|
Abstract:
|
Polysomnography is an overnight study to collect physiological parameters during sleep. It takes several days to score/interpret the raw data from such a study and confirm a diagnosis of, say, Obstructive Sleep Apnea (OSA). The presence of artifacts (anomalies created by malfunctioning sensors) makes scoring more difficult, potentially resulting in misdiagnosis. In infants using pacifiers during sleep, the act of sucking on the pacifier causes artifacts in the oro-nasal sensor (thermistor) monitoring respiratory airflow. The resulting inaccuracy leads to an under-estimation of the Apnea Hypopnea Index, the basis for an OSA diagnosis. So researchers are now exploring two other information sources (blood oxygen saturation readings from a pulse-oximeter and occurrence of arousal events). They first look for statistical association between the thermistor and the pulse-oximeter/arousal data and then statistically predict a modified AHI using the latter whenever the former is corrupt. This project aims at developing several competing probabilistic models for these data-types. They include naïve Bayes, Beta-Binomial, correlated homogeneous Poisson Process and double-chain Markov models
|
