Spectral Properties of ECG Signal

Abstract

Spectral analysis belongs to widely used method for biomedical data exploration. In order to interpret spectrum correctly, prior knowledge how various phenomena affect the spectrum of analyzed signal is desirable. Spectrum of ECG signal is affected by waveform component shapes, their time positions within cardiac cycle as well as by regularity of heart period. This paper deals with effect of heart rate variability on ECG signal spectrum. Three representations are proposed to model variable heart timing in the case of normal sinus rhythm (pulse process with time jitter, random inter-pulse interval and integral pulse frequency modulation model). Properties of corresponding spectra are theoretically studied and compared with real-world signals from PhysioBank data archive.