A naturalistic visual task for studying distance estimation

PROJECT SUMMARY Vision facilitates navigation through the world by providing sensory information about the environment, such as the distance to relevant objects. A key feature of visual perception is the active exploration of the visual scene through translation of the eyes, head, and body. Visual cortex has been proposed to

combine these self-generated motor signals with visual input to compute information about objects in the environment. While recent studies have shown that a significant fraction of neurons in mouse V1 encode movement information and do not simply act as visual feature detectors, models of V1 function

have largely ignored motor efference and sensory reafferent contributions. We aim to elucidate the neural mechanisms underlying active vision by investigating depth perception from motion parallax - a fundamental visual computation that combines observer self-motion and retinal image displacement to

calculate the distance to objects in the environment. We will first adapt an ethological, freely-moving gerbil/rat distance estimation task to mice in order to determine the types of visual cues mice use to gauge depth when jumping across a gap. We will then manipulate the activity of visual cortex and its inputs from brain regions conveying movement-related

signals, in order to test their roles specifically in distance estimation from motion parallax. The experiments proposed in this R21 application provide the foundation for future studies at the neural circuit level, to determine how visual and movement signals are integrated for computations such as

distance estimation, particularly in a natural context.