RT info:eu-repo/semantics/article
T1 Multivariate Analysis of Blood Oxygen Saturation Recordings in Obstructive Sleep Apnea Diagnosis
A1 Álvarez González, Daniel
A1 Hornero Sánchez, Roberto
A1 Marcos Martín, José Víctor
A1 Campo Matias, Félix del
AB This study focuses on the analysis of blood oxygensaturation (SaO2) from nocturnal pulse oximetry (NPO) to helpin the diagnosis of the obstructive sleep apnea (OSA) syndrome.A population of 148 patients suspected of suffering from OSAsyndrome was studied. A wide set of 16 features was used tocharacterize changes in the SaO2 profile during the night. Ourfeature set included common statistics in the time andfrequency domains, conventional spectral characteristics fromthe power spectral density (PSD) function and nonlinearfeatures. We performed feature selection by means of a stepforwardlogistic regression (LR) approach with leave-one-outcross-validation. Second and fourth order statistical momentsin the time domain (M2t and M4t), the relative power in the0.014 – 0.033 Hz frequency band (PR) and the Lempel-Zivcomplexity (LZC) were automatically selected. 92.0%sensitivity, 85.4% specificity and 89.7% accuracy wereobtained. The optimum feature set significantly improved thediagnostic ability of each feature individually. Furthermore,our results outperformed classic oximetric indexes commonlyused by physicians. We conclude that simultaneous analysis inthe time and frequency domains by means of statisticalmoments, spectral and nonlinear features could providecomplementary information from NPO to improve OSAdiagnosis.
PB IEEE
SN 0018-9294
YR 2010
FD 2010
LK https://uvadoc.uva.es/handle/10324/82788
UL https://uvadoc.uva.es/handle/10324/82788
LA eng
NO Alvarez, D., Hornero, R., Marcos, J.V. and del Campo, F., 2010. Multivariate analysis of blood oxygen saturation recordings in obstructive sleep apnea diagnosis. IEEE Transactions on Biomedical Engineering, 57(12), pp.2816-2824.
NO Producción Científica
DS UVaDOC
RD 17-feb-2026