Research Areas

Acoustic Sleep Apnea Diagnosis

Sleep apnea syndrome is a common and serious respiratory disorder.  Apnea is a cessation of airflow to the lungs (usually during sleep) which lasts for at least 10 seconds. Polysomnography (PSG) during the entire night is currently the accepted Gold Standard diagnostic method of sleep apnea. The standard PSG consists of recording various physiological parameters including EEG, ECG, EMG of chins and legs’ muscles, nasal airflow, electro-oculogram (EOG), abdominal and thoracic movements, and blood oxygen saturation (SaO2) and usually snoring sounds. However, the high cost of the system, discomfort of the electrodes connecting to the body and the high amount of information required to be analyzed are the main disadvantages of this method. To obtain a full sleep study in Manitoba there is currently a long waiting list (3300 patients) and long waiting times (3.4 to 8.3 years). Hence, healthcare providers and payers are seeking alternative methods, portable devices, and automated/intelligent systems in which sleep apnea testing can be done in the patient’s home. We have been one of the first groups who started to use respiratory sounds for apnea/hypopnea detection. Our work on this topic has been recognized and funded by industry (TRLabs and Sasktel, and BCC) to develop a prototype of a system including a portable device that has the capability to record respiratory and snore sounds as well as SaO2, and transmit data to a center, in which a smart expert software detects apnea/hypopnea events and present the apnea index and other clinical information to a physician. This system has been patented in US and Europe. The advantages of our proposed system to other portable devices are that it is as accurate as PSG but without the need for supervision, is very convenient for patients and also has a fast and user friendly analysis.

Since 2010, we started to use tracheal breathing sounds and advance signal processing with machine learning to develop a quick screening tool to detect OSA in patients whilst awake. That has resulted a technology that we call it AWakeOSA (patented) and published in several journal papers. AWakeOSA technology records a few breathing sounds during wakefulness and within a few seconds can identify people with OSA and in need of treatment with an accuracy above 90%. To learn more about our acoustic OSA  tecnology you may click on this link and see the slides.

Here are the list of our publications related to this topic:


Selected Full Papers Published in Refereed Journals