A temporal-difference model of perceptual stability in color vision

We consider the problem of how humans can maintain a stable perception of object color across saccades in spite of the changes in sensory input caused by the spatially non-homogeneous receptor spectral sensitivities. We propose a method, based on a temporal-difference reinforcement learning scheme, for constructing associations between pre- and post-motor stimuli and which yields a constant color percept across saccades.