Monitoring and assessment of obstructive sleep apnea and its outcomes through the processing of physiological signals of wearables

Sleep apnea-hypopnea syndrome (SAHS) is a sleep disorder characterized by repetitive periods of interruption or considerable reduction of respiratory flow, with serious consequences for the health and quality of life of the patients, such as daytime sleepiness and cardiovascular risk. These consequences can be reduced by adequate treatment but most of SAHS patients remain undiagnosed (>80%), since SAHS diagnosis relies on polysomnography (PSG), which requires an overnight stay in a sleep laboratory and supervision by experts. This also causes high costs for the health systems due to the complications related to non-treated SAHS. Moreover, SAHS severity assessment is currently based on apnea-hypopnea index (AHI) which correlates poorly with SAHS complications.

The main objective of this project is to improve diagnosis and monitoring of SAHS focusing on: i) alternative markers, derived from the processing of physiological signals during night, that correlate better than AHI with two of most common SAHS complications: daytime sleepiness and cardiovascular risk; ii) wearable devices which can reach a broader population, reducing underdiagnosis, and allow long-term monitoring (months/years) to follow SAHS evolution and treatment efficiency.

On one hand, signal processing techniques will be developed to assess autonomic nervous system (ANS) in SASH patients, which should be robust in the presence of apneic events. These techniques will include: heart rate variability analysis guided by respiration and cardiorespiratory interactions, derived from the joint analysis of electrocardiogram (ECG) and respiration; ventricular repolarization dynamics, derived from ECG; and model-based pulse morphologic features, derived from the pulsephotoplethysmographic (PPG) signal. On the other hand, three wearable devices will be considered: a custom-made ECG-based armband, a PPG-based wrist device and a PPG-based forehead device. Biomedical signal processing algorithms will be developed for obtaining a long-term cardiorespiratory monitor based on our custom-made ECG armband. Algorithms to estimate cardiorespiratory parameters will be developed, including respiratory rate, tidal volume and ANS biomarkers, analysing and maximizing their coverage during sleep.

On the other hand, PPG signal processing algorithms will be developed, including methods for identifying and attenuating artefacts , and for deriving robust respiratory and ANS markers from wearable PPG signals recorded at wrist and forehead.Different approaches will be proposed and validated for sleep disordered breathing events detection and classification (central/obstructive apnea/hypopnea). To achieve these goals, the recording of different databases will be conducted in the context of this project, including ambulatory recordings as well as in-hospital recordings with simultaneous PSG, at different time points, to monitor SAHS evolution and treatment efficiency. Finally, the proposed methods will be implemented and validated in the wearable devices, bridging the gap between research and technology transfer to society.

The project is expected to make major scientific and technical contributions to SAHS diagnosis and monitoring, improving health and quality of life of SAHS patients and reducing the high social and economic burden that SAHS represents to our health system. The project relies on a multidisciplinary and international team, including collaboration with companies.

PID2021-126734OB-C21