The many abilities of cuttlefish
Wardill: Cuttlefish [are] kinda described as the masters of camouflage. They can change their skin texture. They can change the coloration. They can charge their pattern, and change their color. So, there’s a lot of features there to make them hide very well. And also what’s interesting about cephalopods is they are extremely good predators. They will hunt after prey by ballistically shooting out – or sort of extending their tentacles – to capture prey. And they have a kind of thin mucus coating all over their skin. And so one of the interesting things about cephalopods is, I guess because they’re soft and squishy, they’ve evolved to regenerate extremely well. So you can cut off at least two-thirds of the arm and it’ll grow back. You know, their neural system – in like, for example, an octopus – you can cut half of their spinal cord and it’ll completely regenerate millions of neurons back to being as though it was normal. You know, you wouldn’t be able to tell. Which, to me, is insanely remarkable. But, when you think about evolution, it’s like, “Hang on. If everything’s wanting to eat you, you better have a way of at least recovering from some of the smaller injuries.” And so, the first ones we were using were almost two years old, and they don’t live much longer than that, to be honest. In fact, in the wild, they probably wouldn’t make it that far. That’d be eaten by something. And the size is probably about maybe 8 – 10 inches from the tip of the mantle to sort of where the arms are sitting. So about let’s say 20 centimeters, something like that. But they can grow much bigger in the wild. Like they’ll get easily more than a foot long, you know, maybe even you know, a foot and a half long. European cuttlefish, which is the ones where we using in the study, are very big.
[ Back to topics ]
Origins of the study
Leigh: Trevor’s lab investigates the neural processing of visual information by invertebrates, the knowledge of which is particularly scarce. Much of their research focuses on visual circuitry of the fruit fly, as well as in other species of flies. But he and his team are also interested in the visual system of cuttlefish, which – like squid and octopuses – are cephalopods. So Ryan and I were curious what interested him in studying these animals.
Wardill: Because this is a very visually driven process, we were fairly certain, you know, the first time we saw cuttlefish in Roger Hanlon’s lab, we were like, “That has to be judging distance. There’s no way that would work without it.” And he’s one of the world experts on cephalopods, if not the world expert on cuttlefish, in particular, about their camouflage and their behavior in the wild. But, you know, John Messenger had looked at that, you know, some while ago and found that if you surgically deactivate one eye, they still can capture prey. And we were shocked. But it turns out, you know, there was still hanging questions of why this would be the case. And, around this time we were speculating about, you know, it should be able to judge depth. And a paper came out from Jenny Read’s lab on praying mantis. And you know, they’d known before that that it had 3D vision or stereopsis, but they wanted to do some more sophisticated experiments. And so the idea was: if you use, you know, those colored 3D glasses that, you know, we used to have in the cinema, they would be a good way of investigating how the brain processes depth information. And so because that paper came out and we read it, and we’re all like, “Wow, this is so cool. Maybe we can do this with cuttlefish,” was the crazy idea. We sort of looked at each other and were like, “I don’t know if this is gonna work,” but, we’re like, “It would be very cool if did.” And so that was sort of the inspiration, you know: just watching animals, and then seeing another paper, we sort of joined the dots and thought, “Let’s try.”
[ Back to topics ]
Eye convergence, divergence, and yoked eye movement
Watkins: Unlike us, the eyes of cuttlefish are on the opposite sides of their heads. So they obtain 3D vision through a process called “vergence,” in which their eyes rotate towards each other when focusing on closer objects, and away from each other when looking at objects that are farther away. We asked Trevor to explain how this process operates.
Wardill: Eye convergence is when – you know, in the case of cuttlefish – the eyes rotate from looking laterally to looking forward. And so, they’re converging to the same point. And to get some sort of scientific term, it’s called “yoked” when our eyes are locked together. And we’re permanently that way. And so, our eyes have a certain distance between them. And because they’re always moving this yoked [way] together, then our binocular overlap doesn’t change. And then “divergence” is pulling your eyes away. We can’t do that because ours are as said yoke together. cuttlefish can. And there’s a really good reason, probably, that cuttlefish have this independent eye movement, as well as the stereopsis part. It’s because when they’re looking laterally, you know, each eye has about a 180 degree view. Which means the two eyes together can see almost 360 degrees around the animal. Which means if a predator is coming, they will see some motion and be able to get the hell out of there or adjust their camouflage so that they don’t get seen. But when they hunt they rotate their eyes around to the front and then they can’t see predators as well behind them. And we don’t really know – like we know this specialization in the retina is similar to us where we have a fovea. We know from looking at, you know, retinas of cuttlefish that they have a fovea. It’s a little bit different, it’s sort of more elongated. But what we don’t know for sure that fovea is the bit that’s focusing on the pray they’re going to capture. But you would anticipate that’s what’s going on.
