Hi list!
While the radiologists in our hospital are doing spinal fiber tracking
with Neuro3D (Siemens inhouse software), am I trying to do so with FSL.
Hereby I had several problems and I wonder if anybody has made some
similar experience.
First of all the images I have (Siemens 3T, 25 directions, 3
acquisitions) are very inhomogenous. The 3 volumes of the same
direction never look similar and in the fslview movie mode I have the
impression the spine shakes. I took the mean of the 3 aquisition for
every direction...
I nevertheless arrived at dtifit and bedpostx.bedpostX and the dyads
and f1 and so on look not too bad, the spine beeing a nice blue line
within some colour noise.
But the result of probtrackX are just ugly vegetables, growing around
my several masks. There were no fibers longer than 1cm, nothing along
the spine, but in any direction...
Even drawing one line along the spine and using this line as a mask
didn't give any good output, rather a kind of mossy cucumber.
I imagine the problem might be that eddy_correct uses flirt which uses
registration fitted to BRAINS (using a sphere etc.) I saw, that my
meaned and eddy_corrected data got some empty volumes.
Without applying eddy_correct, however, I think the direction volumes
are to heterogenous for a reasonable fiber tracking.
I thus would need a registration tool that works for non-brain scans.
Could this be?
Any suggestions are greatly appreciated!
Thanks a lot!
Markus
--
Dr. med. Markus Gschwind, M.D.
Laboratory for Neurology and Imaging of Cognition
Dept of Neurosciences
University Medical Center (CMU)
1 Michel-Servet - 1211 GENEVA - CH
Tel 0041 (0) 22 379 5324
Fax 0041 (0) 22 379 5402
email: [log in to unmask]
http://labnic.unige.ch
|