 Let's get this workshop started. It's a pleasure to be chair. I want to start with some thanks, first of all. I'd like to thank Porick for helping to organize this session and of course the speakers for coming to tell us about their work and obviously the audience. But I thought I could start the workshop on open collaboration in computational neuroscience by just giving a little bit of a motivating argument for why we need more open collaboration in computational neuroscience. So many of you will be involved in developing models of neurons and synapses. And then perhaps fewer of you will be involved in modeling more large scale networks. And then very few of you will be involved in modeling the whole brain. And this illustrates a scaling factor that in the long term, we want to be able to have very detailed models of brain function with this sort of detail. But we want to be able to model both networks and groups of networks and systems in order to really understand how the brain represents information of sensory systems, motor systems, and cognitive systems. Now, traditionally, this model development has proceeded in this rather linear fashion where you'd have a new PhD student in your lab or a postdoc who may use existing data or may generate experimental data themselves and then start to build a model to fit that data. And this whole process may take one or two years, depending on their ability and the difficulty of the problem for building, for example, a single cell model. Now, this sort of approach doesn't scale very well, because often, once the PhD students finish, the code may be released on a database like ModelDB, but is frozen and may not be reused. If we really want to build a sort of modular approach to modeling the brain, we need to come up with a better system for reuse of model components and really building on what people have done before. So then the idea is to actually to collaborate together to build models and use open source type approaches in order to do that. So this shows a sort of cultural change in this type of approach for model building in collaborative development, and that's what we'll be talking today and the associated neuroinformatics tools. And there are many benefits to getting together to build these complex models, not only continuity and building in a modular fashion, but also, as we know from open source, you in the end get a much more robust product and potentially, by having a lot of more people looking at a model, it's going to do its job much better. But what do we need to go into a more collaborative mode? Well, today we're going to have some examples, just some of the things that are out there that can be used. So if we start with the data, the data sets out there that you can download, and we're going to hear about neuroelectro today. There are tools for collaborative model development, both the open source software development tools, but also simulators, which have been around for a while, applications, and again, we're going to hear some of these today. But also, crucially important for building the complexity of models and having this modular approach is to actually standardize the description of the subparts of the model. So if you have a SNAPTIC model or a neuronal model, you can then plug it in. So you need a common language, and there are a number of initiatives to develop a common language for computational neuroscience, and some of them are mentioned here. And this also, we have the example from Systems Biology, which has shown us that this is actually a very, very effective way forward. And finally, we may be thinking about neuroinformatics, but we mustn't forget about the experimentalists, because ultimately, the experimentalists know how the system works. They're intimately versed in the behaviors. And unless we can get them into this cycle of model development, we're never going to get really realistic, robust models. So we need to combine these standardized model descriptions with repositories that enable us or enable people who aren't familiar with coding to actually go into the model and see the details of the model and see how it behaves. And these two initiatives, Open Source Brain and Open Worm, are trying to do that, as well as model DB as well now. And Biomodels is from Systems Biology. So what are the benefits of collaborative modeling? Well, obviously, I've mentioned scaling up model systems, continuity of projects, improve community resources if we can reuse things, and the development of an integrated tool chain and making it much easier to build models of, for example, neurons. But I think the most important thing is that by taking this approach, we'll be able to actually improve the scientific value of computational models of the brain. And we can do that by improving reproducibility of the models, accessibility of the models, allowing experimentalists to really see whether the model is doing what the biology is doing, this portability, and obviously transparency. And reproducibility, unless we have reproducibility, it isn't science. So what are the difficulties, issues that may come up today? And I think we should, when the talks go through, we'll have some specific questions. But at the end, we can discuss some of these more general issues. Neuroscientists is a competitive system. Neuroscientists don't share. So your parents will have told you to share throughout your childhood. But when you become a neuroscientist, you go back to being about three years old. Vested interest. Can we change the prevailing culture? There's a big investment in the tool chain. And it's quite an arduous job to build the tools. How can we ensure quality of models? How can we grade them? And so well, the efforts of the developers who are putting in all this effort into the tool chain, for example, be rewarded within an academic or mainly academic system. And one for the actual modelers themselves, are you ready to expose all of the warts of your model in a way that anybody can see them right away? So those are the things that are slowing us down. These are the benefits, and I think the benefits outweigh by far most of the difficulties and issues are actually things that shouldn't be an issue. OK, so as I mentioned, we've got a great lineup today. I don't need to go through all of them. I think we should get down to the people who are actually doing the work. And we're going to have about 15 to 20 minutes each. And then hopefully that'll allow maybe five or 10 minutes discussion, especially if people want to get involved in how to push collaborative modeling forward.