Meeting Abstract

S1-1.7  Wednesday, Jan. 4  Detecting motion while moving: a simple visual reflex revealed by animal-robot interactions DICKINSON, M.*; ZABALA, F.; POLIDORA, P.; ROBIE, A.; BRANSON, K.; PERONA, P.; Univ. of Washingon; Univ. of Washingon; IORodeo; Janelia Farm, HHMI; Janelia Farm, HHMI; Caltech

Ever since the Cambrian, there has been selective pressure on visual systems to develop mechanisms for detecting predators, prey and potential mates. Here, we used a tiny robot to show that walking fruit flies exploit an elegantly simple visual mechanism for detecting the presence of nearby moving objects. When an animal is at rest, an object moving in any direction may be easily detected by motion-sensitive visual circuits. During locomotion, however, this strategy is compromised by the spatially complex optic flow fields on the retina created by an animal’s own motion through the environment. In particular, because stationary visual features in the environment move front-to-back (progressively) on the retina as an animal translates forward, moving objects that also create front-to-back optic flow are difficult to distinguish from the background. However, objects that move so as to create back-to-front (regressive) optic flow, may be unambiguously distinguished from the static background. Thus, animals ought to exhibit enhanced sensitivity to such patterns, a hypothesis we term ‘regressive motion salience’. We explicitly tested this hypothesis by constructing a fly-sized robot that was programmed to interact with a real fly within a large behavioral arena. Our measurements indicate that whereas walking flies freeze in response to a regressively moving fly-sized object, they ignore a progressively moving one. Because the assumptions underlying the regressive motion salience hypothesis are general, we suspect this mechanism to be quite widespread among eyed, motile organisms. Our experiments provide further evidence for the utility of behavioral robotics as a method for analyzing the sensory basis of social interactions.