STRAW, Andrew D.; O’CARROLL, David C.; University of Adelaide, Australia; University of Adelaide, Australia: Adaptation of Wide-Field Movement Detecting Neurons in Hoverflies Produces Velocity Constancy to Natural Scenes

Adaptation is a ubiquitous phenomenon of sensory neurons in which the initial response is altered by prolonged exposure to a stimulus. Adaptation emphasizes changes in an animal’s surroundings at the expense of encoding absolute quantities. In general, this is a useful principle that maximizes efficiency of processing information. However, it is unclear why neurons involved in movement detection exhibit such adaptation, since an algorithm devised to estimate self-motion, environmental structure and movement of objects within the environment might require an absolute measure of velocity. To test the role of adaptation, we recorded the response of wide-field movement detecting neurons of hoverflies to panoramic images taken from their habitat. As in earlier work with more stereotyped stimuli (e.g. sinusoids), we find significant adaptation within the neural responses. We find that adaptation to natural images decreases the sensitivity of the neuron to visual contrast and leads to response invariance with respect to the velocity of the scene. It appears that adaptation helps encode the absolute quantity of visual velocity, so we call this phenomenon velocity constancy. Earlier theoretical and experimental studies highlight the inability of the fundamental movement detection operation to provide a signal directly related to velocity. Our results suggest that motion adaptation functions to provide the animal in a natural setting with compensation for the shortcomings of the basic algorithm employed for movement detection. This investigation and others from our lab increasingly suggest that a key function of wide-field movement detecting neurons is to provide a robust measure of visual velocity in real-world conditions.