Hi Mike
we expect the cortex to have the topology of a sheet, and pathologies that change that will cause unpredictable results. If you can segment the pathology you would typically want to fill it in in the wm.mgz. As for the deep segmentation, again it will be a case-by-case basis. The nonlineear warp used in the segmentation is extremely flexible so it should be able to handle relatively large deformations, but at some point of course it will fail.
cheers Bruce
On Tue, 1 Jan 2019, Michael Sughrue wrote:
External Email - Use Caution
I see... what does the fix topology step expect from the wm.mgz meaning how is it testing it and what are the values it would accept... for example if we were to script an auto fill step which calls a program which works on wm.mgz between steps what would it have to do to make this work With the second point about the location mattering in gliomas...I assume this has an impact on areas like the cingulate ....and hippocampus where the hemispheres and brainstem are separated ....if so can this be addressed....also are there other tricky areas....does the deep segmentation work if the brain is shifted ....and the basal ganglia are not near their atlas positions.....
Thanks in advance
Mike
On Tue, 1 Jan. 2019, 3:18 am Bruce Fischl <fischl@nmr.mgh.harvard.edu wrote: I see. Getting FS working is going to be a case-by-case thing then. For holes extending into the ventricles you will probably need to fill the hole in the wm.mgz to make the topology closer to what we expect it to be. For the gliomas it depends on where they are cheers Bruce On Tue, 1 Jan 2019, Michael Sughrue wrote:
> > External Email - Use Caution > > In some cases this is a hole in the cortex extending to the ventricle...in others it 8s a low grade > glioma which substantially enlarges ....and distorts the cortex > > On Tue, 1 Jan. 2019, 2:54 am Bruce Fischl <fischl@nmr.mgh.harvard.edu wrote: > Hi Matt > > yes, we have done this a number of times but of course it depends a lot > on the details of the abnormality so it is hard to give generic advice > > cheers > Bruce > > On Fri, 28 Dec 2018, Michael Sughrue > wrote: > > > > > External Email - Use Caution > > > > Has anyone had success extracting surfaces of brains with obviously abnormal > > surfaces....post-surgical or very distorted cortices from tumor etc....even if the > extraction is > > imperfect, is it possible to do with freesurfer and get something imperfect but > useable. > > ....typically this gets hung up at the fix tesselation step..... > > if this isnt possible presently, any suggestions about how to create modify the source > code to > > create such a tool? > > > > On Fri, Dec 28, 2018 at 8:49 PM Matthieu Vanhoutte <matthieuvanhoutte@gmail.com> > wrote: > > External Email - Use Caution > > > > Dear Freesurfer’s experts, > > > > I used mri_coreg command to coregister PET image onto T1 volume. Basically, this > command > > coregister PET image onto orig.mgz volume (default target). However, I wonder if > T1.mgz > > wouldn’t be a better target since it has been corrected for inhomogeneities and > > intensity normalized ? > > > > Thanks for helping ! > > > > Best, > > Matt > > > > _______________________________________________ > > Freesurfer mailing list > > Freesurfer@nmr.mgh.harvard.edu > > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer > > > > > >_______________________________________________ > Freesurfer mailing list > Freesurfer@nmr.mgh.harvard.edu > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer > > >_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer