Hi Emily, there may be a simpler way to get at what you want. You can use mri_label2vol to create a mask of the label in mni305 2mm space, something like
mri_label2vol --label yourlabel --temp $SUBJECTS_DIR/fsaverage/mri.2mm/orig.mgz --regheader $SUBJECTS_DIR/fsaverage/mri.2mm/orig.mgz --o labelmask.mgz --labelvoxvol 8
Make sure the mask looks ok tkmedit fsaverage orig.mgz -overlay labelmask.mgz -fminmax .5 1
Then run mri_segstats --i ces.nii.gz --seg labelmask.mgz --segid 1 --avgwf ces.label.dat
where ces.nii.gz is the output of isxconcat-sess for the analysis/contrast you want to extract the data from
doug
On 12/12/2013 02:04 PM, Emily Boeke wrote:
Hi Doug,
Thanks for your response.
I have another question related to functional ROIs. The method below produces ROIs based on individual subjects' activation for a contrast. I would also like to make ROIs based on averaged group activation for a contrast and extract individuals' percent signal change from that ROI. In the past I have done this by creating a label in tkmedit out of the activation and using the mri_label2label command to fit it to the individual subjects' spaces:
mri_label2label --srclabel ${SUBJECTS_DIR}/fsaverage/label/${label} --srcsubject fsaverage --trglabel ${SUBJECTS_DIR}/${subject}/label/${label} --trgsubject ${subject} --regmethod volume
Then I would run func2roi, which is no longer used in current versions of freesurfer.
When I was running this label2label command in previous analyses, I did not have the -mni305 2 flag in my mkanalysis command. In my current analysis, I do have the -mni305 2 flag. Does this change how I have to run mri_label2label? Since the individual level analysis is in mni305 space, do I not even need to run mri_label2label at all? Could I just run funcroi-config with -label and then the name of the label file instead of seg?
Thanks!
Emily
On Dec 12, 2013, at 11:36 AM, Douglas N Greve wrote:
On 12/12/2013 10:08 AM, Emily Boeke wrote:
Hi Freesurfers,
I would like to make sure I am using the funcroi commands correctly for anatomical and functional ROI analyses (using the aseg to anatomically constrain), and I have some questions about the commands.
If I want to extract the average percent signal change in the whole left amygdala for a particular contrast, is this the correct version of the commands to run?
- funcroi-config -seg aseg.mgz 18 -roi <name>.cfg -analysis <analysisname>
- funcroi-sess -roi <name>.cfg -sf sf.txt -df sesspar
- funcroi-table-sess -roi <name>.cfg -sf sf.txt -df sesspar -o <name>.txt -c <contrast> -m cespct
If I want to extract the average percent signal change from an anatomically constrained functional ROI (e.g. an ROI representing activation in a certain contrast within the lh amygdala), is this the correct version of the commands to run:
- funcroi-config -seg aseg.mgz 18 -roi <name>.cfg -analysis <analysisname> -contrast <contrast> -sign pos -thresh 1.3
- funcroi-sess -roi <name>.cfg -sf sf.txt -df sesspar
- funcroi-table-sess -roi <name>.cfg -sf sf.txt -df sesspar -o <name>.txt -a <analysisname> -c <contrast> -m cespct
My questions are (regarding the second set of commands):
-in the right hippocampus example under funcroi-config --help, the anatomical constraint is specified with --annot aseg.mgz 53, but it should be --seg aseg.mgz 53 like in my examples, correct? Otherwise you get an error.
Oops, you're right. I've made the change in our tree.
-Is it ok to use one analysis name in the funcroi-config file and another in funcroi-table-sess? Would that allow me to make an ROI from one analysis and use it to extract signal from another analysis?
Yes
-If I get a message saying some subjects had "empty ROIs" and the value in the table for them is 0, does this mean that subjects had no activation (meeting the threshold specified) in the first contrast I specified (in funcroi-config) so no ROI could be created from their activation?
Correct
-If I want to look at activation in more than one anatomical region (for example, bilateral amygdala), is there any way to do that with the -seg option?
The way to do this is to create a new segmentation for each subject that contains only the amyg, eg, mri_binarize --i aseg.mgz --match 18 54 --o amyg.mgz This creates a mask of the amygdalae in which the amygdala voxels = 1 and everything else is 0 Then specify --seg amyg.mgz 1 Use 1 here because the amygdala voxels = 1
doug
Pardon the long message. Thanks for the help!
Emily
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://gate.nmr.mgh.harvard.edu/filedrop2 www.nmr.mgh.harvard.edu/facility/filedrop/index.html Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/