Listener supported. WNYC Studios. This is Science Friday. I'm Flora Lixman. Today on the podcast, what if you could see a color no one has ever seen before? It was so much more saturated than any teal I've seen before. In fact, more saturated than any teal I could see in the natural world. Think about the colors of the world around you. The blue of a cloudless sky, the green of a new leaf, the blazing red of a tulip's petals. We see these colors because of the way our eyes work.
But what if we could change how our eyes respond to light? presenting them with light in a form they'd never encounter in the natural world. Well, researchers tried that, and it turns out that trick might make the eye see a color it's never seen before. Joining me now to talk about it are James Fong. He's a computer science PhD student at UC Berkeley and one of the authors of a paper describing this research in the journal Science Advances.
and his advisor, Dr. Ren Ung, a professor in electrical engineering and computer science at UC Berkeley. Welcome to you both. Good to be here. Hey, Floyd. Thanks for having us. Wren, I want to start with you. You made yourself a study subject, so that means that you are one of the few people on the planet to have seen this color. Can you just walk me through what that experience was like?
Absolutely. I'll start by saying it was a pleasure to see this color in many ways, many dimensions, no pun intended. First, because it was very recognizable as something new. because it looked like a different color than the color of the laser, the natural color of the laser that we're using to stimulate the retina. So it's very obvious when it happened. I could name the color right away, and it was beautiful.
The other thing was, of course, when I realized that it was a different color, it was a sign, a visual sign that the system was starting to work precisely enough. And the system can do many things. So we've been working on it, especially James, for many years. So there was also the delight in realizing that all of that hard work was starting to pay off. The experiment work, too. That's right.
OK, I feel like this is a tough question because it's a new color, so it's hard, maybe hard to describe it. Is there a way to describe it so that we might be able to. Picture it or approximate it? Absolutely. You know, that's one of the things about this particular study I think that's been fun is that it's very easy to describe and I think can be quite relatable for folks.
Before I describe it, though, maybe I'll mention that the notion of new colors absolutely encompasses colors that you wouldn't be able to name, that wouldn't look like colors of the rainbow, at least in theory, right? So that's kind of... what this system is enabling us to do going forward today. But for the study that was just published, it was the first example of a color that showed the system is working. And that color is, we named, well, James actually named it Olo.
And it looks blue-green. It's very nameable. Teal, peacock green, peacock blue. And the thing that distinguishes it is that it was so much more saturated than any... teal I've seen before. In fact, more saturated than any teal I could see in the natural world. So Olo's super teal. But you said that you're planning to make new colors that aren't the colors of the rainbow, that are sort of beyond our imagination of what a color is?
That's right. I should rewind a little bit on the concept of colors that you don't see already and couldn't name without a new term. Let's start with colorblindness, okay? Someone that's completely colorblind. And vision scientists often say that a completely colorblind person goes around the world and sees everything in shades of blue and yellow only.
Pause for a moment. And just imagine if you're looking around, you look around your room right now. And if you have normal color vision, of course, you see all the colors of the rainbow. But imagine that everything looked just blue and yellow. And your whole life, it just looked blue and yellow. And then imagine that one day you open your eyes and you see all the colors of the rainbow that have been missing in your perception up to that point.
That's the kind of color novelty that I'm describing. That's a new dimension of color, okay, from color blindness to full color. And this is on the horizon? With this experimental setup? Exactly. This is research that we're doing right now in the lab using this platform on which we saw Olo. That platform is called OzVision. We're trying to boost...
colorblind individuals to full color and see what the perception is and what kind of new color functionality they have, what color tests they can pass now. And then about new colors beyond the rainbow. The idea there is to boost someone from having three photoreceptor types in the retina, normal color vision, to four. And if you do that, there's another student in our group, Jessica Lee.
She has a paper about tetrachromacy and the theory of what tetrachromatic color experience would be like, like what would it be like to try to walk around the world and see the world filled with this four-dimensional color instead of three-dimensional color. The brain would be receiving a dimension of color beyond the rainbow and trying to determine whether the brain can receive that, make sense of that, create a visual percept out of that is the core of the reason.
Wow. James, talk me through the setup for this experiment. Okay. So... The idea is that the end goal, I think, of this platform is that we want to have programmable control over the activity of every cone cell in the retina at every point in the time. And if you don't already know, the cone cells plus the rows in the retina are... the light sensitive cells in your eye that eventually lead to your vision.
