As to the cause of this, there are several possibilities:
1) Subject motion could be different on one day to the next leading to
differing qualities of uncertainty in diffusion direction estimation and
different tractography results.
2) The fewer directions you acquire, the more variable your results are
likely to be, as there is more uncertainty in your estimation of the fiber
orientations.
3) If you are not correcting for EPI distortion, and the head is oriented
differently from session 1 to session 2, the EPI distortion (which is large
in the brain stem) could affect the tracked pathway differently.
Thus, this is not necessarily unexpected and not the result of something
wrong with your tractography method.
Peace,
Matt.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of Torsten Ruest
Sent: Monday, June 07, 2010 7:33 AM
To: [log in to unmask]
Subject: [FSL] Repeadability probtrackX
Hi there,
I have a question regarding probtrackX. We acquired several scans of the
same participants using the same sequence, 2 in one session on one day with
repositioning, and 2 on another also with repositioning. I convered the
DICOM into niftii with MRIconvert, then I ran bedpost / probtrackX (linear
registration, see command below) on all of them using a unilateral BA4 seed
mask obtained from the wfu_pickatlas and the medulla termination/waypoint
mask plus a midline exclusion mask. I then thresholded the resulting
fdt_paths at 50 using fslmaths and then extracted the mean FA of the
resulting corticospinal tract (see command below). Unexpectedly, I get very
different values using this technique and was wondering whether I did
something wrong.
Example of one participant:
Day1:
FA voxels volume
0.444591 461 3688
0.429658 2166 17328
Day2:
0.399608 2970 23760
0.404466 3113 24904
Thanks very much for your help.
Cheers,
Torsten
ProbtrackX command:
for i in $Bedpost_Dirs_1; do
/usr/local/FSL/fsl-4.1.5-centos4_64/bin/probtrackx --mode=seedmask -x
/home/tr83/MY_ANALYSIS/Tracts/BA4l_dil2.hdr -l -c 0.2 -S 2000
--steplength=0.5 -P 5000 --xfm=$i/std2S0.mat
--avoid=/home/tr83/MY_ANALYSIS/Tracts/midline_mask.nii
--stop=/home/tr83/MY_ANALYSIS/Tracts/Medulla_dil2.hdr --forcedir --opd -s
$i/merged -m $i/nodif_brain_mask
--dir=$i/BA4l_2_Medulla_waypoint_termination
--waypoints=/home/t/tr/tr83/MY_ANALYSIS/Tracts/medulla_dil2_waypoint.txt
done
FSL stats command to extract values:
for i in $DTI_Dirs; do
cd $i
fslmaths $i.bedpostX/BA4l_2_Medulla_waypoint_termination/fdt_paths.nii.gz
-thr 50 -bin $i.bedpostX/BA4l_2_Medulla_waypoint_termination/fdt_paths_50;
fslstats dti_FA_2_std.nii.gz -k
$i.bedpostX/BA4l_2_Medulla_waypoint_termination/fdt_paths_50.nii.gz -M -V;
done
|