A mouse running through a virtual tunnel reveals how the brain estimates depth using nothing but motion and memory. The visual system must reconstruct three-dimensional space from flat retinal images, and while depth from binocular cues is well understood, the neural circuits behind monocular depth perception—specifically motion parallax, where nearby objects appear to move faster than distant ones as the observer moves—remain largely unknown. This project uses calcium imaging and electrophysiology in mice navigating virtual reality and freely exploring real environments to trace how visual motion signals and the animal’s own speed combine across the visual cortex to create depth selectivity. This is fundamental science: it asks how a brain builds spatial awareness from self-generated movement, a question with no immediate practical application. However, understanding these circuits in a key model organism—the mouse—could eventually inform the design of autonomous systems that navigate without binocular vision, such as drones or planetary rovers, and help neuroscientists interpret decades of head-fixed vision experiments by revealing how those results relate to natural behaviour.
View original technical description
The visual system seamlessly processes visual information to parse the spatial organisation of visual scenes. While the world and our perception of it are three-dimensional, the images formed by our eyes are two-dimensional. The missing depth information must be estimated by neuronal circuits within the visual system using a variety of cues, including binocular cues, which depend on vision in both eyes, and monocular cues, which rely on information available to one eye. While neural circuit mechanisms underlying depth estimation from binocular disparity - the difference between images formed by the two eyes - have been studied extensively, much less is known about mechanisms supporting monocular depth perception. An important but poorly understood monocular depth cue is motion parallax - the apparent movement of objects within the visual scene resulting from movement of the observer. Previous work has demonstrated that many animals, including rodents have an innate capacity for depth perception that does not depend on binocular vision but likely relies on visual motion that animals experience as they move to explore their environment1. This ability relies on the visual cortex, suggesting that this region acts as a part of a circuit for estimation of depth from motion parallax2. Our preliminary results using calcium imaging in mice navigating in virtual reality (VR) demonstrate that neurons in the mouse primary visual cortex are highly selective for the depth of visual cues from motion parallax. This selectivity emerges as the result of integration of visual motion signals and top-down modulation encoding the animals' speed of locomotion. These result raise new questions about where in the visual system this depth selectivity arises, how it is transformed across the visual cortical hierarchy, and how it relates to neuronal responses under naturalistic conditions in freely moving animals. To address these questions, this project will use multiphoton imaging and electrophysiological recordings in virtual reality environments and during free exploration to determine how locomotion and visual motion signals are integrated in different stages across the visual pathway. This research will advance our understanding of how animals process visual signals as they actively explore their environment and identify the neuronal mechanisms that enable them to estimate depth using motion parallax. As experiments in visual physiology have traditional relied on recordings in head-fixed animals, understanding how visual responses during head fixation relate to those during free exploration will have a wider impact by helping the interpretation of results of other researchers using head-restrained animals and virtual reality environments to study vision. Over the last decade, the mouse visual cortex has emerged as one of the key models of cortical function, with large-scale efforts dedicated to characterising its cell type composition and connectivity3-6. However, our understanding of how these circuits support vision under naturalistic conditions remains limited. Therefore, revealing the computations carried out by visual cortical circuits in mice exploring three-dimensional environments will help explain the function of the circuits identified by these projects.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know