Comparative analysis of different fractal methods in studying post-ictal ECG signals of epilepsy patient

Abstract

Measuring EEG signals as a diagnostic tool for Epileptic patient is a well traveled way. Different non-linear statistical approaches have been applied over the past years on this signal to reveal the nature of connection between epilepsy and EEG. In this work we study the post-ictal ECG signals of epileptic patient collected from MIT-BIH database using mono-fractal methods as well as multifractal approach. We compare the results of the same statistical methods with healthy normal group. Result from monofractal analysis such as Rescaled range analysis indicates that the ECG signals of epileptic patients are anti-persistent in nature whereas for healthy normal people it is persistent. Detrended fluctuation analysis also confirms the same fact and declares that ECG signals of healthy normal people are more persistent and more correlated than epileptic patients. Finally we use the multifractal approach on the ECG signals. Result from the Multifractal detrended fluctuation analysis confirms that healthy normal people have higher degree of multifractality compared to epileptic patients.