Hi! I’ve encountered another problem while using our data.
We’ve done a probabilistic tractography and got an incomplete one compared to the tractography that we’ve done in the tutorial.
Here’s a comparison between the two. (left: tutorial tractography, right: our wrong one)
Hi Sujin,
This may happen when the x-component of your gradients is inverted (see this thread). Depending on your data, you could use dwigradcheck to try to revert it. This solution worked in my data but I let the rest of the community to check whether this is a correct way to check the x, y and z bvecs:
Thanks for your reply
I tried dwigradcheck as u have shown and there was a slight improvement, but guess that wasn’t a cause.
Anyway, I really appreciate ur help.
It’s often very difficult to diagnose data quality issues if only the end result of a pipeline is shown. Showing some greater detail about the series of steps and the intermediate data may help.
One hypothesis I do have from the image on the right is that you may be using ACT with tissue information having been derived from a very poorly aligned image. This can crop pathways due to that misalignment causing the application of inappropriate tissue priors. I would suggest checking the alignment of multi-modal image information using mrview's Overlay tool, rather than flipping back and forth between images (which can be misleading).
Thanks for your suggestion! I checked the alignment and found out there was a severe problem in the process. The problem was triggered because I used mp2rage image rather than mprage image.
Thank you so much for your help
I just have to say that your reply has always helped me so much.
Suzan.