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
