Bipolar disorder (BD) affects a significant portion of the population of the world, yet there has been limited success in developing novel treatments for the disorder. One of the major reasons for this dearth is the absence of suitable animal models for BD. Traditionally, animal models of human phenomena have been evaluated based on similarity to the human syndrome, response to appropriately corresponding medications, and the degree to which a model supports a common mechanistic theory between the human disorder and the model itself. The following review emphasizes the use of 'reverse translation', drawing on patient-based findings to develop suitable animal models for BD. We highlight some examples of this strategy, emphasizing their construct validity as a starting point. These studies have produced informative models that have altered the expression of genes/pathways implicated in BD, including the point mutation D181A of mouse mitochondrial DNA polymerase (POLG), glutamate receptor 6 (GluR6), Clock, extracellular regulated kinase 1 (ERK1), glycogen synthase kinase-3beta (GSK-3beta), B-cell lymphoma 2 (Bcl-2) and Bcl-2-associated athanogene (BAG-1). These studies demonstrate that this method is useful, viable and deserves attention in new efforts to generate animal models of BD.