What do babies and fish have in common?

Patrick Cavanagh, Stuart Anstis, Daphne Maurer, Terri Lewis

Fishes and babies can neither read nor speak, yet we persuaded them to tell us what colors they see. We showed them special moving colored patterns on a computer screen. If they saw the colors, the babies followed the movement with their eyes, and the fishes followed it by moving their whole bodies, swimming after the patterns in an innate optomotor response.

We made a movie only four frames long which looped repetitively. Each frame filled the whole computer screen with vertical colored stripes. In the left hand column, the stripes at Time 1 were light red and dark green, and at Time 2 they were light & dark yellow. The yellow stripes were shifted sideways by half a bar width, so the stripes appeared to jump sideways — but which way? The brain could not pair up succcessive stripes on the basis of color, because all the stripes at Time 2 were the same color (yellow). So the brain had to pair them up on the basis of luminance. If the red stripes were lighter than the green, the stripes appeared to jump to the right toward the nearest light yellow stripe (left-hand column). If the red stripes were darker than the green, the stripes appeared to jump to the left toward the nearest dark yellow stripe (right-hand column). The movie translates lightness into motion.

We titrated red against green luminance until, at equiluminance, perceived motion disappeared for adult obsesrvers, babies stopped making pursuit eye movements, and fishes stopped swimming in circles. Red-deficient observers rquire more red to make an equiluminous mathc, and green-deficients require more green. We were able to identify individual color-defective babies. Conclusions: Outputs from cones into the luminance pathways were in place within the first months of life. Also, guppy fish are more green-sensitive and less red-senstivie than humans.