In recent years there has been renewed interest in the importance and analysis of the clinical P wave in the development of cardiovascular disease. Current research has focused upon P wave morphology and its diagnostic importance in the assessment of atrial disorders and cardiac conditions in general. The analysis of P waves in clinical electrocardiogram recordings presents several problems arising from the poor signal to noise ratio of the P wave. The variability of the P wave, its susceptibility to influences from anterior processes and physiological alterations in elderly patients all affect the conspicuousness of P wave features. The current interest in the morphology of the P wave is due to advances in morphological analysis of other physiological signals and the inadequacies of classic cardiological measures of the clinical P wave.

Processing of the ECG has been a research topic within the signal processing field for several decades, accelerated by the more recent widespread acceptance of digital recording techniques in clinical cardiology practices. Improvements in digital processing and recording technologies coupled with reduction in their costs have seen a proliferation of signal processing concepts being applied to clinical cardiology for both practical and exploratory causes. A recent area of interest in biomedical signal analysis problems is the application of wavelet transforms.

The use of the wavelet transform has gained popularity in time-frequency analysis because of the flexibility it offers in the choice of analysing basis functions. This is particularly advantageous in the analysis of characteristic physiological signals where wavelet basis functions can be chosen to highlight the idiosyncrasies of signal’s morphology.

Proposed in this dissertation is a novel application of the wavelet transform to the analysis of P wave morphology recorded in clinical cardiology settings. These characteristics are sensitive to subtle changes in the morphology of the P wave which is indicative of variations in the electrical wavefront propagation through the atria. This novel method of characterising the P wave morphology uses both temporal and frequency domain aspects of the ECG signal.

This novel technique was applied to the problems of classification of P wave morphology and estimation of physical atrial dimensions. The investigations presented in this dissertation show that this novel analysis method compares well against traditional ECG descriptors in identifying abnormal P wave morphology. Additionally, several wavelet transform characteristics of the P wave proposed here provided significantly better indications of physical atrial dimensions than traditional scalar descriptors of the ECG P wave. This result has great potential benefit for rural and regional cardiovascular healthcare in areas where access to more elaborate cardiac assessment technologies is limited or non-existent.