 First of all, thank you very much to the organizer for inviting me. So today I'm not going to talk exactly about collective behavior, because I don't know if we can treat slime mold as a collective, but we will see. So I will talk about slime molds. So first of all, just let's place slime molds in the acryl tree. So here you have like the, nope, I don't think this thing is working, oh yeah. So here you have like the general big kingdoms, like the fungi kingdom, the animal kingdom, and here's the plants. And you can see that most of the eukaryotes are unicellular organisms, and very valuable. Huge diversity of unicellular organisms. Slime molds are, oops, sorry, I will use this one, that would be easier. Slime mold belong to a kingdom called amebozoa, I think the most famous species in amebozoa is ameboa proteus. So usually when you see slime mold in the field, they sit on trees like that, in forests, they can also sit in your garden. There's more than 1,000 species of slime molds, they can have all kinds of colors and shape. And this is the life cycle, usually you have a huge cell called the plasmodia, and the plasmodia will transform into spores to reproduce. So what I'm more interested in in this slime mold is the plasmodia stage, so the huge cell. So here's a life cycle, so you have the plasmodia stage and it will go around like that with a sexual reproduction, like classical reproduction, except that there's not two sexes but 720 different sexual types in slime molds. So it's quite a bit more complicated. So this cell, as you can see in the video, is giganteous and in fact it was even recorded in the Guinness, the largest cell, it's 10 meters square. So this is a cell that is particular because it's polynucleated, so in fact inside the slime molds you have billions and billions of nuclei. So what is interesting for somebody studying animal behavior when you come to slime molds is in fact the fact that they're moving and in fact they're moving all the time. The life of a slime mold is going forward. So you can see a slime mold here exploring leaves in a forest and it's always very strange to think that you are looking at just one cell. So this cell is quite particular because it moves and it moves using a network of veins that you can see here on this green marking. You can see the vein of the slime mold and it's using these veins to move. How does it do it? If you look inside the vein you will see the protoplasm moving and the protoplasm is moving back and forth inside the cell. So in fact by doing this movement the slime mold is pushing the membrane that you can see here at the bottom and it's pushing the membrane forward, take it back, pushing forward. It's like the slime mold is going two steps ahead and going back one step. So if you look from outside, this is how it looks like, you have this kind of oscillation mechanism. And so it's pushing and you can even see by the eye the contraction of the veins. If you look a bit further away you can see all slime molds oscillating and you can't see that in nature, just to tell you. It's time lapse. So in fact the oscillation, so the protoplasm is moving every one minute and a half but this oscillation is like around one hour cycle. So you won't see that. You have to be very passionate to see this kind of thing. So this is for example a follicose septica oscillating in the forest. So slime molds are moving in the forest and they're a huge predator of bacteria that they also feed on fungi and when they feed on fungi they really take them all. So nothing is left from the fungi. So you will see this video is 24 hours video in the lab. So you see nothing is left. So problem when you have slime mold in the lab, it's the fact that they grow exponentially. So in fact a slime mold would double size every day so you need to feed them every day. So you have to go to the lab all the weekends, Saturday and Sunday. If not they would escape. It's a bit of a nightmare. It's a nice animal but it's not an animal. Anyway there's a good way you can make the slime mold slip. So it's good. You can drive them and keep them like that. It's called a dormant stage, it's called a sclerosia and you can keep them like that for more than two years. And you can wake them up just by adding some water. So it's good. You can still go on holiday or in conferences. So slime molds are pretty interesting when you want to study decision making because they sense a lot of different stimuli such as light, humidity, temperature, osmolarity, mechanical stress, chemical signal, whatever, lots of stimuli that they can sense. So it's a very nice organism to study decision making. So the pioneer studying slime mold was Toshiyaki Nakagaki by these two very famous work. One was published in Nature and he shows that slime mold can solve a maze and can take the shortest path in this maze to link to old flakes, for example. And the second experiment he was famous for was the network experiment where he shows that slime mold can build network that are smarter than railway network in Japan. So he got two ignoble price for these two experiments and he was very proud. So I will present some of our work we've done so far in terms of decision making. In different contexts I will talk about nutrition, navigation, phenotypic plasticity and social interaction and lastly