Physically motivated projection of the electrocardiogram

We present a physically motivated projection of the 12 lead electrocardiogram, based on a deep learning model trained on 21,799 recordings from the PTB-XL database, that allows the separation between left and right ventricular activity. We show that the observed dyssynchrony is in accordance with ot...

Full description

Saved in:
Bibliographic Details
Published inarXiv.org
Main Authors Wildowicz, Sebastian, Gradowski, Tomasz, Figura, Paulina, Olczak, Igor, Sobiech, Judyta, Buchner, Teodor
Format Paper
LanguageEnglish
Published Ithaca Cornell University Library, arXiv.org 15.10.2024
Subjects
Algorithms
Electrocardiography
Electrodes
Imaging
Inverse problems
Population distribution
Wave fronts
Online AccessGet full text

Cover

More Information
Summary:We present a physically motivated projection of the 12 lead electrocardiogram, based on a deep learning model trained on 21,799 recordings from the PTB-XL database, that allows the separation between left and right ventricular activity. We show that the observed dyssynchrony is in accordance with other experimental methods. The PhysECG method may be considered a reduced solution to inverse problem of ECG imaging, which reconstructs the epicardial potentials and enables a good qualitative assessment of pathological changes. The foundations of the model are based on a molecular approach to biopotential propagation. We present the quality metrics for the reconstruction based on the model on testing set selected from the PTB database. The heart's activity in view of the method is decomposed into two disjoint processes: the passage of the electric activation wavefront and the response of cardiomyocytes. We introduce the idea of the electrode-resolved activity function, which represents the mass of the ventricle in Phase 0 of action potential within the lead field of each electrode. We also show the effects of the population distribution of the anatomical position of the septum with regard to the direction of the V3 electrode. We present and discuss four healthy controls: male and female, right bundle branch block, and anterior myocardial infarction. We demonstrate that the results obtained using PhysECG are in accordance with the functional changes evoked by pathology. The algorithm may be used to reconstruct the activity on any 12 lead or Holter recordings. The method offers limited power of ECG imaging based entirely on sole ECG, which presents an affordable solution accessible for low and middle-income countries where access to resource-demanding imaging modalities is limited.
ISSN:2331-8422