Yu, Long2019-05-132019-05-132017-02https://hdl.handle.net/11299/202903University of Minnesota Ph.D. dissertation. 2016. Major: Biomedical Engineering. Advisor: Bin He. 1 computer file (PDF); 140 pages.Non-invasive cardiac electrical imaging techniques aim to directly visualize the intra-cardiac electrical activities and promise to assist in clinical diagnosis and treatment of cardiac arrhythmias, a family of highly dangerous disease leading to a hundreds of thousands of deaths and disabilities in the United States alone. In this dissertation, a line of investigations is included regarding cardiac electrical sparse imaging - a novel three dimensional cardiac imaging technique – from mathematical formulation of the imaging problem to validations studies covering numerical models, animal healthy and pathological models and patients with ventricular arrhythmias both during and before cardiac ablative procedures. With its spatiotemporal sparse problem, the novel imaging method incorporate cardiac electrophysiological features into the imaging process in order to achieve improvements in spatiotemporal resolution and, consequentially, general performance of the imaging technique. Simulation studies were conducted using a cardia automaton based heart-torso numerical model to verify the performance of the technique against various disturbances resembling the clinical challenges. Based on the numerical studies, rigorous animal studies using intra-cardiac simultaneous mapping technique were conducted to further validate the technique in biological systems such as canine and swine under healthy and pathological conditions such as myocardial infarction and congestive heart failure. Moreover, to evaluate the performance and compatibility of the technique in real life clinical challenges, further in-procedural and pre-procedural clinical studies were carried out on patients with ventricular arrhythmias. High accuracy and strong robustness can be observed by comparing the imaged activations with the mapped ones. The imaging technique achieved good performance not only in numerical simulations, but also in animal models with complex pathological conditions. Strong correlation was observed from the comparisons on ventricular arrhythmias with both focal and reentrant patterns. In further clinical studies, the technique also achieved good performance in localizing the arrhythmia foci and imaging the 3D activation pattern during the arrhythmias. The promising results shown in the studies indicate that the technique has good capability in visualizing the whole heart electrical activities and in providing key information such as arrhythmia foci and reentry pathways to assist in clinical practice in various pathological conditions.enArrhythmiaElectrical tomographyElectrophysiologyImagingThesis or Dissertation