In-Home Sleep Apnea Severity Classification using Contact-free Load Cells and an AdaBoosted Decision Tree Algorithm

Abstract

We present a method for automated diagnosis and classification of severity of sleep apnea using an array of non-contact pressure-sensitive sensors placed underneath a mattress as an alternative to conventional obtrusive sensors. Our algorithm comprises two stages: i) A decision tree classifier that identifies patients with sleep apnea, and ii) a subsequent linear regression model that estimates the Apnea-Hypopnea Index (AHI), which is used to determine the severity of sleep disordered breathing. We tested our algorithm on a cohort of 14 patients who underwent overnight home sleep apnea test. The machine learning algorithm was trained and performance was evaluated using leave-one-patient-out cross-validation. The accuracy of the proposed approach in detecting sleep apnea is 86.96%, with sensitivity and specificity of 81.82% and 91.67%, respectively. Moreover, classification of severity of the sleep disorder was correctly assigned in 11 out of 14 cases, and the mean absolute error in the AHI estimation was calculated to be 3.83 events/hr.

Design and Evaluation of a Non-Contact Bed-Mounted Sensing Device for Automated In-Home Detection of Obstructive Sleep Apnea: A Pilot Study

“…This paper makes several contributions beyond what we had previously published [19]. First, we present the analysis of the accuracy of our sensing device against gold standard PSG and HSAT tests, which shows that our sensing system and the proposed automated signal processing pipeline achieved high sensitivity in detecting the presence and severity of OSA in-clinic and in-home settings.…”

“…The distribution of the subassemblies within the aluminum plate was such that two cells would register the weight on either side of the body across the shoulders and the other two cells would target the weight in the hip. A more detailed description of the design and fabrication of the LC-based acquisition system was provided in our previously published work [19]. The data collected from the LCs was relayed to an Odroid-U3 processor (Samsung Electronics Co. Ltd., Suwon, Korea) via a serial peripheral interface (SPI) bus.…”

“…Next, the resulting signal was scaled to have zero mean and unit standard deviation. The Fast Fourier Transform (FFT) was used to calculate the power spectral density (PSD) of the processed LC signal and statistical moments within consecutive non-overlapping frequency sub-bands of 0.02 Hz width were computed in order to characterize the distribution of the magnitude of Fourier coefficients [19].…”

“…We developed a two-stage classifier to automatically detect OSA and classify its severity. Our previous work [19] discusses the architecture of the classification system. In order to assess the generalization performance of the developed algorithms, we used leave-one-patient-out cross-validation.…”

“…In the current study, we assess the diagnostic accuracy of a novel AHI measurement system designed in our laboratory at Oregon Health and Science University (OHSU) [19]. We present accuracy results obtained both within a monitored sleep clinic and under free living conditions in patients’ homes.…”