Also to make things even easier-- if you can figure out the angulation (if any) for your subject... and then do the same for a couple of "uncorrupted" subjects who had the same scan parameters-- you may be able to simply copy the BVECS file from another subject in your study who had the same FOV angulation and the identical scan parameters..
On Sun, Aug 22, 2010 at 11:38 PM, Sarina Iwabuchi
<[log in to unmask]> wrote:
Hi David,
Thanks for your reply. I've gone back to the original scans and tried MRIConvert as well as mricrons dcm2nii and its always the same with the third repeat being different. But it looks the first two aren't correct either. I'm thinking its corrupt (although not sure how it becomes corrupt since all the acquisition parameters have been identical)? So I guess there's nothing I can do with it?
Cheers,
SarinaOn Wed, Aug 18, 2010 at 9:09 PM, David Flitney
<[log in to unmask]> wrote:
This does sound suspicious. If one of the bvecs files is different from the others, either it's corrupt or the acquisition for that sample was different. The bvecs file records the orientation of each separate DWI acquisition which is then used to calculate the tensor image. If the orientations for that sample were simply different one would still expect the resulting tensor calculations to be valid and so yield valid measures of FA, V1, etc. Since you don't believe the results you're getting for that sample I'd say your bvecs file is probably wrong. Any chance you can re-check it against your original scan records?
On 18 Aug 2010, at 06:21, Sarina Iwabuchi wrote:
Hi,
Hopefully someone can help me? One of my subject that I ran through dtifit came up with very strange colourations of FA and V1, which I figured was something to do with the bvec file. We have 3 repeats of the sequence, and for whatever reason, the third one has completely different values than the first two, which obviously isn't taken into account when analysing the data as it's assumed each run would have the same bvecs? So... is there any way to fix the bvecs, or is the only way to cut out the last run and analyse with 2 runs? If so, would having 2 runs for only one subject be problematic when it came to comparing results in a reliable way?
Cheers,
Sarina
--
Dave Flitney, IT Manager
Oxford Centre for Functional MRI of the Brain
--
David A Gutman, M.D. Ph.D.
Center for Comprehensive Informatics
Emory University School of Medicine