A. S. Krasichkov

The results of the ECG statistical structure investigation presented here is based on a wide bank of Holter monito-rograms recorded by the «Kardiotechnika» ECG monitor kindly provided by the Pavlov State Medical University. Using enclosed original software the provided monitorograms have been converted to text files that allowed post-processing. The discretization rate was 128 Hz. A large number of available records allowed to preselect solely high quality ECG records for further analysis. On the basis of the physiological argumentation and ECG processing results, the following statements have been confirmed: (i) the Q-T interval duration is proportional to the R-R interval by a Bazett formulae, with an individual proportionality coefficient; (ii) under the same heart rate the deviations in the duration of the intervals is comparable with the R-wave duration; (iii) when the heart rate changes, the ECG shape remains, but is rescaled in time; (iv) the durations of all intervals and waves are approximately linearly related to the Q-T interval duration with individual amplitude- and temporal-rescaling parameters; (v) wave fluctuation amplitudes are descibed by Pearson-type distributions and can be to a certain extent approximated by the first three components of an Edgeword series; (vi) no significant persistent correlations between wave amplitudes can be observed. The outlined properties of an individual ECG have been experimentally confirmed using the local correlation coefficient analysis.