Project Summary Understanding the relationship between function of distinct cell types and behavior is a major challenge of modern neuroscience and is being extensively investigated in various model systems. The concept of functionally distinct cell types has been relatively well defined in the early visual system, especially in the mammalian retina. However, our understanding of the contribution of different retinal neurons responsible for the initial visual processing to specific behavioral responses remains rather limited. Here we propose to identify looming stimulus-sensitive retinal circuits using specific behaviors as readouts. Threatening visual inputs, such as approaching objects, trigger universal defensive behaviors in animals and humans. Recently, several brain circuits that respond to the looming stimulus have been identified using optogenetic techniques. Nevertheless, we still have a very limited knowledge about specific retinal circuits important for defensive behaviors and on their potential contribution to such behavioral outputs. To identify circuits for looming-triggered behaviors, we will focus on the retinal ganglion cells (RGCs) that receive visual inputs from interneurons and send axons to the brain. Each RGC type is supposed to respond to specific visual features, and it has been suggested that there are ~46 distinct RGC types in the mouse retina. Several studies, including our own, have generated multiple transgenic mouse lines that label specific RGC types and characterized them using morphological and physiological methods. Specific RGC types in those transgenics were marked by expression of either fluorescent proteins or DNA-modifying recombinase (e.g., Cre) allowing further genetic manipulations. To define which types of RGCs specialize in detecting approaching objects, we will selectively express a genetically encoded toxin or optogenetic regulators in specific RGC types and examine behavioral consequences of such manipulations. Based on previous findings from the looming-trigged behavioral paradigms and our preliminary data, we will start our analysis focusing on the candidate RGC types (i.e., W3 and OFF alpha RGCs) later extending to other RGCs. Our study will identify distinct RGC types that are necessary and sufficient for the defensive behaviors to approaching objects. Identification and characterization of specific retinal neurons regulating visual-threat related behaviors should allow us to deconstruct the circuitry involved in detection and interpretation of fearful stimuli. It will also help us understand an overall structure and functional properties of fear-related circuits. Moreover, the genetic methods utilized in this proposal could be expanded further to investigate the roles of a wide range of cell types in multiple behavioral outcomes.