Point images in the visual system: new interest in an old idea

Abstract A central problem of vision research is to characterize the patterns of neural activity evoked throughout the visual system by retinal stimulation. These patterns represent not only the information about a retinal image available to the rest of the brain, but also the transformations of that image that are appropriate to the functions of the system's various parts. Some recent studies have approached this task by asking first how the system deals with information from a single retinal point. This requires that the traditional question of receptive-field analysis be turned on its head: instead of asking ‘Here is a cell; where are the points that it sees?', the question becomes, ‘Here is a point; where are the cells that see it?'.