Are fish smarter than we thought?

This is Maurice the goldfish. A real life variant of Nemo, Dory, Cleo, Bubbles and the rest of our fishy heroes. Maurice is a poster child representative of his animal group, the fish. Most people think they aren't that intelligent. Goldfish like Maurice, for example, are said to have a memory span of only three seconds. But is that really true? From the ocean depths to the shallowest ponds, wherever there's water on the planet, fish are present. And they've been around much longer than us humans. But what do we really know about them? Researchers worldwide have conducted a range of experiments on fish to test facets of what we call intelligence, cognitive abilities, social organisation, adaptability and even emotions. The results are not only surprising, but verge on the spectacular. Maurice and co are apparently capable of much more than we previously thought. A team of Japanese scientists have demonstrated that goldfish were able to tell the difference between Bach's Toccata and Stravinsky's The Rite of Spring. The test for the fish consisted in having them swim to one corner or other of the tank. According to the piece of music being played in order to get a food reward. This extraordinary feat allows only one conclusion. Maurice the goldfish and his fellow species seem to be able to hear. And that in turn prompts the question, if fish can hear, can they also produce sounds? But fish don't have vocal cords. So how could they possibly do that? The lagoon of Morea in the South Pacific offers the best conditions for exploring the soundscape of a coral reef. Eric Pementier and David Leckini are bio-acousticians. Equipped with a hydrophone, the two scientists are searching for sounds. This waterproof microphone converts sound waves into electrical signals. First, they encounter a kind of cooing sound. It comes from the damselfish that live here. You can hear these cooing sounds quite clearly. It's a pretty cute mechanism. So you've got to picture the upper and lower jaws, and then a ligament connecting the tongue to the lower jaw. Whenever the fish pulls its tongue back, its mouth closes. It can do this 10 times in a row, really fast. That's what produces this cooing sound. Damselfish make sounds even without vocal cords. With their trained ears, the researchers identify another sound that's produced by a different mechanism. It's not a damselfish, it's a soldierfish, also known here as ee. The sound is produced when muscular action causes the swim bladder to vibrate. This is an organ that helps fish stabilize their position in the water. Soldierfish have two muscles that run from the skull along the ribs, which in turn are connected to the swim bladder. The speed of muscle contraction determines the frequency of the sound. Different species make different sounds, so the cause can be used to identify the fish. Clattering jaws, drumming muscles. Turns out fish aren't so quiet after all. So what's the purpose of these sounds? What could these fish be saying? The typical sound of a coral reef is an organized cacophony, with dozens of species communicating with each other for various reasons. What we're hearing right now are damselfish. During the mating season, the males emit acoustic signals to attract females. Two rivals may also communicate. When another fish comes too close, a clownfish and its anemone generates a sound to say, this is my home. They can also emit warning signals when a predator approaches, and they use communication to orient themselves in the lagoon, which is very large. Whether courting a mate, defending their territory, or warning off enemies, fish apparently communicate with each other in their own unique way. This acoustic communication can certainly be described as language. They're constantly communicating, always with a purpose, and it's no different with Maurice and his aquarium. They hear everything at different frequencies and at different times of the day and night. But can fish also remember what they've just communicated? The myth of the goldfish's three-second memory is persistent. But what if reality is quite different? Memory would be another sign of intelligence in fish. To test Maurice's memory, we're going to try to train him. Will he be able to learn a few tricks? Over several weeks of training, our goldfish will learn to navigate different routes within its aquarium. The first step is to follow this rod. It hides a reward, providing a real incentive for the fish to take part in the experiment. Maurice quickly realizes that the rod contains food. But the big question is, can he remember this? And for how long? At the University of Oxford, ethologists Kate Newport and Anna Garcia research the behavior of animals in their natural habitats. They are investigating the spatial memory of various fish species. To prove that goldfish have more complex cognitive abilities than previously thought, they have designed various experiments. One of them takes place in this black-and-white striped tank. The question is, can goldfish remember how long a certain route is? So what gets rewarded is traveling that learned distance and then returning back to their home area. Once we've trained the fish to be able to swim the learned distance, they'll actually swim just to here, stop swimming, not go all the way to the end, and then turn around and come back. But how do the fish know when to turn around? Are they simply acting on instinct? Or, hard to believe, could they be using the number of stripes they've passed as a kind of internal measure? To find out, the researchers halve the width of the stripes. If the fish truly understands the distance traveled, it will now cross twice as many stripes before turning. But if it's counting stripes to gauge distance, it will turn back halfway through. So they won't keep going, they'll just go to here and then turn around. Exactly, because they're getting the same frequency of spatial information at this point that they would be with double the width stripes at this point. So they think they've traveled the distance, then turn around and come home for a food reward. This would mean that fish can memorize visual clues in their environment to use for orientation. In this case, the stripes in the tank. But why do we assume fish have memories like sieves? Where does that misconception come from? People used to think that fish weren't very intelligent. And I think a lot of this actually comes from our own human bias. We thought that fish were very, very different from humans, and only humans are intelligent. And so clearly fish and anything kind of below the primates had no intelligence. And if we look at the structures of the brain, the brain of humans is really large and covering the outside of it is the highly folded neocortex. And this is the site of advanced processing. And this is where emotions and abstract reasoning and memory are being processed. But in fish, we can't say specifically where memory is occurring. We think it could be in parts of the hippocampus. But there's still a lot more research to be done to know for sure if that's the entire story or if some parts of memory, for example, short term might occur there and longer term memories might be formed elsewhere in the brain. So little is known about how the brains of fish work. And that's precisely what makes them so interesting for researchers. After less than a week of training, Maurice follows the rod even through obstacles that a fish would normally find too risky. This hoop, for instance. Or this narrow tunnel. Without a reward, he'd probably avoid such perils. But Maurice associates the stick with a positive experience. So fish can be trained. They're capable of learning something if it promises a positive return, i.e. food, just like any other trained animal. A sign of intelligence. So why are fish so underestimated? At first glance, they're not in the least like mammals, let alone us humans. But fish are vertebrates just like humans. The collections in the basement of the Paris Museum of Natural History reveal more about the evolution of the so-called fish. This is a genetic term encompassing a vast range of species. Science doesn't recognize fish as a category, though of course fishmongers still do. What we traditionally refer to as fish is a huge group of animals. 33,000 species, which is almost 60% of all modern day vertebrates. So the most diverse group of vertebrates on earth is fish. There are three classes. The first one being bony fish. Cartilaginous fish such as rays and sharks, whose skeletons are made of cartilage, and then the somewhat smaller class of jawless fish. In this model showing the evolutionary relationships between different organisms, humans are found in the tetrapod branch of the vertebrates. And here it becomes apparent the ancestor of the tetrapods split from the other bony fish much more recently than the bony fish diverged from the cartilaginous fish. In other words, from an evolutionary perspective, goldfish are closer to humans than sharks. Man is terrestrial, fish are aquatic, and so it was considered that everything that hadn't left the water was primitive. Leaving the water was equated with a superior level of development. But 420 million years of fish evolution can't be ignored. That's enough time to specialize and develop the intelligence needed to survive underwater. Maurice's next task is to kick the ball into the goal. The first steps are the hardest, they say. But who needs legs when you have a mouth? Well, that didn't work out too well. But for some species, such as the orange-dotted tuskfish, the mouth does play a decisive role in procuring food. Biologist Giacomo Bernardi was the first to film this type of wrasse in action. He digs up the clam, then swims with it more than 10 meters to a rock, his anvil, the only place in the reef where he can crack the clam open to then eat it. I noticed that there were already several shell fragments here, so he comes there regularly to use this tool and feed himself. Just like crows and monkeys, some fish species are apparently capable of using tools. Unlike their land-based relatives, however, it is almost impossible to document this behavior. The biggest problem with underwater research is that you have so little time. A dive lasts an hour, sometimes an hour and a half at best, which is not enough for studying behavior. It was one of the most wonderful moments in my life as a marine biologist when I saw this fish using a tool. I had been waiting 25 years for that. It was just amazing. And Maurice too, as it turns out, is a pretty good shot. But he's not the only one with a nifty technique. Archerfish are at home in tropical coastal mangroves. Using a targeted jet of water, they can shoot down insects near the shoreline and devour them. For scientists like Kate Newport, these fish are revealing. They make precise decisions and are therefore suitable for visual and cognitive tests. We chose the archerfish because they're visual predators, and we know that they can recognize objects above the water surface. So their ability to hit something very accurately means that as scientists, we know exactly what choice they were making. In a remarkable study, Kate Newport and her team discovered that archerfish can actually recognize human faces. This experiment was divided into two sections. So the first step was training, and in that case they had just two images of faces that they saw all the time. They can spit at either of the faces, but only one of them will result in a reward. Otherwise they get no food. And the next step was when we tested them and we said, can you still recognize or discriminate the face that you're originally trained to when there's a whole lot of new faces mixed in? The answer is yes. Archerfish are able to recognize a familiar face among more than 40 unfamiliar ones. Kate Newport then decided to go one step further. In the second experiment, we trained the fish to discriminate two faces, and all they ever saw was the frontal views of these faces. And then we rotated the faces, and we tested whether the fish were still able to discriminate those