there is pretty extensive documentation on the (always confusing) coordinate systems we use
- we do *not* resample into tal coords - all calculations are done in subject native space, after resampling to 256^3, 1mm iso (which we call "conforming" for no very good reason)
- The talairach.xfm file is a linear transfom that takes an RAS coordinate in the subject space into the MNI space I believe (Doug can confirm this). Not a voxel coord.
- The talairach.xfm is used for some rough positioning of things like where the brain is, but mainly is for estimated total intracranial volume (eTIV) via the Buckner method of using the determinant of the xform.
- Label files are stored by default in "tkreg ras" which is an RAS coordinate system originally used for surfaces. It's documented on our wiki, but it is in individual subject space, not an group space.
-- tksurfer reports tkreg ras.
I hope that helps. Read the wiki, Doug has put a lot of time into documenting this stuff because it is endlessly confusing unless you are really careful cheers Bruce
On Sun, 7 Apr 2013, ZhiLiangLong wrote:
Hi all: I am confused with coordinate systems in FS. I have several questions bellow. 1. when we run recon-all -s subjid -all, Is the individual native space transformed to talairach coordinate system?. when run recon-all -s subjid -qcache, does the talairach system go into MNI system (fsaverage space)? 2. what is the tranformation matrix within the file of <subjid>/mri/transforms/talairach.xfm?, what is it used for? 3. Does the coordinates within .label files under <subjid>/label/ belong to the MNI coordinate system? 4. In TkSurfer Tool, does the coordinate of "Ve
rtex RAS" belong to
MNI coordinate? Hope someone gives me suggestions. best