 Okay, this is for us to start the second steps and Cornelia talk about and navigation. Thank you. Yeah, hello everybody. It's, yeah, it's really nice to be here today. And I will start with some early bark that Mike has done together with Tom in the mid 70s. Once they have done experiments on hope of life, and they were actually one of the first people who have shows that those who applies the spatial memory in order to return back to two points they came from. So I'm not going into detail because Tom will talk more about later on, but they have been those hope of life and have shown that they use visual information in order to return back to where they started the chase. And this is one of the people from the paper and they show her pink and different parts and those nice always end up in a similar location. And this depends as well on how the visual cues are around this place so if there are lots of clothes by visual cues they're more accurate in order to return, then if it's in the more open open area. And, yeah, in a similar spirit later on, work was done on on ends and we see similar behavior and spend what we have seen and on the previous slide for her price so those and also remember places and they go back to important places as for instance here. And we see a dozen of these comments where they learn to associate the location of the feeder with them. And with those three cylinders and they use this external visual information to learn that place and if those cylinders together with them is used to a test is where there is no no feeder. We see that they are searching around the place that we should be so we get this really nice search piece. And if the cylinders are much further away, they appear to be small on the answer, you know, so there is no space, no location in this space that is correct. So we don't get a location that actually searching for it. But if that, if the cylinders are much further away and they are enlarged in size, you can appear correctly. So again, here you have search and you can pick again and so searching for the feeder in the center of those cylinders. And, yeah, so this is what that was done in decency in the field, but over the last few years we got lots of knowledge on and navigation on wood and some work that was done in the lab. And we see similar so this is similar experiment that was wouldn't learn to recognize a feeder in the center of two cylinders. And what we have seen before as well if for instance the cylinders in the past, have a different size so one is smaller, and one is bigger. So those ends, they fixate a lot on the small one and they are approaching the small one much more than the bigger one because this is the space where the visual views that have been up here correctly. So we see how the parts are then shaped towards the smaller one. So there's a series of experiments that was done in the lab. And so for the last few years, and by looking at the past it is, we have learned a lot, how they use their eyes for navigation and how they are looking at visual cues, and how those visual cues are controlling their course, and as well, how, how they control for them. So we have learned for errors by hand really quick rotations and so on. So, we learned quite a lot about their visual navigation mechanisms in this lab conditions but what we know as well is how they actually navigate in natural conditions. And that's why we went to the first last year and we did an experiment where we looked at how they actually navigate through cloth natural habitat. Here this picture we see the habitat of the Buddha and so they have this nice mouse and they are navigating along those shared trails and shared, yeah, along those shared trails to trees where they go up and get honey to you from Aphids. So we have those quite complex habitats and we have those fortune trees they are navigating on and looking at other animals we know from different animals that they can follow. They can learn and follow, it doesn't work the goods. So animals lose and learn environmental information in order to guide their path. So here for instance on the top we have an example of pigeons. So we have a start and an end. And if you look at those blue parts there are a couple of parts from the same pigeon. And this pigeon doesn't like the direct feeling in start and end but it really learns and follows a specific route and it repeatedly does the same over time, even though it could fly on a direct line to the target. And similar behavior we see in desert ends. So here we have a desert end that inhabits a semi open desert environment. So here we have an S and DL feeder and those are two different ends. And we have two parts from both ends. And they are almost falling exactly the same route. And in those ends they don't have their own trails or in this experiment they also don't have partial equation anymore because they bring this back to the feeder again. So they need to learn and feature for the environment in order to be able to follow again and again the same route. So we are asking the same question if those ends that inhabits these really complex environments and navigate a lot of those shared trails, if they do the same. So we were taking ends that just came down from the fortune tree. And then we fit them with a handheld camera up to the moment they were at the nest and before they entered the nest, we moved them back to the, to the bottom of the tree and within them a second time. We moved them back to the tree and recorded them as a search time. So we go for every end and three homing parts and the videos that we recorded from those. And this is the video how they look when we get them. So, as you can see it's quite hard to actually see the end. I think it was not before. But I gave them a little color dog so that we didn't lose them by following the end. But luckily, there are some recent developments in tracking algorithms, and we were collaborating with people from anywhere and the University of Minnesota in Germany. And they have recently developed and published a tracking algorithm with which you can track animals and we construct their, their environment. And this is one example of one of those ends. So, down here, there will be the tree, and then we have three parts of the same and up to the moment she's in the nest. And, as you can see, there's lots of overlap in this individual past as a victim and radiation but still is and follows the same similar pattern. And this is another example that and then almost identically three times follow the same pass. And here we have a few more examples so those are five different different ends that we have tracked and feeling great those are the parts from all of the ends we have tracked so we have tracks 20 and 20 and 20 and 20 and 20 and 20. which then leads to the question that, or to the... So that makes us realise that those points actually need to deal with which amounts of memory if they are able to have these ideas into our good roots. Yeah. So you said that the cars look very