On 07/23/2011 01:31 PM, Colin Reveley wrote: > strictly, no tissue is missing - it's just compacted, and the remaiing > tip is angled down. > > so one strategy would be: just forget it. > > you've seen the surface I've made in prior mails (the structural scan > has always had the tip missing, only the first DWI scan was whole. Now > that the new DWI scan is missing the tip too, the surface and data I'm > owrking on are identical morphologically) > > I'm attaching two pial surfaces, made just by hitting "3d view" in fsl > view. > > one is a deskulled, upsampled F99 (I trust you will recognise that) > > the other my data, the brain YE11. > > the question: shall I bother making the very considerable effort, or > shall I just register the damned brains and be done with it and do my > science. If you are considering using the surfaces from 3dview, then please don't. These are useful views for many purposes, but not for surface based registration. Sulci are fused, defects abound, and they are probably not topologically accurate. They are not intended for this purpose. This is my take; if others disagree, by all means speak up.
I'd try the posterior olfactory sulcus span border trick I recommended earlier and register. You're not missing a temporal lobe here. (Thank goodness.) > > thanks for help. No rush on reply. Good luck with AT&T. apparently I > owe them money. this surprises me since I've not been within the > united states since december 2009. I'm sure they know best though. Don't worry. Your secret is safe with me and the other 600+ caret-users subcribers. > > Colin > > On 23 July 2011 18:00, <[email protected] > <mailto:[email protected]>> wrote: > > Today's Topics: > > 1. Re: partial data (Donna Dierker) > > > ---------- Forwarded message ---------- > From: Donna Dierker <[email protected] > <mailto:[email protected]>> > To: "Caret, SureFit, and SuMS software users" > <[email protected] > <mailto:[email protected]>> > Date: Fri, 22 Jul 2011 09:43:04 -0500 > Subject: Re: [caret-users] partial data > Surface surgery seems pretty drastic, and my hunch is it will > cause bigger problems than it will solve. > > Ditto on the remaking the surface from a truncated volume strategy. > > We're in sync in thinking a border-based solution is the best of > all evils. > > My hunch would be to find something -- not necessarily a sulcus or > gyrus, but something that could be clearly identified in both > surfaces that would delimit/contain the difference. I haven't > looked at your data, but the idea that came to mind is Drawing a > border across the posterior terminations of the medial orbital > gyrus and gyrus rectus (i.e., border is perpendicular to those > gyri, connecting them where they meet the medial wall). Something > like that. I have no illusions this could totally pen in your > problems, but this is my gut reaction to your problem. > > (I'm at home today getting AT&T U-verse installed, so I may not be > able to check email for a while. Perhaps never, I fear.) > > On Jul 22, 2011, at 7:23 AM, Colin Reveley wrote: > > > I have a data set (the same brain you have seen donna, but a new > data set) in which the far rostral tip is missing. > > > > I'm not expecting a positive answer, given my inter-mediate > level skills in caret, but: > > > > would it be possible to create a surface from this data, and > perform surface registration of F99 and associated data onto it > BUT have the registration take into account the missing tissue at > the rostral end? > > > > i.e., it knows that the tip of the brain being registered to is > missing, but is present in the brain being registered. > > > > basically, regions are missing from my data. Could surface > registration take that into account, and ignore that tissue? > > > > maybe the only way is to convert all F99 data to volumes, cut > the tissue off the F99 mri, remake the F99 surface, put the volume > data back on the surface? > > > > But I'm hoping for a border-based solution. i.e. cut the tip off > F99 in the surface domain. > > > > It's unfortunate that the tip is missing. It is due to > compaction due to the brain sitting in the fridge for maybe 6 months. > > > > There are two reasons to use this data over previous data from > the brain which was taken when it was fresher. mainly we have a > new DWI set with more gradients, and that data is really good > despite the age of the brain. secondly, the new data has a > structural scan that's perfectly registered to the DWI, which was > not true before. > > > > any help appreciated. > > > > Colin > _______________________________________________ caret-users mailing list [email protected] http://brainvis.wustl.edu/mailman/listinfo/caret-users
