Color vision strategy defies textbook picture

Color vision may actually work like a colorized version of a black-and-white movie, a new study suggests.

Cone cells, which sense red, green or blue light, detect white more often than colors, researchers report September 14 in Science Advances. The textbook-rewriting discovery could change scientists’ thinking about how color vision works.

For decades, researchers have known that three types of cone cells in the retina are responsible for color vision. Those cone cells were thought to send “red,” “green” and “blue” signals to the brain. The brain supposedly combines the colors, much the way a color printer does, to create a rainbow-hued picture of the world (including black and white). But the new findings indicate that “the retina is doing more of the work, and it’s doing it in a more simpleminded way,” says Jay Neitz, a color vision scientist at the University of Washington in Seattle who was not involved in the study.
Red and green cone cells each come in two types: One type signals “white”; another signals color, vision researcher Ramkumar Sabesan and colleagues at the University of California, Berkeley, discovered. The large number of cells that detect white (and black — the absence of white) create a high-resolution black-and-white picture of a person’s surroundings, picking out edges and fine details. Red- and green-signaling cells fill in low-resolution color information. The process works much like filling in a coloring book or adding color to a black-and-white film, says Sabesan, who is now at the University of Washington.

Sabesan and colleagues discerned this color vision strategy by stimulating about 273 individual cone cells in the eyes of two men from the lab. The technological accomplishment of stimulating single cone cells in the retina is akin to getting people to walk on the moon, says Neitz. “It is a super technological achievement. It is an amazing thing.”

Sabesan’s team first used a microscope that could peer into living human eyes to map hundreds of light-detecting cone cells in the two volunteers. In order to get a clear picture of the cells through the distortion of the lens and cornea, the researchers borrowed techniques that astronomers use to compensate for disturbances in the atmosphere.
With the blur from imperfections in the eye corrected, the researchers had to precisely target individual cells to hit with the laser. Because the eye is constantly jiggling, the researchers had to determine the pattern of the eye movements to predict where cone cells would be several milliseconds in the future. Over about two years, the researchers repeatedly stimulated 273 red or green cones one by one. After a flash of laser light was delivered to the cone, the men would indicate on a keyboard what color they had seen.

Of the red cones the researchers stimulated, 119 made the men see white, while only 48 flashed red. Similarly, only 21 of the green cones tested actually signaled green, while 77 registered white. Each individual cone probably signals only white or color, the researchers say. “It’s a rather inefficient arrangement,” says Donald MacLeod, a vision scientist at the University of California, San Diego. All the cone cells are capable of detecting color, but few actually seem to do so.
Cells surrounded by cones that detect a different color were more likely to send white signals, the researchers found. That finding is unexpected and runs counter to a popular idea that cones ringed by cells detecting other colors would be better at color detection, MacLeod says.

These findings could be good news for people with color blindness. The results suggest that gene therapy that adds red or green cones could work even in adults, Neitz says. Although his group gave a monkey full color vision (SN: 10/10/09, p.14), many researchers thought human brains would never be able to incorporate additional color information even though the eye could detect it. The new findings indicate the brain needs to learn only that there is one more color needed to fill in to a basically black-and-white picture, a task it should accomplish easily, Neitz says.