Temporal and Spatial Regularization for Imaging Cardiac Electrical Function

Noninvasive electrophysiology has attained more interest during the last years because of the availability of multichannel cardiomagnetometer and body surface potential mapping systems. The main focus is the determination of the impressed electrical sources in the human heart from body surface potential (BSP) and magnetocardiogram (MCG) mapping data. The noninvasive procedure should help to shorten invasive surgery time e.g. in the case of catheter ablation and help to decide whether an invasive procedure is necessary or not. The objective of this project is the development of inverse algorithms for the estimation of the activation time (AT) map including temporal and spatial regularization considering an inhomogeneous boundary element (BE) torso model. Electrical anisotropy is neglected in the BE approach. The estimation of the AT map (phase 0) enables the imaging of the propagation pattern on the endocardium and epicardium. The applied analytical description of the surface transmembrane potential for phase 0 is based on an analytical function and includes the following parameters: activation time, rise time, action potential amplitude and resting membrane potential. Each of these parameters, except the activation time are assumed to be constant at the BE nodes of the surface of the heart. The AT map is estimated applying the critical point theorem and a subsequent nonlinear optimization approach.