 Okay, so thank you for a really interesting talk, so Daniel introduced this idea that brain areas in the default node network are on the highest level of hierarchy, and they are connected just because the areas on the highest level need to communicate with each other. But how do these areas actually, during development, know that they need to connect with each other? Development is in your wheelhouse I think today, so it would actually make a lot of sense. They don't have to know that, they just happen to be in the right place in the right time when they develop, and so the natural thing is to look for the next bodies to arrive and to just make friends with each other. But they are far away in space? In the adult brain, but in development they are very close together, and so what you call the higher order areas, so areas that are of low neuronal density and of relatively low akitontine differentiation in terms of their laminar structure, they tend to develop early on, or at least they finish early in their development, and so these are the first bodies to arrive, so one of the explanations also for why they have so many connections in the adult brain, may be that they are just the first arrivals for the party. And so they are in a very good position to form many, many links with all the other guys. And using data on birth dates in C. elegans for instance, Marcus Kaiser and colleagues have demonstrated that this birth date directly relates to the number of connections that the neuron in C. elegans picks up, and similar processes may help to explain the degree of connectivity in the adult brain. But in terms of being here in the room, I mean if you went to the right party, right place, right time, you formed the right connections, and these are the first ones to arrive, and so they can form very dense interconnections early on. And of course what you see in the adult brain is not exactly what you have in the developing brain, but something that goes through several processes of pruning and excise expansion and modification, and of course very specific mechanisms of cell to cell targeting and signalling, and of course to experience-based plasticity and postnatal plasticity processes. So it's somewhat modified and different picture in the adult, but this is maybe a basic developmental setup that helps to explain a certain scaffold of connectivity that you can still recognize in the adult brain. Okay, so thank you. That's a fascinating answer. But these areas are evolutionary newer, but you are saying that in the development they originally are in these higher level areas? I know relatively little about evolution, and I'm not sure what was there earlier on or later. I think there are various hypotheses on which areas are more fundamental evolutionary or not, at least developmentally, they develop quite early. If you had asked that same question 30 years ago, people would have said it's very straightforward. Everything protects to everything, and then pruning comes, and magically you end up with adult patterns of connectivity. I don't believe that for a moment. I think that's ridiculous. One of the reasons why I believe that these long-distance, weak connections which are going to be linking associational areas are highly significant is because exactly for the reason you pointed out, the challenge for the formation of these long-distance connections during development is huge. I think there is pathway selection. I think there's directed growth, and I think these are the factors which are going to come into it. The point is that when we look at these very, very weak, very long-distance connections, they're highly systematically, statistically speaking, representative. That's to say they are consistent across individuals. They don't look like anything to do with noise. I mean, I want to deroot that line of questions. If there's more on that, because my question is very different, so if there's more comments first on that. Okay, I'll deroot them. I was wondering in the various talks how much of the non-interprobability aspect of the various platforms was a problem of how much of the mapping of terminology between, you know, because you're spanning such a large and, you know, variable space. What was the hurdles and the difficulties in terms of recognizing those principles and where are these more like informatics problems coming in in those research? So I guess I'm going to take this one. I can tell you what we were trying to do with the Marmoset connectivity atlas, because this is sort of the newest guide to the party. And, you know, the first thing that we did is to look around and see, you know, if we want to be, because of course we would like to be interoperable and provide the data that I've used by many for different purposes, is to just go and look around what's on the market in what form it is available and how you can cater to this. And then you have to just design whatever you are doing so it works this way. And there are obviously unnecessary trade-offs, right? So you were yesterday on the meeting and there was a question, why we used retrograde connectivity and supposedly anti-regrade could be or should be better. Well the problem is that if you want to be compatible, you probably would like to use compatible methods or the methods that could be quantified and compared with results of other studies. You have to employ the definitions of the quantities that others employed and, you know, and stick to it. Especially when it comes to areas and homologs, I think we address this in a clever way. So as I was trying to show, except for allowing for this parcelation-based analysis, we were like, okay, let's drop the concept of areas altogether and just let's express it in a purely geometric way. So that's why I think it's very important that we provide a stortastic location of every labeled cell. Therefore you can take any of the parcelation of the marbles at the cortex that you would like to draw and impose