Ruth,
I would put your data into radiological orientation, then use DTIFIT to get
the gradients in order, and then rerun BEDPOST. Your probtrack results
should work correctly then. I think Tim's point was the FSLView is not
handling neurologically oriented images correctly, so you can't trust what
you see in FSLView with neurologically oriented V1 or dyadic vectors images.
Peace,
Matt.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of Ruth Carper
Sent: Thursday, August 02, 2007 3:27 PM
To: [log in to unmask]
Subject: Re: [FSL] mismatch between DTIFit and BEDPOST results
Hi guys,
Matt's suggestion was very helpful and shows that BEDPOST and DTIFIT are
getting the same results after all (ie. V1 and dyadic_vectors look the
same). So the failure seems to be at the ProbTrack stage. Tim is correct
that I'm in neurological format, but, if I understand you correctly, I
already did fix #2. Before I ran BEDPOST and DTIFIT I changed x to -x in my
bvecs. How is it changing it back at the ProbTrack stage? That is, why
does ProbTrack give one interpretation of the data and dyadic-vectors a
different one?
I'll try putting it into Radiological space, but I'm still a little confused
as to the descrepancy.
Thanks for your help,
Ruth
|