similar. I agree, but the scale bar is like one meter, right? So it's sometimes half a meter and five. And I guess from the end, wouldn't that look completely different? So we will... I will show some pictures how it actually looks in the different places. So yeah, I agree, they have some variation, but if we compare how similar the cars are from one end, it's more similar than if you compare with all the other ones. But yeah, you sometimes have some... I mean, this is 12 meters long and... So this can be 40 centimeters between... Yeah, there's some variation. So they are not perfectly followed the same route, but they are more similar than when looking at all of the cars. So that means they somehow need to remember all those places. And here we have an example of one end. So this specific end, this is a sectional of the normal cars. She always ends up here. Here there's a bit of variation. Here in this case, this end three times ends up in the same location. And if we now look at the images or the panoramic pictures, we have a lot of looks at those places. We can start to think about why this is the case. So here, in yellow, we have highlighted the pictures we can see. At this point, and in red, the pictures we will see at the other point. So we put the panoramic camera on the ground. And here in the middle, it's where we have always ends up. This is how the velcro looks. 26 centimeters to the left or to the right. And when we now focus on the yellow location, so we have here the top, the left location. Here we have the middle location. And here we have the right location. And this is sampled down a bit so that it's more similar. And we haven't done any further analysis on that yet. But as you can see, it looks rather similar. If we go to the other location, we again have the, so in the same place, the location, the end always ends up. Here it looks this, how it would look, 20 centimeters to the left. This is how it would look 20 centimeters to the right. And again, it's looking rather similar, but then we also see some differences. So like, there are some differences in some parts of the, of the panorama. So that, and how different or how similar, similar across all in a specific location might depend on how similar the velcro looks a bit to the left or to the right, or if it's really distinctive. And knowing that there is need to deal with so much, so many memories we can also ask where this memory is actually stored. So we had the different experiments where we investigated where those ends actually stored and which memory. And we were focusing on the central complex and the mushroom bodies. And I give you a quick summary of the projects we have done on the mushroom body. So the mushroom bodies are known from other insects that they are important in mainly in papers, what we have learned from that is that they are important for olfactory associations. And that's about how they are actually involved in, in long-term visual noise. And those ends, they both get olfactory and visual input in the mushroom bodies. So we were asking about the role of those mushroom bodies in the visual memory of those ends. And so doing that we did chemical lesions. So we had an experiment where we could train our ends in the lab, in this, on this platform from the center to a feeder that was both set for a similar visual cue. And then we recorded the past before and after the lesion. And before any head training is happening, we only do these ends that are not going to be set up. So this is the behavior. We see that most of the ends are ending up at some point at the visual cue. So they have this innate attraction to a visual cue. But if we train them to a feeder, that is a way in the visual cue, we get past that look like that after one. So they learn to use this visual information in order to navigate to a feeder location. And with those well trained ends, we then can go to the next step, which is doing a chemical lesion in the mushroom body and then see how this affects their visual memory. And this is what we found. So those were the ends that we're able to aggregate to the feeder location. We took them, we injected a small amount of cocaine with Bocloide, which is a local anesthetic that can be injected and it's coupled with a fluorescent disolator. We can localize where the infection actually was found. And if those ends are released again in the familiar, we see that they are no longer able to go to the feeder. So in this case here, we only had two out of almost 20 ends that were still going to the feeder location. And most of the other ends, they switched towards the left, which is the innate response. So we were also asking if this is actually a motor bias or if this is caused by the loss of the visual memory. So we had a new set of ends that was now trained to the other side of the visual cue. So we started with the training, then we let them learn to use the visual information to navigate to the feeder location. And then we did exactly the same lesion again. And what we found in this set of experiments was the same, so they were not able to navigate to the feeder location. But in this case, they now switched towards the right, again towards the visual cue. So it's not a motor bias of telling more often to the left or to the right, and this is not how the parts are shifting, but it's actually they switch from their learns to their innate response if the visual body is not fully functioning. And an additional test we did was, instead of doing the lesion, we cover the eye with paint. And so the ends are normally trained to the feeder and then we cover the eye, and we cover the parts again. And what we see here is that most of those ends are still able to navigate to the feeder. So that means that the motion of the lesion we have done is not the same and an absence of the peripheral visual input. And yeah, so in this slide, it was summarized by Stanley Kainze and this also nicely shows, so we are doing our free lesions in the input region of the motion body, the policies. And at the same time, there was another team about H.A. Marenka who did lesions in the motion body output region. So they had some deletion here in the vertical zone. And similar to what we have found was that, again, ends that were perfectly able to navigate accurately before the lesion were completely lost after the lesion was done. So the entire motion body needs to be functional in order for the ends to be accurate in visual navigation. In the last few slides, I go and move on to the next question that you can ask them as well. So if they have those ideas in verticals and we know that the visual memory is stored in the motion body, we can also ask how they actually learn those rules. And this was an experiment that was done a couple of years back. And the general setup of this experiment that we have done with Desedans was that we had a nest and we placed a feeder in the field that in this case was eight meters away from the feeder. And then we let the ends establish the woods so they