PROJECT SUMMARY / ABSTRACT Our current understanding of primate motion perception is often lauded as one of the great achievements of computational systems neuroscience. Due to its early successes in explicating the fundamentals of neural coding and relations between brain activity and perception, and further constrained by the practical limitations of the macaque model, it has remained rooted in conventional experimental approaches. Despite the important, influential concepts that this approach has produced, here, we step outside this paradigm and highlight that this knowledge needs to be aligned with the major challenges actually posed to the brain during real motion perception. We propose a comprehensive, data-driven, quantitatively-rooted research program that exploits unique opportunities provided by the marmoset model system. The proposed framework and experiments leverage the smooth brains and relatively small body size of marmosets, as well as their amenability to naturalistic testing paradigms in laboratory environments. This work lays the experimental, conceptual, and technical groundwork for a computational neuro-ethology framework in which real-time closed loop recording and perturbation of both neural activity and behavior can be used to probe the neural computations underlying natural primate behaviors. Specific Aim 1. Acquisition and characterization of real, binocular visual motion (including both optic flow and object motion) in freely moving marmosets and humans, performing natural visually- guided navigation and foraging in real and augmented reality (AR) environments. We will place marmosets in realistic, natural environments and allow them to hunt moving prey and to forage for stationary, partially hidden food items. We will measure their body and eye movements, as well as their views of the visual scene, to build a quantitative model of natural image/movie statistics that captures the effects of self-motion and object motion in the real world, and to assess the task-specific nature of these statistics. We will also acquire and analyze eye, head, body, and natural movie data from humans performing behaviors that map on to the studied marmoset behaviors to assess the commonalities and differences across primates. Specific Aim 2. Large scale, multi-area recording of activity in V1, MT, and MST in marmosets viewing realistic, complex visual motion and optic flow patterns in both freely moving conditions and head-fixed virtual reality (VR). We will record activity in the visual areas of freely-moving marmosets performing natural tasks and behaviors, and apply intuitive but sophisticated statistical analyses to link the neural activity to the visual inputs and the detailed body movements. This will provide the first characterization of primate visual system activity during natural vision and action. We will use a head-fixed VR environment to perform detailed tests of how self-motion modulates visual responses, as well as to perf...