Analyzing late ventricular potentials

The authors show the mechanism of high frequency attenuation introduced by signal averaging and how it can be avoided by applying a high resolution alignment approach, coupled with optimal filtering of the reference lead. Several authors suggest that ventricular late potentials are due to high frequency components caused by abnormal ventricular conduction. To investigate properly this hypothesis, it was necessary to adopt an alignment algorithm that avoids attenuation of high frequency components. Thus, the use of low alignment error techniques is of paramount importance. The proposed algorithm is of particular interest with respect to new trends in late ventricular potential analysis in the frequency domain. More advanced frequency analysis techniques have been recently proposed to gain higher frequency resolution. The wavelet transformation approach was proposed by several authors (among others Meste et al., 1994; Morlet et al., 1991; and Senhadji et al., 1990). The application of advanced frequency analysis techniques is due to the challenge implicit in the characteristics of late potentials, such as reduced amplitude and nonstationarity of these high frequency signals embedded in noise. This problem underscores the need for refined alignment before averaging to guarantee that a negative finding is not related to alignment errors. In conclusion, the optimal filtering and high resolution alignment technique yields improved reliability in detecting late ventricular potentials. This technique overcomes the drawbacks of signal averaging due to the low-pass filtering effect introduced by alignment errors