Short-term analysis of heart-rate variability using wavelet packets: an efficient detector of sleep apnea episodes

Obstructive sleep apnea is a very common condition characterized by over 10 seconds interruption of breathing during sleep. The accurate and fast detection of sleep apnea stages is of great importance for patients´ recovery and treatment. A criterion for the diagnosis of such respiratory disorder based on the characteristics of the heart-rate variability signal in the wavelet packets domain is presented. Preclassified data of 20 apneic patients and 20 controls drawn from two international databases were subjected to analysis. Experimental results show that a wavelet packet-based measure can be formed for noninvasive diagnosis of sleep apnea. This was justified in all cases of apneic patients where all apneic episodes were accurately identified. The technique is rather simple and could be easily used for prescreening apneic patients in sleep laboratories and intensive care units.