The mammalian retina consists of five classes of neurons: photoreceptor, horizontal, bipolar, amacrine, and ganglion cells. Based on cell morphology, electrophysiological properties, connectivity, and gene expression patterns, each class of retinal neurons is further subdivided into many distinct cell types. Each type of photoreceptor, bipolar, and ganglion cell tiles the retina, collectively providing a complete representation across the visual scene. Visual signals are processed by at least 80 distinct cell types and at least 20 separate circuits in the retina. These circuits comprise parallel pathways from the photoreceptor cells to ganglion cells, each forming a channel of visual information. Feed-forward and feedback inhibition of horizontal and amacrine cells shape these parallel pathways. However, the cell-type-specific roles of inhibitory circuits in retinal information processing remain unknown. Here we summarize parallel processing strategies in the retina, and then introduce our viral and electrophysiological approaches that reveal the roles of genetically defined subtypes of amacrine cells in retinal circuits.