Improving Automatic Detection of Obstructive Sleep Apnea Through Nonlinear Analysis of Sustained Speech

Abstract

We present a novel approach for the detection of severe obstructive sleep apnea (OSA) based on patients' voices introducing nonlinear measures to describe sustained speech dynamics. Nonlinear features were combined with state-of-the-art speech recognition systems using statistical modeling techniques (Gaussian mixture models, GMMs) over cepstral parameterization (MFCC) for both continuous and sustained speech. Tests were performed on a database including speech records from both severe OSA and control speakers. A 10 % relative reduction in classification error was obtained for sustained speech when combining MFCC-GMM and nonlinear features, and 33 % when fusing nonlinear features with both sustained and continuous MFCC-GMM. Accuracy reached 88.5 % allowing the system to be used in OSA early detection. Tests showed that nonlinear features and MFCCs are lightly correlated on sustained speech, but uncorrected on continuous speech. Results also suggest the existence of nonlinear effects in OSA patients' voices, which should be found in continuous speech.

Declines in Connected Language Are Associated with Very Early Mild Cognitive Impairment: Results from the Wisconsin Registry for Alzheimer’s Prevention

“…Although poor sleep quality has been associated with cognitive decline ( Lim et al, 2013 ), no such associations have been reported with connected language. However, sustained speech analysis was able to detect participants with obstructive sleep apnea (OSA) with 89% accuracy ( Blanco et al, 2013 ), so it is possible that a more detailed speech and voice analysis may yield more information about the relationship between sleep quality and discourse features. Anxiety has been associated with increases in dysfluency in persons who stutter ( Yang et al, 2017 ), and with pragmatic language difficulties ( Cummings, 2007 ), but research is lacking regarding the relationship between anxiety and connected language in the general population or in persons with cognitive decline.…”

Detection of severe obstructive sleep apnea through voice analysis

“…So far, either due to small samples or subjective judgements, it is hard to quantify up to what extent or under what circumstances we might consider voice as a good discrimination measure between OSA and healthy subjects. Recent efforts such as [6] try to model the upper-airway in OSA subjects as compared to controls by employing computational fluid dynamics models, and they conclude that there is a clear tendency to closure of the upper-airway in OSA. As the upper-way coincides in part with the vocal tract, the thinning of the lumen and tendency to closure experienced in OSA do suggest that there may be an identifiable dysfunction in voice also.…”

Obstrüktif Uyku Apnesi Tanıma için Öznitelik Seçimi

“…They obtain best results with i-vectors and SVR with linear kernel. Blanco et al introduce a novel method to detect OSA based on patient's voices [29] on the manually collected dataset. In [30], Aydogan et al performs visual scoring of 74 patients using two methods, namely morphological filter and ANN-based to diagnose OSA.…”