Simulation and experimental studies of the factors influencing the frequency spectrum of cardiac extracellular waveforms

Abstract

Spectral analysis of electrocardiographic signals has been proposed as a tool to detect features reflecting cardiac diseases, such as ventricular hypertrophy, myocardial infarction, and a predisposition to sustained ventricular tachycardia. The lack of a theoretical basis to address this question prompted the authors to undertake a simulation study using a bidomain volume conductor model of a strip of cardiac tissue, combined with Fourier analysis, and electrograms recorded from an isolated right atrial canine preparation. In the crista terminalis, the bandwidth of the normal electrogram was 840 ± 200 Hz (mean ± SD) during longitudinal propagation and 660 ± 370 Hz during transverse propagation. During premature stimulation, signal bandwidth and propagation velocity increase with the coupling interval. In the model, a linear combination of V̇ max and propagation velocity values allows simulation of the various features of premature excitation. V̇ max is the major determinant of the high-frequency content of the signal. An important decrease in the high-frequency content of electrograms occurs when the recording electrode is moved away from the preparation or the simulation model; at distances larger than 1–5 mm, the bandwidth levels off to a value of 50–120 Hz. Partial blockade of axial current flow in the direction of propagation due to microscopic discontinuities and variable activation delays at these discontinuities may be the cause of fragmented activity in necrotic myocardium, which is associated with a reduced bandwidth. Thus, short- and long-term effects of ischemia followed by infarction, such as decreased propagation velocity, decreased action potential upstroke, and fragmentation, tend to decrease the electrocardiographic bandwidth.