Hi Donna, I'm happy to let you know that we were able to get it!!
Briefly: We applied the deformation map (provided in the HumanAtlas_standard-scenes.164k_fs_LR.spec) to our .paint files (made via caret_command). We were then able to load them into this spec file. I'll put together a more detailed explanation this week. Thanks again and to all those who offered help as well! -Sam On 3/20/12 4:47 PM, "Damien Fair" <[email protected]> wrote: >Quick question Donna (note: our labels look fine in free surfer). To apply >the deformation map created from the script below, do we need to first >convert the labels to paint? > >Thanks, >Damien > > >Damien Fair, PA-C, Ph.D. >Oregon Health and Science University >Assistant Professor, Behavioral Neuroscience and Psychiatry >Assistant Scientist, Advanced Imaging Research Center >3181 SW Sam Jackson Park Road HRC5D30 >Portland, Oregon 97239 >503-418-0995 (o) >503-494-9988 (f) >[email protected] or >[email protected] > >Shipping: >Oregon Health and Science University >Behavioral Neuroscience Department >3181 SW Sam Jackson Park Road HRC5D30 >Portland, Oregon 97239 > > > > > >On 3/19/12 9:19 AM, "Donna Dierker" <[email protected]> wrote: > >>I don't know if you've been in contact with DVE, but something popped >>this onto the top of his priority stack. He's asked me, Erin, etc. to >>review this fs_LR dataset for public release -- soon (within a week?). >> >>What I could do is send you the script I have -- in its current state -- >>along with a rough draft README. Hell, actually it's here: >> >>http://brainmap.wustl.edu/pub/donna/WUSTL/WUNDER/YUNING/freesurfer_to_fs_ >>L >>R.zip >>login pub >>password download >> >>It's using Freesurfer's registration (sphere.reg) to get the surface on >>the hem-specific 164k fsaverage mesh, and then it unprojects it on a >>fs_LR sphere that has the hems in correspondence. It then generates >>deform_maps in both directions, so you can apply a single deform_map to >>get to and from initial mesh to fs_LR 164k mesh. >> >>But before you can apply this deform_map to your labels, you must make >>sure you have things straight on the native mesh. Look at the labels in >>Freesurfer. If they're not right, your problem is upstream of our >>script. >> >> >>On Mar 17, 2012, at 2:12 PM, Damien Fair wrote: >> >>> No problem DonnaŠand thanks. >>> >>> We'll take a look into the headers of the deform maps to see if we can >>>sort this outŠand we'll put a couple more posts up as well. >>> >>> Also, I think we're going to stick with trying to use the fs_LR >>>surfaces as it is likely the most straight forward for what we're trying >>>to do as of now. It's not specific to PALS per se. >>> >>> >>> Thanks for your help! >>> >>> Damien >>> >>> Damien Fair, PA-C, Ph.D. >>> Oregon Health and Science University >>> Assistant Professor, Behavioral Neuroscience and Psychiatry >>> Assistant Scientist, Advanced Imaging Research Center >>> 3181 SW Sam Jackson Park Road HRC5D30 >>> Portland, Oregon 97239 >>> 503-418-0995 (o) >>> 503-494-9988 (f) >>> [email protected] or >>> [email protected] >>> >>> Shipping: >>> Oregon Health and Science University >>> Behavioral Neuroscience Department >>> 3181 SW Sam Jackson Park Road HRC5D30 >>> Portland, Oregon 97239 >>> >>> From: Donna Dierker <[email protected]> >>> Date: Sat, 17 Mar 2012 07:15:20 -0700 >>> To: Sam Carpenter <[email protected]> >>> Cc: David Grayson <[email protected]>, Damien Fair <[email protected]> >>> Subject: Re: [caret-users] .annot or .label conversion to caret >>> >>> Sorry I haven't been able to attend to this properly. I was off with >>>the kids yesterday, and today is a birthday party for one of them, >>>followed by St. Pat's stuff. Maybe Sunday. >>> >>> But here are some more hints: >>> >>> * Try Rouhi's scripts, if possible. He's a post-doc or grad student >>>for Guy Orban who did this last summer. Ask him if you have trouble. >>> >>> * There are complex issues with these atlases. Sounds like you found >>>some fs_LR datasets in sumsdb (probably publications: vanessen / cc 2011 >>>human). Note that things have to be on the 164k standard mesh for