What you have done sounds like the right way to go about this. I think you
should find some consistent way to threshold your pathways before using them
as masks to measure FA or MD (see previous discussions in the list on this
question). Failing that, you should probably binarize your masks before
using them for measurements (fslmaths -bin).
Peace,
Matt.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of James Sheehan
Sent: Wednesday, June 25, 2008 11:01 AM
To: [log in to unmask]
Subject: [FSL] Tractography Masking & FA, MD Thresholding.
So, I apologize for this often asked and answered question. I have found
quite a few responses for find FA values of a tracts using avwstats, but
have not found responses for fslstats. So, I thought it might be worth
addressing again, at least I hope.
I ran ProbtrackX using a str2dif transform. As I understand it I need to use
flirt to transform the tract in to diffusion space and create a mask. I used
the code:
# flirt -in <fdt path> -ref <nodif image> -applyxfm -init
<xfms/str2diff.mat> -datatype float -out <mask name>
which I modified from the FAQ about changing a mask from one space to
another. I then viewed the mask in FSLview, which looked good.
I then used fslstats from the terminal to pull mean FA and MD information of
non zero voxels about the tract using the code:
#fslstats <ref dti_FA or dti_MD> -k <mask image> -M
I would like to know if this is the correct way to do this.
To follow up with a second question, I had a few low intensity tracts
(values of 1 to 4) that have masks with low intensity values of .3. If I use
the Mean FA code it returns: ERROR:: Empty mask image 0.000000. If I
threshold the lower value to .01 it returns an FA of .24, however when I use
the threshold on other tracts they all return values close to .24. So, I am
wondering if I am placing my threshold option in the correct place in the
code:
#fslstats <ref dti_FA or dti_MD> -k <mask image> -l .01 -M
Any help or insight would be greatly appreciated!!!!!
Thank you,
-James
|