Dr. Sujay Datta

University of Akron, Department of Statistics

Modeling Sleep Apnea Events Data from Pacifier Users for Improved Diagnosis of OSA 

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 and corrupts the thermistor readings. In the absence of any alternative, the current standard practice is to discard all the artifact-corrupted portions of the thermistor data. The resulting loss of information leads to an underestimation of the Apnea Hypopnea Index (AHI), the basis for an OSA diagnosis. One remedy would be to try to reconstruct the thermistor signal for the artifact-corrupted portions by somehow removing the artifacts. Researchers are now exploring two other information sources (blood oxygen saturation dips indicated by a pulse-oximeter and occurrence of arousal events) that are statistically associated with the thermistor readings and therefore may be helpful in statistically modifying the AHI. Here we first describe the setup and elaborate on the nature of the problem. Then we briefly describe the ‘engineering’ solution via signal reconstruction. Finally, we focus on a couple of statistical solutions to this problem via Bayesian modeling and discrete stochastic modeling.