Impact of signal preprocessing on the inverse localization of the origin of ventricular pacing

Investigators: ŠVEHLÍKOVÁ, J. – ZELINKA, J. – DOGRUSOZ, Y. S. – GOOD, W. -TYŠLER, M. – BEAR, L.

Within the cooperation in the Consortium for ECG imaging (www.ecg-imaging.org) we participated in the study about the influence of ECG signal preprocessing methods on the results of the inverse solutions. Data were provided by IHU-LYRIC, Bordeaux. ECG signals were recorded  during an in vivo experiment while stimulating the right ventricle of a porcine heart embedded in a torso tank; 128 signals on the torso as well as 108 signals on the epicardium were recorded simultaneously. The ECG signals were processed by combinations of 7 methods for high-frequency noise removal, and 4 methods for baseline drift removal. The influence of the mentioned methods on the accuracy of the stimulation origin localization by the inverse solution using a single dipole was examinated. The highest impact on the localization accuracy had the constant shift of the signal (offset) remaining after application of the high- and low-pass filters. After  removal of the residual offset the localization error decreased considerably from 5.5 cm to 0.7-1.3 cm. In related publications, the same processed signals were used as input data for different inverse methods aiming at recovering of the whole ventricular activation from the recorded body surface ECG signals.

Figure: Left: Root-mean-square value (RMS) of the originally measured ECG signal (30 s). Right: RMS of the averaged signal after application of 4 different methods for baseline wander removal; after application of 3 methods, significant signal shift remained.

 

Partners:

The Consortium for ECG Imaging – CEI, www.ecg-imaging.org  (11 organizations)

 

Related project:

VEGA 2/0125/19 „Measurement and modeling of the cardiac electrical field for noninvasive identification and interpretation of structural changes of the ventricular myocardium leading to ventricular arrhythmias.“

 

Publications:

  1. SVEHLIKOVA, Jana – ZELINKA, Jan – DOGRUSOZ, Yesim Serinagaoglu – GOOD, Wilson – TYSLER, Milan – BEAR, Laura: Impact of Signal Preprocessing on the Inverse Localization of the Origin of Ventricular Pacing. Computing in Cardiology 2018, Maastricht, Netherlands, Sept. 23-26, vol. 45, ISSN: 2325-887X DOI: 10.22489/CinC.2018.315
  2. BEAR, Laura – DOGRUSOZ, Yesim Serinagaoglu – SVEHLIKOVA, Jana – COLL-FONT, Jaume – GOOD, Wilson – vanDAM, Eelco – MacLEOD, Rob – ABELL, Emma – WALTON, Richard – CORONEL, Ruben – HAISSAGUERRE, Michel – DUBOIS, Remi: Effects of ECG Signal Processing on the Inverse Problem of Electrocardiography. Computing in Cardiology 2018, Maastricht, Netherlands, Sept. 23-26. vol. 45, ISSN: 2325-887X DOI: 10.22489/CinC.2018.070
  3. DOGRUSOZ, Yesim Serinagaoglu – BEAR, Laura – SVEHLIKOVA, Jana – GOOD, Wilson – COLL-FONT, Jaume – vanDAM, Eelco – DUBOIS, Remi – MacLEOD, Robert S.: Reduction of Effects of Noise on the Inverse Problem of Electrocardiography with Bayesian