 It's getting a bit difficult because what you want to show you is more difficult while the time is much less So we will guide you more through the process. It will more be a guided Guided walkthrough than really a full exercise, but I would like what I would like to show is a Rather new tool that we have developed to anchor partial volumes to the big brain model And this tool is set up in a very similar way that the Atlas viewer is working that you have seen so You do not download the big brain model on your computer to work with it because it's really huge data set It is a terabyte in size but instead you upload your own image to a server and then you manipulate the Position of your volume in the big print space on online in a web browser And this is this is a quite new tool. So the currently the version that you find online is Has a rather simple set of functionality and it's not as Not yet being used by so many users as you have it for quickening, which is already available for a longer time So the the principle is we will soon have a look at it you can actually Upload and you can select from from some example volume metric data sets But you can also upload your own image volume. This requires you to log in using an arc ID and What you then see is it looks very similar to the to the interactive Atlas here It's basically using many software components many of the same software components But you will then see the uploaded Partial volume on top as an overlay on the of the big brain and the basic idea is that to just manipulate it there So you can simply click it and move it around you can rotate it using the shift button again As just as we did for the for navigating the big brain and you can also do some So there is also a dialogue to manipulate the translation and the scale and rotation, but you can just do it with the mouse immediately and then there's a There's an advanced mode where where you can then add corresponding landmarks to refine the position this will allow you then to to apply an affine some affine components of For the alignment so introducing some shear in different directions We will not do this today because it requires quite some some expertise For neural anatomists it might be quite quite intuitive though Yeah, the tool has an additional mode that that can always be switched where you see the two volumes side by side So in the default mode you see an overlay, but there's on the top left There's a button that you can always switch to a side-by-side view which in particular for adding landmarks is for many people more convenient And it has a it has an undo history browser, which is simply the third button Which actually shows all the manipulations that you have done and you can roll back in time all of those You cannot jump around, but you can just sequentially roll back And there's an output of the tool what you can do is actually you could if the Uploaded volume was a was a nifty file You can download that nifty file with an up with an updated affine matrix being embedded But what you can also do is open the result in the Atlas viewer So that you have all the functionality that we have seen in the last session Right at hand together with your own image date. So let's assume that you that you upload a Whatever, you know an image of the hippocampus what we will do then you would be able to actually see the the maps on top of it and also do spatial queries and so on in the Atlas viewer and A third possibility is you can submit it to the curation teams If you want to actually this this data set to be released in the human brain project to the public We will walk you through it with an example data set you will guide it here just one Mentioning of future directions. So we are currently implementing ways since in the human brain I told you in the beginning you have this This massive differences between different subjects. We are now implementing mechanisms that will allow to also do non-linear Deformations so there's already one Function in development, which is not yet released where if you have a cortical patch Like a 3d volume of a piece of cortex and you have a segmentation of the cortex You can use that segmentation actually to normalize the volume to the to correctly Correctly project into the cortical depth of the big frame So there's actually it goes into the neuroanatomy But there's a there's a model of how the cortical depth is actually affected by curvature And this model can be computed from a segmentation of the of the gray matter and of the cortex And this is actually in the background of light so that you can then hit a button Once you're almost at the right position as you can see here You can actually hit a button and then the deformation is computed which matches as As well as possible the curvature and the cortical depth for now. This is this is not yet released. This is in development Yeah, but let's give it a try. So here's here's the the web address will uber dot apps dot HPP dot EU and You're not you don't need to log in There's an example data set that you can just use without logging in and you will Hi, yeah, do you maybe it's just cashed on my machine Wait see it's also here Let me let me just so is everybody on the website already I Hope you saw the note that you are requiring orchid ID So you just have to authorize it. You're through that already talk to me Hmm exactly and then you select the select over here the hippocampus part Just go through this now you already did log in with your orchid You have to I'm trying to Because start is not available right now. So you have to go back back. I would have just gone through it so and I Hope you see All that so in my in my case when I traded with my laptop it took a while until the data were loaded You can move around the individually the The data yourself and of course you can also again move around the Sections if you are within here and aim for the part you can move around your within your data sample if You go to the brain again Then you move the whole Big brain so the space be a bit aware of this because if you adjust your your data And then move into the wrong spot you move the wrong sample and you don't want to do that once you're adjusted everything So the first thing would be to see if They're actually if the data is actually in the right direction in the right general direction, and I don't know if you Maybe walked around a bit here and had a look at At your data set then you may realize that it's Rotated in the wrong direction, so I must say I'm not very anatomical experience. I didn't see that on the first glance so what you should do is you go over to the Transform incoming volume. So that's on the upper left Corner the top icon You press it and then you have actually here the possibility to already Rotate in the directions and what we should do we should rotate it for 90 degrees in the x direction in order to get it in the rough Correct rotation already So everybody was able to find this Okay, so what we now are doing is so the opacity I don't know if I think Timo didn't show that before the opacity of or the