The Visual Aftereffect That Can Last Three Months

In 1965, Celeste McCollough, a psychologist at Oberlin, published a four-page paper in Science titled "Color Adaptation of Edge-Detectors in the Human Visual System." The procedure was almost insultingly simple. Stare for a few minutes, alternating, at a red horizontal grating and a green vertical grating. Then look at a black-and-white grating.

The horizontal stripes will look faintly green. The vertical stripes will look faintly pink. Tilt the page ninety degrees and the colors swap with it.

That second sentence is the part that broke the textbook. Conventional aftereffects — the cyan blob you see after staring at a red square — depend on the retina, and they fade in under a minute. The McCollough effect doesn't behave like that. The tint is locked to the orientation of the test lines, not to the patch of retina you stared at. Whatever is doing this lives further up the visual stream, in cells that already know which way an edge is pointing.

In 1975, Robert Jones and Dennis Holding tried to figure out how long it lasted. They induced the effect once, then split their subjects: one group was tested every day, the other was sent home and not tested again for nearly three months. The frequently-tested group lost the effect within five days. The untested group, eighty-five days later, still saw the colors. The act of looking for the aftereffect was what wore it out.

McCollough was twenty-eight when she ran the original experiment. The effect named after her is one of the longest-lasting perceptual changes in healthy human vision, and sixty years on, nobody has fully explained why.