Skip to Main content Skip to Navigation

Analyse et modèlisation de l'activité électrique du coeur dans le cas de pathologies ventriculaires

Abstract : This work is situated in the analysis of the arrhythmia rhythm disorders of the heart, and more particularly of those resulting from problems of the ventricles. Two problems were tackled which require specific methods in signal processing. The first relates to the study of episodes of Ventricular Fibrillation (VF) obtained from human beings with endocardial recordings indicated when implantable defibrillation are tested in the department of Cardiology of the university hospital (CHU) of Nice. The process generating the VF not being yet clearly identified, two assumptions were proposed: signal modelled by fundamental and harmonics either stable or time dependant. It is shown that we can highlight using short episodes (5 to 15 seconds) significant fluctuations of the pitch thanks to adaptive or evolutionary algorithms. A practical result, found for the first time in human subjects, is the correlation between the pitch of the episode of VF and the refractory period given by physician for various frequencies of stimulation. The second problem relates to recordings of the ECG high resolution using a system with 64 electrodes (Body Surface Potential Mapping: BSPM), in collaboration with the Institute of Biocybernetic of Warsaw. The original idea is to put forward a measure of the spatial dispersion of ECG wave shapes. We show the relevance of the information brought by this measurement by comparing a group of subjects having had a myocardial infarction with a healthy reference group. The measurement of dispersion is calculated, for each column of electrodes, starting from the differences in shape between each signal and a signal of reference. For this reference, we propose an averaged signal obtained by a technique of averaging, invariant by translation and scaling: Integral Shape Averaging, (ISA). We can allot to this reference an average position on the thorax allowing to define “an average path”. An interesting result observed on the real signals is that this path is an invariant, i.e., independent of the subject and the ECG wave. The practical repercussions of this work are in diagnosis assistance and the modeling of cardiac electrophysiology. One can quote as immediate prospects: (i) Concerning the first problem, the improvement of the technique of defibrillation (moment of the shock and intensity) as well as a finer analysis of the frequency variation if episodes of longer VF can be recorded (in animals). The same algorithms could be applied to episodes of auricular fibrillation and compared with the results of the literature. (ii) For the second problem, the study of the spatial dispersion of the ECG waveforms should be extended to populations having had auricular fibrillations. One can also consider a simultaneous study of the spatial dispersion of the forms (QRS or T) and the correlation between the pitch observed in an episode of VF and the refractory period.
Document type :
Complete list of metadatas

Cited literature [81 references]  Display  Hide  Download
Contributor : Estelle Nivault <>
Submitted on : Tuesday, November 27, 2007 - 2:35:35 PM
Last modification on : Wednesday, October 14, 2020 - 4:23:26 AM
Long-term archiving on: : Monday, April 12, 2010 - 5:16:19 AM


  • HAL Id : tel-00192273, version 1



Balkine Khaddoumi. Analyse et modèlisation de l'activité électrique du coeur dans le cas de pathologies ventriculaires. Sciences de l'ingénieur [physics]. Université Nice Sophia Antipolis, 2005. Français. ⟨tel-00192273⟩



Record views


Files downloads