Hi,
I am trying to run FSL-VBM and have a couple of questions (in order of
processing stage):
1. My raw image intensities are between 2500 – 9500 (using fslstats) so I
haven’t applied any scaling factor but ran “tbss_1_preproc -f 1” to ensure
the data were in the correct format / directory structure. OK?
2. BET with the default intensity threshold (is this 0.4?) cuts a bit too
much gray matter off the frontal and occipital lobe on some subjects. I re-
ran with –f 0.3 and this looks better but some subjects have “too much”
brain included (the outline runs into the skull or eyes). Does BET have to
be perfect on everyone for FSL-VBM to work and is it better to have too
much or too little brain included if one is forced to choose?
3. I ran the “fslvbm_2_template –a” step and it has produced T.hdr and
T.img files and the following .hdr and .img files for all subjects; pve
0,1,2, brain_seg, GM, GM_xflipped but there are no template_GM.hdr or
template_4D-GM.nii.gz files. Presumably the template was not created? I
have waited several days so I guess it must have had time to run. I don’t
think there’s anything wrong with my template_list - I have listed the
names of all my subjects’ (equal number from each group) original data
files, e.g. CON01.0001.hdr – i.e. these are as they appear in the orig
data directory but not in the FAi directory with the CON01.0001_FAi.hdr
suffix. Any ideas? If I re-run this step will it start from the beginning
and over-write all the segmented .img files?
Thanks for your help.
|