Recognition and quantification of sleep apnea by analysis of heart rate variability parameters

This study was performed within the scope of the CinC-2000 challenge on detection and quantification of obstructive sleep apnea from single channel ECGs. Established statistical time domain heart rate variability (HRV) measures as well as parameters based on time delay embedding and correlation analysis are investigated for their diagnostic quality by means of receiver operating characteristics (ROC) analysis on a training set of 35 ECGs and cross validated on an independent tests set of equal size. Moreover, several feature combinations are evaluated with a second order polynomial classifier. The results indicate, that at least most of the information necessary to recognize sleep apnea is contained in the EGG. Recognition rates up to 93% for screening of apnea patients and 85.55% for minute by minute quantification are achieved.