P-wave complexity in normal subjects and computer models
Mark Potse, Theo A. R. Lankveld, Stef Zeemering, Pieter C. Dagnelie, Coen D. A. Stehouwer, Ronald M. Henry, André C. Linnenbank, Nico H. L. Kuijpers, and Ulrich Schotten
J. Electrocardiol. 49:545-553, 2016 doi:10.1016/j.jelectrocard.2016.05.005 (open access)

manuscript PDF


Background P waves reported in electrocardiology literature uniformly appear smooth. Computer simulation and signal analysis studies have shown much more complex shapes.

Objective We systematically investigated P-wave complexity in normal volunteers using high-fidelity electrocardiographic techniques without filtering.

Methods We recorded 5-minute multichannel ECGs in 16 healthy volunteers. Noise and interference were reduced by averaging over 300 beats per recording. In addition, normal P waves were simulated with a realistic model of the human atria.

Results Measured P waves had an average of 4.1 peaks (range 1-10) that were reproducible between recordings. Simulated P waves demonstrated similar complexity, which was related to structural discontinuities in the computer model of the atria.

Conclusion The true shape of the P wave is very irregular and is best seen in ECGs averaged over many beats.


For more than a year we have been trying to build a computer model of the human atria that would reproduce normal P waves, such as seen in a standard ECG or in the literature. However, no matter what we tried the simulated signals were always much more complex than that. Finally we started questioning what we saw in the literature, and in the ECGs we had. Were the measurement techniques really good enough to detect the very complex signals that we simulated, if they were actually there in reality? After all, P waves are very small, close to the amplitude of the noise in the recordings, and therefore the standard approach is to apply filters that remove small-scale details. We used the most accurate methods that we reasonably could, to measure P waves in normal subjects without filtering. As this paper shows, we found that the model was right: indeed the shape of the P wave is very complex.

A preliminary version of this work has been presented at the 40th International Congress on Electrocardiology, Glasgow, August 2013.


This study was supported by the 7th Framework Program of the European Union through a Marie Curie International Reintegration grant (number 256493) to M.P. and through the Collaborative project EUTRAF (number 261057).

A.C.L. was supported by the Dutch Technology Foundation STW under grant number 10959.

Simulations were performed on an SGI Altix 4700 supercomputer at Université de Montréal, operated by Calcul Québec and Compute Canada, and financed by the Canada Foundation for Innovation (CFI), NanoQuébec, RMGA, and the Fonds de recherche du Québec - Nature et technologies (FRQ-NT)