Detection of sleep disorders by a modified Matching Pursuit algorithm

Abstract

Sleep disturbances are, beside headaches, the most frequently articulated problems at general practitioners. Approximately 20% of adults in the western world suffer from sleep disturbances, most commonly sleep apnea (SA), which affects 2-4% of middle-aged adults. Therefore a reliable, ambulant screening test is requested, which is easy to perform and does not necessarily demand profound knowledge of sleep medicine. In this paper a new Matching Pursuit based algorithm is presented, that uses a combination of SpO2 and photoplethysmographically derived pulse wave information to calculate a respiratory disturbance index (RDI). Furthermore an "autonomic arousal index" (AAI) was constructed to reflect the intensity of pulsatile changes suggestive sudden bursts of sympathetic activity associated with arousal from sleep. A signal decomposition algorithm, based on a dictionary of timefrequency atoms (known as "Matching Pursuit method"), has been modified in order to analyse different patterns in the photoplethysmographic signals. The performance of the algorithm was tested on 62 consecutive adult patients with suspected SA, who were referred to the sleep laboratory. In a second step indices of autonomic arousals were analysed and compared in different patient groups. The correlation coefficient between manual scored AHI and automatically calculated RDI, using only pulse oximetry channels, was r = 0.967. Bland-Altmann analysis showed a mean difference of -0.6 between the two parameters. Using a cut-off value of RDI ≥ 15/h for SA classification, a sensitivity of 96.2% and specificity of 91.7% was reached. The mean AAI differed significantly between healthy individuals and people with moderate number of respiratory events, severe SA patients and insomniacs. This novel computer algorithm provides a simple and highly accurate tool for quantification of SA and provides important information about autonomic activity during sleep. Thus such screening system appears to provide important information for the diagnosis of other diseases like autonomic neuropathy or insomnia.