The lh.area in fsaverage is the area of the white surface (as in an individual subject). But the white surface is not like an individual in that many of the folds have been smoothed out (because it is an average). So mri_surfcluster fixes this internally, which is why you see the discrepancy (and the first method is larger than the second). If you want them to be identical, add --no-fix-vertex-area to mri_surfcluster.
On 6/11/2024 10:49 AM, wangzhiwei3233 wrote:
External Email - Use Caution
Dear Freesurfer experts, I got the area of a cluster by two methods, and found the results were inconstant.
Method 1: During group level analysis, the individual surface was normalized to fsaverage. I exported the cluster area and label file by command /mri_surfcluster/. The command ws as followed: / / /mri_surfcluster --in ./sig.nii.gz --subject fsaverage --hemi lh --centroid --annot aparc --thmin ${thp} --sign ${sigsignTemp} --no-adjust --minarea ${tharea} --sum area${tharea}.${sigsignTemp}${thp_infilename}.cluster.summary --o area${tharea}.${sigsignTemp}${thp_infilename}.cluster.nii.gz --ocn area${tharea}.${sigsignTemp}${thp_infilename}.ocn.nii.gz --olab $Dir_Subj_fMRI/$out_dir_group/$analysis/$contrast/label_area${tharea}.${sigsignTemp}${thp_infilename}/label /
Method 2: I calculated the cluster area using the label file outputed by /mri_surfcluster/ via the matlab funciton read_curv. I did this to test the method for calculating the area of a label manually drawn by me. 1, I first got the area of each vertex by the command area = read_curv('../fsaverage/surf/lh.area'); 2, I red the IDs of vertices in the label file and added 1 3, I got area values of vertices in the label file and summed them.
I found the areas calculated by the Method2 were much smaller(except one ROI) than the ones exported by /mri_surfcluster /as shown below.
Area from Method 1： # ClusterNo Size(mm^2) NVtxs Annot 1 *833.94* 1999 postcentral 2 *578.52 * 1300 precuneus 3 *243.78 * 463 middletemporal 4 *184.72* 344 inferiorparietal 5 *111.76* 292 insula
Area from Method 2： ClusterNumber Area(mm^2) NVtxs AnotomicalRegion 1*637.39* 1999postcentral 2*480.44* 1300precuneus 3*140.37*463middletemporal 4*166.33*344inferiorparietal 5*112.76*292insula
What's the reason? Where did I do wrong in the method 2?
Thanks.
Best,
zhiwei
--
Zhiwei Wang（王志伟), Associate Investigator
International Center for Primate Brain Research
Center for Excellence in Brain and Intelligence Technology
Institute of Neuroscience,Chinese Academy of Sciences
Room 215, Building #5, 500 Qiangye Road, Shanghai, 201602,China
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer