The first states that stationary retinal images quickly disappear because the interior of a uniformly luminous object remains unchanging regardless of eye movements. Such a stationary image would become invisible without its edges, which move accordingly with eye movements and keep the object visible. The second piece of evidence is based on physiological studies of retinal ganglion cells. These are the only retinal cells that make contact with the brain, allowing them to provide crucial information about the role of edges. Each retinal ganglion cell is affected only by light falling in a small region of the retina called it "receptive field"". Receptive fields are always circular and subdivided into two zones much like a doughnut. The ratio between the amounts of light falling on these two zones, the "center"" (the hole) and the "surround"" (the doughnut), determine a ganglion cell's response. This simply means that whenever the amounts of light in the center and the surround are equal, the ganglion cell gives an equivalent response. Thus, ganglion cells that are processing interior regions of a uniformly luminous cannot provide the brain with any useful information about the luminance of that object. Consequently, the brain's only source of information about the luminance of an object is the signals it receives from retinal ganglion cells whose receptive fields contain images of the object's edges. Based on these signals, the brain must then be able to deduce what degree of brightness it should assign to the interior of the object.
Not surprisingly, due to the brain's reliance on edge information, we can trick the brain into making errors in its brightness judgments by using stimuli with unusual edges. One example of this is the Cornsweet Illusion, which shows that the brightness of an extended region is determined by the luminance ratios at its borders alone. This is illustrated in the luminance profile shown below:.
