Hi Doug,
I am sorry that I am slow to understand. This is all new to me!
To clarify:
1) I will run my SPM mask (with significant clusters) through the recon-all pipeline. Are there any specific flags that I need to include?
2) How do I map the anatomical clusters to the surface?
Thanks again for your help!
Ola
On Jun 6, 2013, at 12:39 PM, Douglas N Greve wrote:
The relationship of the fsaverage surface to the fsaverage volume is complicated, to say the least. Obviously, the temptation is to think that a volume analysis in MNI space can be mapped to the fsaverage surface just like you would do for an individual. People do this, but we think it is a bad idea and strongly recommend against it. The xyz coordinates for a vertex in the fsaverage surface are the average of the MNI coordinates for that vertex across individuals. This averaging removes all of the small gyri and sulci making the surface have much less area and making the surface be fairly sparse in the volume. This is inevitable when combining subjects using a 12 DOF transform. It simply cannot be done.
The more recent MNI template used by SPM and FSL are created with a nonlinear transform that at least creates an identifiable cortex, but this cortex is not matched to that in fsaverage very well. So if you have something in SPM or FSL MNI space, then you should run the template through recon-all to create a surface that aligns well, map your activation to this surface, then use mri_surf2surf or mri_label2label to map to fsaverage using the surface-based registration between the SPM/FSL template and fsaverage. This will do a much better job than going directly from the volume to fsaverage, but it still might not be perfect.
doug
On 06/06/2013 12:20 PM, McMains, Stephanie wrote:
Here is the original thread (to at least most of it). Maybe I am just an idiot, but I cant' figure out how to easily reply from the web to an entire thread, and pulling it out of the digests I get is a pain!!
They wanted to take an ROI from SPM and put it into free surfer to extract cortical thickness. I thought going through fsaverage was a good way to do this. But as I said, looking at the surface, I have some concerns. I am more confused by when you tell them to run the MNI template through recon-all to get a surface, because I thought fsaverage was a kind of MNI space that has already been run through?
Thanks, Stephanie
what space is your VBM map in? doug
On 06/05/2013 11:06 AM, Ozranov-Palchik, Ola wrote: Hi Doug,
Thank you for your response! I am still a bit confused about the vol2surf command. I have a volumetric mask with voxels of interest based on a VBM analysis in SPM. I would like to extract the cortical thickness measurements based on these voxels. I am not clear what specifications I need to include when using vol2surf. For example, what do I use as my registration matrix?
Thank you for your help! Ola
Ola Ozranov-Palchik M.S. Research Coordinator, Gaab Lab Laboratories of Cognitive Neuroscience Children's Hospital Boston 1 Autumn Street, Mailbox # 713 Boston, MA 02115 olga.ozra...@childrens.harvard.edu http://childrens.harvard.edu phone: (617) 334-5046 main lab: (617) 355-0400 fax: (617) 730-0518 http://www.childrenshospital.org/research/gaablab
-----Original Message----- From: freesurfer-boun...@nmr.mgh.harvard.edu http://nmr.mgh.harvard.edu [mailto:freesurfer-boun...@nmr.mgh.harvard.edu] On Behalf Of Douglas N Greve Sent: Tuesday, June 04, 2013 1:48 PM To:freesurfer@nmr.mgh.harvard.edu mailto:freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] SPM ROI to Cortical Thickness
Make sure that your SPM volume is registered to the anatomical, then use mri_vol2surf to map the volume to the surface, then mri_segstats to extract the mean of the thickness
doug
On 06/04/2013 01:45 PM, Ozranov-Palchik, Ola wrote: main lab: (617) 355-0400 Dear FreeSurfer Experts,
I was wondering if anyone can advise on the best way to use an ROI mask created in SPM (XjView) to extract cortical thickness values in freesurfer.
Thank you in advance for your help!
Ola
Ola Ozranov-Palchik M.S. Research Coordinator, Gaab Lab Laboratories of Cognitive Neuroscience
Children's Hospital Boston 1 Autumn Street, Mailbox # 713 Boston, MA 02115 olga.ozra...@childrens.harvard.edu http://childrens.harvard.edu mailto:olga.ozra...@childrens.harvard.edu phone: (617) 334-5046 main lab: (617) 355-0400 fax: (617) 730-0518 http://www.childrenshospital.org/research/gaablab
Then it goes on:
Hi Ola, this probably will not work. The tal coordinate is notoriously inaccurate in cortex. You are verylikely to totally miss the area you are looking for. A better way is to run the MNI template through recon-all, map your VBM result to the surface of that, then use mri_surf2surf to map the effect to each individual. doug
On 06/05/2013 02:58 PM, Ozranov-Palchik, Ola wrote: Hi Doug,
Thank you, again! I decided to try to use the Talairach coordinates from my analysis for ROI extraction. I created an average subject and preprocessed the data (mris_preproc). I obtained the vertex index based on the coordinates from the average subject. Any ideas on how I can now find the corresponding vertices in all my subjects?
Thanks very much! Ola
-- Douglas N. Greve, Ph.D. MGH-NMR Center gr...@nmr.mgh.harvard.edu http://nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422