{"id":1700,"date":"2018-12-16T00:00:33","date_gmt":"2018-12-15T23:00:33","guid":{"rendered":"https:\/\/www.um.sav.sk\/en\/?p=1700"},"modified":"2020-11-09T13:56:51","modified_gmt":"2020-11-09T12:56:51","slug":"3-geometric-delimitation-of-sources-in-non-invasive-localization-of-the-ventricular-arrhythmia-lesion","status":"publish","type":"post","link":"https:\/\/www.um.sav.sk\/en\/selected-results\/3-geometric-delimitation-of-sources-in-non-invasive-localization-of-the-ventricular-arrhythmia-lesion\/","title":{"rendered":"Geometric delimitation of sources in non-invasive localization of the ventricular arrhythmia lesion"},"content":{"rendered":"

Investigators: J. \u0160vehl\u00edkov\u00e1, M. Teplan, M. Ty\u0161ler<\/p>\n

In the simulation study, we solved the inverse task of ECG electrocardiography in order to evaluate the possibility of localization of premature ventricular activity (PCA) simulated at 8 different points in a realistic model of the ventricles. We used simulated data provided by the Institute of Biomedical Engineering Karlsruhe, Germany within the cooperation of the Consortium for ECG Imaging. Noise with a signal-to-noise ratio (SNR) of 20, 15 and 10 dB was added to the simulated surface potential maps (BSPM). To solve the inverse problem, we investigated three types of models of the equivalent electrical generator of the heart: transmembrane voltages, epicardial potentials, and dipoles. Instead of regularizing the generally poorly conditioned inverse problem, we used the assumption of a point equivalent source as a representative of the initial PCA. The location error (LE) of individual PCA sources was more position dependent than generator type. If the PCA source was located on the free wall of the left ventricle, or on the anterior wall of the right ventricle, the LE was stable up to 20 mm and robust to noise up to an SNR of 10dB. For SNR 10dB, the results were unstable, and if the PCA source was in the septum or posteriorly in the right ventricle, the LE reached up to 50 mm. The joint evaluation of the results for different types of generators led to a reduction in LE for all simulated positions of the PCA source.<\/span><\/p>\n

\"\"<\/a><\/p>\n

 <\/p>\n

Fig. 1:<\/strong><\/span> Left: Simulated PVC positions (RV – right ventricle, LV – left ventricle, ANT – anterior, POST – posterior, LAT – lateral, APEX – apical, EP – epicardial, CEN – central).<\/span> Right: localization errors [mm] at maximum noise (10 dB) for individual positions of PVC sources and different types of model generators shown by box graphs.<\/span><\/em><\/span><\/p>\n

 <\/p>\n

Related projects:<\/strong><\/p>\n