A virtual sleep laboratory

Cheng Shi, Mehrdad Nourani, Gopal Gupta, Lakshman Tamil

Abstract: A virtual sleep laboratory capable of providing low cost and personalized sleep apnea monitoring using a smartphone has been demonstrated. This system can be used both at home and clinical care settings. This fully automated system derives sleep apnea information from ECG and/or oximeter and uses support vector machine classifiers to classify the events as apena or non-apnea episodes. The system has been tested using Physionet Apnea-ECG database and yields an accuracy close to 90% for both ECG and oximetry sensors. The system is a fully automated internet based system capable of supporting monitoring of hundreads of people at the same time and is being readied for a limited patient trial. This system will be demonstrated at the conference.