the last work we've been doing about learning and social interaction. So first of all nutrition, I will show that slime mold are able to solve complex nutritional challenge and maximize their growth. So first of all you need to know that we don't feed fungi to the slime mold in the lab because we need to rear fungi. So in fact somebody in the 60s found that slime mold feed very well on old flakes. But for these experiments we needed to design a proper diet that we could manipulate. So we designed by adding like protein powders, sugars, mineral salts, vitamins, etc. It was a work done with Steve Simpson, Chris Rayden, Tanya Latte. So we were growing our slime molds on different diets like that and we were looking at the growth, how they developed and we also weighed them at the end of the experiment to see the growth in term of mass. So here are the results. So you can see here's the proportion of carbohydrates in the diet and here's the proportion of protein. So we made a lot of different diets varying in the concentration in the ratio of protein to carbohydrates but not in the concentration. So we made 17 diets and you can see here so the slime mold on the diet and you can see here the results in term of growth. So here it's optimum for the growth and it's twice more protein than carbohydrates. If there's too much carbohydrates in the diet it's very bad for the slime molds. So knowing what the intake target was we looked at different kind of Bivouan also. So here it's landscape, here you have the amount of protein in the food, here you have the amount of carbohydrates and here it's a number of fragments of slime molds. So funny thing happens, it never happened usually in slime molds, they separated on a high protein diet, they make different individuals, they cut themselves in different colonies of individuals. So after we challenged them with first concentration so we manipulate the amount of food inside the diet, not manipulating the ratio and we look at how the slime mold was growing in term of mass and in term of area. So in fact all the slime mold was the same mass on all type of concentration but the difference was in the area. You can see that on a very diluted diet the area was very large in comparison with a highly concentrated diet. So in fact the slime mold was coping with the deletion by spreading a lot but it was not growing more in term of mass. After we played with a binary choice where we offer the slime mold two diets, a high protein diet and a high carb diet, none of them were good for the slime mold, so we had to pick both of them to make like a perfect ratio. And we show that so here you have all the binary choice we tested and you can see that all the slime molds come here at the optimum at twice more protein than carbohydrates. So the slime mold were able to really take a bit of each food to create the perfect diet. So we went a little bit further and we go with a more complex decision making problem where we put lots of different diets and we look at how the slime mold was choosing the diet. So all these diets are equi-caloric but they're just a protein and carbs ratio changing. And we could see that all the slime mold in average focus also their activity here on the protein, on the diet offering twice more protein than carbohydrates. So they were able to regulate the intake and maximize performance. So this was the first example of decision making in slime mold. A second example is navigation. So if you look at the slime mold growing on a negar plate, you will see that he makes all these soda pods and he's moving around so there's nothing to eat there. And you can see that when he has been somewhere and he left, he leaves a trail behind him. So he's here. It's a mucus. It's an extracellular polyionic glycoprotein that you can see here. And we were wondering if the slime mold could use this kind of trail like ants can do for exploration. So first we look at if this trail was attractive. So the setup was simple. You put the slime mold here and here there's a trail and here there's nothing. And we let the slime mold choose. And there's two food source at the end to motivate him. So you will see that most of the time, in fact in most, almost 100% of cases, the slime mold would pick the branch without a trail. So he would avoid the trail. So we looked at if he could make a difference between different trails coming from different strains, same species but different strains. So here we had a Japanese trail against an Australian strain and we tested Australian slime mold and Japanese slime mold. And we could see there was no preference for any. So slime mold would pick randomly one of the trail. So after we changed species, we wanted to see if he could differentiate trail from different species. And we show that if we use these new species called Lepidoderma here. So you put the trail here, here's a trail of fissarum. If you test fissarum, it would always pick almost always, not always, the Lepidoderma trail. So it can make a difference between trails from different species and go for another species. So after, with Chris, we came up with this idea to study how these trails can be used in more complex problems. So this is a common used shaped bridge problem. So here you have the slime mold, here you have the food. So the slime mold is attracted to the food but here there's a maze and it cannot cross this barrier. So here is a video. So you can see the slime mold is falling into the trap of course because it can sense the food but it will avoid it and you can find here the food source. So you can see that the mean distance travel is 20 centimeters and almost all of them succeed in this task. Now next we put trail everywhere to mimic an explorated environment already. So you have the slime mold here but you have trail everywhere, a mucous everywhere. And so here that's what happened. So the slime mold is going here, so you go up first. So I show you the best slime mold. So you go in the trap and it's a bit lost, completely lost in fact. You go up, you go there. So it takes a while and it crosses in average 45 centimeters or so twice the distance of the controlled slime mold. So we were showing by these experiments that slime mold use their trail to navigate. It's like an external memory. Next I wanted to show you like phenotypic plasticity. When you study ants like I was doing before you would always use lots of different colonies but in the world of people working on slime mold they always use one clone lineage. So I wanted to try just to compare different clone lineage to see if there was a difference. I show you just in a simple experiment where slime mold have to explore just a plate of agar and you will see straight away the difference in term of exploration. So you can see that the Japanese strain is far quicker than the other strain. The U.S. strain make these very tiny pseudopods and the Australian strain grows in an isotropic manner for a very long time before doing pseudopods and it's very, very slow. So we measure these differences between strain and we show that irregularity was like, so in fact the U.S. slime mold was very irregular while the Australian slime mold will grow isotropically for most of the experiment and you can see that the speed, the Australian slime mold was very slow and the U.S. slime mold was the fastest one. So we wanted to see if these differences were just in term of exploration so we moved forward and we noticed that there was not just avoidance between slime mold using the trail but there was also attraction between slime molds. So what we did was a very simple experiment where you put the slime mold, there's another slime mold, a cue left by slime mold feeding so it's very happy and here is nothing and you will see that in fact most of the slime mold are attracted to this cue left by a congener that is feeding and this attraction is stronger for Australian slime mold and the attractiveness of the slime mold when it's an Australian is more attractive too. So Australian are more attracted to a congener and they're also more attractive to the others in comparison with the U.S. U.S. strain. And after so we offer them a choice between a congener and food. Do you prefer to go to a congener or do you prefer to go to food? And Australian slime mold would always pick the congener while the U.S. S.S. slime mold would always pick the food. So we were wondering what was the slime mold releasing in the environment that would attract the others. So we found out that in fact when a slime mold is feeding is releasing calcium. So we measure this level of calcium and we show that Australian are releasing a lot of calcium in comparison with the U.S. S.S. H, U.S. H strain. And we also look at the attraction for a source of calcium where we manipulate the concentration and we show that Australian could pick very tiny amount of calcium in the environment while the American strain was very bad at picking calcium inside the environment. So and if you block the calcium in cues left by congener while feeding using EDTA, you will see that there's no attraction anymore. So it was really the calcium that was playing the was responsible for this attraction. So after we wanted to see this difference in signaling and behavior. So the fact that for example the slime mold is releasing a lot of calcium and can pick very tiny amount of calcium in the environment. And the fact that it's slow would play a role in the way they explore an environment with another slime mold. So this is the setup was like that that two food source and two slime molds. And so here you see what the Australian are doing. Usually they would always pick the same food source. So it was kind of collective decision by two slime molds. And they would always find it quicker if you compare when you put a slime mold alone. So there was some sort of cooperation. But if you test the USA strain, you will see that they have a complete random distribution that would pick the food randomly. So there was no effect of social interaction. So by this experiment we were showing that we have like an interaction between the fact that you release a lot, that you attract it to a stimulus. So the difference in attractiveness can be responsible of different social strategy in slime molds. So it was a very simple mechanism. And so we use a model to prove these results. So lastly, I want to spend a bit more time on learning and social interaction. So the question, because I work in a lab, we're working on animal cognition, I was wondering, because we've seen that most of animals can learn something, I was wondering if slime molds, even if they don't have a brain, could learn something. So I came up, I started to study the simplest form of learning because we are with slime mold, no brain. So I think it was a bit easier to start as simple. So we study habituation. So it has been shown in Aplasia and Eric Kandel got a Nobel Prize for showing this behavior. So in fact, if you touch the siphon of Aplasia, she would retract a gill straight away. And if you do that a lot of time, she would stop doing it. So habituation is a phenomenon that is very common. So I will give you another example that is more understandable for us. So habituation is when you enter a room and it smells very bad. So at the beginning, of course, you notice it, you say, oh my gosh, it smells very bad. But after a while, after one hour, you say, well, that's not so bad. So habituation, in fact, is the ability to ignore irrelevant repetitive stimulus. And so this repeated application of the stimulus results in a decreased response. You stop responding to these stimulus because you know it's armless. The second criteria to define habituation is specificity. So imagine you are in this room that smells very bad and suddenly there's a new order in this room that is also very bad but different. You will notice it. In fact, you have to prove specificity to prove habituation because it precludes the fatigue of mortal or sensory response. So in fact, the specificity allows organism to respond to novel and potentially harmful stimulus. So it's very important this criteria to define habituation. And the last criteria is recovery. So it means that if you get out of the room to take a coffee for two hours and you come back, you will smell the big smelly order again. Of course, you're not habituated to bad order for life. And so this is what makes habituation different from adaptation. So we wanted to see if the slime mold could habituate it. Because habituation has been shown in lots of different animals but never in unicellular organism except in this particular science, called stentor, right? So it pieces a unicellular organism. And when you touch it, you will see it will retract up like that. And they show that if you do it again and again and again, it will stop doing it. But in these papers that didn't check for specificity or recovery. So we wanted to check what kind of stimulus we could use to show habituation in slime mold. You cannot touch the slime mold. In fact, you can touch a slime mold but it doesn't do anything. So we use more like chemical repellent to really use a kind of strong response from the slime molds. So the experiment was simple. You add a slime mold here, you have a bridge, and here you have a food source. And so we will look at the first time the slime mold contacts the bridge. The shape of the soda pod when it's contacting the bridge. And after 24 hours, we would transfer the slime mold and offer the bridge again. So this bridge was even a control bridge. So it was just plain agar for five days. And on day nine, we will test the control slime mold with agar, quinine with a very bitter substance, slime molds hate it. And caffeine is another bitter substance like slime or hate. And after we had a resting period and after we would test them again. So and after we had two treatments, the quinine group. So if they had to cross a bridge with quinine five days in a row, they would be tested for habituation with quinine, tested with caffeine for fatigue, for specificity. And after they would be allowed to rest and we would test them for recovery. And we would do the same with the caffeine group. So repulsive behavior in slime mold is quite obvious. First, it took a long time to go on the bridge. As soon as they can smell a repellent, they don't want to go on it. And after when they would cross, they would cross very slowly. And lastly, they would make these tiny, tiny soda pods to cross the bridge. So I will just show you the results about the shape of the soda pods. So here is a control group. So we measure the shape of the soda pods and we calculated a circularity index. More it's close to one, more you like spreading on the bridge, less it's close to zero, more you like very digitated soda pods. So here are the results. So you can see that a slime mold would not have any repellent, will really spread on the bridge. And if you offer them quinine on day six, they like it. You make these tiny, tiny soda pods. And after you can rest and you can show this response again. So the control group was fine. So let's look now at the treatment. So the one with quinine, you can see that first, they don't like it at all. They make these really tiny soda pods in these very slow soda pods. And after you can see that through the day, the soda pods spread more and more and more and you have habituation on day six. You have specificity because the slime molds still react to caffeine. And you have recovery because if you allow the slime mold to rest, it will respond to quinine again. And we have exactly the same results with caffeine, habituation, specificity. So it tricks, it reacts to quinine, and it recovers. So this is in picture because usually picture works better. So you can see here control. So you see spread very widely on the bridge. And when you give him quinine, he makes these very tiny soda pods. And after on the resting period when he gets