We're interested in how neural circuits perform computations. In particular, we study how the vertebrate retina translates the visual scene into electrical impulses in the optic nerve. Recent results have shown that visual processing in the retina has a number of surprising properties, including adaptation to various statistics of visual images and sensitivity to object motion but not eye motion. Our immediate goals are to determine the mechanisms of these and other processes, and more generally to understand how the retina encodes components of the visual scene such as luminance, object boundaries and motion. To do this, we use a versatile set of experimental and theoretical approaches. While projecting visual scenes onto the isolated retina, an extracellular multielectrode array is used to record a substantial fraction of the output of a small patch of retina. Simultaneously, we record intracellularly from retinal interneurons in order to monitor and perturb elements of the circuit as it operates. Additionally, we are combining two-photon calcium imaging with multielectrode recording as a way to access subcellular neural structures while recording the output of the retina. Finally, all of this data is assembled and interpreted in the context of mathematical models to predict and explain the output of the retinal circuit.