Analysis and development of integrated index for diagnosis of coronary artery disease based on heart rate signals

Abstract

Coronary Artery Disease (CAD) is one of the very common type of cardiovascular diseases which is the killer of world’s 7.4 million population. Coronary artery disease is characterized by the narrowing and hardening of arteries supplying blood to the muscles of heart, owing to the deposition of waxy substance called plaque in them. The consequence of coronary artery disease may be a heart attack or heart stroke. Hence, the patient’s suffering from coronary artery disease are always at the risk of death. In the world, where trained cardiologists may diagnose coronary artery disease with manual errors, computer aided diagnostics methods may be of great help. Thus in this thesis we have proposed an efficient way to diagnose coronary artery disease using heart rate signals. We have used Empirical Mode Decomposition (EMD) to decompose the heart rate signal into Intrinsic Mode Functions (IMFs). The features namely: Area of Second Order Difference Plot (SODP area), Area of Analytical Signal Representation (ASR area), Amplitude Modulation (AM) bandwidth, Frequency Modulation (FM) bandwidth and Fourier Bessel expansion (FBE) based mean frequency are extracted from these IMFs of different signals. These features are then subjected to Kruskal-Wallis statistical test to check their statistical significance.In the next part of this work, we have used the same dataset and derived the modes of each signal using Empirical Wavelet Transform (EWT). The same set of features is derived from these obtained modes of the signals. These features are again tested for their statistical significance and the best three features are selected to derive an integrated index for discrimination between normal and coronary artery disease heart rate signals using a single value.