Thomas,
Thank you so much for this input. I tried out both the striatum and the general MNI152 template. For the striatum the match is *SO* much better. The MNI152 template also looks alright - though it's a little fuzzy.
I'm pretty happy with this now and I'll definitely have more confidence in any striatum-based connectivity differences I see now.
Thanks again.
Paul
On Thu, Jun 19, 2014 at 11:07 AM, Thomas Yeo thomas.yeo@nus.edu.sg wrote:
Sorry for the slow reply. Here's a (non-ideal) suggestion:
- Assuming you are quite happy with the freesurfer striatal parcellation
in your AD subjects, then I am assuming freesurfer nonlinear registration (talairach.m3z) is working quite well. Talairach.m3z warps your subject to an internal freesurfer space (kinda like MNI305, but not quite). Let's say the freesurfer recon-all output is at <something>/AD_SUBJECT_FS/
- Run the MNI152 1mm template (the one from FSL) through recon-all.
Recon-all will give you a Talairach.m3z that allows you to map the MNI152 1mm template to the internal freesurfer space. Let's say the freesurfer recon-all output is at <something>/MNI152_FS/
- Then do the following:
a) Use mri_vol2vol to upsample the Choi striatal atlas which is 2mm resolution to the 1mm MNI152 template:
mri_vol2vol --mov Choi_atlas.nii.gz --targ MNI152/mri/norm.mgz
--regheader --o Choi_atlas1mm.nii.gz --no-save-reg --interp nearest
Notice that I use norm.mgz of the MNI template rather than the original MNI template? norm.mgz is the 256 x 256 x 256 conformed version of the MNI template that recon-all puts through.
b) warp the Choi_atlas1mm.nii.gz to freesurfer nonlinear volumetric space:
setenv SUBJECTS_DIR <something> mri_vol2vol_used --mov Choi_atlas1mm.nii.gz --s MNI152_FS --targ
$FREESURFER_HOME/average/mni305.cor.mgz --m3z talairach.m3z --o Choi_atlas_freesurfer_internal_space.nii.gz --interp nearest
c) warp the Choi_atlas_freesurfer_internal_space.nii.gz to your subject:
setenv SUBJECTS_DIR <something> mri_vol2vol --mov $SUBJECTS_DIR/AD_SUBJECT_FS/mri/norm.mgz --s
AD_SUBJECT_FS --targ Choi_atlas_freesurfer_internal_space.nii.gz --m3z talairach.m3z --o Choi_atlas_AD_subject.nii.gz --interp nearest --inv-morph
This is not optimal because of the double interpolation. You might want to use the MNI template instead of the Choi_atlas to test the above, so you can check the goodness of the warp. The final warped MNI template should hopefully look identical to your AD subject. If that works, then use the Choi_atlas. Note that mri_vol2vol does not work properly for talairach.m3z below version 5, so you should use version 5x mri_vol2vol.
--Thomas
On Wed, Jun 18, 2014 at 10:04 PM, Harms, Michael mharms@wustl.edu wrote:
Hi, The recon-all based striatal parcellations are based on the anatomy of each particular subject, guided by a probabilistic atlas. That is inherently likely to be more accurate that just taking a set of ROIs/parcellations defined in some (non-probabilistic) atlas and warping them to each subject via an affine transformation, which is what it sounds like you are doing with the "Choi ROIs".
cheers, -MH
-- Michael Harms, Ph.D.
Conte Center for the Neuroscience of Mental Disorders Washington University School of Medicine Department of Psychiatry, Box 8134 660 South Euclid Ave. Tel: 314-747-6173 St. Louis, MO 63110 Email: mharms@wustl.edu
From: Paul Beach pabeach1@gmail.com Reply-To: Freesurfer support list freesurfer@nmr.mgh.harvard.edu Date: Wednesday, June 18, 2014 8:52 AM To: Freesurfer support list freesurfer@nmr.mgh.harvard.edu Cc: Thomas Yeo thomas.yeo@nus.edu.sg, "Yeo, Boon Thye Thomas -- Boon Thye Thomas Yeo" ythomas@csail.mit.edu Subject: Re: [Freesurfer] Fwd: Improving translation of Choi striatal ROIs to original domain
Hi Bruce,
Thanks for the response.
I actually got this pipeline partly from Thomas a few months back. However, I wasn't sure if anyone had suggestions perhaps for a recon-all based command for going from the MNI152 1mm template-based fit to individual subjects. The recon-all/Freesurfer inherent striatal parcellations are so well fitted to even my severe AD patients, so I was hoping I could somehow adapt this to the Choi ROIs.
On Wed, Jun 18, 2014 at 8:45 AM, Bruce Fischl <fischl@nmr.mgh.harvard.edu
wrote:
Hi Paul
Thomas Yeo (ccd) would be the best person to help you, but he may not be reading email for a while....
cheers Bruce
On Wed, 18 Jun 2014, Paul Beach wrote:
Hi Freesurfers,
My processing stream involves moving the parcellated functional networks of Yeo and Choi to original subject domain to do connectivity analyses. While my process works very well for the Yeo networks I'm rather unsatisfied by the results of the Choi translations.
I'm hoping someone has some suggestions for improving things so that the Choi ROIs map onto individual subjects nearly as well as the general Freesurfer striatal segmentations.
NB - I work with AD patients, so I'm sure part of the problem is atrophy-based. However, I'm sure there are ways to improve things...
My current pipeline involves two steps: mri_vol2vol \ --mov Choi2012_17Networks_MNI152_FreeSurferConformed1mm_TightMask.nii.gz \ --targ $FSLDIR/data/standard/MNI152_T1_2mm_brain.nii.gz --regheader \ --o FSL_choi_17Net_MNI152_tight_parcellation.nii.gz --no-save-reg --interp nearest
mri_label2vol \ --seg FSL_choi_17Net_MNI152_tight_parcellation.nii.gz \ --reg $SUBJECTS_DIR/{$subj}/mri/transforms/reg.mni152.2mm.dat \ --invertmtx \ --o Choi_17Network_tight_striatum_orig.nii.gz \ --temp $SUBJECTS_DIR/{$subj}/mri/orig.mgz
I suspect one way to improve things would be to do a recon-all based procedure, but I have no clue what commands within the recon-all domain that would involve.
Thanks for your suggestions.
Paul Beach DO/PhD candidate - Year VI Michigan State University
- College of Osteopathic Medicine
- Neuroscience Program - MSU Cognitive and Geriatric Neurology Team
(CoGeNT)
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
-- Paul Beach DO/PhD candidate - Year VI Michigan State University
- College of Osteopathic Medicine
- Neuroscience Program
- MSU *Co*gnitive and *Ge*riatric *N*eurology *T*eam (*CoGeNT*)
The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail.
freesurfer@nmr.mgh.harvard.edu
-
Paul Beach