original two faces. And again, the archerfish aimed at the correct face. So facial recognition is apparently not exclusive to humans. A pretty impressive conclusion to come to. But there's more. It seems like a really easy task, but it's not. If you were trying to train a computer to do this, even with deep learning, a lot of computer programs are still failing to do this task. So I found this one really exciting because the fish were still able to discriminate the faces, despite them being rotated, when they had no experience with the new views of them. And it suggested that they were able to extrapolate from one view of something over to another, despite huge changes in what the image actually looked like. That one's my favorite. Everyone's just like, oh, discriminate faces. But that one I thought was really exciting. If there are fish species that recognize faces, could there also be species that recognize themselves? The Mirror Self-Recognition Test, the MSR, was developed in the 1970s. Also known as the Mark Test, it was designed to show whether an animal recognizes itself in its reflection. In the test, a mark is applied to a part of the animal's body that it can only see in the mirror. If it tries to remove the mark, it has clearly recognized itself in the mirror. Only a few species have this talent. Great apes, for example. Elephants or the crow family. And of course, humans, but only from the age of 18 months. So how do fish behave? Behavioral ecologist Alex Jordan chose his favorite fish to test. The Pacific blue-striped cleaner wrasse. A true cleanliness fanatic that is constantly removing parasites from itself and others. If there is one animal that reacts to a spot on its body, it's the cleaner wrasse. So in the test, the animals typically pass through various phases. And the first phase is like this. Mouth fighting in this case of the fish, but aggressive behavior where they're attacking the mirror. You can see the fish is trying to bite continuously what it sees as an enemy. After that, you get this kind of strange behaviors. Now, if a fish is swimming upside down, normally it's not a good sign. But here it's this contingency testing behavior where the animal is doing something strange over and over again and trying to see the response of the mirror. If they go through these phases of the test, then we can do the final step, which is to give them a mark. And then we look for this scraping behavior in an attempt to remove it. Amazingly, this little fish with a very small brain could pass a test like the mirror test, which is held as the top, top measure of animal intelligence. And this was incredible. If the blue striped cleaner wrasse recognizes itself, does it also recognize others? And if so, does it change its social behavior depending on who it's dealing with? Léonore Bonheur from the Laboratory of Ecoethology in Neuchâtel is researching the cleaner wrasse in its natural habitat, the atolls of Polynesia. Could there be a shrewd strategist hidden inside this industrious cleaner? A cleaning station is a simple structure such as a rock cave or a coral head where cleaner wrasse fish gather. Reef fish seek out these places to get rid of annoying parasites and dead skin tissue, which the cleaners feed on. The reef fish know the cleaning stations. They come by and wait their turn. In the daily cleaning operation, the scientists recognized a hierarchy in the treatment of the clients, which can be divided into three large groups. First, the peaceful fish, both visitors and permanent residents. The latter, such as clownfish, have a tiny territory connected to the cleaning station. Since they're always there, they're obliged to wait till the cleaner wrasse has time for them. Visitors, on the other hand, can change stations if the cleaner is too busy. The cleaner seems to treat them preferentially so that they don't move on to the competition. And then there are predatory fish. Cleaner fish are much more cooperative with a predator than with other visitors. Take the moray eel, for example, which is quite common here. Since the cleaners are potentially dinner for these predators, they make every effort to treat them satisfactorily. The cleaner wrasse has a voracious appetite. Sometimes it simply needs more than parasites and dead skin flakes, even if that scares off potential regular customers. Cleaner fish have to constantly suppress their appetite because they actually have a preference for the mucus layer of fish. Sometimes they cheat, as it is called, and stray from the cleaning symbiosis. It can be quite painful for the customers who fall victim to the snack attacks. To ensure that they come back anyway, the cleaner pampers them with a soothing anti-stress massage. These little animals have amazing strategies. With around 2,000 interactions per day, they are constantly faced with choices. How they interact is also influenced by the clientele. The cleaner is much more cooperative with prestigious visitors, meaning they are less likely to cheat. They massage more and simply provide a quality service. This is called the audience effect. In other words, cleaners are able to analyze their environment and adapt their behavior accordingly. So cleaner fish only cheat when conditions are favorable, their behavior rarely observed in the animal world. But are fish capable of real emotions? Since animals cannot report verbally their internal states, we will never know if they feel what we feel. What we can do is measure behavior and measure physiological endpoints and try to see if there are homologues in animals to human emotions. In Portugal, biologist Rui Oliveira and his team are researching pain perception in fish. They are equipped with the cellular and molecular mechanisms to experience noxious stimuli and to respond to them in an aversive way, avoiding them. And this is adaptive, because if you don't feel the stimuli and if you don't respond to them, then you are killed very easily. This raises the question, if fish can perceive their own pain and react to it, can they also perceive the pain of their