Hello, I am conducting a group analysis on brains that have been
Freesurfed (including parcelation), and then processed in SUMA
following the guidelines under " Surface-Based Cross-Subject Analysis"
in the manual.

I have created the standard-mesh surfaces for each subject using the
MapIcosahedron. Now I would like to take the information stored in the
output of each subject's Freesurfer ROI parcellation (?h.aparc.annot) so that I could use that information for statistical analyses on the surface. That is, my goal is to generate a suma-readable "dset" in which each vertex in the
standard-mesh space (the output of MapIcosahedron) would be annotated with a number that represents the ROI the vertex belongs to (grabbed from Freesurfer's parcellation).

I first generated a text file from the binary “.annot” file using FSread_annot –roi_1D
That text file marks, for each vertex, the ROI identifier that the vertex belongs to in the parcellation scheme. This is the information I would like to import to the SUMA domain, and this processes is not working as I thought.

Given that the original space has 132394 vertexes, and the
standard-mesh has 196002, I want to use the most precise
interpolation method.
I have used SurfToSurf with a command line such as this:

SurfToSurf \
-i_fs $subj_$hemi_mesh140_std.smoothwm.asc \
-i_fs $subj_$hemi.smoothwm.asc \
-output_params Data \
-data FSread.output\[1]

where Fsread.output\[1] points to the column that holds the ROI identifiers for the vertexes.

The results seem OK, in general. However, in SurfToSurf, when the “data” is mapped from one surface to the other (in this case, the “data” are integers standing for ROI identifiers) some of the values change due to the interpolation (i.e., the ROI identifiers sometimes become numbers such as 54.16 etc’ rather than INTs).

I tried overcoming this by using the “NearestNode” option in the SurfToSurf command line, (which I thought would disable the interpolation) but the outcome is exactly the same as in the above example.

Any ideas – much appreciated.


Thanks,

Oori