I tried to train my color vision. Here’s what happened.

One afternoon during my PhD, I took a break from lab work to snack on a banana. I grabbed a seat in the office, slid off my headphones, and peeled open my treat. Just as I bit in, I noticed a labmate staring at me. 

"Max," she told me, holding back some laughter. "That banana is a 4."

I instantly knew what she meant because I'd made this mistake before. The banana was days from being ripe, and I had misjudged the color. A laughably green banana.

Even before I knew that I had mild deuteranomaly (so-called red-green colorblindness), I struggled with cryptic color schemes on spreadsheets and graphs. Whether in spite or because of this, color theory fascinated me. I had neurological, practical, and philosophical questions. Why do we call the retina’s longest wavelength cone “red” when it actually best absorbs yellow-green light? Why does mixing paint obey different rules than mixing light? If I could see through your eyes, would your mental images match mine — does your blue match my blue? And I had questions that blended all three, like what the hell is brown??? ^‡

Imagine my delight, then, when I found a Wordle-style color guessing game called Hexcodle.

The name is a portmanteau of Wordle and “hexadecimal code,” the 16-digit language used to program digital colors. Your phone and TV display millions of colors by mixing precise intensities of red, green, and blue light. The code represents each intensity with a range of 16 digits — 0,1,2,3,4,5,6,7,8,9,A,B,C,D,E,F. Each color usually runs from 0 to 255 — 00 to FF. #FF0000 codes for red, #00FF00 for green, and #0000FF for blue. And mixing all the lights (#FFFFFF) or none of the lights (#000000) in RGB gives white and black, respectively. 

We don’t understand color vision as well as you’d think

Hexcodle is meant to appeal to graphic design nerds. Yet as someone who struggles to tease out subtle elements in color — a dash of red in a sea of blue, or a sprinkle of green in a stone gray — I clung to Hexcodle. In doing so, I posed a provocative question: Could I train myself to improve my color vision? 

I wanted to excel in this game despite my limits. And I wondered whether I could combat those same limits. Over weeks of playing it, the simple game challenged my color vision in ways I never expected. And it made me rethink what I — and everyone else — are seeing when we look at colors.

Plastic eyes

We don’t understand color vision as well as you’d think from decades of academic study. We have classic theories of how red, green, blue, yellow, bright, and dark encode in the brain, Mike Webster of the University of Nevada, Reno, told me. “That's still the theory you get in textbooks. But it's a very naive theory,” he said. “There's a whole mystery of how the brain really represents color.”

I began my quest by phoning the color blindness expert Jay Neitz. Neitz has led studies with his wife, Maureen, for more than. In a pivotal study from 2009, their team at the University of Washington cured color blindness in monkeys — a feat so surprising that it earned them TIME Magazine’s 3rd-best scientific discovery that year. 

Neitz’ Nature study exploded minds for a few reasons. Monkeys (and people) usually have three cones that absorb short (blue/violet), medium (green), and long (yellow-green) wavelengths. Most color deficiencies come from anomalies in the cones, but one form called “dichromatism” is a genetic condition where one cone is entirely missing. The Neitzes treated dichromat monkeys with this categorically severe version of color blindness. And not only did they replace the missing type of cone with a first-of-its-kind gene therapy, they did this in adult monkeys, raising entirely new questions about sensory plasticity.

“Most people believed that you'd have to do it in a very young monkey to get them to be able to take advantage of the new cone,” Neitz said. 

Now … let me just speak for myself, I’m not getting this treatment anytime soon. The therapy is far from approval and, more importantly, I’m not so deficient that I want an eye injection that temporarily detaches the retina. (Neitz’s lab currently investigates more palatable delivery methods.) 

“We learn through our whole entire lives. That's also a kind of plasticity”

What instead compelled me about this work was the sort of neuroplasticity here. Neuroplasticity is basically the idea that the brain forges new connections when existing ones leave it deficient. It’s an innate flexibility that runs wild in your developmental years and wanes with age. Yet something was happening in these adult monkey’s brains that meant they could go from lacking a cone for “normal” vision to quickly accepting a new anatomical input. 

