Non-invasive epicardial imaging of human ventricular fibrillation

The spatial distribution of ECG torso potentials during ventricular fibrillation (VF) might provide useful information about underlying electrical dynamics. We used an inverse solution technique to non-invasively construct images of epicardial activity of human VF. A 120-lead mapping system was used to record body surface potential maps (BSPM) from eight anesthetized patients during VF induction following implantable defibrillator placement. Epicardial potential maps of VF were derived mathematically by inverse solution using Tikhonov regularization and L-curve method, assuming a homogenous bounded torso. To assess accuracy, VF was simulated in a large-scale numerical anisotropic heart model incorporating ionic currents. Potential fields were simulated within the torso volume conductor and on the body surface by forward solution to assess the degree of spatial information attenuation. Calculated inverse solution was compared with epicardial activity on the heart model. Spatial features of VF attenuate with distance from the heart due to the volume conductor; however, the model results demonstrate that inverse solution can resolve epicardial VF patterns to a limited, but potentially useful, degree with larger spatial scales being preserved.