Hello, I need to determine the transform between the surface coordinates between FS and the surface generated using some of the tools from JHU. I think that I have a handle on this, but just wanted to confirm a couple things.
1) First, the "voxel to ras transform" returned by 'mri_info' is the transform between a 0-based volume index DEFINED IN THE AXES OF THAT VOLUME'S ORIENTATION (as also returned by mri_info) to a "ras" space with origin at the center of the physical space spanned by that volume, correct (provided that you have not specified an alternative "ras" center as part of an 'mri_convert' command)? That is, say you have a rawavg.mgz volume created with a plain vanilla 'mri_convert' command on an Analyze volume (.img, .hdr) having the default Analyze orientation (hist.orient = 0, axial slice direction). This rawavg.mgz volume will have an 'mri_info' orientation of "LAS". And you have the corresponding orig.mgz volume with a (conformed) orientation of "LIA". THEN, the "voxel to ras transform" indicated for these volumes is NOT to an identical "ras" space, correct? (Namely, there will be a translation between the two "ras" spaces, due to the different orientations of the rawavg.mgz and orig.mgz, correct?)
2) The "SurfaceRAS" space of the vertices of the actual FS surfaces is always defined relative to the "conformed" volume with "LIA" orientation, correct?
3) Given (1) and (2), it is necessary to account for axes orientation of any non-conformed (i.e., non-LIA) volumes as part of the transformation between that volume representation and "SurfaceRAS" of the actual surfaces, since the "ras" center of the volume will shift by 1 voxel in physical space when you reverse the orientation of a given axis (given the use of a 0-based volume indexing scheme). Correct?
Hope my questions are clear. I know that these various transformations can be tricky to get just right.
thanks, Mike H.