Hi Paul,
On 01/16/2013 10:27 AM, Paul Beach wrote:
Hello,
I have two issues, currently.
First, I was wondering if it would be possible to create separate FSAverage brains for different subject groups. I am working with healthy seniors and various stages of Alzheimer's disease, so doing so would be very helpful from the perspective of not conflating the differences in neuroanatomy between the these groups. One idea we had was to create separate HOME directories for each group, but we wanted to know if there was an easier way. I can't find anything in the interwebs about this.
You can create your own fsaverage subject using make_average_subject. Note that when you run recon-all it registers to the standard fsaverage space. If you want to register to a different space, you will need to re-run mris_register. Once you do this, you will not be able to compare groups.
Second, I utilize some ROIs that are not specified by FS (such as secondary somatosensory cortex and anterior vs. posterior insula). My means of getting around this was to use FSAverage and tksurfer to hand-draw them, save them as '.label' files, map them to individual subjects, and export them to AFNI since much of our correlation analysis is done on that platform. However, I've noticed that when I view these hand-drawn ROIs in AFNI they only show up at the border of the gray and white matter, rather than filling the entirety of the gray matter. Note that this problem does not occur for ROIs in the look up table (like the caudal ACC). Exporting these to AFNI gives you a nice ROI that fills all the gray matter in the defined region.
Try using --proj frac 0 1 .1 with mri_label2vol. Or you can create a new annotation and then run mri_aparc2aseg. The 2nd method works a little better since the 1st method can leave some holes.
doug
_Code I'm using_: For the latter problem, what I first use is mri_label2label to map the fsaverage-based ROI onto individual subjects, for example with the left anterior insula, before using mri_label2vol: *mri_label2label --srclabel lh.aIns.label --srcsubject fsaverage --trglabel lh.aIns.label --trgsubject [subj] --regmethod surface --hemi lh*
*mri_label2vol --label ../label/lh.aIns.label --regheader aparc+aseg.mgz --o gmroi_volume.nii.gz --temp orig/001.mgz*
Then I copy the new volume file to their AFNI folder, rename it, and run my correlation analysis on them.
Note, that for those ROIs that are in the look up table (those parcellated by freesurfer) I simply use mri_label2vol before extracting them using 3dcalc (for example, the caudal anterior cingulate)... *mri_label2vol --seg aparc+aseg.mgz --regheader aparc+aseg.mgz --o gmroi_volume.nii.gz --temp orig/001.mgz*
*3dcalc -a gmroi_volume.nii.gz -expr 'equals(a,1002)' -prefix ROI_FS_lh.caudalACC.nii.gz*
tl:dr Is it possible to "easily" create multiple FSAverages based on selected subjects (in my case, subject groups separated by disease state)?
and
Does anyone have any advice on how to get hand-drawn surface-based ROIs from Freesurfer's tksurfer into AFNI that full cover the entirety of the gray matter?
Cheers and thanks!
Paul Beach
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer