 Welcome, everybody, to this first workshop session, the neuroinformatics of neuroanatomy. I'm Mary M. Martone from UC San Diego. And I told Mary I would introduce myself. But for a long time, I've run the program on ontologies for neural structures for the INCF and have been long involved in trying to figure out what a next generation platform for neuroanatomy might look like. So the reason that this session has the title that it does is that it's become apparent over many, many years. We've been building scaffolds that use neuroanatomy as a way to place information both spatially and terminologically with reference to the neural space. So anatomy is very foundational to how we build information systems in neuroinformatics. Neuroanatomy is also a science unto itself. And you have the people who are still working to try to figure out exactly how the brain is put together, probably more than any organ. It has a wealth of parcelation schemes that keep coming and coming and coming. Yet as you saw that quote from the editorial George Geoscolli and I wrote in 2013, it's not that we mind people having disagreements. In fact, science will move forward. But if we can have those disagreements in a place where it's easy to relate information or one point of view to another, it's a much more powerful platform because we actually can have some validation for some of the things that we do. In fact, some of the very first neuroinformatics projects, going back to my early years in graduate school, we had little digital atlases that came on a floppy disk and you could look at. So we've been doing this for almost two decades. But informatics also requires that the information is made machine-processable. And right now, in terms of digital atlasing, that means putting discrete boundaries around things for ontologies. It means creating discrete categories. And we know the brain's not to put together like that. We know that it's a giant network and so many of these things are artificial. But nevertheless, you have to have some reference system. And it's been very difficult to get researchers to agree on what that sort of global reference system is. And registering from one representation to another still remains a big challenge. But I think that Linda put it very well in her opening session. At some point, we have to get to the users. And that means we have to get to those who are practicing the science of neuroanatomy. And it's very much still performed by individual who's published their data as figures and text inside of papers. And that's a very unprocessable mode of producing this data. And it makes it very difficult for us to have these arguments, as we say, in a way that machines can help us. And it's also quite remarkable that most parcellations, I think because of the way we're used to things, most terminologies reflect a very two-dimensional view of the brain. But the minute you look at full renderings of MRI, you realize a lot of these parcellations that we have are largely artificial. They are imposed by our own viewpoints as to how the brain is organized. And do we need to consider that anatomy needs to be redone in light of three dimensions? So we've got four excellent speakers who have been working on informatic systems at various different approaches for a very long time. And most of their talk is going to be about their particular system and letting you know what it does and what it doesn't do. But I wanted to, I also gave everybody a set of questions that said, you know, think about what assumptions need to be made to create effective neuroinformatic systems based on neuroanatomy. There are always assumptions that go into making these systems. And some of those relate to actual neuroanatomy. Some of them are requirements that are imposed by the machine. Does the requirement to have machine-processable anatomical information require neuroanatomists to change the way they perform neuroanatomy? If so, can we recommend best practices for those who produce or will produce neuroanatomical information in the future? And more importantly, do we have the tool and data infrastructures required to allow them to do this? These questions have become very important in light of the HBP, the Brain Initiative in the US where large amounts of data is going to be produced and we would like it to be referenced appropriately. And so I just wanted to show this slide, again, for those of you who maybe aren't neuroanatomists, but this is an excellent paper that came out in 2014. This is the way neuroanatomy is practiced. This is a nistelstain through the prefrontal cortex of the mouse. Turns out there's at least six different parcelation schemes of the prefrontal cortex of the mouse alone. And at least these authors, even though a lot of their information was published in text, they did an interesting thing. They took the original plates from these different atlas and superimposed their perspective on top of it. So at least you have something, whoops, which says, oh, what happened? Which says this is comparable to this, but this table here, which lists all of that, is very difficult for a machine to parse. So we've got the four speakers and we don't have a lot of time in this session for a lot of discussion. We wanted to give people time to present, but we are gonna have a little neuroinformatics of neuroanatomy lunch and tomorrow where we can continue the discussion and also see if we can pull together a perspective piece to see where we are. So, with no further ado, I'd like to introduce our first speaker, Trigve Lirgaard from the University of Oslo. And Trigve.