So if one had full programmable control over all of those individual cells, then in principle, you should be able to recreate any visual stimulus that you could possibly imagine, at least the ones that come from the real world. To be clear, that's not something that we've totally achieved in this study, but that's sort of the main goal of what we're trying to do here. So the current platform right now, the way it works is that we use this.
wonderful piece of hardware called an adaptive optics scanning laser ophthalmoscope, which is a big mouthful, but the idea is that we're able to... project this tiny little point of light that is about the size of a cone cell onto your eye and scan it such that we can collect a live feed of what the retina looks like. So it looks like this black and white image, which is a bunch of bright spots that indicate where all the cone cells are.
And at the same time, in that same little bright spot of light, which is, I should say, is in infrared, so it's nearly invisible to the subject, we inject a visible wavelength beam, so in this case, 543 nanometers. and then modulates the intensity of that beam so that as it passes over each cone cell of interest, it... is increased to the level necessary to activate that cone cell to the degree that we're interested in having it activate, right? So you're shooting a laser into the eye.
And you can shoot very specific wavelengths of light. And you, importantly, can actually target the exact cone cells you want to hit. Yes, that's the idea. And normally humans have... Three types of cones, right? Yes. So there's three types of cone cells. They call them the L, M, and S cone cells for short, which are short for being sensitive to relatively long, medium, or short wavelengths of light.
All right, so the long wavelength sensitive cones are more receptive to the red end of the rainbow spectrum. And then the short wavelength cones are more receptive to the bluer end. And then the M cells, the medium wavelength cone cells are somewhere in between.
What's really interesting about the M cells is that the spectral sensitivity is very highly overlapped with the L and the S spectral sensitivities. So as you sweep through different wavelengths of light, you'll find that there's no wavelength of light. that will only activate the M's and neither the L's or the S's. So this is interesting, right? This also means by extension that no natural light, which is a combination of these individual wavelengths, can stimulate the M cells in isolation.
So Olo is interesting because that corresponds to us taking this programmable platform and programming it to send light, like to target the light only to the M cells. If done correctly, then only the M cells would fire, and this would correspond to a totally new color percept. Okay, so when we're talking about giving people access to new colors, colors that aren't part of the rainbow we've never seen before.
tetrachromatic vision, how do you actually do that? Is it more than just exposing them to certain wavelengths of light? It totally is. And in fact, in our group, we're trying to do this in two very different ways, Flora. So the first one is with this OzVision system where people have seen Olo. The idea is that we take a subset of the cells on the retina and we say, hey, we're going to call you a virtual new clone type.
And we pretend that they're filled with a fourth photopigment that they're not. But we stimulate them as if they are filled with that fourth photopigment, and thereby we send a fourth channel of color to the brain. Then it's all up to sort of can the brain make sense of that or not. The idea is how many distinct cell types there are sending color information to the brain. In a colorblind person, a hard colorblind person, that's two. In regular color vision, it's three.
And one way to send that third color channel of or additional color channels of information into the brain is to program OzVision. Another way might be to change the eyeball itself. that isn't being done in people, but studies have been done in squirrel monkeys. In the Knight's lab of University of Washington, they use gene therapy to add a third photopigment type onto the retina physically of colorblind squirrel monkeys.
Those squirrel monkeys through into adulthood have failed color tests, right, their whole life. In adulthood, they get this gene therapy. The eyeballs turn from two cone receptor types to three. And as far as the researchers can tell right away, those monkeys start passing these color vision tests full speed that they failed their whole lives into adulthood. So it's just a remarkable study. So that'd be another way to do it.
You know, Ren, this study makes me think about a thing that I think, you know, people often talk about, about perception, where this question of like, is your experience of a color, you know, or generally experience of the world the same as my experience of it? Does this work? help us understand that question better. Is that a question that's interesting to you as someone who studies this?
I love that you went there, Flora, because that is one of the big questions, right? Philosophically, or when we ruminate about what we know. the philosophy of what we can know about our experience or what we know about the world from our perception. Those are core questions. And we love those questions in our research group because we have to literally confront those philosophical questions every day and trying to think about how to do this science in a rigorous way.
So absolutely, we have to really ask ourselves, hey, how do I know if the yellow that I'm seeing is the same as your yellow? or if the teal that I'm seeing is the same as your teal, or if we manage to boost a colorblind person to something that functionally seems like all colors of the rainbow or full color vision. How would we really convince ourselves of that? You know, what is a science experiment to do that?
So we can test a person that's been boosted from colorblind vision to full color and see if now they need three colors in order to match all the test colors that we can show them. So that's one. And vision scientists love that kind of test. I think everyday people have a different experience of color that is more unusual to think about theoretically, and that is we can order the colors of the rainbow.
Right. If I ask you, Flora, think of the rainbow. And what are the neighboring hues for orange? You would say yellow. Red. Red and yellow. Most people with a little bit of thought will say, yeah, absolutely. Orange is neighboring by yellow and red. It's definitely not green. It's definitely not purple. Those would be weird neighbors for orange.
