Azadeh Yadollahi

Obstructive sleep apnea (OSA) is a prevalent yet underdiagnosed condition characterized by repetitive upper airway obstruction during sleep. Current gold standard diagnostic standards rely on polysomnography (PSG), which is resource-intensive. Since upper airway characteristics impact both OSA and speech production, speech processing has emerged as a promising alternative for OSA screening. However, prior work has focused primarily on acoustic features. This study aims to develop a speech-based screening and severity estimation pipeline for OSA using self-supervised learning (SSL) and multimodal acoustic features. We proposed a novel fusion framework combining SSL-derived speech representations from pre-trained neural networks with traditional acoustic features and time-frequency representations of speech phase and magnitude. Elongated vowels recorded during wakefulness were used to screen for OSA at two apnea-hypopnea index (AHI) thresholds (10 and 30 events/hour) and to estimate AHI. Data were collected across three research sites, comprising participants of varied sex, race, and OSA severity. For OSA screening, the models achieved balanced accuracies of 0.79 (AHI $\geq$10) and 0.74 (AHI $\geq$30) in females, and 0.80 and 0.78 in males, respectively. AHI estimation yielded mean absolute errors of 12.0 events/hour (r = 0.63) in females and 14.7 events/hour (r = 0.52) in males. Our results demonstrate the feasibility of using speech, especially vowel phonation during wakefulness, as a biomarker for OSA risk and severity estimation. The approach generalizes well across diverse demographic groups. This study presents a significant step toward accessible, low-burden, and cost-effective OSA screening, with broad implications for scalable sleep health assessments.

Cardiopulmonary exercise test (CPET) is widely used for assessing exercise capacity of individuals through ventilatory anaerobic threshold (AT) and peak oxygen uptake (VO _2peak ). However, CPET requires specialized equipment, resources, time, and trained personnel. In this study, we investigated the feasibility of estimating ventilatory AT and submaximal VO _2peak using tracheal respiratory sounds recording with a wearable sensor ( The Patch ) during incremental cycle ergometer testing. Twenty-four healthy adults (18 females, age: 31.0 ± 9.1 years, body mass index: 23.5 ± 3.3 kg/m ² ) underwent CPET with simultaneous tracheal sounds recording. Acoustic features (sound energy, sound intensity, respiratory rate, and ventilation) were extracted. AT was identified from breakpoints in acoustic features and compared with reference AT from ventilatory gas analysis. VO _2peak was estimated using random forest regression models with subject-specific and subject-independent training models. Agreement with CPET was investigated using Bland–Altman analysis. Ventilatory AT was detected in 84% (sound energy) and 96% (sound intensity) of participants using The Patch . Workload stage agreement with CPET was 88% and 100% for sound energy and intensity, respectively, with AT timing within 20 s in 62% and 92% of cases. The subject specific model estimated VO _2peak with minimal bias [confidence intervals] of − 1.01 [− 3.04 to 1.02] (mL/kg/min), while the subject-independent validation model exhibited wider variability of − 1.28 [− 12.39 to 9.82] (mL/kg/min). Our results shows that tracheal sound analysis can estimate ventilatory AT and provides a promising approach for non-invasive exercise capacity assessment. On the other hand, the estimation of VO _2peak shows promise but requires further validation in maximal effort protocols.

Heart failure is a leading cause of morbidity and mortality, and fluid congestion complicates its management. Conventional home-monitoring methods, such as

Obstructive sleep apnea (OSA) represents a prevalent condition impacting over 9% of the general adult population. Various treatment options have been clinically proposed and utilized, with a particular focus on continuous positive airway pressure (CPAP) and oral appliances due to their overall effectiveness and higher adherence rates. CPAP therapy has demonstrated greater effectiveness but lower adherence compared to oral appliances. However, treatment success of oral appliances is not always guaranteed, hence sleep physicians are more cautious in their prescriptions unless they can reasonably estimate the chance of responding to oral appliance therapy. Prior studies often rely on invasive or inconvenient methodologies such as drug-induced sleep endoscopy (DISE), cephalometry, multisensor catheters, or full polysomnography (PSG). In this prospective study, we collected data with a home sleep apnea test (HSAT) device from 50 participants (38 using mandibular advancement devices [MADs] and 12 using tongue stabilizing devices [TSDs]). We used a simple yet informative data source: snoring vibrations extracted from a nasal pressure sensor with a low sampling frequency (125 Hz). Using spectro-temporal analysis of the snoring signal, we successfully predicted therapy efficacy with accuracies of 88% for MAD and 91% for TSD. Our proposed methodology presents a promising approach that can be utilized without further need for PSG or integrated within PSG testing for accurate prediction of oral appliance efficacy.

Purpose: Cardiovascular disorders are the leading cause of mortality worldwide with obstructive sleep apnea (OSA) as the independent risk factor. Heart sounds are strong modalities to obtain clinically relevant information regarding the functioning of the heart valves and blood flow. The objective of this study was to use a small wearable device to record and investigate the changes in heart sounds during respiratory events (reduction and cessation of breathings) and their association with oxyhemoglobin desaturation (hypoxemia). Patients and Methods: Sleep assessment and tracheal respiratory and heart sounds were recorded simultaneously from 58 individuals who were suspected of having OSA. Sleep assessment was performed using in-laboratory polysomnography. Tracheal respiratory and heart sounds were recorded over the suprasternal notch using a small device with embedded microphone and accelerometer called the Patch. Heart sounds were extracted from bandpass filtered tracheal sounds using smoothed Hilbert envelope on decomposed signal. For each individual, data from 20 obstructive events during Non-Rapid Eye Movement stage-2 of sleep were randomly selected for analysis. Results: A significant increase in heart sounds’ intensities from before to after the termination of respiratory events was observed. Also, there was a significant positive correlation between the magnitude of hypoxemia and the increase in heart sounds’ intensities (r> 0.82, p< 0.001). In addition, the changes in heart sounds were significantly correlated with heart rate and blood pressure. Conclusion: Our results indicate that heart sound analysis can be used as an alternative modality for assessing the cardiovascular burden of sleep apnea, which may indicate the risk of cardiovascular disorders.