Dear All,
Gwenaëlle:
Thank you for your reply. w.r.t. to the question about AAL template. I
followed your advice and overlay the AAL template on top of
GM_mod_merg.nii and other files of interest, and they do match. An
additional check on the header information shows that the sform and
qform transform parameters are identical for both nifti files. I am
satisfied that they have one-to-one correspondence on voxel space.
As a sanity check, I looked in to the voxel values of
GM_mod_merg.nii. I always read voxel values for GM_*merg*.nii as
probability values so I was expecting its value to be [0,1]. However,
occasionally the voxel value in this file exceed 1. For my particular
case the voxel values are [0, 1.2.4). Smoothed version GM_mod_merg_s2.nii
was not strong enough to bring the voxel values back to [0,1] but
GM_mod_merg_s4.nii was.
I think the voxel values of [0,2.4) is OK because
- It is
only to be expected as non-linear transformation was performed which
means some voxel values can be bump up higher than 1.0 and,
- 2.4
is not to far away from 1.0. If I accept that non-linear transformation
will bump some voxel values higher than 1.0, 2.4 is not outside the realm of possible
values, and,
- noise contributed to the observed value.
Is my view reasonable?
Many thanks in advance and best regards
Cinly