it on the results and have your connectivity results piped through any parcelation you wish. Yes, good morning. So thank you very much for the talks. It was interesting to learn, since I'm a little outside of the area. But my question is addressed to Professor Kennedy to expand on the comment you made in the beginning about the mouse. Knowing a lot about the mouse is not going to help us about the human. But presumably the principles that are coming out will lead there. So I was just, I guess, a little confused by the comment. So, you know, so if you could expand. Thank you. Yeah. Well, a provocative comment, perhaps. Yes. The experiments that we can do in the mouse are much more sophisticated. And so we can test ideas. There's a recent study I saw showing that long distance connections are specifically targeting groups of cells which have a close lineage. That's going to massively increase the clustering at the single cell level. Now I don't know how you would do this in the macaque. I don't know how you would do that experiment. I don't think the technology exists to be able to ask the question if the connection of the projection from a long distance going from area A to area B is specifically targeting a group of cells which have a close lineage in that way. Now the consequences of that in the mouse are very, very, very fundamental. They're going to completely change how we think about how individual cells communicate across the cortex. So we have to battle and try and answer that question. Is the macaque different in that respect? I think anybody working on macaque actually has an ethical obligation because if you can do it in the mouse and if it gives you the same result, then you do it in the mouse and you get the same result. And if it's different, then you need to know about it. I think that we have as a community to work in parallel on the mouse and the macaque. I think that the tools will improve. There's transgenesis in producing macaque models. Mooming poo is pushing for this in Shanghai and that's going to be an important factor. There's the work in Marma set which is very, very exciting and there's definitely going to be and people like Pascal Fries is using optogenetics very successfully now in Marma set. So I think there's light at the end of the tunnel but I think that at the end of the day, the other point I wanted to make is there's not going to be a single model. The work of Frank Pollux and Cécile Charié looking at this gene which is expressed in the spines and showing that they exist in humans and this neotinic property of the spine in humans is going to be very different from any other macaque, any other primate. So there's not going to be a single model. I think we have to triangulate. I think we have to work on different non-human primate models. If you're interested in socialization, the Marma set is a prime choice. If you're interested in, so that was the meaning of what I wanted to say. I think mouse work is, I think there should be a much tighter interaction between work done in the mouse and the macaque. Does that answer you? More or less? More or less? What about the less then? I'll repeat the question briefly. Is the terminology we use sufficient for what we do, or should we also discuss to have a more precise terminology to better reflect the methods that are used? Just a short comment and I think we need to continue discussions outside this session. Maybe let's go down the line. I would just say yes. I think we would need better terminology. Just a connect term sounds sexy, but it's not capturing all the implications of what we want to do. And if you want to have precisely defined interpretations and precisely defined focus on what you're doing, you need to precisely define your terms. I definitely agree with that. In practice it's a matter of having a mental shortcut during a very time constrained presentation. Use the term connect them and the context is supposed to point to the details. Yes, I agree that the more precise we are, the better in this context. Yes, I think there's a big problem with vocabulary. I mean, functional connectivity, which I think is going to be the root into understanding human cortex and its relationship and the understanding of what we understand in macaque and taking it forward, but it's a terrible terminology. I mean, there's nothing functional about it and there's no connection. But it's used in a way that might, in a big audience, not everybody's up to date on what exactly you're talking about and people going away thinking, oh, he's talking about functions and connections and he's not. He's talking about something quite distinct. So I think that there is a problem. I think the word connect term is not bad at all. It's the idea that you really want to work with a square matrix. So in the good old days you would stick and inject one area and you'd say, oh, it's got this and that. You'd work in a sort of plan about how this was all functioning on the back of an envelope, but you wouldn't be knowing about the full connectivity of a subset of areas. I think connect term is correct. Just to keep this fun, I'm going to go ahead and say that, no, no, I think there is good in having some of the mix of terminologies, I agree with, you know, but the value of having these terminologies that overlap inappropriately at times is that we find a way to be able to talk to each other and even though there might be almost strategic misunderstandings that are taking place, it allows us to cross, to step out of our disciplines that we're comfortable in and to believe that we're at least sharing in a similar dialogue to extend beyond our domain limits. So there's also value to that. Okay, thank you very much. I think we have had a lot of food for thought. There will be an important message about all those means of sustaining ourselves. But first, let us give a big applause for the participants here who have given us a very exciting session on connectivity.