In this thesis, we offer a detailed description of the Koivumäki model, a computational model for the electrophysiology of the human atrial cardiomyocyte. The model is examined to see to what extent it exhibits cardiac alternans, an important dynamical substrate for atrial fibrillation. We show that the Koivumäki model reproduces calcium-driven alternans through refractoriness of ryanodine receptor release channels. We observe large differences in alternans behavior between the 2014 and 2015 formulations of the model, and show that these differences can partly be attributed to different formulations of the L-type calcium current in the two model variants, especially so for the maximal L-type channel conductance. In addition, we give a review of four proposed models for cardiac myofibroblasts. We use the Maleckar model for the human atrial fibroblast to extend the Koivumäki model to include fibroblast-myocyte coupling. The effects of the coupling on the virtual Koivumäki myocyte are examined as the coupling strength and number of fibroblasts are varied. We show that fibroblast-myocyte coupling leads to significant changes in the myocyte action potential, which we attribute mainly to a depolarization of myocyte resting membrane potential. Action potential amplitude and upstroke velocity are both reduced, following fibroblast-myocyte coupling. We show this to be caused by decreased availability of fast inward sodium current due to depolarized resting potential of the myocyte. We see a close-to-linear decrease in action potential amplitude and upstroke velocity of the action potential with increasing RMP, up until the point where the myocyte fails to excite entirely. Having explored the cardiac alternans behavior of the model, as well as extending it to include fibroblast-myocyte coupling, will be helpful for further development and use of the model in atrial fibrillation research.