[ Back to topics ]
Outfitting cuttlefish with 3D glasses
Leigh: With these “preliminary” matters out of the way, Ryan and I were eager to learn: just how do you go about putting 3D glasses on a cuttlefish anyhow?
Wardill: Well, cephalopods in general are colorblind. And, you know, when you first read that or hear about that, it’s like, “Whoa, that’s ridiculous.” I mean, they produce color on their skin. That’s their daily business. But we had a little bit of experience from a previous project where we wanted to blindfold the cuttlefish. Because there was an idea that the opsin that’s expressed in their eye, which is also expressed in their skin, may give them some ability to see color. But anyway, long story short, we figured out something ages ago about blindfolding cuttlefish. And one of the ways of attaching blindfold was using super glue. And the nice thing about super glue – you know, anyone who’s had an accident with super glue – is that if you put it on the skin, it will fall off. You know, it might take a few days, but it will fall off. And so, that was how I think of well, that’s how I’m going to put them on. I’m going to use super glue to put it on their head. And the other thing we kind of knew is that the animals were going to play with it. So, you know, they have eight arms that are very flexible. Whatever you put on the animal, they’re going to try and take off. So, the second sort of thing that we guessed was, “If I super glue on a piece of Velcro, then the glasses could be attached with Velcro. But, equally, the cuttlefish could take its glasses off if it just didn’t like it.” And that, you know, was a really good idea in hindsight because, yes, as you might guess, some cuttlefish – when you put glasses on their head – don’t want to wear them and just take them off. And it’s like, “All right, fair enough. You are a cuttlefish.” But what we discovered is that if we put the glasses on the animal, and put it in the tank, and immediately gave it shrimp – something to do, something to think about – most of the time – and by most I mean probably around 60% of the time – they would be off wanting to chase that snack than worry about the glasses. And the glasses were designed such that they had a similar luminance on either side. So, for their visual input, it would just look like the light intensity was less than it was before. But in marine environment, light intensities change a lot. And so, being less, they like it better. They actually don’t like bright light. And so, once they’re on and they’re not sensing that, you know, they’re annoying – then they’re probably not a bad thing.
[ Back to topics ]
How blue/red 3D glasses work (even with colorblind cuttlefish)
Watkins: While it’s more common today for 3D movies to be viewed with grayish polarized glasses, the earliest system used one red and one blue lens to form a stereoscopic 3D image. As these were the sort of lenses that Trevor and his team used as well, we asked him to remind us how 3D glasses create the illusion of depth, as well as how red/blue lenses could work with cuttlefish, since they lack color vision. We’ll hear what he had to say after this short break.
ad: Altmetric‘s podcast, available on Soundcloud, Apple Podcasts & Google Podcasts
Watkins: Here again is Trevor Wardill.
Wardill: The whole idea of the 3D glasses for the cinema is that one eye is getting one stimulus and the other eye is actually getting a different stimulus. So, if you go to the cinema – well, if you saw these old films and you took the 3D glasses off, what you would see is that there’s two images, but they’re offset. There’s a disparity, there’s a difference between the two images that they’re projecting onto the screen. But by you wearing the glasses you’re filtering out the wrong image in one eye and vice versa in the other eye. And then what your brain is doing is taking those two images, which are different, and figuring out what the disparity is; what the difference is. And that difference tells your brain whether, you know, the Tyrannosaurus Rex in the movie is going to jump out and bite you, or whether it’s a tree a long way away in the distance. And so it gives you that 3D effect. In the case the cuttlefish we can generate two shrimp images, and make them a certain distance apart, and run those on the screen. That’s pretty trivial. Well, there was a bit of work behind it, but once you figured that out, it was trivial. But as it turns out, a cuttlefish being colorblind is actually even better. We use the filters – the colored filters, in this case – to block out the wrong image going to the eye of interest, and vice versa, so that again there’s a difference in image between the left and the eye. We could use a human-sized version of these glasses that we used for the cuttlefish to check our stimulus, but we have color vision, which meant it wouldn’t look as effective to us. That the depth information from the cuttlefish movie – well, in this case the shrimp movie I suppose – as it would to the cuttlefish, but the same principle applies. So us humans can figure out the depth, and and so can the cuttlefish.
[ Back to topics ]
The experimental arena
Leigh: Since studying the eye movements of cuttlefish – whether donned in 3D glasses or not – has to be a non-trivial task, Ryan and I were curious about the experimental set-up that allowed Trevor and his team to capture detailed information about their vision.
