Neural pathways and computations that achieve stable contrast processing tuned to natural scenes

Natural scenes are highly dynamic, challenging the reliability of visual processing. Yet, humans and many animals can perform accurate visual behaviors in nature, whereas computer vision devices struggle with changing environments. How do animal eyes achieve this? Here, we reveal the algorithms and mechanisms of rapid luminance gain control in Drosophila, resulting in stable visual processing. We identify the dendrites of specific third order neurons as the site of luminance gain control. The circuitry further involves wide field neurons, matching computational predictions that local spatial pooling can drive optimal contrast processing in natural scenes in which light conditions change rapidly. Experiments and theory argue that a spatially pooled luminance signal achieves luminance gain control via a divisive normalization process. This normalization relies on shunting inhibition using a glutamate-gated chloride channel, GluClα. Our work describes computationally, algorithmically, and mechanistically, how visual circuits can achieve robust visual processing in dynamically changing, natural visual scenes.