人眼为何以四种色调感知颜色

A sample of the photos used to assess the human eye’s response to natural scenes. [Image: Alexander Belsten, Bruno Olshausen/UC Berkeley]

Last year, researchers from the University of California, Berkeley, USA, manipulated the human eye to see a new color, which they called “olo” (see “A New Color Comes into View,” OPN, May 2025). Now, scientists from the same team have developed a theory for understanding why humans generally perceive color in combinations of four “pure” hues—red, yellow, green and blue—despite having three kinds of cone cells in the eye (J. Opt. Soc. Am. A, doi: 10.1364/JOSAA.598897).

“[The hues’] special status remains a mystery,” said study author Alexander Belsten. “No known property of light—the cone photoreceptors in the eye or neural representations in the brain—can explain why these four hues, rather than some other set, should occupy this privileged position.”

The new theory, the authors say, can help reconcile the two historically debated explanations of color vision: Hermann von Helmholtz’s theory based on the three photoreceptor cells in the eye and Ewald Hering’s theory that focuses on the opposing nature of the four hues.

Color across cultures

Most people in most cultures have a tendency to see and name colors as a combination of red, yellow, green and blue. Orange, for example, is commonly accepted to be a combination of yellow and red, but most would not say that red is a combination of orange and purple. People also tend to perceive these colors as opposing: Red is the opposite of green, and blue is the opposite of yellow.

“We tend to think of red as just red, not as a mixture,” said Belsten. “This same unique property holds for green, blue and yellow: Each appears perceptually pure rather than as a mixture of neighboring hues.”

Linguist Paul Kay and the late anthropologist Brent Berlin, both associated with UC Berkeley, studied descriptions of colors and found that these four hues were at the center of color categorization across dozens of languages, both written and unwritten, which they believe points to a unique characteristic of how humans experience color. However, scientists have not been able to find a connection between human perception of these four hues and the three kinds of cone cells in the eye that detect different wavelengths of light, nor have they been able to explain it by studying the neurons that process vision in the brain.

Simplified perception

In the new work by the Berkeley team, Belsten and neuroscientist Bruno Olshausen propose that the human brain’s preference for simplicity is at the heart of color perception. They found that the natural world has a restricted color palette that the brain represents in combinations of the four opposing colors: red versus green and blue versus yellow.

“Think about this like north, south, east and west. If you say you're going north, that means you're not going south,” said Olshausen. “That’s what’s kind of striking about this study. It captures exactly this idea of opponency that physiologists described back in the 19^th century, which is that blue and yellow appear to be opposites and that red and green appear to be opposites. This has puzzled color vision researchers for a long time.”

Belsten said that the new theory could possibly resolve the two historically competing accounts of color vision that emerged from von Helmholtz and Hering in the late 19^th century. "We now know that both are valid accounts for how we see color. The resolution is that while the former speaks to physiological sensory mechanisms in the retina, the latter describes a psychological basis for describing our subjective experience of color … that is grounded in the structure of the natural visual environment."

[Enlarge image]

The distribution of hues in natural images is displayed in a color space where distance from the center indicates more saturated color and the angle within the space corresponds to different hues, as indicated by the color wheel. The center represents gray. If all colors appeared at random in nature, you would expect the distribution indicated by the dashed concentric circles. Instead, nature’s palette tends heavily toward red and somewhat less toward yellow-green and blue-green. [Image: Alexander Belsten, Bruno Olshausen/UC Berkeley]

An asymmetrical world of color

To test their theory, the researchers leveraged recently accumulated datasets of the spectral distribution of natural scenes (they focused on natural colors rather than the human-created saturated pigments in display technology). Using the data, they simulated how the cones in the retina respond to the spectral distribution in nature and found that the variation is mostly based on brightness and most of the world’s pixels are colorless.

When they looked only at color, they found that the eye’s sensitivity to color is asymmetrical: The eye is most sensitive to red and then yellow-green and blue-green. The team also found that most hues in nature are much less-saturated than red—the sky is a pale blue and vegetation is a light yellow-green, which contrasts with the reds found in soil, rocks and the bark of tree trunks.

“One major finding from the data is that the color distribution in the environment is very asymmetric and non-uniform,” Olshausen said. “Nobody’s ever looked at this before because the availability of these very large datasets is relatively new. You need to have millions and millions of pixels to start to see the dominant colors because some color pixels are only one in 100,000 or one in a million.”

From three to four neurons

They then applied a concept called sparse coding, which explains that the brain efficiently processes large amounts of sensory information using a small number of active neurons, to see how humans can perceive color distribution while minimizing neuronal activity. Belsten and Olshausen learned that while three types of neurons that are responsible for perceiving red, yellow-green and blue-green were able to encode all colors found in nature in sets of two, a fourth neuron (breaking the colors into the four hues) enabled the same representation of colors in sets of twos and fours and, importantly, the perception of opposing pairs of color. "Sparse coding inference introduces mutual exclusivity … and yields a color representation that mirrors the phenomenology of color perception,” explained Belsten.

“Humans do it this way because neurons can now describe the world in terms of something and its opposite. Red is not green, just like north is not south. You don’t have this parsimonious representation in terms of opponent colors if you have only three colors to encode the world,” Olshausen said. “What I'm really excited about is that we have an explanation now for the first time. All these pieces just fall together. It's almost too good to be true.”