Dynamic retinal coding in natural environments

Project Summary Natural environments present numerous challenges to the visual system, including the presence of large and frequent changes in light intensity. Such changes occur when an animal moves from sunlight into shadow, or when it shifts its gaze from a bright to a darker area in a scene (e.g., sky

versus shaded ground). These changes can occur frequently, e.g. many times per minute, and the intensity changes can be as large as 50-100-fold. These sudden, frequent, and large changes in light intensity present a challenge to the retina, which must transmit a reliable visual signal as it dynamically adapts to the new intensity: this adaptation is likely to be only partial, because the

intensity is likely to change again within the next few seconds. Furthermore, adaptation mechanisms triggered by these changes in light intensity alter the spatial and temporal integration of retinal ganglion cell (RGC) receptive fields, which is tantamount to changing the neural code sent to downstream brain circuits. How can reliable visual signaling be achieved when the code

is in perpetual flux? These dynamic lighting conditions differ substantially from those typically probed in laboratory experiments, where the mean and contrast of stimuli are often held approximately constant. The overarching goal of this proposal is to understand how naturalistic dynamic intensity conditions impact retinal function and visual signaling. We hypothesize that by

examining retinal coding under these dynamic intensity conditions, we will learn how diverse adaptation mechanisms work in concert across multiple cell types to provide a reliable signal to downstream brain areas in natural environments. This work is significant because it will advance our understanding of how the visual system copes with rapid and naturalistic changes in light

intensity. Aim 1 will determine the role of adaptation in phototransduction in shaping RGC responses. Aim 2 will probe the contribution of RGC spike generation. And Aim 3 will determine the impact of adaptation in intermediate circuitry, with a focus on the role of AII amacrine cells. These mechanisms will be linked using a CNN-based framework that permits study not only of

how the individual mechanisms work, but how they interact and collectively shape RGC responses. These aims will reveal how adaptation in distinct loci of retinal processing (photoreceptors, interneurons and RGCs) shape the encoding of visual features in dynamic environments. Furthermore, they will reveal how key RGC types in the primate and rodent visual

systems deal with naturalistic fluctuations in light intensity as they visually scan and move through the environment.