Hello all,
We have run TBSS on our neonatal cohort using FSL version 4.1.4. Due to
the age of the subjects we have identified the "most representative" FA
image and used this as our target image. After the registration and
post-registration of our subjects and running the prestats we are facing
a dilemma of how strict we should be about excluding subject FA images
that are warped. I understand that this step has a certain level of
subjectivity involved.
When comparing each FA image (in the all_FA folder) with the overlapping
thresholded mean_FA_skeleton the areas of interest are the major white
matter tracts. When we have misalignment there, it is straight forward
that the specific FA image should be excluded from further analysis.
However when we have alignment of the major white matter tracts and the
mean_FA_skeleton but warping in the anterior-posterior regions how does
one proceed? In some cases the amount of warping is exceptionally a lot
and it makes sense to exclude those but what about the cases where we
have slightly/moderately warped images and alignment. How does one judge
in these circumstances?
Kind Regards,
Georgios Alexandrou, MD
Neonatal Research Unit,
Karolinska Institute
|