Graduation Year

2005

Document Type

Thesis

Degree

M.S.I.E.

Degree Granting Department

Industrial Engineering

Major Professor

José L. Zayas-Castro, Ph.D.

Committee Member

Tapas K. Das, PhD

Committee Member

Kendall F. Morris, Ph.D.

Committee Member

A.N.V. Rao, Ph.D

Keywords

Principal curve, Intermittent hypoxia, Multichannel signal monitoring, Factor analysis, Wavelet transformation

Abstract

Obstructive sleep apnea is a potentially life-threatening condition characterized by repetitive episodes of upper airway obstruction that occur during sleep, usually associated with a reduction in blood oxygen saturation. In US population, 9% of women, 24% of men, and 2% of children have been diagnosed with obstructive sleep apnea, suggesting that 18 million people may suffer from the consequences of nightly episodes of apnea. One of the most significant symptoms of obstructive sleep apnea is profound and repeated hypoxia. The analysis of the interaction between cardiovascular and respiratory signals has been a widely-explored area of research due to the significance of the results in describing a functional relationship between the underlying physiologic systems; however, statistical and analytical approaches to analyze the changes in these signals before and after hypoxia are still in their early stages of evolution. A major motivation for this research has been the lack of methodologies to detect mean and/or variance shifts and identify root sources of variation in time-frequency characteristics of multichannel data.

The contributions of this thesis are twofold. First, multiscale energy distributions based on wavelet transformations of the analyzed physiological signs are analyzed. This is followed by the development of an online multichannel monitoring approach based on principal curves that detects changes in the wavelet coefficients extracted from the analyzed signals.

Share

COinS