Wu, Kanai, & Shimojo (2004 Nature 429 262) described a compelling illusion demonstrating a steady-state misbinding of color and motion. Here, we took advantage of the illusion and performed psychophysical and fMRI adaptation experiments to explore the neural mechanism of color-motion misbinding. The stimulus subtended 20 deg by 14 deg of visual angle and contained two sheets of random dots, one sheet moving up and the other moving down. On the upward-moving sheet, dots in the right-end area (4 deg by 14 deg) were red, and the rest of the dots were green. On the downward-moving sheet, dots in the right-end area were green, and the rest of the dots were red. When subjects fixated at the center of the stimulus, they bound the color and motion of the dots in the right-end area erroneously–the red dots appeared to move downwards and the green dots appeared to move upwards. In the psychophysical experiment, we measured the color-contingent motion aftereffect in the right-end area after adaptation to the illusory stimulus. A significant aftereffect was observed as if subjects had adapted to the perceived binding of color and motion, rather than the physical binding. For example, after adaptation, stationary red dots appeared to move upwards, and stationary green dots appeared to move downwards. In the fMRI experiment, we measured direction-selective motion adaptation effects in V1, V2, V3, V4, V3A/B, and V5. Relative to other cortical areas, V2 showed a much stronger adaptation effect to the perceived motion direction (rather than the physical direction) for both the red and green dots. Significantly, the fMRI adaptation effect in V2 correlated with the color-contingent motion aftereffect across twelve subjects. This study provides the first human evidence that color and motion could be misbound at a very early stage of visual processing.
