A key function of the visual system is to extract behaviourally relevant features about the visual scene from the barrage ofincoming photons. This process begins in the retina, a powerful image processer that recent studies have revealedseparates the incoming information into ~20 distinct neural representations. Within central visual brain regions specificdetails of the visual scene are often neatly organized, remaining spatially segregated from each other. Despite our detailedknowledge of how information is organized within the visual system, progress in gaining a mechanistic understanding ofhow the brain is able to extract salient features from the visual scene, or how this information is used to guide behavior hasstalled. Too often each brain centre is studied in isolation, meaning the details of its inputs and function of its outputs arefrequently a point of conjecture. In particular, the input to central visual areas, despite the accumulated evidence, arerepeatedly assumed to simply be a relayed representation of the visual scene from the retina.In order to provide mechanistic insight into how visual information is processed in central brain regions and understand howit is used to direct behavior this proposal will delineate the inputs, originating in the retina, to specific behaviouralcomputations in central brain regions. This will allow us to determine which retinal channels are assigned dedicatedcomputational tasks and how information from the retina is recombined to highlight key aspects of the visual scene used todirect behavior. To accomplish this conditional expression systems and trans-synaptic viral based circuit tracing will becombined with targeted patch-clamp recording and calcium imaging to link genetically identified cell types of the retina withbehaviourally relevant computations in central brain regions. This work will provide insight into the circuit mechanisms usedby the brain to process the visual world.