 Give you a specific example of how the cave 2 has accelerated some neuro science research and in fact it's in a project which has already been mentioned by Wojtek into Huntington's disease. Okay. I hope you can see that. Okay. So the cave 2 facility is a landmark ultra scale visualization facility at Monash. It runs in both 2D and 3D modes and at the time of construction, and we still believe this is true, we believe it's the largest hybrid 2D, 3D visualization facility in the world. It is 84 megapixels, 84 million pixels, and there's a computing cluster which powers this facility. Essentially it's dedicated to just displaying graphics and that runs at 90 teraflops. Or about one teraflop per screen. There are 80 screens in total. It features a very high resolution head and wand tracking system comprising some bicon cameras around the top of the screens and for motion capture. And there's also a 22 channel sound system which is a 3D sound system which enables us to play really neat tricks with the audio and play sounds in certain locations within the facility itself. You'll notice that it's circular. It's eight metres in diameter. And there's a really good point why circular systems are important to us and I'm going to show you that in my example. Now, Boitech briefly alluded to this and I wanted to point out how we think about this type of facility in the context of an overall strategy at Monash University. And we call this the strategy for the 21st century microscope and it underpins our thinking when we operate with a new research centre. So a microscope which has been at the centre of scientific discovery for hundreds of years really has three parts to it. It has a light source down the bottom. It has focusing knobs in the middle and it has a viewfinder at the top. We believe the modern microscope isn't a single instrument but in fact the seamless orchestration of capabilities and facilities that perform those roles. So down the bottom the light source can be a beam line at the synchotron or an MRI scanner, you know, some data producing instrumentation. In the middle here we have the focusing knobs. Now, when a researcher uses a microscope, they are interacting with the microscope. And this is a theme that's come up in all of these talks in this afternoon's session is it's really important to make that supercomputer interactive because that's how researchers want to engage with the modern microscope. And so we build this around the massive project which Voitec spoke about and that represents the kind of the core of the engine of the microscope. And then the computer screen is the viewfinder and there are obviously many different types of viewfinder but to us immersive visualisation and the cave 2 facility offers a unique opportunity. And in fact we like to describe it as the highest quality lens for our microscope. Now all of this is obviously linked up with high end IT to manage the data. We know we're on a data deluge, exponentially growing amounts of data need to be managed. And then at the top of the microscope we could think about having others actually view and look down our own microscopes and so we regard that as the data dissemination part. So over five years our e-research programme at Monash has been built around this concept and I'm going to talk to you about just one part of that. Voitec's already talked about the middle part and in some sense Gary and others represent the lower part. So the cave 2 facility at Monash was built in a brand new building and this is David Barnes, he's the manager of the facility. And what he's looking at here is tractography data and here he is now looking at some data from an archaeological site. But in many senses the tractography demo is one of the more spectacular ones. Now this is Jason Lee, who's at the Electronic Visualisation Lab in Chicago and they worked with us in building our cave 2 facility. They built the prototype and they kindly allowed us to build ours bigger, so it was a nice of them. But that's their facility, they were responsible for the original cave. Actually the lab you might know is also responsible for the original computer graphics in the very, very first Star Wars movie. In case you're wondering cave is a recursive acronym. Now we do do a lot of demonstrations and the cave facility does provide an opportunity to showcase Monash research. What you've seen there in a time lapse video is one demonstration which runs through a number of different use cases for the cave and as I pointed out the very top, the connectome example is always a stunning example. The one example that gets all of the audience going, ooh, ah, is actually standing on the surface of Mars and seeing Mars as though you were in a spacecraft on the surface at the limits of human visual security. Now the cave facility is both a 2D and a 3D facility. The 3D nature is very important to us and in fact one thing you appreciate very quickly is that objects when you see them in the cave appear as though they're in front of the screens. They're almost as like they're occupying the space that you're standing in. This is very important to us to help us understand spatially what's going on. When you go to a 3D movie most of the action appears to your brain to occur behind the screen so you're like looking through a window, but in the cave it's actually happening all around you. The architectural firms love it because they can make you feel like you're walking through a room. But it's also the highest quality lens for discovery and a number of our biomedical scientists love it because it enables them to do things they've never been able to do before. See things in incredible detail but also within the context of the image. We have histology images for example that 100,000 pixels by 100,000. Researchers in the past could not actually view those on a regular computer screen but they can in the cave. Some of you might have seen this particular image. It's always a good slide to talk about when we talk about visualisation. We think it's underdone in general and we are working hard to promote its use because to us visualisation actually helps you discover the unknown unknowns. There's a famous quote from Donald Rumsfield which I'll let you read there. He was talking about Iraq but it also applies to research. Computers, if you're applying them in a data mining context, only find things that you know about. They only find things that you know how to find. Visualisation provides you with an opportunity to see things that you'd never seen before and in fact you could think of it like the cave facility as a machine for generating a new hypothesis. The reason for that image there is half of you I imagine can see what's in that image. Put up your hand if you see what it is. Yeah, it's about half of you. Some of you are difficulty finding it because you don't know what it is. But now I'll tell you that it's a dog. It's a Dalmatian dog. I think probably those who couldn't see it before can now see it. My little point about finding the unknown unknowns. To us neuroscience represents the exemplar for big data visualisation. We've seen it in other domains. Our e-research centre is dedicated to all of the disciplines but neuroscience is one of the most important disciplines for us. But the point here is the humble desktop no longer works. The screen is too small. The bucket is too small. In terms of the desktop computer is being dwarfed by the big data that's coming off the instrumentation. And when you want to collaborate around the data you tend to huddle around a little computer screen. It's also typically in 2D. So