Effects of analyzed signal duration and phase on the results of fast fourier transform analysis of the surface electrocardiogram in subjects with and without late potentials.

A microcomputer-based system capable of performing time- and frequency-domain analysis on the same set of acquired and signal-averaged raw data was used to correlate late potentials detected in the time domain with the results of frequency-domain analysis. Ten patients with spontaneous or inducible sustained ventricular tachycardia (VT) known to have abnormal late potentials in the time-domain signal-averaged electrocardiogram and 10 normal subjects without late potentials were studied. Fast Fourier transform analysis was performed on a segment that comprised the last 40 ms of the QRS and the ST segment up to the onset of the T wave as identified visually. The high-frequency signal content, expressed as the area ratio or the peak magnitude, was found to be markedly dependent on the length of the analyzed QRS-ST segment. A change of as little as 3 ms in the duration of the estimated QRS-ST segment changed the results of the frequency analysis across proposed boundaries of normalcy in normal subjects and in patients with VT. This resulted in both false-negative and false-positive conclusions. Similar results were obtained when the effects of varying analyzed signal length or phase were studied using a pure synthesized sine wave signal. For frequency analysis to be clinically useful and reproducible, standards of normalcy must be established for a signal region of fixed duration or the technique must be modified so as to be insensitive to duration of signal sample.