Unobtrusively Detecting Apnea and Hypopnea Events via a Hydraulic Bed Sensor

Abstract

Disordered breathing during sleep impacts sleep quality and the perceived amount of rest obtained while also serving as a potential indicator of other health conditions or risks. Apneas and hypopneas are leading indicators of disordered breathing, often quantified by an apnea-hypopnea index (AHI). Polysomnography is the gold standard for detecting apnea and hypopnea events (and thus calculating a subject’s AHI), but despite the inconvenience of sleeping in a strange place with numerous instruments attached, polysomnography delivers only a snapshot in time and is not practical for long-term monitoring. In this work, we describe a method of detecting apnea and hypopnea events during sleep using a hydraulic bed sensor, which has proven valuable for other dimensions of long-term monitoring and early detection of illness. We compare our results to those produced by a polysomnography lab, including calculation of respiratory disturbance indices. We successfully detect 73.6% of apneas with 77.2% precision, and our calculations for apnea index (AI) and respiratory disturbance index (RDI) are precise enough to indicate the appropriate severity of sleep apnea-hypopnea syndrome (SAHS) for each of our subjects.