that >>>deform_map to work, and they can be on that mesh one of two ways: the >>>hem-specific fsaverage.L or fsaverage.R, where vertex i could be in the >>>occipital lobe on the left hem, frontal in the right hem. Or fs_LR >>>where they are in correspondence (really DVE's mesh). If you more the >>>deform_map, the header will give clues as to what that deform_map wants. >>> >>> * On the Caret download site, there's a freesurfer to caret pipeline >>>that imports freesurfer into PALS, but it won't import any labels. >>>You'll need to do that, and apply not one, but two deform_maps -- a >>>downsampling one followed by a registration one. There is no script to >>>do that, but you might be able to figure it out. (Rouhi could help; he >>>did both the fs_LR route and the PALS route, for comparison purposes.) >>> >>> So decide which atlas you like the most, and focus on that stream. >>>Keep your progress/questions on the list, because now people like Rouhi >>>and Colin are providing real value-added to the support VE lab can >>>provide. >>> >>> Between you and me, I'd rather be looking at your problem than dealing >>>with hyper, ADHD kids, but duty calls. >>> >>> >>> From: Sam Carpenter <[email protected]> >>> To: "[email protected]" <[email protected]> >>> Cc: David Grayson <[email protected]>; Damien Fair <[email protected]> >>> Sent: Fri, March 16, 2012 8:27:31 PM >>> Subject: Re: [caret-users] .annot or .label conversion to caret >>> >>> Hi Donna, >>> >>> Thanks for your quick reply, your recall was more than enough for us to >>>piece it together. Damien and I worked through the conversion and were >>>able to load it into Caret. But unfortunately, it doesn¹t appear to >>>have converted correctly. I've outlined our steps below with a few >>>select screen shots. >>> >>> First step: caret_command >>> $: caret_command -file-convert -fsl2c Yeo_labels/ ../surf/lh.pial >>>output_file.paint (I've invluded this output file as well) >>> >>> **you were correct with the reference surface files being a potential >>>issue, we were able to use compatible reference file for the >>>conversion** >>> >>> Second Step (where things started to go wrong): Loading the .paint and >>>viewing in caret >>> We loaded the spec file that you mentioned (it was available on the >>>website) HumanAtlas_standard-scenes.164k_fs_LR.spec >>> (see images for options selected) >>> >>> We loaded the output_file.paint that we created and then adjusted the >>>options within DC/Paint Main as well the drop down Attributes/Area Color >>>File/EditŠ >>> >>> I've included screen shots of what happens when we try and make the >>>regions visible. It appears as the conversion did not work correctly as >>>the each region; when selected appears as a random cluster (the included >>>image has only one .label file selected) >>> >>> Thanks so much for your time and assistance. If you have any other >>>questions or comments; all input is greatly appreciated. >>> >>> Thanks again! >>> >>> -Sam >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> From: Damien Fair <[email protected]> >>> Date: Fri, 16 Mar 2012 14:32:10 -0700 >>> To: "[email protected]" <[email protected]> >>> Cc: Microsoft Office User <[email protected]>, David Grayson >>><[email protected]> >>> Subject: Re: [caret-users] .annot or .label conversion to caret >>> >>> Hello Donna, >>> >>> FYI, I think I found it on SumsDB and we are running some tests. If we >>>run into trouble, we'll send a post. Thanks, >>> >>> Damien >>> >>> >>> Damien Fair, PA-C, Ph.D. >>> Oregon Health and Science University >>> Assistant Professor, Behavioral Neuroscience and Psychiatry >>> Assistant Scientist, Advanced Imaging Research Center >>> 3181 SW Sam Jackson Park Road HRC5D30 >>> Portland, Oregon 97239 >>> 503-418-0995 (o) >>> 503-494-9988 (f) >>> [email protected] or >>> [email protected] >>> >>> Shipping: >>> Oregon Health and Science University >>> Behavioral Neuroscience Department >>> 3181 SW Sam Jackson Park Road HRC5D30 >>> Portland, Oregon 97239 >>> >>> From: Damien Fair <[email protected]> >>> Date: Fri, 16 Mar 2012 12:53:31 -0700 >>> To: "[email