An integrated signal processing environment for detection of sleep disordered breathing in children using spectral and nonlinear dynamic measures of heart rate variability signal.

摘要

This thesis emphasizes an application of novel signal processing techniques to detect Sleep Disordered Breathing (SDB) in children using their Heart Rate Variability (HRV). The HRV has been derived using an Enhanced Hilbert Transform (EHT) algorithm with a missing peak correction capability. Various signal processing techniques have been implemented on the HRV signal to obtain sensitive measures used in detection of SDB. All these techniques have been integrated into a user friendly interface.; The algorithms were implemented in MATLAB 6.5 and the Graphical User Interface (GUI) in LabVIEW 7.1. The GUI provides a complete patient report with the summary of all the analyses performed. All the algorithms were developed, validated, and implemented on data from Physionet's ECG databases and children data obtained from Adelaide Women's and Children's Hospital (WCH). The final results demonstrated that the EHT derived HRV yielded 100% accuracy when checked manually. The results obtained from the data files JT and NS have shown 80% sensitivity, 70% specificity giving an overall accuracy of 75%. The analyses performed were integrated into a friendly and easy to access software and the features obtained demonstrated needed sensitivity to detect SDB.