Dear FSL experts,
I have a question about the registration of a very limited fov functional
dataset to standard space. I can register the dataset to standard space, but
not in a three step approach, as can be set up with the feat GUI. Instead I
use FLIRT with the following steps:
Part 1 register structural datasets to standard space.
# A) t2_mean (initial hi-res image)-> t2_sag_brain (hi-res image)-> t1_brain
(this is all intra-individual)
# B) t1_brain -> MNI152_T1_1mm_brain
# C) t2_mean -> MNI152_T1_1mm_brain (initial hi-res image -> to standard)
Part 2 apply the obtained transformation to the functional data.
# D) Resample functional data to the same resolution of the t2_mean image
(the functional and the t2_mean had the exact scan orientation, so no
additional registration is needed)
# E) Resampled_functional -> t1_brain
# F) Resampled_functional2t1_brain -> MNI152_T1_1mm_brain
My first question is, how to apply this transformation to my first level
(filtered_func) data, so that I can use it in the higher level analysis?
Which transformation matrices do I need to place where, so that the higher
level feat will be able to use it?
My second question is if step E and F can be combined by concatenating the
transformation matrices? I tried this, but ended up having a very different
result then when using E and F as seperate steps. Would there be a way (to
use convert_xfm) to get the proper transformation in one step?
Thank you,
Niels van Strien
VU university medical center
|