Statistical analysis of fluctuations in the ECG morphology

We investigate ECG recordings of healthy subjects (H subjects) and those with heart failure (HF subjects) based on a new measure derived from the shape of the ECG signal during each heart beat. From this measure we construct sequences, called "morphograms", which are complementary to sequences of successive interbeat time intervals, commonly called "tachograms". The morphograrns show long-range anticorrelations between the increments of their successive elements, analogous to the behavior found previously in the tachograms. By studying the average strength of the fluctuations of the increments on certain scales, we find that it is possible to completely separate the two groups of H and HF subjects. The separation is comparable, and in certain cases even more pronounced than that found earlier based on the tachograms. (c) 2005 Elsevier B.V. All rights reserved.