You are here: Home > For Researchers > Projects > OSA+ - Multimodal Signal Analysis for Unobtrusive Characterization of Obstructive Sleep Apnea (OSA+)

OSA+ - Multimodal Signal Analysis for Unobtrusive Characterization of Obstructive Sleep Apnea (OSA+)

From 01-01-2016 to 31-12-2020

Description

Obstructive sleep apnea (OSA) is a sleep-related breathing disorder, characterized by intermittent disruptions of normal breathing patterns during sleep. Its prevalence is increasing [1], [2], and around 100 million people suffer from OSA worldwide [3]. OSA is an important cause of increased cardiovascular risk [4], [5] and even all-cause mortality [6], [7].

Polysomnography (PSG) is the gold standard for diagnosis of OSA [8], [9]. Yet it requires costly sleep center facilities, overnight hospitalization, and instrumentation that may disrupt the normal sleep pattern of patients. PSG is not suitable as a population screening tool; while it is estimated that nearly 90% of OSA patients remain undiagnosed, in particular for mild cases unaware at an early stage [10]. Also, PSG has limited predictive value of cardiovascular risk and correlates poorly with OSA therapy outcome.

Taken together, there is a clear need for better characterization of OSA patients to enable early detection, cardiovascular risk stratification, and treatment selection and monitoring. To this end it is essential that measurements are unobtrusive so that patients can be observed for a prolonged period in an unsupervised home environment and without disturbing normal sleep patterns.

We aim to (1) enable reliable OSA screening with unobtrusive measurements, and (2) characterize OSA patients beyond existing markers for improved cardiovascular risk assessment and treatment guidance.

We will use photoplethysmography (PPG) [11], respiratory sounds (such as snoring) [12], as well as ballistocardiography (BCG) [13]. The analysis of these unobtrusive measurements is challenging due to their intrinsic uncertainties and artefacts, and due to the unsupervised settings in which they originate. To deal with these challenges, we will use a probabilistic signal analysis approach that exploits pathophysiological prior knowledge, and where needed combine multiple measurement modalities.

The ultimate goal is to develop a reliable sleep platform that can be used in a home environment, which will herald a new era in OSA diagnosis and treatment for a large population. Precise characterization of OSA patients may also improve our understanding of the underlying pathophysiology; and can help prioritize patients for treatment, ultimately reducing their morbidity and mortality.

The proposed project will build on the unique expertise of the project team in unobtrusive sensing, signal processing, biomedical data analysis, physiological modeling, and OSA diagnosis and treatment.