How do animals use vision to control behavior? When light falls in the mammalian eye, the incoming photons are sensed by photoreceptors in the outer retina and transduced into electric activity of neurons. Neurons process the electrically encoded information and pass it on to neurons in the inner retina, subcortical brain regions and, finally, to cortex. Each of these steps adds a layer of complexity to the resulting computation. Ultimately, the light-induced activity reaches muscles and causes behavior. While we have a general understanding of the neural circuitry within the retina, many of the details remain unknown. Furthermore, we lack understanding of the path from vision to behavior as soon as the signals leave the eye via the optic nerve. Consequently, we understand the functional roles of retinal computations within the local context of a retinal circuit and the output of the retina, but not with respect to their ultimate goal: behavior. Direction-selective retinal ganglion cells, for example, send information about the direction of visual motion to subcortical brain regions, including the nucleus of the optic tract (NOT). The NOT is necessary for the optokinetic reflex, a visual behavior that helps stabilizing the image on the retina. The circuit in the retina that computes direction selectivity is one of the best-studied neural circuits in the mammalian brain. Yet, we have little understanding of how the brain uses the acquired information to control behavior. We will extend our understanding of computations within the retina, bridge the gap to the subcortical projection targets one synapse further downstream, and put the computations in the behavioral context of the optokinetic reflex.