Wardill: We have the cuttlefish in a fairly square tank. It’s about a foot and half square, and about a foot deep of water. And the cuttlefish is in there looking at the screen, which is up against the side of the tank. And from above we have a high-speed camera. And by high-speed, I mean it’ll, you know, run in a thousand frames per second. And it’s high resolution. So it’s 2,000 pixels by 2,000 pixels. It’s camera from Photron. They’re very expensive because they are the best in the world, as far as we know. We’ve tried lots of different manufacturers, but we love them. For good reason. They have amazing software, but the cameras are also very, very sensitive to light. And what that means is you can get, you know, exquisite detail even at low light. And because that camera covers most of the tank, we can zoom down and we can see the actual eyes of the cuttlefish. And cuttlefish have a W-shaped pupil, basically. We have a round pupil. Theirs has a slit that forms a W-shape, and that W-shape is thought to reduce the light down considerably in bright daylight, but at night time when it’s very dimmed, then they have a round pupil like ours that just lets in lots and lots of light. It’s thought that it helps work with underlying retina. But because there’s a W-shape in bright light, which is when we do most of our experiments, then from above you can actually see the sort of – in this case it’s an inverted W. So you can see the bottoms of the W from an overhead shot. And so, we can track those two points over time, and that gives us the direction in which the eye’s viewing. So we were quite lucky, in this case, that they have these funny-shaped pupils, because that meant in any one particular frame from the video we can extract the position of where the eyes are looking. And that’s how we got this change of eye movement over time.
[ Back to topics ]
Stereopsis and independent eye movement among cuttlefish
Watkins: Trevor’s article reports three surprising findings regarding cuttlefish vision: first, they see in stereo; second, they strike at shrimp more quickly when they’re able to perceive depth; and third, they can extract 3D cues even when the information received by each eye isn’t correlated in the way that’s required by humans. We asked Trevor what these findings suggest about cuttlefish vision, and how that’s different from what we’ve previously known about how it works.
Wardill: 3D vision – or stereopsis – was originally, you know, not thought to be likely because of the fact they could hunt this one eye. And, over the course of the study, we discovered that their behavior is different when they only have stimulus provided for one eye. They’re not as fast. So they’re much more cautious when hunting with stimulus provided to one eye. And that makes sense, because they don’t have the depth information. But the other thing that was kind remarkable that came out of the study – which I guess we hadn’t appreciated from the literature – is that animals normally that have independent eye movement … like, you can think of a chameleon is a great example of this. Their eyes just sort of crazy it go everywhere. They don’t tend to have stereopsis, or the ability to judge depth from their inputs from the left and the eye. And the reason being, at least the way scientists thought about it, was that when you have independent eye movements, for them to join together, you’ve got a map in your brain somewhere of those two respective fields of view. But then fuse them together to get a sort of perception of the whole world. And that fusing together is a rather complicated problem because – you know, in the case of chameleon is good one – if you think in your mind, their eyes rotate around as well as moving, you know, up and down. So you can imagine for them trying to fuse that together would be really tricky. And the way they judge depth is by using the focal length of the lens to do what they call “accommodation” to judge distance. And they do that with one eye. And so, the eye movements of cuttlefish are somewhat independent. So what that means is when we move our eyes around, we have them locked together. So with both eyes are looking at the same point. And if we move our eyes, the eyes move together. In the case of the cuttlefish, they don’t always move together. So, in fact, most of the time they’re looking laterally and looking at their own kind of thing. And then when they go to hunt, they rotate the eyes forward so that they have more binocular overlaps and more vision from both eyes. And that’s when they’re more locked together so that they can get the best resolution of the target that they’re about to attack. So the cuttlefish to have stereopsis with independent eye movement was really quite remarkable because they can do both.
[ Back to topics ]
Correlated, uncorrelated and anti-correlated backgrounds
Leigh: Trevor and his team’s study tested three different manipulations. In the first, the pixels comprising the left and right portions of the shrimp were correlated in such a way that both humans and cuttlefish are able to properly determine the image’s depth. In the second condition, the pixels were anti-correlated such that humans wouldn’t properly detect their depth, but – like praying mantis – they found that the cuttlefish could. Lastly, when the pixels were uncorrelated … a condition in which praying mantis are still able to detect depth, both humans and cuttlefish can not, as Trevor explains next.
