Neuroscientists can now predict what color you’re seeing. The secret is surprisingly black and white.

A new study published in The Journal of Neuroscience suggests that the human brain responds to colors in a way that is surprisingly consistent across different people. By comparing patterns of brain activity, researchers were able to predict what color a person was seeing using a model trained on brain activity from other people. The findings indicate that color perception, while subjective, may be tied to shared structures in the visual system.

Color perception often feels personal. People frequently wonder whether the red they see looks the same to someone else. While this philosophical question remains open, neuroscience provides tools to explore whether the brain processes color in similar ways across individuals. Previous studies have shown that it is possible to decode what color a person is seeing based on their own brain activity. However, it was unclear whether such decoding could work across different brains.

Michael Bannert and Andreas Bartels, researchers at the University of Tübingen and Max Planck Institute for Biological Cybernetics, set out to explore whether color evokes similar patterns of brain activity in different people. They also wanted to understand how the brain’s organization based on visual space contributes to color processing.

Their study used a technique that aligned patterns of brain responses across individuals, not based on color itself, but on responses to black-and-white stimuli. This alignment allowed them to test whether information about color could be transferred from one person’s brain to another’s.

“We had been working on color vision for quite a while already when we discovered a method (developed by others) that made it possible to align neural responses from different brains to one another,” explained Bannert, a postdoctoral researcher. “‘Aligned’ here means that given a pattern of brain activity in one observer, you could compute what pattern of brain activity corresponds to that in another observer. Those methods were typically applied to brain data collected while people watched very complex, naturalistic input like movies. They had not been used to align isolated single visual features.”

“In particular in the context of color we believed that this was an important applications because it has been a long-standing philosophical question if people see the same colors differently or similarly. This matters because colors do not objectively exist in the external world but instead are the product of our minds. Color experience is just the result of our minds making sense of the sensory input.”

“Although we were aware that fMRI data would not tell us how different observers would consciously perceive a color (in the sense of qualia, i.e., in the sense of ‘what it is like to see red’) we felt that it would be the closest thing we could do to address this question by asking a slightly easier one: Are the patterns of brain activity in some way similar when two different people see the same stimulus?”

The study involved 15 adult participants who underwent brain scans while looking at visual stimuli. Each participant took part in two types of scanning sessions. In the first, researchers used a method called retinotopic mapping to identify how different parts of the visual field are represented across the cortex. This process involved viewing a black-and-white checkerboard pattern that moved systematically through the visual field. Because the stimuli contained no color, these scans captured brain activity related only to spatial location.

In the second part of the study, participants viewed colorful visual patterns consisting of expanding concentric rings. These rings came in three colors—red, green, and yellow—and were shown at either a high or low level of brightness. While participants focused on a central dot, their brain responses to these color stimuli were recorded using functional magnetic resonance imaging.

To align data across brains, the researchers applied a computational method known as shared response modeling. This technique allows the researchers to compare patterns of brain activity from different people by projecting them into a shared neural space. Importantly, the model was trained using only the data from the black-and-white retinotopic mapping scans. In other words, the alignment of brains was based solely on how they responded to the location of visual stimuli, not their color.

Once the brains were aligned, the team tested whether a machine learning classifier could use the brain activity from one group of participants to correctly predict the color or brightness being viewed by another participant. The results suggested that this was indeed possible.

When the model was trained on data from multiple participants and tested on a new person, it was able to identify the viewed color at levels significantly above chance. The success of this cross-brain prediction implies that color triggers shared patterns of neural activity in specific areas of the visual cortex. This outcome suggests that some aspects of color processing are consistent across individuals, at least in early and intermediate visual areas.

“When different observers see the same color stimulus, their brains responds in ways that are to some extent similar,” Bannert told PsyPost. “They are so similar in fact that it is possible to predict from activity in one person’s brain what color they are seeing if one already knows what the brain responses look like in other people’s brains. So, to their brains the same color looks in some sense similar.”

