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 > _______________________________________________ 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
