Quantitative comparison of pre-mortem ECG's with those reconstructed from activation data of a revived heart

The pioneering work of Durrer and associates has made available unique data showing activation isochrones in human hearts. To be certain that the complex experimental procedures do not seriously affect the activation sequence and also to validate accepted basic premises concerning the genesis of the electrocardiogram (ECG), it appears to be essential to show a reasonable correspondence between the subject's ECG before death and ECG's reconstructed from the activation data and heart-lead models. This paper reports the results of such a study. The comparison is evaluated in quantitative terms using several types of correlation coefficient. The significance of these is further evaluated through comparisons of similar coefficients from a random sample of normal records. The correlation coefficient considered to have the most general applicability gives a value of .80; only one subject from the random group showed a higher correspondence. A particular set of leads chosen to minimize the effects of inhomogeneities not accounted for in the modelling process, clearly selected the subject's records from all others with a correlation coefficient of .91 compared to .87, the highest from the random set. Considering the practical limitations on the accuracy of the models underlying these tests, the results are considered, with certain exceptions, to be supportive of the accuracy of the activation experiments and the assumption of a uniform double layer as the ECG source.