Spatio-temporal cardiac pacing sites localization and time varying pericardium potential maps projection using ECG precordial leads and a single moving dipole model

A novel non-invasive method for the spatiotemporal localization of the sites of strongest cardiac activity, and for the creation of time varying Pericardium Potential Maps (PPM), with the use of patients\u27 ECG precordial leads, is proposed in this Thesis study. Compared to previous studies, which analyze electrocardiograms in either time domain or spatial domain, the proposed method has the advantage of a simultaneous spatiotemporal electrocardiograph analysis and a 3-D visualization of pericardium potentials maps, as well as, the pericardium surface polarization patterns during the cardiac cycle. The spatial properties added to the electrocardiogram allow for the analysis of specific regions inside the human heart where potential cardiac malignancies are suspected to occur. In this thesis, the concept of electrocardiography was extended using advanced bioelectromagnetism theory. The propossed MATLAB-based software makes use of the single-moving dipole model, optimized in location and magnitude with respect to the measured precordial leads, and of a realistic Finite Elements Method torso model. The use of the single-moving dipole model allows the specific localization of cardiac pacing sites, i.e the regions within the human heart in which the electrical activity is being generated. The PPMs are displayed simultaneously with precordial leads to allow a 3D visual synchronization between the time varying color coded potential map and the ECG waveforms, which may indicate potential cardiac malignancies. In this Thesis study, the addition of a time dependent visualization of the PPM to the single-moving dipole model resulted in the exact localization of cardiac pacing sites in both space and time. This spatiotemporal analysis is useful for clinical applications since it characterizes the source of the human heart\u27s electrical activity to a specific region or dipole location inside the heart at a specific moment on the cardiac cycle. Then for each time sample, a single dipole is chosen to be responsible for the generation of the set of precordial signals a function of its moment magnitude. Therefore, the dipole moment location changes in magnitude and origin over time. The proposed software was implemented in the analysis of 15 normal patients and 15 patients with cardiac abnormalities. For each case, 20 different sites inside the heart were considered as a possible origins of cardiac activity at each instant of time during a complete cardiac cycle. The hypothesis behind this study is that the dipole with the greatest dipole moment magnitude generates the entire set of precordial signals at a specific time moment. Moreover, the location of the dipole with the greatest dipole moment also indicates the origin of the cardiac activity inside the heart. The logic used to determine whether a patient presents an abnormality relies in the creation of a normal population which provides the necessary parameters for the creation of normal ranges. Those normal parameters establish the expected location of the cardiac activity origin at each one of the sections of a cardiac cycle: the P-wave, QRS-complex, and T-wave. Data are analyzed patient by patient; if a patient is found to have the strongest dipole location out of expected range in one or more cardiac cycle sections, the patient is considered to have an abnormality. Analysis is then fully complemented by the visual inspection of the actual location of the origin of the cardiac activity and the inspection of the polarization and depolarization patterns at the pericardium provided by the PPM projection. Results showed consistently that for normal patients, sources of strongest cardiac activity were located in the atrial region for Q-wave, and the in the ventricle region for both QRS complex and T-wave, whereas for abnormal patients there was no consistency in such locations. The software identified succesfully potential cardiac malignancies and their possible locations inside the heart in 93.33% of the abnormal patients\u27 ECGs. Additionally the inclusion of spectral analysis of the original signals aimed for preliminary arrhythmia detection, signal period calculation and wave segmentation, allowed for robust patient application, since the software parameters needed not to be modified when analyzing different patients. The concepts proposed and explored in this Thesis work provided expected results. Therefore the goals of this research study were successfully achieved