So I've gotten some datasets run through probtrackx and now I'm working
on getting the data from them. What I'd like to do is get the average
FA of fdt_paths. Way back when, this is what I was told to do (for
probtrack). Now that it's a newer version, is this still the best way
to do this?
1) Copy dti_FA to probtrackx directory and move it to structural space
using ApplyXFM_gui and diff2str.mat (fdt_paths is in structural space)
2) Create a binary mask of fdt_paths, apply a threshold (I was advised
by the list to use 50 when I was running probtrack a while ago), and
multiply by FA image
fslmaths fdt_paths -thr 50 -bin -mul dti_FA_realigned FAmask50
dti_realigned = DTI in structural space
FAmask50 = output
3) Calculate average FA of mask
fslstats FAmask50 -M
Also, would it make sense to limit voxels to FA values above a certain
value to account for boundary voxels that might have lower FA due to
white/gray matter overlap? If so, what value would you choose and what
rationale would you use (any refs)? Thanks!
Kristen
P.S. Thanks for your help on the probtrackx issues I was having. When I
reran bedpostx on the subject in question, probtrackx no longer gave me
erros and fdt_paths looked as expected.
|