Why are you using classification masks if you have only one target of
interest? Classification masks are for comparing the strength of the
connection to several targets of a seed region. If you want to generate a
path distribution of the connections between BA10 and BA40 you might as well
use BA10 as your seed and BA40 as your waypoint. Are you actually wanting
to compare the strength of connectivity of the bilateral BA10 region to left
and right BA40? If this is the case you could use two classification masks,
one for right BA40 and one for left BA40.
I am not sure if waypoint and exclusion masks save computational time or not
(i.e. are paths checked to see if they satisfy a waypoint mask or need to be
rejected by an exclusion mask while they are being generated or once they
have completed). For the exclusion mask, I could see this happening while
they were being generated, but this would be harder to do with the waypoint
mask. I'm sure Saad knows the answer. :)
To make sure you are using the right masks and the right types of mask, it
would be helpful for you to say what PATHWAYS you are trying to track and
compare.
Peace,
Matt.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of Torsten Ruest
Sent: Wednesday, June 09, 2010 7:35 AM
To: [log in to unmask]
Subject: Re: [FSL] probtrackx options
Hi there,
just for clarification:
I'd like to segment the BA10 (bilateral), ie I'd like to find those voxels
in BA10 that connect to BA40 unilateral.
I'd use the BA10 mask as seed, and BA40 unilateral as classification mask.
There's no termination or waypoint mask necessary to reach that objective,
is it? In case I would only be interested in particular tracts (that pass
through certain areas) that reach my classification mask, I would use a
waypoint mask, is this right?
In which cases would I need a termination mask (in combination with the
classification mask)?
At the moment I'd run left and right BA40 masks separately although I
realise that I could use both as classification masks. Would the output
seed_* images be the same, ie when run the classification targets separately
or together?
How could I save computational time: I suppose exclusion masks do, but do
waypoint masks as well?
Thanks very much in advance for you invaluable support!
Cheers,
Torsten
|