The analysis of heart rate variability using wavelet transform and artificial neural networks

Wavelet Transform that is used for analyses of non-stationary signals as biomedical signals is one of the most important methods. Heart Rate Variability (HRV) signals having discontinuities and very small frequency ranges are one of the most appropriate for Wavelet Transform.This study contains HRV analyses which are applied on ventricular tachyarrhythmia database that isn?t analyzed as detailed in the literature. This scope consists of HRV analyses with ectopic and without ectopic using Discrete Wavelet Transform (DWT), the determination of Sympathovagal Balance (SB), the detection of frequency bands energy values and compare of its results using DWT and Wavelet Packet Transform (WPT), to optimize the frequency band shifts in DWT using WPT, the automatic detection of SB using DWT and Artificial Neural Networks (ANN), the identification of domination sub-bands using WPT and ANN, analysis of the Very Low Frequency (VLF) band that is defined occasionally in the literature, and the evaluation of all of the obtained results in the Ventricular Tachyarrhythmia database.This thesis is the first study including specifications that HRV analysis with WP, interpretation of VLF band, automatic detection of SB and identification of dominant frequency sub-bands. Obtained results, proposed actual methods and evaluation of Ventricular Tachycardia (VT) and Ventricular Fibrillation (VF) resolves an important drawback.