleep Apnea (OSA) is traditionally diagnosed using multiple channel physiological signal. This often leads to incorrect apnea event detection and weakens the performance of OSA diagnosis. Furthermore, there is a dire need of an automatic OSA screening system in order to alleviate the burden of the clinicians and to make a portable home sleep monitoring system feasible. In this work, an algorithm that uses single lead Electrocardiogram (ECG) to detect OSA events is propounded. The contribution of this work is twofold. First, it proposes an automatic OSA detection algorithm using Empirical Mode Decomposition, higher order statistical features and Extreme Learning Machine (ELM). Second, ELM is introduced in this work and this is the first time ELM has been applied to OSA detection. Experimental outcomes backed by statistical validation evinces that the proposed algorithm is superior to existing ones in accuracy.