Two eyes, one vision: binocular motion perception in human visual cortex

Abstract

An important aspect of human vision is the fact that it is binocular, i.e. that we have two eyes. As a result, the brain nearly always receives two slightly different images of the same visual scene. Yet, we only perceive a single image and thus our brain has to actively combine the binocular visual input into a single, cyclopean view of the world. Using both behavioral and neuroimaging approaches we studied where in the brain this process of binocular integration might take place and what can happen if the integration of information from the two eyes fails. Using computational neuroimaging techniques we demonstrate that the first cortical area that is involved in processing visual information (primary visual cortex, V1) represents a visual scene as it appears on the retina: with two slightly differing images. Starting from the second visual cortical area (V2) the representation of visual scenes is mostly integrated into the single, cyclopean view that we experience in daily life. We also found that when the integration of binocular information does not function properly people can develop a particular deficit that we termed motion agnosia. This deficit results in the observer being unable to discriminate between different directions of motion, essentially making them unable to correctly judge which way objects are moving. However, this deficit occurs without any other abnormality in the eyes or vision in general and therefore we concluded that it is best described as a form of visual agnosia: a disorder where the visual information is still available (the eyes function normally) but the observer is unable to use that information to make inferences about the environment. Using behavioral methods (psychophysics) and fMRI we demonstrate that the origin of motion agnosia is the binocular combination of motion signals from the two eyes.