The researchers also found that color preferences were tied to specific regions in the visual field. For example, certain parts of the brain that processed areas near the center of gaze were more responsive to particular colors, such as yellow, while more peripheral areas showed stronger responses to other colors, like red or green. These patterns varied across different visual areas, suggesting that the brain’s color responses are shaped not only by the color itself but also by where the color appears in the visual field.

To test the reliability of these findings, the researchers ran additional analyses to see if they could predict, across individuals, which specific color each small part of the brain (known as a voxel) would respond to most strongly. These predictions also succeeded above chance, especially in early visual regions like V1 and V2, as well as in mid-level regions such as hV4 and LO1. This finding provides additional evidence that large-scale patterns of color sensitivity are shared across brains and are mapped according to visual space.

“It is important to understand how we predicted what color a person was seeing across different observers: The method involves a computer model that learns for each brain how it represents the visual field, i.e., what parts of the brain light up in response to a stimulus that appears in the center of what you’re seeing, to the left of it, to the right of it, top/bottom etc… basically by systematically stimulating all locations in the sensory array,” Bannert explained.

“For this systematic stimulation, people are shown flashing black and white checkerboard stimuli that move slowly across the field (called retinotopic mapping). Crucially, this stimulus does not involve any color! But nevertheless the computer model finds a mapping between the ways that different brains represents the visual space. And as it turns out, this mapping allows us to decode color from patterns in brain A, although the computer model only ever had access to color patterns observed in brains B, C, D, etc.”

“The fact that color decoding across brains is possible in the described way shows that it is at least partially driven by the fact that there is a systematic relationship between the representation of visual space and the representation of color and that this relationship is preserved across brains. Put simply: ‘Let me measure how your brain represents your visual space (with standard retinotopic mapping) and I will tell you what color you are seeing.'”

While the results suggest a shared neural basis for color perception, the researchers caution that this does not mean all individuals experience color identically. The study did not address how colors are subjectively perceived, only how the brain processes visual input. In other words, while the neural code for a particular color may be similar across people, this does not confirm whether their internal experiences are the same.

Another limitation is that only three colors were tested, and the stimuli did not include blue, a color known to be processed differently in the visual system. The study also focused on relatively simple patterns in controlled conditions. Real-world visual experiences are often more complex, involving interactions of color with motion, texture, and meaning.

Future studies could expand this approach to a wider range of colors, test for responses in more naturalistic settings, or explore how personal experience and learning shape neural responses to color. Researchers may also investigate whether these shared patterns extend to other aspects of visual experience, such as shape, depth, or motion.

“It is important to understand that we are not solving the ‘hard problem’ of consciousness. We do not know what it feels like to observers to see a particular color,” Bannert said. “Nor do we know how that relates to the experience of color in a different person. We just solve the neuroscientific equivalent of that problem: Does seeing the same color elicit the same brain pattern in different observers. It is still possible that, subjectively, people’s subjective color experience will drastically differ even though neural representations may share some similarities. The mapping between neural processes and visual consciousness may still differ wildly across people.”

“We are not trying to solve the problem of color qualia,” the researcher added. “The hard problem of consciousness will remain beyond our methods in the foreseeable future and is this imposing a bound on what we can achieve in this respect. The last 25 years of consciousness research with the most modern neuroscientific methods has failed to show a clear neural correlate of consciousness. If I had to bet, I would say that it’ll stay this way in the next 25 years at least.”

“But there are certain aspects that can be addressed using experimental paradigms that do tell us important lessons about color perception and at least about small aspects of consciousness. (‘Access consciousness’ is an example, i.e., the extent to which perceptual content can be used by observers to provide report about their visual experience. Note that this is something that a ‘zombie’ could do in principle — a person without any consciousness… or: a clever AI.) We use such experimental paradigm for instance to investigate how people’s knowledge about specific object/color associations determine how easily certain stimuli enter our consciousness in the first place? It turns out that congruent object/color pairs (red strawberry) are more likely to be perceived than incongruent ones (red broccoli).”

The study, “Large-Scale Color Biases in the Retinotopic Functional Architecture Are Region Specific and Shared across Human Brains,” was published October 15, 2025.