A bootstrap procedure for classifying QRST integral maps

The analysis of beat-to-beat variability of QRST integral maps (V-IM) enabled the authors to detect specific changes of the cardiac potential field in hypertrophic subjects with high arrhythmia vulnerability. A classification of QRST integral maps was attempted by splitting the variability of maps into various components, submitting experimental data to linear discriminant analysis (ADL) and estimating the relative bias by means of a bootstrap procedure. The definition of V-IM makes it possible to consider time- and space-dependent effects. ADL was applied to two components of V-IM relating to the scale and shape of the signals, and a correct classification was achieved in 100% of the cases.<
>