Automatic sleep-disordered breathing detection using a single channel record in patients with sleep apnea hypopnea syndrome

Abstract

This dissertation investigates the feasibility for automatic detection of sleep-disordered breathing (SDB) such as sleep apnea/hypopnea and snoring using a single channel record in patents with sleep apnea hypopnea syndrome (SAHS). Nasal pressure (NP) data, photoplethysmographic (PPG) data, and piezo snoring sensor data were employed as a single channel record. SDB is internationally underdiagnosed despite of importance of SDB in clinical diagnosis domains since a conventional SDB analysis using polysomnography during full-night sleep is labor-intensive, time-consuming, high-cost and technically complex. Thus, main objective is to present signal processing methods for automatic SDB detection, which appropriate to the portable monitoring for home-based SDB analysis system or continuous positive airway pressure (CPAP) system.Totally four works were presented for automatic SDB detection. Firstly, a method to detect apnea/hypopnea using five rules based on the ways of the AASM guidelines from NP data was proposed. Secondly, a method to detect apnea/hypopnea based on discrete wavelet transform from PPG data was offered. These two methods showed good performance and are comparable to the methods of previously existed literatures. Thirdly, a method to detect snoring from NP data with low sampling rate was investigated. Lastly, a method to detect snoring based on hidden Markov models from piezo snoring sensor data. These two methods have been not reported in previous literatures. The methods provided reliable performance for portable and real-time snoring detection in comparison with the microphone-based methods reported previously in literatures.The proposed methods can provide several interventions which are easy-to-use, portable and real-time SDB monitoring system during overnight sleep since the methods offer good performance, robustness and insensitivity to the

apnea-hypopnea index severity in patients with SAHS. Moreover, the methods could provide sleep experts with the methods to more accurately analyze the sleep quality and to more objectively monitor SDB in PSG study or CPAP system.