Transparency of the data set can be changed in the upper right corner So you can turn that down in order to make it completely transparent You can also turn it up and the best case is to have it first to realize what is actually in the data set you Turn the opacity completely up to a hundred percent What you then should do is that in the lateral you should search for the lateral end of the hippocampus within your data That you upload it. So within your sections, please search for the Lateral end of the hippocampus. I will try to do that as well As far as I can So you might have to switch between the windows a bit I'm really not good in that to me. I think that's fine, right roughly at least So it look should look somehow like this. I think some of you have more experience and Actually adjusting it So you see here in the lateral view in the lateral window somehow here the edge of the hippocampus and if we now move the Move out our data set. So you just move it to the side We can search in our space actually roughly the same spot of the hippocampus So please try to find the lateral one in here as well This happened to me the last time too. No, this is again. Yeah, you see It's hard to see it here What do you think roughly correct? I think it's fine And then you can move your your Area back here. So I have to readjust. I hope this didn't happen to you. I don't know why this happened Yeah, you don't do that Took some time So if we search a bit the opacity again You see that of course our data set is much too large on the one hand and on the other hand That's still it's just in a 90 degree angle and it should be tilled it again So what we want to do if we want to change this we go back to the transform incoming volume And then we go to rotation first and we could adjust the x-axis a bit for example to To 60 degrees Yeah, I don't know I don't want to do 60. I wanted to do 60. Thank you Yeah, so this is our volume and you see that it it's tilted a bit and It kind of roughly matches if I turn off the opacity the tilt of the brain here And then what I want to do is I would like to adjust the scale and You can do this here and then You want to do it smaller? So you just go down and the rough part should be around maybe 94 or a bit less You can play around a bit and try then to adjust your brain So not you're you're now touching not yours your data anymore, but you're adjusting the space to the right position I Think I have to deselect that first in order to calculate So the next step would be so now we have the rough direction we have the rough position of it The next step would be then what Timo showed in the slides very quickly As you go into details you select certain landmarks like for example sample the interior and posterior end of the hippocampus And then you start realigning doing the fine-tuning based on the landmarks and If you want to you could try and play around with this in the in the last 15 minutes Or just play around with the data set itself try to readjust it better than I did Which I think is not that difficult and See if you can maybe figure out some some spots where you can set some landmarks I had a landmark so there's the landmark editor is here So you can add that much you can save the list of landmark and restore it then here's the undo Yeah, we can as long as we don't do anything we can even now You can go backwards and see what you have done and you can go forward again The red one would take you all the way back to reloading the data. It's asking though before doing it and Then there is the possibility to to actually save A jason file with the fi matrix to see your result in the Atlas fewer And this is yet great out But but soon you will be able to submit this into the curation process So it is a case my data set is already uploaded I have already worked a little bit on the positioning and then you can submit it to curation teams and they will they will Support you in releasing the data on the platform You can get them the map of the coordinates based on the original volume to M&I space Sorry, so if you download it you mean it over the nifty file in M&I space I'm just wondering what coordinates you can get out. So you have your original 3d file. Yeah, and then Can you map it to this and then also get the map with M&I The big brain Yes, yes, so it's It is like how you work with all these transformations in practice if you take it to your own machine. That's a that's a difficult issue In the system. This is this is exactly one. So if once you're in the big brain space We are in a system. We have stored transformations from big brain To the M&I spaces that that you can use this is currently not connected yet to this workflow So it's but that that is something that the system will will provide The the difficult problem is how to properly communicate this because if you in you pray if you go from the big brain to M&I They are different choices of how you compute these transformations with which can have really different Different results, you know what we are currently doing is mostly a non-linear information that takes into account the cycle constraints But this is not for everybody and every every experiment what you exactly want So I the difficult part is really how you know how you properly document this to the user these mappings For for this part that we have done now. It is quite transparent, you know, you know yourself what you have done Scheme it can be documented And so on but as soon as you go into non-linear projections across the template spaces We have them, but we are currently thinking about what is the proper way to actually Communicate this with the user Document it But that's the idea. Yes, this is currently the this is a very fresh release It just has the big plan as we fall into reference space But it's it will be very simple to just put the M&I reference space there So that you can already directly say that that is just a switch on our side Mm-hmm Well, there is There is the landmark rack tool, which is also for volumes But so on on details, I think one of you could maybe refer a bit more on it And this is just possible here. So it's it's just we don't have a selector for selecting the mouse brain yet, but it's just There's basically nothing we would have to do. It's just you could simply switch to the mouse brain and use the tool It's just that the first release we only provide the big brain, but it's then nothing speaking against Extending it. Yeah So you can integrate L mRNA atlas Then you can fit most Space That the point is actually this has been mostly written with the big brain in mind because there you have this particular problem That is very very difficult to download the target data set to your own machine You know, they're like you have some interactive Alignment there are interactive alignment tools that you can install in your computer available But they wouldn't support such a big reference volume. So that is that is what why we primarily build it It's that there are ways to do it on your local machine as well, but it's Support here. It's just that you cannot select it yet