If you're a colorblind person that sees the rainbow for the first time, one thing you could ask is, hey, can that person perceive the rainbow in a way where that ordering also comes through? Is that like a natural perceptual thing that occurs? Oh, that's fascinating. Yeah. Like, is it intuitive that those are neighboring color? That's right. Another one is, does the color pop out to you?
If you look at a tree that has a red apple on it, a person that has full color vision will instantly say, oh yeah, there's an apple on that tree. But a colorblind person can't because they can't differentiate red and green. So they have to look at it and scan it. And then they would recognize the shape of the apple, but it would take time. It's a slow process. And you can say, hey, if a colorblind person got boosted to full color, now they can differentiate red and green. OK, great.
Certainly more than colorblindness, but does a red apple pop out on a tree to them instantly the way it does for someone that is, you know, that was born with normal color vision? After the break, the challenge of coming up with a name for a color no one has ever seen before. So we kept calling it LMS010, but that gets kind of boring. And so almost for my own amusement, I was just trying to come up with, oh, what's a better name for this color? Stick around.
Support for Science Friday comes from the Alfred P. Sloan Foundation, working to enhance public understanding of science, technology, and economics in the modern world. We sometimes hear about language being important for understanding color. Like if you don't have a word for the shade of a color, you might not see it. I think that is such an interesting idea, Flora. And I really don't know the answer to that. I'm so curious about it. And I think we have a unique way to get a peek into that.
At the moment where a colorblind person in this OzVision system begins to have this ability to differentiate colors, red, green, order the rainbow, if that's possible, we're testing right away. Can they learn to name the color? And then we want to ask them what the appearance of this is.
And I think there is an opportunity in trying to learn what the names of these colors are in watching video. OK, like, you know, TV shows, for example, and trying to understand or to think about what the colors that the person is seeing and what their names are. Does the experience of viewing those colors, we call that the qualia, right? Does the qualia of those viewed colors change?
in connection with language and the cognitive task of trying to think about what those color names are and how one would communicate it to others. We don't know yet, but if I had to bet, I bet it does have an effect. The language, I'm particularly connected with doing tasks that require you to use the color information. We will see. James, why did you pick the word Olo for this color?
Oh, yeah. So that's a fun story. So for a long time, like way before we were anywhere close to being able to show Olo for real, we were just sort of talking about it as a... coordinate in ambient three-dimensional color space. So you can imagine that there's a three-dimensional color space that includes all natural colors as well as these hypothetical imaginary colors. that you can just label the axes L, M, and S corresponding to the activation levels of the L, M, and S photoreceptors.
And so OLO would be corresponding to the point which has 0L. 1M and then 0S activation. So 0, 1, and 0. And so, yeah, maybe you can start to see this, right? So nerdy. I love it. Yes, yes, yes. So we kept calling it, oh, LMS 010, LMS 010. That gets kind of boring. And so almost for my own amusement, I was just trying to come up with, oh, what's a better name for this color? And I came up with this short list of ideas. And one of them was, why don't we just take 010?
kind of write it out on the whiteboard and kind of squint your eyes. And so, yeah, it looks kind of like the word Olo. Originally, it was just for entertainment, but people really, really loved the name. Hell, that was so great. It's this XOR background, but it's, you know, it just sounds great. And someone pointed out to me that, you know, it's the middle three letters of color, Flora, and I love that too. Yes, yes, that's a total coincidence.
But yes, the name comes directly from just the coordinate and color space. which actually has a lot going for it, right? So for instance, it doesn't presume any particular appearance of the color. Okay, so I think when we were originally, we were trying to think about what this might look like.
Maybe colloquially, we might have called it the greenest green you've ever seen. But we soon realized, well, actually, it's not good to try and name it before anyone's actually seen it. Because who knows, maybe it's something totally different. And in fact...
we would have been wrong. Like a child, you know, like when you see it, you might think the name doesn't fit at all. Yeah, yeah, exactly. So I remember very fondly that someone had come up to me and then said, oh, I really love the name because not only is it cute. But it rhymes with YOLO. And this was sort of, you know, this is during the time when people said, you know.
YOLO culture was sort of still in a tie. It's a meme now. YOLO is a meme now. And you came up with that name. Yeah, thank you. James Fong. Yes, thanks, friend. It's not every day that you get to name a new color. So I'm glad that. I got the chance to do that. James Fong and Dr. Ren Ung, both in electrical engineering and computer science at UC Berkeley. Thanks to you both for joining me today. Laura, thanks so much for having us. Yeah, thanks so much for having us.
And that is about all we have time for. Lots of folks helped make this show happen, including... I'm Flora Lichtman. Thanks for listening.