agar, he spread again and here. And so he has a quinine group. So he gets quinine the first day, he really tried to avoid it. And after you can see the soda pod growing and growing and a full habituation on day six and a recovery on day nine. So this experiment was one of the first time we were showing a simple form of learning in a unicellular organism by following exactly the criteria given by the neuroscientists, meaning decrease of respite, specificity, and recovery. So after, we move to social interaction with learning because there's a particular things in slime mold that is quite interesting. You can take a slime mold and cut it into pieces. And if you put these pieces next to each other, it would fuse again and form a new slime mold. So it's really difficult to define what is an individual in slime molds. So we took this advantage to study how slime mold could transfer what they learned to another congenial. So first, we changed repellent. We didn't use quinine for this experiment, but we used salt. So we had to show first that a slime mold can show habituation to salt. So here, it's the meantime to cross the bridge. And here you have the day, here's the control group. And here you can see that the slime mold habituated quite quickly to the salt. And we also did a test for specificity in this case with quinine. And the slime mold was still reacting to quinine. And we also did the test for recovery. So it was just to show you that they can habituate to salt. Now we can move to the proper experiment. What we did was to train more than 2,000 slime molds to habituate to salt. And we trained 2,000 to be habituated to nothing, just to cross another bridge like that. And that for five days. And after, what we did was to mix the slime mold. So we would take an habituated one with an habituated one, let them fuse and test them together with a bridge with salt. So we will look at the time to reach the food and who reached the food first. Because we did a lot of experiments, we had to compute an habituation index to be able to compare all the experiments together. So in short, when the habituation index is zero, it means that you are repelled by the salt, so it means no habituation. And if it's far larger than zero, it means that you habituated to salt. So here are all the combinations we did. So we had either two slime molds and our control was the two unhabituated slime molds. And so we had either three slime molds and also a control with three unhabituated slime molds and here four slime molds and also a control. So we did all the combination possible and we tested them for habituation. So here you have the results for two slime molds. So if you put two unhabituated slime molds, of course they don't know anything and they're repelled by the salt. So that was good. And if you put like two habituated slime molds, they're still habituated. So the fusion with the congenials didn't affect habituation. And if you put an habituated with an unhabituated one, the entity is habituated to salt. We did the same with three slime molds and you can see that all of them, all the groups show an habituation index. That is far bigger than zero. But all habituated are a bit more habituated than the other one. But still all entities, all merge entities show habituation. And if you look at four, it's exactly the same results. You have all the configuration here. So you can see that all the entities are habituated to salt. Even if there's just one little habituated slime mold in the merge entity. If you put these results as a proportion of habituated slime mold in the entity and the habituation index, if it was just a dilution of the information, you would have the yellow line. But in fact, you don't have, it's a bit like that. So it means that there's something happening when the information is transferred from one slime mold to another. Something that was interesting in this experiment, it's when you put like three slime molds, for example, habituated, unhabituated and habituated one. It was most of the time the unhabituated one would cross the bridge and touch the food source. And not the habituated one, we would have expected the opposite. But it was in 72% of case, the unhabituated slime mold that was crossing the bridge first. So we looked a bit closely what was happening when the slime mold was fusing. So it is this really little video at the bottom. So you can see here, I don't know if you will see it properly, but there's a vein forming between the two slime mold quite quickly. So in fact, this vein appeared three hours after fusion. And here, there's nothing. So we decided to do a fusion, but for a short time. So we put an habituated slime mold with an habituated one and we leave them in contact for one hour here. So when there's no veins, or three hours when there's a vein forming. And we separated them and tested them individually. So these are the results. When you leave them in contact just for one hour. So the two unhabituated slime mold of course show no habituation. The two habituated ones show habituation, so that's fine. But in the merge entity, if you take just an habituated one and you test it alone, it show no habituation. But if you take the habituated one, it still show habituation. So one hour of contact, the information was not transferred from the habituated