External Email - Use Caution
Hello FreeSurfers, We use a custom surface template, generated according to the instructions in the SurfaceRegAndTemplates https://surfer.nmr.mgh.harvard.edu/fswiki/SurfaceRegAndTemplates instructions for inter-subject co-registration. We then do cortical labeling and extract surface and volume stats. In the past, I would re-run mris_register, then jump to labeling using the resulting ?h.sphere.reg (using either a custom atlas or the labeled template surface), and then extract stats. However, I now see that there are steps in -autorecon3 stage of the recon-all workflow, including the pial surface generation step, that are done after the -surfreg step, but prior to stat extraction, that depend on ?h.sphere.reg. So it seems like I should be running the essentially the entire -autorecon3 stage (minus -sphere and the default surface labeling and stat extraction steps) with the custom surface template as input. Is that the case? If so, is there an "expert flag" to run these steps of -autorecon3 with a custom surface template (analogous to the -custom-tal-atlas flag for early stages?)
Thanks, Jason
Hi Jason
we use the parcellation to figure out where we don't need to deform the pial surface. I expect our standard parcellation is good enough for this, so maybe you can just run your own posthoc?
cheers Bruce
On Wed, 19 Dec 2018, Jason Tourville wrote:
External Email - Use Caution
Hello FreeSurfers,We use a custom surface template, generated according to the instructions in the SurfaceRegAndTemplates instructions for inter-subject co-registration. We then do cortical labeling and extract surface and volume stats. In the past, I would re-run mris_register, then jump to labeling using the resulting ?h.sphere.reg (using either a custom atlas or the labeled template surface), and then extract stats. However, I now see that there are steps in -autorecon3 stage of the recon-all workflow, including the pial surface generation step, that are done after the -surfreg step, but prior to stat extraction, that depend on ?h.sphere.reg. So it seems like I should be running the essentially the entire -autorecon3 stage (minus -sphere and the default surface labeling and stat extraction steps) with the custom surface template as input. Is that the case? If so, is there an "expert flag" to run these steps of -autorecon3 with a custom surface template (analogous to the -custom-tal-atlas flag for early stages?)
Thanks, Jason
-- Jason A. Tourville, Ph.D. Research Assistant Professor Department of Speech, Language, and Hearing Sciences Boston University 677 Beacon St. Boston, MA 02215 Phone: (617)353-9484 Fax: (617)353-7755
External Email - Use Caution
Thanks Bruce. I was considering playing around with doing it both ways...seeing if the parcellation based on our template makes a difference, but sounds like it would be small. If I do and find any appreciable difference, will let you know.
Always appreciate the speedy replies.
Best, Jason
On Wed, Dec 19, 2018 at 2:49 PM Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
Hi Jason
we use the parcellation to figure out where we don't need to deform the pial surface. I expect our standard parcellation is good enough for this, so maybe you can just run your own posthoc?
cheers Bruce
On Wed, 19 Dec 2018, Jason Tourville wrote:
External Email - Use CautionHello FreeSurfers,We use a custom surface template, generated according
to the instructions in the
SurfaceRegAndTemplates instructions for inter-subject co-registration.
We then do cortical labeling
and extract surface and volume stats. In the past, I would re-run
mris_register, then jump to
labeling using the resulting ?h.sphere.reg (using either a custom atlas
or the labeled template
surface), and then extract stats. However, I now see that there are
steps in -autorecon3 stage of
the recon-all workflow, including the pial surface generation step, that
are done after the -surfreg
step, but prior to stat extraction, that depend on ?h.sphere.reg. So it
seems like I should be
running the essentially the entire -autorecon3 stage (minus -sphere and
the default surface labeling
and stat extraction steps) with the custom surface template as input. Is
that the case? If so, is
there an "expert flag" to run these steps of -autorecon3 with a custom
surface template (analogous
to the -custom-tal-atlas flag for early stages?)
Thanks, Jason
-- Jason A. Tourville, Ph.D. Research Assistant Professor Department of Speech, Language, and Hearing Sciences Boston University 677 Beacon St. Boston, MA 02215 Phone: (617)353-9484 Fax: (617)353-7755
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
sounds good. Bruce
On Thu, 20 Dec 2018, Jason Tourville wrote:
External Email - Use Caution
Thanks Bruce. I was considering playing around with doing it both ways...seeing if the parcellation based on our template makes a difference, but sounds like it would be small. If I do and find any appreciable difference, will let you know. Always appreciate the speedy replies.
Best, Jason
On Wed, Dec 19, 2018 at 2:49 PM Bruce Fischl fischl@nmr.mgh.harvard.edu wrote: Hi Jason
we use the parcellation to figure out where we don't need to deform the pial surface. I expect our standard parcellation is good enough for this, so maybe you can just run your own posthoc? cheers Bruce On Wed, 19 Dec 2018, Jason Tourville wrote: > > External Email - Use Caution > > Hello FreeSurfers,We use a custom surface template, generated according to the instructions in the > SurfaceRegAndTemplates instructions for inter-subject co-registration. We then do cortical labeling > and extract surface and volume stats. In the past, I would re-run mris_register, then jump to > labeling using the resulting ?h.sphere.reg (using either a custom atlas or the labeled template > surface), and then extract stats. However, I now see that there are steps in -autorecon3 stage of > the recon-all workflow, including the pial surface generation step, that are done after the -surfreg > step, but prior to stat extraction, that depend on ?h.sphere.reg. So it seems like I should be > running the essentially the entire -autorecon3 stage (minus -sphere and the default surface labeling > and stat extraction steps) with the custom surface template as input. Is that the case? If so, is > there an "expert flag" to run these steps of -autorecon3 with a custom surface template (analogous > to the -custom-tal-atlas flag for early stages?) > > Thanks, > Jason > > -- > Jason A. Tourville, Ph.D. > Research Assistant Professor > Department of Speech, Language, > and Hearing Sciences > Boston University > 677 Beacon St. > Boston, MA 02215 > Phone: (617)353-9484 > Fax: (617)353-7755 > >_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer-- Jason A. Tourville, Ph.D. Research Assistant Professor Department of Speech, Language, and Hearing Sciences Boston University 677 Beacon St. Boston, MA 02215 Phone: (617)353-9484 Fax: (617)353-7755
freesurfer@nmr.mgh.harvard.edu
-
Bruce Fischl -
Jason Tourville