Identifying the Site of Upper Airway Collapse in Obstructive Sleep Apnoea Patients Using Snore Signals

Abstract

This study aimed to assess the utility of snoring signal analysis in determining the predominant site-of-collapse of the upper airway during natural sleep in patients with Obstructive Sleep Apnea (OSA). The audio signals of 58 OSA patients were recorded simultaneously with full night polysomnography. The probable site-of-collapse was determined by manual analysis of the airflow signal shape during hypopnoea events, which has been shown to correlate well with the gold-standard methods and corresponding audio signal segments containing snore were manually extracted and processed. Machine learning algorithms were developed to automatically annotate the site-of-collapse of each hypopnoea event into three classes (lateral wall, palate and tongue-base) using linear discriminant analysis. The predominant site-of-collapse for a sleep period was then determined from the individual hypopnoea annotations. The predominant site-of-collapse annotations were then compared to the manually determined annotations. Additionally, k-means clustering was used to group the audio features into clusters and the agreement of the clusters with manually determined site-of-collapse was assessed. The classification model achieved an overall accuracy of 81% for discriminating tongue/non-tongue collapse and accuracy of 64% for all site-of-collapse classes. Cluster analysis showed that the data tends to fit well in two clusters with a mean silhouette coefficient of 0.79 and with an accuracy of 68% for classifying tongue/non-tongue collapse. Our results reveal that the snore signal during hypopnoea can provide information regarding the predominant site-of-collapse in the upper airway. Therefore, the audio signal recorded during sleep could potentially be used as a new tool in identifying the predominant site-of- collapse and consequently improving the treatment selection and outcome .