on last thing: I've found the smoothing of the medial wall to be the toughest bit. VE devotes quite a bit of tutorial to that. notice that the lanmark issues on the shpere I attached are to do with medial wall landmarks, and their overlaps and overlap with cing-sulc (green) landmark.
I have issues because my MRI has two classes (no CSF) and the internal capsule and thalamus actually come out really well. this may be true also in your data if ex-vivo. I take the time to edit the segmentation so that the surface is less complex in the medial wall area and is easier to smooth. I carefully change the segmentation and re-run it till it has *few convexities*. otherwise medial nodes can get confused with lateral surface nodes for whatever algorithmic reason when you draw a border around medial wall. edit in sagital plane. edit often and re-make surface often. then I smooth the medial wall in the GUI (border, paint, roi ooeration, select nodes by paint, smooth nodes) trying to make it as similar to F99 as possible before I register. this issue doesn't seem to trouble others as much as me. hope helps.... On 20 March 2012 18:06, Colin Reveley <[email protected]> wrote: > Tristan - > > by exposure, I've learned: > > 1) you can use any borders you want. > > 2) It's going to work best if there's a sulcus, > > 3) and it's going to work best if the border on the source and target are > reasonably close in the first place (each pair of local landmarks, viewed > on a sphere of atlas or individual. that's important. the source and > targets landmarks can get confused otherwise, and you and up with a mess. > see images) > > 4) sometimes, you can get mileage with land marks that follow ridges, like > parts of the midline or by hippocampus, if sulci are hard to come by. > tricky, but can help. > > 5) I highly, highly recommend Vanessen's 2010 freesurfer to caret > registration turial and data set (to which of course you can add or remove > landmarks - e.g. I think you might want some more frontal landmarks.) > > 6) I've messed with those so that I have a caret-caret registration > version (i.e just got rid of freesurfer importing stuff, pretty much). It > works great. you can have them if you want. > > I don't take credit - I'm advising taking advantage of these scripts from > VE as a base for what you do. I was lost in the wilderness before..... ;) > > I can go from MRI to registered data in a day or two now. The first time > took a month. > > adapt and use the scripts > > whats really imprtant is that source and target landmarks not overlap. > > see images. red shows crossovers in arewas where the green and medial wall > landmarks are badly aligned locally. the fiducial shows consequences. > > On 20 March 2012 17:00, <[email protected]> wrote: > >> Send caret-users mailing list submissions to >> [email protected] >> >> To subscribe or unsubscribe via the World Wide Web, visit >> http://brainvis.wustl.edu/mailman/listinfo/caret-users >> or, via email, send a message with subject or body 'help' to >> [email protected] >> >> You can reach the person managing the list at >> [email protected] >> >> When replying, please edit your Subject line so it is more specific >> than "Re: Contents of caret-users digest..." >> >> Today's Topics: >> >> 1. Open vs closed borders for registration (Tristan Chaplin) >> >> >> ---------- Forwarded message ---------- >> From: Tristan Chaplin <[email protected]> >> To: "Caret, SureFit, and SuMS software users" < >> [email protected]> >> Cc: >> Date: Tue, 20 Mar 2012 14:54:13 +1100 >> Subject: [caret-users] Open vs closed borders for registration >> Hi, >> >> I just wondering, is there advantage/disadvantage to using open or closed >> borders in registration? e.g. why is the medial wall two borders and not a >> single closed border? I had heard a while back that originally borders >> could only be drawn on flat surfaces so this must have been necessary then, >> but now you can draw borders on the 3d surface. >> >> Is there something "simpler" and thus better about using open borders for >> registration? >> >> I was wondering, if you were placing a border on a cortical area with >> that has a topographic organization, e.g. V1, is it better to use two >> borders, one for the upper field boundary, one for the lower field >> boundary, to give it some indication of the topology? I figure the closed >> border has a specific start point, end point and order, so if you do it the >> same in both source and target then it should match up in the same way. >> >> I have been reading Van Essen et al. (2011) Cortical Parcellations of the >> Macaque Monkey Analyzed on Surface-Based Atlases, but I can't seem to work >> it out from based on the description of the algorithm there. It seems like >> it would make no difference if closed or open borders are used. >> >> Any insight or practical experience would be greatly appreciated. >> >> Cheers, >> Tristan >> >> _______________________________________________ >> 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
