You can do it by total number of samples sent out (#samples X #voxels in
mask). This should result in a normalization that is sensitive to
differences in strength of the pathways within a subject independent of the
number of samples sent from the mask. Just keep in mind that if your mask
is considerably larger than the pathway you are tracking from it you will
overpenalize that pathway.
Peace,
Matt.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of Karl Doron
Sent: Thursday, May 28, 2009 12:53 PM
To: [log in to unmask]
Subject: [FSL] tractography: normalize for seed volumes
Hello,
I'm running a tractography study with ProbtrackX in which I'm testing
multiple cortical seed masks. I expect these to converge on a single
subcortical target, but I want to see the spatial arrangement in cortex and
the relative connectivity strengths.
I'm trying to come up with the best way to scale all the results for seed
mask volume differences so that I can make relative comparisons. I used at
waypoints mask and read another thread about using the waytotal data, but
this still seems to overcorrect.
Thanks for any feedback.
-karl
|