Mathematical modeling and reconstruction of the electrical activity of the heart using neural networks

The talk is dedicated to modern approaches for non-invasive study of the heart electrical activity based on the solution of the inverse electrocardiography problem using mathematical modeling methods and neural network algorithms. The principles of constructing the three-dimensional maps of myocardial activation based on a standard 12-channel ECG are presented. The key stages of reconstruction are considered, from the modeling of heart and torso geometry to electrical sources identifying in the heart and the construction of isochronous activation maps. The stages of technology integration into clinical practice are described.