יובל אהרונוביץ

תלמיד המחלקה לתואר שני ירצה בנושא:

Reconstruction of Retrospective Cardiac Activity using Inverse Cellular Automata Model - a simulation study

Introduction: ectopic activity in the cardiac muscle can initiate irregular electrical activity and is one of the major sources of arrhythmogeneity. A possible surgical procedure to eliminate ectopic foci is tissue ablation; however, in order to successfully ablate these sources and to avoid ablating vital tissue, their exact location needs to be identified. Here we propose a novel method for assessing the temporal and spatial characteristics of ectopic foci by reconstructing the cardiac electrical activity back in time. Such retrospective reconstruction may provide the clinician with a visual tool that can unveil the source of any irregular activity. Methods: we employed a phase-map based cellular automata (CA) model for tissue excitation and propagation properties, and established a set of inverse rules to reconstruct the past activity. CA models describe the excitation state of the cardiac tissue based on a set of pre-defined and usually deterministic rules, which govern the transition between a finite numbers of discrete states. These models, despite their simplicity and abstractness, have been demonstrated to successfully simulate complex electrophysiological phenomena in cardiac electrical conduction. We tested our model using three configurations: 1) a 3D cube surface simulating the endocardial layer of a human heart chamber using a CA model for both the forward and the inverse solvers; 2) a 2D matrix simulating human atrial tissue using a detailed biophysical model for the forward solver, and a CA model for the inverse solver; and 3) application of the inverse CA solver for data obtained from an optical mapping video of a ventricular neonatal rat monolayer. Results and conclusions: retrospective propagation patterns were successfully reconstructed in most cases, except for the appearance of minimal boundary artifacts in the 2D models, or impaired reconstruction when the sampling time was significantly delayed relative to the stimulation time. Point stimulations were detected with low temporal error (~0.2-12ms, depending on model parameters) and with low spatial error (~0.07-1.6mm, compared to a 30-mm long geometry). We also demonstrated the capability of the model to reconstruct stable re-entry activity; however, future research should be conducted to account for relative refractory period in the inverse solver so that more complex propagation patterns (e.g., meandering spiral waves) could be reconstructed.

העבודה נעשתה בהנחייתו של ד"ר שרון זלוצ'יבר, המחלקה להנדסה ביו-רפואית, אוניברסיטת תל-אביב

ההרצאה תתקיים ביום ראשון 28.10.2012, בשעה 14:45,

בחדר 315, הבניין הרב תחומי, אוניברסיטת תל אביב