Wardill: So first, rather than having the stimulus being correlated – so what that means is what the left eye sees in the eye sees in luminance and in position is the same – we can do tricks because we’ve got a computer monitor. So you can give one eye one color and the other is a different color, but now instead of having the same luminance in the other eye, you can have the inverse luminance. So, you know one eyes bright one eyes dim. If you ask a human to judge depth from this type of stimulus – where there’s a difference in luminance between the left and the eye – we can’t figure out depth from that type of cue. So, basically, the cuttlefish can do something a little more advanced than what humans can in judging depth using, you know, stimuli that vary. Then we tried the sort of fancy version where the left and eye differs in luminance, but also the stimulus differs in location. And so that’s kind of mad computation that the brain would have to do to figure out, you know, how those get stitched back together and judge depth. Humans definitely can’t do this. But what was remarkable in Jenny Read’s study was that the praying mantis can figure out depth in this scenario. Which, when I read the paper, I’m like, “How?” And I think, you know, there we don’t know the neural basis, but the idea is that the mantis has to capture prey out of midair. It’s fixed, you know, it’s attached to a leaf or a stick or whatever, but it’s got to get that prey. But behind the prey there could be clutter that could be all sorts of confusing stimuli nearby. And so they have to sort all that out and still capture their lunch. And you know, I guess my assumption here is that that’s particularly complicated to do for a praying mantis visual system. So it’s evolved advanced image processing in its brain to extract in and out of that the appropriate signals. Time will tell the, you know, science will eventually tell us how this works. But for now it’s a pretty complicated problem, and that problem the cuttlefish can’t solve. So they’re better than us, but not as good as a praying mantis in judging depth from complicate stimuli.
[ Back to topics ]
Just how smart are cuttlefish?
Watkins: Trevor’s study appeared not once – not twice – but three separate times on CNN’s homepage … at the same time! This led Doug and I to surmise that what’s so compelling about the study isn’t that you can put 3D glasses on a cuttlefish to fool it into thinking that a video of a shrimp is real: it’s that you apparently have to, suggesting to us that there must be a lot going on in their minds. So we closed out our conversation by asking Trevor his thoughts on the matter.
Wardill: Well, I don’t necessarily like to think of them as smart, you know, like the way we think of, you know, other fellow humans could be very smart. But I do consider them very good problem solvers and very good at learning. So, with cephalopods you can give them a task to do, and straight away they catch on. And you know that because the next thing they do is exactly what you’d want to, and then they repeat that forever. So, if that’s the definition of intelligence then yeah, they’re doing pretty damn well. But it’s hard to be sure what’s going on. You know, there’s other studies that I know are going on at the moment which, when they get published, Will further confirm that they’re able to solve pretty sophisticated puzzles. But not all of these studies I totally agree with. So, for example, octopus aren’t taught by their parents how to open shells and yet, you know, they can easily find some sort of shellfish, figure out how to drill a hole in it, pull it apart and eat it. So they evolved to automatically figure these things out, and so some of the puzzles that people give to cephalopods are like, “Yeah, well, they know how to manipulate stuff all day long [and] what you’ve given them is fairly trivial compared to what their deal with in the wild.” It’s relative to what they’ve evolved to handle, you know, over there whatever 500 and something million years of evolution. And so, we have, you know, quite obvious reasons why understanding stereo vision and understanding how the brain works are quite essential for, you know, the existence of humans in a way. Like, we’re all getting older, and living longer, and we have no idea really about how to deal with an aging brain. And we have, you know, sort of basic knowledge of how the brain works. So Inspiration from other animals, I think, is is quite important for forgetting the principles of how brains work. And that’s kind of why we look at in vertebrates. They’re generally simpler. And so for anyone that sort of might say, “Oh, well, that’s just pointless.” Maybe they should consider why we did it, rather than what we found, as such. Because that tends to inform about the usefulness of what the discovery is.
[ Back to topics ]
Links to article, bonus audio and other materials
Leigh: That was Trevor Wardill discussing his paper “Cuttlefish use stereopsis to strike at prey,” which he published along with multiple co-authors, including his wife Paloma Gonzalez-Bellido, on January 8, 2020 in the journal Science Advances. You’ll find a link to their paper at parsingscience.org/e69, along with transcripts, bonus audio, and other materials we discussed during the episode.
Watkins: We hope that Parsing Science helps you hear what you might not have the time to read. And if you’re new to the show – or just missed a few of our recent episodes – then head over to parsingscience.org to check out our entire catalogue. There, you’ll find our conversation with the guest from our previous episode, Royel Johnson, who spoke with us about what factors best predict the college success of youth formerly in foster care … as well as the episode previous to that, in which we spoke with Temple Grandin about how research can improve the wellbeing of livestock.
[ Back to topics ]
Preview of next episode
Leigh: Next time, in episode 70 of Parsing Science, we’ll be joined by Jeremy Gunawardena from the Department of Systems Biology at Harvard Medical School. He’ll discuss his research replicating an experiment originally conducted over a century ago, confirming that at least one single-cell organism – the trumpet-shaped Stentor roeselii – is capable of surprisingly complex decision-making behaviors.
Jeremy Gunawardena: So I started digging into it, and I, you know, read Jennings papers, and found that he was a, you know, really wonderful scientist. And I was very excited about this. And then I discovered – once I started asking around about, “So, who followed up with these experiments? And what’s the modern story?” I was really horrified to discover that, among those people who think about these things, the experiments were thought not to be reproducible. So that was really the starting point for our work.
Leigh: We hope that you’ll join us again.
[ Back to topics ]
