Ventricular Late Potentional Analysis by Adaptive Filtering and Cross-Spectral Technique

This paper presents a novel technique devoted to the detection of late potentials in frequency domain which consists in an adaptive filtering of the ST segment and in a cross-spectral analysis of the filter output. The approach is based on the evidence that late potentials are related to frequency components higher than those present in the high-resolution electrocardiogram of healthy subjects. We also outline the correlation between late potentials and the abrupt modification of the spectro-temporal representation of the ST segment by adopting a cross-spectral technique. We propose an index to measure the presence of anomalous frequency components which was proved effective by means of the analysis of synthetic signals as well as of real recordings. Particularly, we show the effectiveness of the proposed index in the clinical assessment of the arrhythmogenic right ventricular disease