Effect of background sound and impact of snore episode length normalization on snore-based apnea diagnosis

Abstract

Obstructive Sleep Apnea (OSA) is a serious sleep disorder. OSA is commonly associated with snoring but it is not fully utilized in diagnosis. Snoring contains pseudo-periodic ("pitch of snoring") packets of energy that produces the characteristic vibrating sounds familiar to us. Our hypothesis is that the pitch of snoring carries information on the state of the upper airways enabling us to characterize it. Snore Related Sounds (SRS) have the advantage as they can be acquired via non-contact measurements cheaply. However, in practice, it is likely that SRS may be corrupted by background interference such as bed sounds, duvet sounds and speech sounds (collectively referred to as "Other Sounds", OS). In this paper, we explore Intra Snore Pitch Jump probability which captures and quantizes pitch jumps as a feature to diagnose OSA. In particular, we focus into the: (i) the effect of other sounds on the performance of pitch-jump based OSA diagnosis, and (ii) propose a snore episode length-normalization as an efficient way to improve the pitch-jump algorithm. The proposed method was tested on a database of 39 subjects (training set: n=10; validation set: n=29), and ROC curves were drawn. Our method led to improved diagnostic performance (sensitivities of 95% at specificities 86%.