protected]" <[email protected]> >>> Cc: Sam Carpenter <[email protected]>, David Grayson <[email protected]> >>> Subject: Re: [caret-users] .annot or .label conversion to caret >>> >>> Hello Donna, >>> >>> How are you?!? It's been a while. >>> >>> Sam (below) is actually one of my RAs here in Oregon. Is the fs_RL >>>available? Or is there a way for us to get our hands on it? Essentially, >>>we have several surface regions from the fslaverage surface we want to >>>be able to visualize and manipulate as paint files/area color files (as >>>caret give us some viewing flexibility that we are familiar with). This >>>is likely to be easier to do on the fsl_RL surface in caret such that >>>the deformation process would not be necessary and we might even be able >>>to use the .label files directly. >>> >>> Thanks, >>> Damien >>> >>> >>> Damien Fair, PA-C, Ph.D. >>> Oregon Health and Science University >>> Assistant Professor, Behavioral Neuroscience and Psychiatry >>> Assistant Scientist, Advanced Imaging Research Center >>> 3181 SW Sam Jackson Park Road HRC5D30 >>> Portland, Oregon 97239 >>> 503-418-0995 (o) >>> 503-494-9988 (f) >>> [email protected] or >>> [email protected] >>> >>> Shipping: >>> Oregon Health and Science University >>> Behavioral Neuroscience Department >>> 3181 SW Sam Jackson Park Road HRC5D30 >>> Portland, Oregon 97239 >>> >>> From: Donna Dierker <[email protected]> >>> Reply-To: "Caret, SureFit, and SuMS software users" >>><[email protected]> >>> Date: Fri, 16 Mar 2012 08:00:26 -0700 >>> To: "Caret, SureFit, and SuMS software users" >>><[email protected]> >>> Subject: Re: [caret-users] .annot or .label conversion to caret >>> >>> I don't have the answers at my fingertips, but here are a couple of >>>clues that I recall from recent experience doing this: >>> >>> * Enter "caret_command -help-full > /tmp/caret_command.txt" at the >>>command line, and then use a text editor to read /tmp/caret_command.txt. >>> Search for "file-convert" and look for an example that shows conversion >>>from label to paint. >>> >>> * Make sure when you write your .label from Freesurfer that you use the >>>appropriate reference surface. I don't remember the appropriate >>>Freesurfer commands for this, but if you google Freesurfer aparc you >>>might find it. I don't understand the annot and aparc formats as well >>>as I'd like, but there is some command that writes this out vertex-wise >>>relative to a reference surface, and it's important that this be the >>>right surface. >>> >>> To use that deform_map file, your source labels need to be on fsaverage >>>*right* standard mesh. This might affect the reference surface you use >>>when writing your labels. I haven't digested all of this: >>> >>> http://surfer.nmr.mgh.harvard.edu/fswiki/CorticalParcellation >>> >>> We do have a script that makes it easy to get stuff on the new fs_LR >>>atlas, to which we're transitioning from PALS. Some parcellation >>>schemes have been moved to this new atlas already, so it would be >>>helpful to know what is already on PALS that you want to view with your >>>data. >>> >>> >>> On Mar 15, 2012, at 7:54 PM, Sam Carpenter wrote: >>> >>>> Hi All, >>>> >>>> We currently have surface regions generated in FreeSurfer and are >>>>trying to get them over to the PALS atlas in caret. They are in >>>>.annot and .label format currently. Once in Caret I plan to apply the >>>>Reg-to-PALS_fsaverage.RIGHT.RegWithPALS-B12.163842.spec.deform_map to >>>>map them to the PALS atlas. >>>> >>>> The problem: currently I am unable to load or view them in caret. >>>>Does anybody know of a conversion method or way to load them into >>>>caret? (Caret apparently accepts .label files, but they are not >>>>loading correctly) >>>> >>>> All help/suggestions is appreciated. >>>> >>>> Thanks All, >>>> >>>> -S >>>> _______________________________________________ >>>> caret-users mailing list >>>> [email protected] >>>> http://brainvis.wustl.edu/mailman/listinfo/caret-users >>> >> > _______________________________________________ caret-users mailing list [email protected] http://brainvis.wustl.edu/mailman/listinfo/caret-users
