There were no acquisition differences nor pre-processing differences. As you suggested, I used fslhd to check the voxel dimensions and there were a few discrepancies in the header. Below, I have cut+pasted lines in the header that are different between the two groups of FA maps. Would these discrepancies affect TBSS's ability to coregister each FA map to each other (during tbss_2_reg -n)?
Thank you!
GROUP OF SUBJECTS A (n=50)
dim1 122
dim2 122
dim3 68
qto_xyz:1 2.000000 0.000000 0.000000 -120.000000
qto_xyz:2 0.000000 2.000000 0.000000 -120.000000
qto_xyz:3 0.000000 0.000000 2.000000 -66.000000
qto_xyz:4 0.000000 0.000000 0.000000 1.000000
sto_xyz:1 2.000000 0.000000 0.000000 -120.000000
sto_xyz:2 0.000000 2.000000 0.000000 -120.000000
sto_xyz:3 0.000000 0.000000 2.000000 -66.000000
sto_xyz:4 0.000000 0.000000 0.000000 1.000000
GROUP OF SUBJECTS B (n=26)
dim1 81
dim2 106
dim3 76
qto_xyz:1 2.000000 0.000000 0.000000 -80.000000
qto_xyz:2 0.000000 2.000000 0.000000 -120.000000
qto_xyz:3 0.000000 0.000000 2.000000 -60.000000
qto_xyz:4 0.000000 0.000000 0.000000 1.000000
sto_xyz:1 2.000000 0.000000 0.000000 -80.000000
sto_xyz:2 0.000000 2.000000 0.000000 -120.000000
sto_xyz:3 0.000000 0.000000 2.000000 -60.000000
sto_xyz:4 0.000000 0.000000 0.000000 1.000000
On 15 Apr 2010, at 11:41:14 +0100, David Flitney <[log in to unmask]> wrote:
Not sure but some things to try: look at the headers using fslhd to double check the voxel dimensions and extents reported by FSLview; re-check that there were no acquisition differences; has any pre-processing occurred which may explain things (e.g., co-registration if different reference images were used)?
On 15 Apr 2010, at 08:25, Mark Shen wrote:
> Hi FSL experts,
>
> I am inspecting the FA maps for 76 subjects on FSLview. When I click the overlay information for each fA map, 50 of the FA maps say voxels=122x122x68, but 26 say voxels=81x106x76. FSLview will not let the different FA maps overlay on each other because of "incompatible dimensions." However, the overlay information has the correct voxel dimensions for all FA maps as 2x2x2mm.
>
> Why the discrepancy, even though all subjects were acquired identically? More importantly, would this affect the ability to include all subjects in a TBSS analysis, specifically for the tbss_2_reg -n flag to coregister each FA map to each other?
>
> Thank you,
> Mark Shen
>
> M.I.N.D. Institute
> University of California, Davis
>
|