to the unhabituated one. But if you look at the result now for three hours, this is very different. You can see that the slime molds unhabituated one this time, gets the information and show habituation to salt. So this shows that there was a proper transfer of information, of learning information between the two slime molds. So yeah, so this adaptive like behaviorist response via self fusion provides a rapid and efficient means for slime molds to adapt to the environment. Because the recipient slime mold become pre habituated to the environment repellent before the repellent are even encountered. So fusion can confer resistance to naive slime mold that otherwise would be susceptible to this repellent. You have to think that slime mold get separated often in the field, so this behavior could be very efficient. So lastly, in the last five minutes, I wanted to show you the last result we got in the lab. So we wanted to show if there's long term habituation in slime molds. So in fact, what we did was to train slime mold to salt. So here is, we have a group of slime molds that is trained to salt, and here you have a control group after five days habituation. And what we did this time, we drive them. We put them in a dormant stage and we take a peek and we work them up. But we work them up one year after doing the habituation. And so what you can see that first, there's lots of slime molds that didn't work up, but that's normal in slime mold. But 50% work up in the habituated slime mold, but just 10% on the control slime mold. So this was first our results. And second, if you look at the latency to explore an arena full of salt, you would see that the control won't go out of this paper. Usually they would stay there and they would explore after a very long time. But the habituated slime, although one was habituated before going to dormant stage, explores the area quite quickly. So they could keep this information for a dormant stage for one year. So we were wondering, we are now in trying to find the mechanism behind this habituation. What we did was to, in fact, just a little story. In fact, we put an habituated slime mold to queen in with an habituated slime mold to salt. We make them fuse and they didn't want it to fuse. So we found it strange. They ended up fusing that there was something going on. So we were wondering if the slime mold was not in self-repellent. So what we did, we trained slime mold to salt like before. And this time, we grained them and we dosed the level of sodium inside the slime mold. And this is what we found. In fact, the control slime mold has a very small amount of sodium inside the cytoplasm, but the habituated slime mold is full of sodium. So in fact, it's taking up the sodium while it's habituating to salt. And if you look at the amount of sodium, and you look at its performance in term of habituation, you can see that it's completely correlated. More it has sodium, more its performance in exploring or crossing a bridge with salt. So the level of sodium was a very good predictor of its performance. So we wanted to see, so what happened during recovery? It take up the salt, but what is it doing with this salt? So what we did was to train slime mold. And after we put them in a piece of a humid paper with distilled water, we let the slime mold explore this environment. And after we cut this paper and we analyze it for sodium. And what we found is that a paper explored by a control slime mold would have a very small amount of sodium, but the paper explored by an habituated slime mold would be full of sodium. So the slime mold during recovery process was just spitting out the sodium. And of course, to understand why habituation could be kept for such a long time, we look at the level of sodium in sclerosia. In sclerosia habituated to salt in sclerosia, in control sclerosia, and we saw that the amount of sodium was twice more important in an habituated slime mold. So the slime mold was keeping the sodium inside. And lastly, a very tiny experiment we did like two weeks ago. We applied topic in kind of an injection, but it's a passive injection. So you put salt, in fact, a droplet of salt. We control the concentration on top of the slime mold. And we looked at how this slime mold was reacting in two substrates. In a control substrate where there's no sodium, and you can see that the control received a drop of distilled water. And the habituated one explore the dynamic of exploration is exactly the same. But if you put them in a salty substrate, you can see that the control slime mold is very affected by the salt. But the ones that receive just a droplet of sodium on top of his head two hours before the experiment can explore quite quickly the substrate with salt. So we are putting this idea out there that perhaps the slime mold is using kind of circulating memory by keeping the repellent inside this vein network as a memory of the habituation. So I will leave you with this last slide. I hope that I'll show you that unicellular organisms, even if they are not collective, are quite interesting to study cognition and decision making. They're perhaps not collective, but they're really distributed, especially the slime molds. And I'm moving very soon to this kind of organisms to study how they deal with learning protocols. Thank you very much.