[show abstract][hide abstract] ABSTRACT: Optical mapping has become an indispensible tool for studying cardiac electrical activity. However, due to the three-dimensional nature of the optical signal, the optical upstroke is significantly longer than the electrical upstroke. This raises the issue of how to accurately determine the activation time on the epicardial surface. The purpose of this study was to establish a link between the optical upstroke and exact surface activation time using computer simulations, with subsequent validation by a combination of microelectrode recordings and optical mapping experiments. To simulate wave propagation and associated optical signals, we used a hybrid electro-optical model. We found that the time of the surface electrical activation (t(E)) within the accuracy of our simulations coincided with the maximal slope of the optical upstroke (t(F)*) for a broad range of optical attenuation lengths. This was not the case when the activation time was determined at 50% amplitude (t(F50)) of the optical upstroke. The validation experiments were conducted in isolated Langendorff-perfused rat hearts and coronary-perfused pig left ventricles stained with either di-4-ANEPPS or the near-infrared dye di-4-ANBDQBS. We found that t(F)* was a more accurate measure of t(E) than was t(F50) in all experimental settings tested (P = 0.0002). Using t(F)* instead of t(F50) produced the most significant improvement in measurements of the conduction anisotropy and the transmural conduction time in pig ventricles.