Hello Dr. Greve, I used the bbregister tool to register a BOLD EPI volume to the surface with the --init-fsl option. The result looks good in tkregister2. Using the resulting .dat file, I am trying to express the vertex coordinates of the surface in EPI voxel space (EPI volume: 128 x 128 x 36 matrix, limits: -64 to 64, -64 to 64 and -18 to 18).
Given the following 4x4 matrices, vox2ras_epi : vox2ras transformation matrix from the EPI volume header bbregMat : transformation matrix from the .dat file created from bbregister command
Given the vertices of the subject's surface (nVert x 4) : srfVert
I used the following transformation to express the surface in EPI voxel coordinates: srf_epiVox = srfVert * inv(bbregMat) * inv(vox2ras_epi)
When I view the results of this transformation in MATLAB, I can see that the above expression is incorrect. I used imagesc to show one slice of EPI and scatter to plot of all the vertices in that slice.
Can you please let me know what the mistake is? Am I applying the transformation in an incorrect coordinate space? Am I missing any transformations?
Looking forward for your reply. Thank you.
Yours Sincerely, Praneeth
freesurfer@nmr.mgh.harvard.edu