to us the cave 2 is the natural place to go to visualise. As I said the neuro-tractography visualisations are very good ones for us. It's immersive, it's colourful or memorable. The spatial comprehension dramatically improved over the desktop. For example as you are swimming through the tractography you really do appreciate the space and the gaps. And the sizes of the clusters of fibres and so on. But the one example I'm going to show you now is actually a combination of both 2D and 3D. And it's really a good example of how the brain, our human brain is actually the best pattern processing machine we have. And it really is applied in very good detail in this particular example. So we call this comparative visualisation and this is in the context of the image HD project. It's a nice little pun but HD here stands for Huntington's Disease. So as Vojtek mentioned this is a study of 80 subjects. We've got some controls around 30 of those, pre-symptomatic around 30 and some sufferers around 20 of those. But what we do is we visualise their tractography and we visualise all 80 simultaneously. I'm going to show you a video of this actually running. But I'll point to the research, the paper. As Vojtek mentioned, nearly Georgio Caristianus is the principal investigator on this work. This paper was actually published in 2014 and was selected by HD Highlights as one of the most influential in Huntington's Disease. And as I mentioned, the idea is in the context of the cave is to investigate the structural connectivity in the pre-manifest and the symptomatic subjects. And there's some control in there as well. There have been 80 patients that maps very naturally to the 80 screens within the cave. Now the first point I want to make about the circular nature, when you're in the middle of the cave, you don't see the pixels. You're seeing things at the limits of human visual purity. So it's like a retina display. But you're also equal distance from all 80 of them. So you don't tend to naturally favour one particular group over another. And so that's why I believe circular facilities are actually important. We can sort the brains in any order. We often sort them by the onset of the disease or the controls and so on. Or we could sort them by age and so on. And David Barnes, who led a lot of the visualisation work here, prepared this nice video of it running for me. What I'll do is I'll run it, it takes about three minutes and I'll talk to it as it runs. I want to highlight, first of all, you're not seeing a video of the cave in action. You're seeing, just in order to give us enough detail in the video, three columns of the cave. There's a cast of thousands involved. These are the names of all of the people that were involved in the work. And I really do hope this is going to run. It did run in the test. This is not my computer, which doesn't help. I did test it. Oh, I've got it. OK, so we've seen three columns of the cave. And then on the left, you're seeing the web interface of the... That's running on an iPad that you take with you into the cave. So here we can... It's a bit hard to see, but here we can choose how many track samples to reveal. There's a single structural brain, but what we've done is we've mapped on the tractography of the study subjects onto the brain. And so we can choose how many samples. There are millions of tracks, but we can choose how many of those we wish. Now, what we'll do is now we'll reduce the density of the structural brain so that it doesn't get in the way. And then we can start to rotate. Now, when you rotate, you end up rotating all 80 simultaneously. So here we are moving from the top and then moving to a rear view. Now, at the moment, and it's a bit hard to see, it looks like all of the brains are the same. So what we'll do is we'll pick a particular region and we'll just expose the tracks that are coming from that region. Now, it's pretty clear that we're dealing with different people. So we'll just, again, we'll rotate to get a rear view of the 80. And as you can see, it's quite clear. Now, what I'm going to do is I'm going to show you how the main conclusion in that research is actually observable within the cave. So the main conclusion in the research is that there is a difference between the controls and the symptomatics when you isolate two particular regions and you look at the tracks. Now, here what we've done is we've now chosen two regions and we're looking just at the tracks between those two regions. And we are going to push the number of samples right up because, of course, we're dealing with a much smaller set of links. But now, and it's probably not very clear to you, there are clearly identifiable differences that occur at a group level between the controls and the symptomatics in that particular case. And so this has opened up a number of opportunities for us. We've validated previous research using the cave, but there's other things that the cave's done. It's actually accelerated the research process. In fact, the group talk about how it accelerates the data quality control work time. It used to take three months in terms of cleaning up the data. It's now down to two hours. It is just so much quicker to do it when you see everything at once. So as I pointed out, the technology and the approach has been validated by the same discovery. So this provides us with hope that we're going to see new observations and new discoveries within the cave. And that discovery was essentially group level differences in the pathways between the two regions. Now I know I've got one minute, so I'll wrap up. As I said, the cave 2 is a hypothesis generation machine and it's the highest quality lens to look at digital data. But it's also an effective communication tool for understanding spatial aspects. Now I'm just going to throw out there this final slide, maybe some future work. For example, what about using the cave for some sort of neurofeedback type training? Having a number of people within the cave with EEG headsets navigating and interacting with what they're seeing. And so we can have them interact with a virtual environment for a group of patients and not just a single individual. OK, with that, any questions? So Paul, that was fascinating. And my question is a bit more genuine around the principle of visualisation led scientific discoveries, which is something I know that David Barnes has espoused and I think it's really fascinating. I guess it's a question about the relevance of the given of that approach to different disciplines. And after the work you've done over the last year or so, how do you see it in terms of that approach to applying to the science, as you mentioned, including neuroscience? Yeah, so in our microscope analogy, it's clear that we tend to work closely with disciplines that use instruments that produce data. And as I said, discoveries often results by simply just looking down a microscope and going, ah, now that's interesting. I wonder what causes that or wonder why that's like that. And that drives a hypothesis which then derives, hopefully, the discovery. So disciplines which operate in that way work particularly well with the cave. And so the other example is, of course, structural biology and understanding, I guess, how things are coming together from a structural point of view at the molecular level. But also, again, they work with instruments that produce a huge amount of data. And rather than mine it for stuff they know about, they want to use visualisation to discover things they don't know about. Okay, as chair of the session, I think we're going to wrap it up because it is...