Screening of sleep apnea with decomposition of photoplethysmogram to extract respiratory rate

Abstract

Sleep Apnea is a life threatening sleep syndrome, more prevalent among the adult population and has several serious health issues associated with it. It is a silent killer disease often left unnoticed by the individuals affected by the syndrome. Sleep apnea is mainly due to some obstructions in the upper airway and therefore named as Obstructive Sleep Apnea (OSA), where the normal breathing stops completely for a specific period of time and starts again with a loud snore or cough. Individuals affected with OSA are likely to experience such episodes several times during sleep and therefore suffer from an incisive lack of oxygen. The Photoplethysmogram (PPG) signal obtained from pulse oximeter has the unique characteristics of monitoring this cursory lack of oxygen in the blood. This is a simple non-invasive technique from which the respiratory effort signal can be extracted by applying suitable algorithms. The ultimate goal of this work is to analyze the strength of different frequency components of the PPG signal and to estimate the respiratory rate (RR) by decomposing the signal into several intrinsic mode functions (IMFs), thereby extracting the respiratory modulation. The technique involved is empirical mode decomposition, which applies the Hilbert Spectral Analysis (HAS) method to the IMFs to obtain the instantaneous frequency data. The instantaneous frequency peaks extracted by decomposing the PPG signal reflects the respiratory rate. The dataset of sleep apnea patients, downloaded online is analyzed using the algorithm and the error rate obtained for measuring the respiration rate by using the above stated methodology is about ± 8%.