Donna, the link appears to be password protected, thank you Traci http://brainmap.wustl.edu/pub/donna/FREESURFER/SCRIPTS/2009_09/
-----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Donna Dierker Sent: Thursday, October 08, 2009 1:42 PM To: Caret, SureFit, and SuMS software users Subject: Re: [caret-users] do spherical registration using freesurfer generated surfaces Hi Traci, No, I don't. *Someone* -- YOU KNOW WHO YOU ARE -- has me working on similar scripts, but different enough to be both a fair amount of work and not as applicable to what I had in mind. I hope to wrap that up this week (but we'll see). Meanwhile, I posted interim versions of what I do have: http://brainmap.wustl.edu/pub/donna/FREESURFER/SCRIPTS/2009_09/ freesurfer2caret.Ver2a.preborder.mni2wustl.sh - imports Freesurfer to Caret and creates inputs for next script draw_core6_landmarks.tim.sh - draws borders border_variability.sh - generates quantitative measures of landmark quality generate_scene.check.sh - generates captures, scenes of borders [manual step: Layers: Borders: Border Update] postborder.sh - smooths medial wall, morphs and aligns sphere, registers If you have more than a handful of subjects, it will probably be worth getting these to work. Realistically, I hope to have something applicable to more typical Freesurfer input by 10/31. The script above had more steps like this than the typical stream would have, due to strangeness of the input data and inadequate talairach.xfm (not imgreg's fault): caret_command -surface-apply-transformation-matrix $CASE.$HEM_FLAG.MIDTHICKNESS.coord $CASE.$HEM_FLAG.topo $CASE.$HEM_FLAG.MIDTHICKNESS_orig.coord -matrix $MATRIX Donna On 10/08/2009 01:15 PM, Sandoval, Traci I wrote: > Hi Donna, do you have these cool caret scripts ready ie... > > Next week, I'll adapt these import scripts to a dataset more like what yours > is likely to be. When I am done, I can post the resulting scripts. > >> Can you ping me in two weeks? >> >> Donna >> > > Thanks for all your work and help. Traci > > Traci Sandoval > Research Assistant > NeuroPsychometric Research Lab > School of Behavioral and Brain Sciences - University of Texas at Dallas > Center for BrainHealth - University of Texas at Dallas > Department of Psychiatry - University of Texas Southwestern Medical > www.utdallas.edu/research/nprlab > 972-883-3270 > > > > -----Original Message----- > From: [email protected] > [mailto:[email protected]] On Behalf Of Donna Dierker > Sent: Friday, September 18, 2009 11:20 AM > To: Caret, SureFit, and SuMS software users > Subject: Re: [caret-users] do spherical registration using freesurfer > generated surfaces > > Traci, > > Yes, one lab here at Wash U is doing that. A few comments: > > * Since you don't have a flat map, the reference distance for the > posterior sylvian starting point changes from 9mm posterior to the > SF/SF2 intersection (relative to the flat map) to 12mm posterior to the > SF/SF2 intersection on the spherical map. > > * It's tougher to discern the full sylvian trajectory on the spherical > map, as well as center the sylvian on the main window, such that you can > see the whole length. Fortunately, the absolutely fun-to-use Layers: > Borders: Border Update makes this easy to fix, if you can't draw the > whole length in one go. (If your older Caret version doesn't have this > feature, you really MUST upgrade.) > > * You have to draw the medial wall dorsal and ventral in separate > segments, and you have to pay attention to their orientation (dorsal > starts anteriorly, ventral posteriorly). Calcarine starts posterior. > Flattening took care of this for you. > > * The lack of a compressed medial wall view makes it harder to draw the > medial wall ventral. > > Donna > > On 09/18/2009 10:19 AM, Sandoval, Traci I wrote: > >> Donna, That’s sounds great, I have scripts that take my freesurfer brains >> all the way to flattening in caret, but I flatten and draw borders manually. >> I noticed you mentioned you guys are doing the auto-landmarks on the >> spherical surface, and not flattening at all. Can I do that manually as >> well? And skip flattening? Thanks Traci >> >> -----Original Message----- >> From: [email protected] >> [mailto:[email protected]] On Behalf Of Donna Dierker >> Sent: Friday, September 18, 2009 9:26 AM >> To: Caret, SureFit, and SuMS software users >> Subject: Re: [caret-users] do spherical registration using freesurfer >> generated surfaces >> >> Traci, >> >> I do have some scripts I could make available now, but in two-four >> weeks, I'll have more general purpose ones. The most complete set were >> actually developed for a data set segmented in Caret. But they include >> the auto-landmarks, border variability (QA), generating scenes/captures >> showing the borders, and running registration via command line. There >> also are several versions of a "preborder" script that covers importing >> from Freesurfer, but most of these used anatomical data that was in >> 711-2* space before feeding to Freesurfer. Your data is unlikely to fit >> that category. >> >> Next week, I'll adapt these import scripts to a dataset more like what >> yours is likely to be. When I am done, I can post the resulting scripts. >> >> Can you ping me in two weeks? >> >> Donna >> >> On 09/17/2009 04:07 PM, Sandoval, Traci I wrote: >> >> >>> Donna, I would love to use these scripts that aren't publicized yet. How >>> can I attain said scripts?!! Thank you for all your work and questions >>> answered. Traci >>> >>> >>> Traci Sandoval >>> Research Assistant >>> NeuroPsychometric Research Lab >>> School of Behavioral and Brain Sciences - University of Texas at Dallas >>> Center for BrainHealth - University of Texas at Dallas >>> Department of Psychiatry - University of Texas Southwestern Medical >>> www.utdallas.edu/research/nprlab >>> 972-883-3270 >>> >>> >>> >>> -----Original Message----- >>> From: [email protected] >>> [mailto:[email protected]] On Behalf Of Donna Dierker >>> Sent: Thursday, September 17, 2009 8:44 AM >>> To: Caret, SureFit, and SuMS software users >>> Subject: Re: [caret-users] do spherical registration using freesurfer >>> generated surfaces >>> >>> Hi Jidan, >>> >>> See inline replies below. >>> >>> Donna >>> >>> On 09/17/2009 02:06 AM, z丹丹 wrote: >>> >>> >>> >>>> Hi, >>>> >>>> I already have fiducial surfaces generated by Freesurfer, which are >>>> all in the MNI space. I want to do spherical registration to an atlas >>>> spherical surface. The steps i need to do are: >>>> >>>> 1, generate the ellipsoid, spherical surfaces, and flat maps. >>>> 2, create the 6 landmarks on both individual and atlas surface. >>>> 3, do spherical mapping. >>>> >>>> Are the procedure right? Thanks! >>>> >>>> >>>> >>> Funny you should ask. We (the Van Essen Lab) are using Freesurfer for no >>> less than five projects, and we have shell scripts that streamline much >>> of the import, border drawing, QA, and registration. There is a feature >>> we call "auto-landmarks" that draws the borders reasonably well. Many >>> borders need touch-ups, but a single rater fixed 72 hemispheres' borders >>> in 1.5 days. If you have only a handful of subjects, it's not clear >>> whether it would be worth it for you to use our methods. But if you have >>> dozens, then they would save you a lot of time. >>> >>> We haven't broadly publicized these features/scripts, because they're >>> still in the final stages of development and testing. But they have been >>> working well for us. >>> >>> Even if you don't use auto landmarks or these scripts, one question >>> stands out in your steps above: Do you need a flat map? No, not really. >>> It certainly is easier to draw landmarks on the flat map than the >>> sphere; however, I'm not sure it's worth the hassle. Then again, if >>> you're not using the scripts, it might be the shortest path to registration. >>> >>> >>> >>>> In fact before I move to the 2nd step, I was stuck in the flat map >>>> generation. I met two problems: >>>> 1, As my surface is in MNI space, the surface orientation is not right >>>> in the CARET software. I found that the anterior and posterior are >>>> reversed with superior and inferior. Is there any method to roate it >>>> into the right orientation in CARET? >>>> >>>> >>>> >>> Yes, but this is not consistent with my experience of MNI space nor >>> Freesurfer surface orientation. While MNI volumes are sometimes stored >>> using left-handed coordinate systems (what some call radiological >>> orientation), the coordinates are always reported with -x being left, +x >>> being right; -y being posterior, +y being anterior; -z being inferior, >>> +z being superior. I've never seen a Freesurfer surface with the Y and Z >>> axes flipped -- not even ones written in MNI space. >>> >>> It's easy to flip a surface from LAS to LPI >>> (http://brainmap.wustl.edu/OLD/SureFit/orient.html) like this: >>> >>> caret_command -surface-apply-transformation-matrix $COORD $TOPO >>> $FLIPPED_COORD -matrix "1 0 0 0 0 -1 0 0 0 0 -1 0 0 0 0 1" >>> >>> But I really would not brush this off. I would get to the bottom of why >>> the supposedly MNI surface is not in the expected orientation. The >>> evidence suggests it is not in MNI space, as asserted. >>> >>> >>> >>>> 2, The other problem is , I found that the origin of my surface is not >>>> AC. So when I flatten the surface, it can't find the correct medial >>>> wall to cut. Do you think there is a way to solve this too? >>>> >>>> >>>> >>> This is more evidence that the surface is not really in MNI space. How >>> did it get transformed to MNI? >>> >>> Note that the MNI origin is not exactly the AC; see >>> http://imaging.mrc-cbu.cam.ac.uk/imaging/MniTalairach for details. But >>> it is certainly close enough to the AC that the compressed medial wall >>> view should be centered over the medial wall. The fact that it is not >>> indicates a problem with your transformation to MNI space. >>> >>> >>> >>>> Thanks a lot! >>>> >>>> Jidan >>>> >>>> _______________________________________________ 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