One one hand, this discovery revealed a stunning degree of plasticity. But on the other, I had no plans to change my retinas — I wanted to squeeze more out of the parts I’ve got — and this study highlighted just how important it is for color vision to have all the parts in optimal working order. Gulp.

So I asked Neitz more directly whether he thinks color vision is plastic enough to train. He hesitated for several seconds, then affirmed that our visual systems constantly adjust to what our eyes experience. 

“We learn through our whole entire lives. That's also a kind of plasticity,” Neitz said. 

Neitz demonstrated two decades ago that study participants who wore red-tinted goggles for a few hours every day became less sensitive to red light. Here’s an excerpt from the results, bold emphasis is my own.

“... each subject's unique yellow shifted progressively further away from his or her baseline. Initially, the size of the shift was small; however, after many days of exposure, for example, to the red alteration conditions, the shift had grown so large that wavelengths previously called red came to be consistently called green in appearance.”

These were adult humans, and in another stunning display of plasticity, it took about a week for participants’ color vision to return to normal. But this evidence of lasting changes with repetition fanned the embers of my optimism.

One route for improving my color vision, albeit temporarily, could be an over-the-counter remedy. Commercial colorblind glasses block out confusing colors between (often long-medium) cones that overlap more than normal, but there’s limited evidence of long-lasting perceptual shifts caused by frequent use.

It’s by design that our visual systems recalibrate incessantly, correcting for changes in the world and changes in ourselves. The same neurological “settings” that let you see on a bright day don’t work when you step into a matinee movie. Even as eye lenses yellow with age, the visual system compensates to produce a clear picture. Our perception of colors and ability to discriminate between them changes remarkably little with age, according to John Barbur, a professor of optics and visual science with City St George's University of London. “Color vision is actually arguably the most robust attribute,” he said. 

“As you age, all kinds of horrible things are happening to your visual system — it’s falling apart,” said Webster, , who wrote a scientific review about compensation in color vision. “You don't want to see the world falling apart.” 

So while color discrimination varies from person to person, including among people without deficiencies, the one constant is an individual brain’s desire to extract as much information as possible from a finite number of cells, and keep that information consistent.

That’s where neurological adjustments and compensation shine. Think of it like tweaking contrast on a photograph or focusing a camera lens. Proper adjustments reveal more edges and details — more useful information. Similarly, when color differences are subtle, our adjustments can help discern finer differences, correcting for a yellowing filter or muted colors, for example. The adjustments are volume knobs on a neuron’s activity levels. 

We know that something like this happens because people with anomalous trichromacy can discriminate colors better than their cones may suggest, and not just because they’ve learned which colors “should be” most different. When Webster’s team displayed colors to people in brain scanners in a 2021 study, they reported stronger associated brain activity in people with deficient cones compared to people with functional ones. “It really is a sensory magnification and not just a learned way of reporting the world,” Webster said.

There’s no evidence that we can train away color deficiencies but, theoretically, practice can improve vision perception in general. “There's no question, you can train yourself to become better.” Webster said. “By forcing yourself to make these fine changes and getting the feedback on what's correct, you can really learn to see better.” 

Hearing that excited me, but I couldn’t help wondering whether I could notice any gains.

Gaming my vision

Within my first few days playing Hexcodle, I noticed my color deficiency sneaking up on me. 

Some of my supposedly incorrect guesses looked indistinguishable (to me) from the correct target. On Hexcodle #438, for example, I saw a bright blue just slightly off from pure. I guessed “0A” for red, “39” for green, and “E3” for blue. (Two hexadecimal digits cover 0 to 255. So this RGB guess translates to “10”, “57”, and “227” in conventional base-10 numbers). 

My guess seemed quite close. But it was way off. So I guessed again based on the feedback: #3535FA. Still wrong but now I really couldn’t distinguish my guess from the target. It took me two more guesses to find the right answer: #5336F7, or “Meteor Shower.” Put differently, my first guess of red intensity was about 4%, but the correct red intensity was 33%. 

It felt like the only way to maneuver towards a correct answer was to use hints or guess wildly. How could I expect to learn the lesson when I couldn’t see the lesson?

I should caveat for a sec. There are 16,777,216 unique six-digit hex codes. If you feel it’s unrealistic to pick the exact code on the first or second try … join the club: That’d be a bonkers color acuity. Borderline sorcery. I agree. 

