Predictive position coding

点击翻译按钮获取中文摘要

This project addresses how we know where things are, specifically the how the brain codes the positions of objects that we see in front of us. The location of an object is explicitly represented by the cells in the visual system that respond to it, as each cell responds to only one small region of the visual world, linking its position to a location on the retina at the back of the eye. However, this explicit coding cannot explain the many dramatic deviations of perceived location that arise when the eyes or the target are in motion. Here I will build on the proposal that location is not just a simple registration of retinal location with compensation for head and eye movement but is instead a sophisticated, predictive system that constructs probable target locations based a variety of inputs, including most importantly, the target's current motion. We have demonstrated two distinct processing streams for this high-level location coding, one that is perception-based and one that is linked to eye movements. To do so, we have used two powerful illusions: the double-drift (Lisi & Cavanagh, 2015) and the flash-grab (Cavanagh & Anstis, 2013). The first illusion produces a very large displacement in perceived position orthogonal to the motion of the target. Remarkably, eye movements are unaffected by this illusion, they still land on the physical rather than perceived position. Our pilot results in brain imaging suggest surprisingly that this perceptual illusion emerges in the frontal areas of the brain, raising the possibility that visual perception itself is constructed at this very high level rather than in lower level, specifically visual areas. The second stimulus is a flash presented on a moving background where the flash is seen displaced in the direction of motion. For this stimulus, eye movements show the same illusion as perception, suggesting that the flash-grab stimulus can isolate a correction for target motion, compensating for the delays in neural transmission. This prediction of the present is used in the eye movement system and then mirrored in the perceptual system. The position deviations created by these two stimuli offer a new window into the neural representation of location. I propose to use these stimuli to examine the bases of the position code used by perceptual and motoric systems. Numerous studies have addressed the position coding underlying spatial navigation, work that recently won a Nobel prize (O'Keefe, Moser, & Moser, 2014). In contrast, much less is known about the immediate coding of position that is the main input into this navigational system, at least for humans. The goal of the proposed research is to work out the mechanisms of this visual location coding. This will help in the clinical area to address general location disorders affecting spatial navigation and wayfinding as well as providing guidance for biomedical applications in constructing artificial and prosthetic vision systems.