Dear Lulu,
> I finished doing the processing of TOPUP and eddy with fsl. And also reoriented the bvec file after your reply.
>
> I checked our diffusion direction by the code in fsl manual, and found that it's a a single shell high angular resolution diffusion imaging scheme (full sphere).
>
> I usually processed the diffusion data with dsi studio, so this time I also loaded the eddy_corrected_data and re-oriented diffusion gradient table into dsi studio as before. But this time I found that no matter what reconstrucion scheme I use (GQI, QBI, or even DTI), generated fibers are all nearly straight lines (in the attachment). But when I imported the raw diffusion data, the resulting fibers were right.
>
it is very hard to tell what has gone wrong here, especially since I don't know dsi studio. The output of eddy should be quite similar to the input, basically the same images corrected for some distortions and a little movement. First of all check that that is the case. Secondly, your rotated bvecs should be very similar to your unrotated bvecs so check for that next. If both of those are true and you still get massively different results I think you need to contact the ddi people.
> So now I am checking if there was some thing wrong with my processing. Most parameters I used were set as default except the acqparam.txt and the number of volumes.
>
> Our acquistion was used with three same phase encoding directions (three b0 images, R>>L, withour blipping direction I think), so I wrote the acqparam.txt as below.
I am a little confused here. If you haven't acquired data (not even b=0) with different blip-directions, how can you run topup on it? I think I need to understand that before I can help you.
Jesper
>
> 1 0 0 0.108
> 1 0 0 0.108
> 1 0 0 0.108
>
> The contents in the index.txt file are 63 ones. I wondered that if this was the correct way to write these files, and if it is enough to represent all information?
>
> Many thanks for all your reply.
>
> Best wishes,
> Lulu
>
>
>
|