Real-time detection of sleep breathing disorders

D. Feuerstein , L. Graindorge , A. Amblard , A. Tatar (1) , G. Guerrero (1) , S. Christophle-Boulard , C. Loiodice , A.I. Hernandez (1) , J.-L. Pépin

1 LTSI - Laboratoire Traitement du Signal et de l'Image

Diagnosis of sleep-related breathing disorders (SBD) usually relies on manual retrospective analysis of the signals recorded during a whole night. We evaluated a novel detector, implemented in a cardio-respiratory Holter device, and capable of automatic event-based detection of SBDs in real time. Events (hypopneas and apneas) detected in real time were compared to those scored on the simultaneously recorded gold standard polysomnography (PSG). 4240 events were recorded by the PSG in 30 severe obstructive SBD patients. The sensitivity and positive predictive value of the detector were 86.2% (CI. 85.2-87.2%) and 60. 7% (C.l. 59.5-61.9%) respectively. The performance of this novel detector suggests that it could be used to trigger and/or adjust an event-based SBD treatment. © 2015 CCAL.

Mots clés: Cardiology; Breathing disorders; Event-based; Gold standards; Polysomnography; Positive predictive values; Real-time detection; Retrospective analysis; Sleep-related breathing disorders; Diagnosis

Domaines: Ingénierie biomédicale

D. Feuerstein, L. Graindorge, A. Amblard, A. Tatar, G. Guerrero, et al.. Real-time detection of sleep breathing disorders. 42nd Computing in Cardiology Conference, CinC 2015, Sep 2015, Nice, France. pp.317--320, ⟨10.1109/CIC.2015.7408650⟩. ⟨hal-01372343⟩