Hi,
The most likely problem is that something went wrong in step 5.
What happens if you apply the result of step 5 to the original B0
image? Does it end up well registered to the T1? I assume that
you've checked the results of step 3 and 4 carefully and that these
are fine.
I'm afraid I don't quite follow your last paragraph so can't help
you there without more information.
All the best,
Mark
Philip Koester wrote:
> Hello,
>
> I am pretty new to FSL and want to preprocess structural (t2/t1) and DTI
> data for TBSS analysis. As I don`t know where and why the problem occurs I`d
> like to shortly describe the steps I have taken so far:
>
> 1.eddy corrected DTI data
> 2.Cut out the first (unweighted B0) image
> 3.Register this B0 to T2
> 4.Register T2 to T1
> 5.assemble the resulting matrices in concatxfm
> 6.Apply Flirt Transformation with the result of „5.“, all 11 (1B0 + 10
> B1000) DTI images as input and the B0 as reference.
>
> The resulting images are blurred and incorrectly oriented. In step 3. and 4.
> I already used the option „images incorrectly oriented“ for Flirt, otherwise
> the result of the B0 to T2 registration was also incorrectly oriented. The
> resulting images were correct, the problem in step 6 nevertheless occurred.
> (Through trial and error I already noticed that this problem does not occur
> if I make use of the „Identity Transformation“ in Flirt, therefore I
> consider, that the matrices might still be based on the incorrectly oriented
> image.)
> What is also striking is that the structural images were shown in horizontal
> slices in the top left of FSLView whereas the diffusion data showed the
> sagittal slice in the top left. The orientation of the initial structural
> and DTI images was correct, whereas the orientation was not correct on the
> blurred and distorted images (result of 6.) .
>
> Do you have any suggestions?
>
> Thanks for your help,
>
> Philip
>
>
|