Mechanisms of defibrillation remain poorly understood. Defibrillation success depends on the elimination of fibrillation without shock-induced arrhythmogenesis. We optically mapped selected epicardial regions of rabbit hearts (n=20) during shocks applied with the use of implantable defibrillator electrodes during the refractory period. Monophasic shocks resulted in virtual electrode polarization (VEP). Positive values of VEP resulted in a prolongation of the action potential duration, whereas negative polarization shortened the action potential duration, resulting in partial or complete recovery of the excitability. After a shock, new propagated wavefronts emerged at the boundary between the 2 regions and reexcited negatively polarized regions. Conduction velocity and maximum action potential upstroke rate of rise dV/dt (max) of shock-induced activation depended on the transmembrane potential at the end of the shock. Linear regression analysis showed that dV/dt(max) of postshock activation reached 50% of that of normal action potential at a V(m) value of -56.7+/-0.6 mV postshock voltage (n=9257). Less negative potentials resulted in slow conduction and blocks, whereas more negative potentials resulted in faster conduction. Although wavebreaks were produced in either condition, they degenerated into arrhythmias only when conduction was slow. Shock-induced VEP is essential in extinguishing fibrillation but can reinduce arrhythmias by producing excitable gaps. Reexcitation of these gaps through progressive increase in shock strength may provide the basis for the lower and upper limits of vulnerability. The former may correspond to the origination of slow wavefronts of reexcitation and phase singularities. The latter corresponds to fast conduction during which wavebreaks no longer produce sustained arrhythmias.