were walking back and forth. And after they have established the woods, we recorded their homing paths. And then later on we have introduced a trap. So this was a channel that was hidden into the ground. So it was in the ground which meant the ends didn't see it up to the moment they actually settled in. So ends normally walked back home and at this point they suddenly settled into the channel that was perpendicular to their homing woods. And because this channel only had a single exit, they spent quite a while in this channel until they finally found the exit and then were able to walk back home. So it was a negative experience they had because they spent a lot of time in this channel finding the exit. If we then kept this setup for another day, we saw if we looked at the homing paths and again that they have developed new woods. So some ends have now sufficiently go into the channel somewhere close to the exit and then go out quite quickly. While others have learned woods then either leads them along the left side or along the right side of this channel. We then went into more detail because what's known in ants is that they often show this unscathed behavior which is a moment they unstop and rotates on the spot. And often this is done when ants are alarming or when we are unfamiliar with the situation. So we were looking at this scanning behavior in those paths which is an indicator where learning is actually happening. And this is what we found. So here on the left we see a schematic drawing of the setup so we have the feeder here and so walking home this is the first part of the route here we have the track and then in right we have the second part was the homing route until the end of the nest and then we counted the number of scans they are doing in those different parts of the route which is shown here. So here we have the number of scans and this is the frequency it happens and those two graphs show if there is no track so they are normally walking back most of the ants usually they don't tend to scan so most of the ants have serious scans in both parts of the route. If we now open the track for the first time the first half of the homing route is normal as always and here we see they are not scanning and then they are in the track for a while if they are coming out of the track the behavior in the second part of the route is not different but interestingly if they are now back at the feeder later on in them so they have their second homing path so they will fall the second part into the track we see that the behavior is changing so in this first half of the homing route when they walk again towards the track they start to do lots of scanning so this is the moment they start to relearn the route once they are in the track and they are out of the track we see again the second half is the same so that means that ants do remember the loss used their hand before they fell into the track previously and when they are encountering those use again they change their behavior which is a kind of trace conditioning because it happens before the negative event has happened and then over time the new routes become positively reinforced again and over time those new routes will emerge so we have here like it's a combination of repetitive learning and aggressive based learning so here we have all these areas that are labelled between repetitive learning is happening but if they are falling into the track for the first time this aggressive based learning so they have this occasion that is just before the negative event and this is the area where the learning is happening so this is the area where all these scans are happening and then over time they start to develop a new route and the scanning is stopping again and they are kind of in a situation as they were previously but in a new location and you have to finish up this is way how this could all work so we have ants that experience a new and the projection viewers they then project onto the canyon cells and we have the motion body output new ones and what usually happens is that the views are either associated with positive or negative balances and this can happen through modulation of the synoptic race between the canyon cells and the motion body output new ones which happens through the dopaminergic new ones and in the case of the ant being trapped this is a negative event so in this case the dopaminergic new ones they can decrease the connection strength and the canyon cells and the motion body output new ones and then this balance between a person motion body output new ones and controls the steering which then controls the traction or impulsion and so that's all from my side and I would like to thank all the people I was working on in this project both here from Sussex and the collaborators from all the universities in minstrel and to us and thank you thank you thank you thank you thank you thank you thank you thank you thank you thank you thank you thank you thank you thank you thank you thank you thank you thank you thank you thank you thank you thank you thank you thank you So, after four years of the trap, the first nexity, they start scanning, we call it, some distance before. It's one of the timescales. I'm trying to get an idea of the timescale of the memory engram, whatever you call it, the answer establishing. We don't have the exact time, but it's a couple of seconds, at least like one second or a bit more. Okay, right. Okay, this is the goldfish. So, about the scanning, do you actually look whether they scan after they've come, because, you know, it's a big event. Maybe they want to figure out where did this happen, so maybe look back. So then recognize it again, because it's quite amazing that they would recognize it the next time around, which is much in the future. Yeah, if we look at that, we looked at the scanning all the way, but they didn't change the behavior once they came out of the channel. Can they see stuff in the channel? It's relatively deep, so they have mostly sky above them, so they use different to what they experience on the surface of the ground. Could you bring about other information that the engram may have, that could be an instance, for instance, in that spot that you put in the engram, whether the texture of the ground, the landscape, or whether the sense, pheromones, and other elements, Yeah, so I mean, so other work has shown they can use all the sensors as well, for instance, on factory tools, but it is quite, like the ground is relatively the same all over the place, so it's hard to say if they have learned anything else than the visual tools, because they have the visual panorama, but then, for instance, the ground structure is quite identical. And I guess, yeah, so it's possible that they have some other information, but yeah, we think that vision is the main factor. One more. Yeah, it's mainly like matching the core view with my voice, from that point of view. All right. Oh, once you both, you guys have our next speaker, and he has a life of vision.