Hi!
I have asked similar questions before but now we have run the processes and just want to understand the output.
We want to use anatomical scans for classification but using raw data from different projects proved difficult since they are so different. Preforming normal normalization (e.g. warping NL warping to MNI space) would remove any interesting differences. That is why we turned to fslvbm! We want 2 data sets (used separate to compare which one classifies best and together to see if they together help each other):
The Jacobians: since they should tell us, per voxel, how much each brain structure had to be "altered/scaled" to fit the tempalte and thus would indicate differences across subjects.
The final GM output maps: Because GM maps should also track e.g. age.
Q1) The Jacobians I'm after I presume are the ones produced by fslvbm_3_proc? Here:
${FSLDIR}/bin/fsl_reg ${g}_GM template_GM ${g}_GM_to_template_GM -fnirt \"--config=GM_2_MNI152GM_2mm.cnf --jout=${g}_JAC_nl\";
Which generates this file:
${g}_JAC_nl\
So the file would be ....JAC_nl.nii.gz?
Q2) The final GM output per subject, which file is that?
I see that this part
${g}_GM_to_template_GM -mul ${g}_JAC_nl ${g}_GM_to_template_GM_mod
and
\$FSLDIR/bin/fslmerge -t GM_merg \`\${FSLDIR}/bin/imglob ../struc/*_GM_to_template_GM.*\`
\$FSLDIR/bin/fslmerge -t GM_mod_merg \`\${FSLDIR}/bin/imglob ../struc/*_GM_to_template_GM_mod.*\`
The documentation says that the final output is GM_merg and GM_merg_mod so I assume we want the individual GM_to_tempalte_GM files or GM_to_template_mod files? I got these files in the stats folder
fslvbm3b GM_mask.nii.gz GM_mod_merg_s2.nii.gz template_GM.nii.gz
fslvbm3b.e104144 GM_merg.nii.gz GM_mod_merg_s3.nii.gz
fslvbm3b.o104144 GM_mod_merg.nii.gz GM_mod_merg_s4.nii.gz
(where I assume s are for different smoothing)
Q3) The mod-files are GM_merge * Jacobians. This is supposed done as:
"a compensation (or "modulation") for the contraction/enlargement due to the non-linear component of the transformation: each voxel of each registered grey matter image is multiplied by the Jacobian..."
If this multiplication is to compensate for non-linear warping, why does the GM_to_template_GM and GM_to_template_mod perfectly overlap in fsl eyes and only differ in intensity? I would expect them to look different?
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
|