fellow fish? To find out, the research team conducted experiments with zebrafish, a species known for its showling behavior, and devised a very special experiment. In the first part of the experiment, the test fish imitates the fish in the video, which has been exposed to a painful substance. Just like the fish in the video, it becomes restless, swims frantically, and finally freezes at the bottom. So zebrafish do indeed experience emotional contagion. What's more, when the test fish has the choice between the video with the calm fish and the one with the stressed fish, it swims towards the zebrafish that's clearly in trouble. How can this behavior be explained? We tried to see if there is emotion recognition involved, which suggests that they have a representation of the emotional states of others. What we are observing in our experiment is that they swim towards the fish that are expressing fear and they stay near them. So I think this could be seen as the very early sign of something that later on might have evolved into what we call empathy in mammals and humans. Even salmon farms have reacted to Rui Oliveira's experiments. In many aquaculture facilities, salmon are fed antibiotics because the cramped conditions and lack of exercise make the animals sick. But now, in Norway in particular, several studies are focusing on the fish's emotional well-being, specifically on depression. It seems that some farm salmon simply lost the will to live and subsequently die. If fish can feel negative emotions, can they also feel joy or happiness? Maurice from the aquarium is no longer the same Maurice. It's Mauricette now. And there was another new addition. This is Nemo. The pair will soon grow accustomed to one another. We know that fish are capable of social interaction. They're not solitary creatures, but rather social beings that can display remarkably complex behaviors. However, this raises a deeper question. Beyond mere interaction, are fish also capable of forming lasting mutual bonds? Chloé Lobu and François-Xavier Deschaumes have looked into it. For their research, they scanned monogamous fish species, searching for one that formed committed and lasting partnerships. The cichlids were an obvious choice. We choose them because they exhibit behavior just like birds. They choose their partners, care for their young and build nests. Instead of birds, which don't interest me, we looked for fish that exhibited pronounced monogamous behavior. Once they've found each other, most cichlid couples protect their territory and raise their young together. The better the teamwork, the greater their reproductive success. If there's bonding, a separation would be very upsetting. But how do we measure the impact of separation on a hypothetical emotional state of cichlids? The research scientists adapted a well-known experiment from psychology, the ambiguity test. For several days, they exposed females from cichlid pairings to a range of different situations. First case, positive feedback. A box with a black lid is placed in the tank. When the fish open the box, they find tasty worms inside every time. The black lid promises a reward. Second case, the negative signal. This time, the box has a white lid, and when the fish open it, they find nothing. They quickly learn to distinguish the rewarding black lid from the unproductive white lid. And then comes the third case, the ambiguous signal. To test the emotional state of the female cichlids, we confront them with an ambiguous signal, namely a box with a grey lid. The fish must decide. Is grey like black? Then a reward awaits me. Or is grey more like white? Then it's not worth the effort. How it assesses this depends on its mood. Does the fish see it more optimistically or pessimistically? Here we actually have a quantifiable measure of the emotional state of the fish. Optimistic females always open the grey lid very quickly. The researchers now wanted to know whether they retained their optimism when separated from their partner. To this end, they isolated the females for a long period of time and then measured their reaction time again in all three cases. Black, white and grey. The result? All of them opened the ambiguous box much later than before the separation. Apparently, the female cichlids had become more pessimistic and were now behaving differently. The absence of the male seemed to slow down their reaction. Can we conclude from this that they are unhappy without their partner? Or conversely, that they are happier with their partner? So, I like to talk about rose-coloured glasses. Being in a positive emotional state means seeing the world through rose-coloured glasses. And when your partner is gone, you are sad and everything looks darker. Yes, this suggests that there is an emotional bond between the partners. If there's this emotional bond, it could be adaptive. Because in this species, it's crucial that both parents take care of their young. On an evolutionary level, this emotional bond would ensure that they stay together. When separated from their partner, they feel bad, so the bond keeps them together. They end up raising their young together, which in turn increases their reproductive success. The advantage of such an evolutionary adaptation is obvious. The experiment on the emotional life of cichlids has sparked lively debate among experts. Emotional attachment to a partner is not a unique property of our species, since we know it has evolved in at least one other species. Now it's about researching this in other ones. Are fish capable of emotions? We can no longer assume that they're just cold, unfeeling creatures. Like many of their fellow fish species, Morissette and Nemo may be a pair of sensitive souls. There's a long road ahead before we can truly understand these amazing creatures, their abilities and their intelligence. More than 90% of the oceans remain unexplored. Much remains to be discovered in this unique underwater world. One thing we can say for sure, Morissette, Nemo, Maurice and company are capable of so much more than previously believed.