1) The values in each voxel represent the number of streamlines that left
the seed mask and made it to the waypoint mask that also crossed through
each voxel.
2) The best way to get a probability in this case is to divide the image by
the waytotal (which is in the same folder as fdt_paths). This is valid if
you know there to be a connection between the seed and the waypoint. Thus,
if you know that there is a connection, then the probability of that
connection in any given voxel is the number of samples in that voxel divided
by the total number of samples that reached the waypoint mask. You are
inferring the probability of the location of the one true connection. In
practice, the connection is probably spread over several voxels (in cross
section) and thus the maximum probabilities in each voxel may seem a little
low.
Peace,
Matt.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of Michael
Sent: Tuesday, July 27, 2010 8:10 PM
To: [log in to unmask]
Subject: [FSL] Probtrackx samples and probabilities
Hello,
I've been recently running probtrackx in fsl and have a question about the
value in each voxel in the resulting tractography map. I used a seed mask in
the thalamus and a waypoint mask around the area of the central sulcus.
The masks that I drew are in MNI space at 2x2x2mm^3. I'm using 5000
samples.
Here are a few questions:
1) The value in the each voxel represent the number of samples that pass
from the thalamus to my waypoint mask, correct?
2) How does this relate to the probability of connection? Do you divide the
number of samples in each voxel by {5000 samples * #of voxels in Thalamus
seed mask (650 voxels) }. If this is the case, my highest maximum intensity
is roughly around 0.001, which isn't even %1. Is this correct? Is this
normal? Is there something that I'm missing here?
Your help would greatly be appreciated.
Thanks,
Michael
|