Unfortunately, you and I are wrong. Some people can do it. Hexcodle developers, Ekim Karabey and Hannah Larsen added user-requested “Hard” and “Expert” modes that limit and eliminate hints. And TikToker Jared Cross has gone viral for his skills nailing colors immediately. In one video, he appears to play the game with inverted colors. 

I assumed that as the eyes behind Hexcodle, Karabey and Larsen would excel at the game. But Larsen confessed that’s a common misconception. “We're just as clueless as everyone else,” she said. 

Larsen and Karabey both prefer the “Mini” version of the game, which only requests one digit for each red, green, and blue channel. “I just feel like my eyes aren't well trained enough. I'm literally not able to tell the difference to that granular extent,” Karabey said.

In more neurological words, what he might mean is that his brain can’t amplify a signal that it can’t detect in the first place. 

The duo behind the internet’s best daily color-guessing game

Meet Hannah and Ekim, the minds (and eyes) behind the nerdiest graphic design game around.

SequencerMax G. Levy

But how much do we really know about how our brains process color? Increasingly, the inner-workings of color vision are being questioned. 

Scientists’ classic model for human color vision is opponent process theory, which proposes that we build color information from different scales: light vs. dark, blue vs. yellow, and red vs. green. If you imagine each quality existing on an axis, you wind up with a multidimensional space of limitless color combinations. It’s intuitive from retinal anatomy because these colors correspond roughly to our cones. But it also implies that pure red and green, for example, are singular characters inside the brain. That’s where the theory starts to fray. 

Scientists have yet to find cells in the brain that directly represent the red-green and blue-yellow comparisons predicted by the theory. “Yellow, red, green, and blue should be special,” Webster said. “But the more we look for them, we can't find a place in the brain where those colors are any more represented or in a different way than other hues like orange or purple.” 

Even the standard roles of your hordes of cones, rods, and photosensitive ganglion (which contribute to circadian rhythms) get muddled in real vision.  These different “channels” of vision don’t perform in isolation. When one fails, the other signals become more important. It’s a sort of compensation, or adaptation, or plasticity. “To me, it's just making best use out of what there is,” Barbur said. 

By extension, there’s no way to drill one of these players without engaging the whole system. 

My takeaway was that you can train day and night on color vision tests — really devote attention toward piecing together camouflaged numbers — and you may improve at that task. But the system is so complex that we don’t yet know what is improving biologically. 

And, for now, we can’t say whether training in color tasks can improve color vision in general. In Webster’s 2020 review, he and his co-authors propose that the limits of plasticity are still equally mysterious: “Understanding why compensation is incomplete is as important as understanding why it occurs at all,” they wrote.

These different “channels” of vision don’t perform in isolation. When one fails, the other signals become more important.

My Hexcodle performance has improved slightly, but I can’t say whether that means I have better color acuity. I want to believe I’m more perceptive. But is it because I’ve improved some aspect of my biology, or am I just more attentive? When I stare at a color, a conveyor belt of questions rolls into my psyche: Is there green in this blue? Red? Is it pale? I ask myself whether gray is an absence of color. 

Just as the yellow you see on your phone is actually a sneaky combo of red and green pixels, gray is what happens when red, green, and blue intensities match. So colored light comes from imbalance. Migrating from #777777 (Lucky Grey) to #7777BB presents a blue-tinged Stormy Horizon. Move down to #777733, and you’ll find Garden Weed. 

This renewed perspective helps me imagine what happens in my cones, deficiencies and all. (Bless their little hearts). It may help as I continue seeking gains. “The only way you know what the rules are about — what plasticity there is — is to do the experiment,” Neitz told me. 

Neitz’s latest research explores what “extra dimension” may arise when people (mainly women) have a fourth cone.The benefit of tetrachromacy remains a mystery since trichromatic vision has guided so much of our world, from paints to LEDs to language. Though we lack the vocabulary to describe that extra dimension, tetrachromats may find it the boundless spectra of nature: floral pigments, soils, and of course underripe bananas. 

Maybe tetrachromats can see something in a banana peel that you don’t. Then again, maybe one day I will too.