Negative Afterimage (Stare-and-Shift Illusion)

A ghost in the opposite colour

Stare at a strongly coloured shape for about twenty seconds without moving your eyes, then look at a plain white or grey surface. A faint copy of the shape appears, the same size and outline, but in a different colour. Stare at cyan and the ghost looks reddish; stare at yellow and it looks bluish; stare at black and it looks white. The timer in the demo above guides you through it: keep your eyes on the dot for the full count, then keep looking at the same spot on the blank square.

This is a negative afterimage. Unlike most illusions on this site, it isn't a matter of interpretation. Something in your visual system has changed over those twenty seconds, and the ghost is the after-effect.

Adaptation, not just tired cones

Colour vision starts with three types of cone in the retina, sensitive to short, medium and long wavelengths. Staring at cyan drives the cones that respond to blue and green light hard for a long time, and the visual system adapts: its sensitivity to that colour turns down. When you then look at white, which stimulates all three cone types about equally, the adapted signals are weaker than normal while the rest are at full strength, so the blank patch looks tinted towards red.

The textbook story puts all of this in the cones, but the adaptation that matters most seems to happen a step later. Qasim Zaidi and colleagues compared the colours of afterimages with recordings from the retina and found that they were explained by adaptation in the retinal ganglion cells, the neurons that carry signals from the eye to the brain, rather than in the cones themselves (Zaidi et al., 2012).

The brain contributes too. Shinsuke Shimojo and colleagues showed people a figure in which a coloured surface was only perceived, filled in by the brain between real coloured edges, and found that it left an afterimage of the whole illusory surface (Shimojo, Kamitani & Nishida, 2001). An afterimage of something that was never on the screen can't come from the eye alone.

Why the opposite colour

The colour of the ghost comes from how the visual system codes colour after the cones. Rather than keeping three independent channels, it compares them in opposing pairs: red against green, blue against yellow, and light against dark. Ewald Hering proposed this in 1878, largely to explain afterimages and the fact that nobody can picture a "reddish green" (Hering, 1878). Leo Hurvich and Dorothea Jameson later measured the opponent channels (Hurvich & Jameson, 1957), and cells coded this way have since been found in the retina and the brain. Adapting one side of a channel doesn't just weaken that colour; it tips the balance towards its opposite, which is why the afterimage is a definite colour rather than a faded copy.

The effect shows up best on a plain surface. A coloured or patterned surface floods the same pathways with strong new signals that swamp the small imbalance, which is why the demo switches to an empty square.

From Aristotle to Hering

People have noticed afterimages since antiquity: Aristotle described the coloured patches left after looking at the sun. Johann Wolfgang von Goethe devoted part of his 1810 Theory of Colours to them, with careful descriptions of the coloured "ghosts" left by staring at coloured panels. The explanation that still holds came from Hering's opponent theory, worked out decades before anyone could record from a single neuron. Afterimages are one of the clearest signs that colour is computed, not simply recorded, the same lesson as simultaneous contrast and the dress.