Dear experts,
I have come across a problem that I am having some difficulty resolving in the DWI dataset I am working with. I would like to have your opinion on it.
Providing a brief overview:
- My sequence consists of:
- 1 acquisition of b0 AP (7 volumes)
- 1 acquisition of b0 PA (2 volumes)
- 1 acquisition of 1 b0 AP, 32 b1000, and 32 b2000 (65 volumes)
In my case, I am only using the b0 that was acquired with the other b values for reasons related to the scanner (the separately acquired b0s had different intensities from the other acquired b0).
-
I denoised and preprocessed the data with PCA, Gibbs unringing, topup, and eddy.
-
I calculated a DKIfit, and in some of the subjects, such as the one I am sending attached, there were very high FA values (0.9999, I even found FA values > 1), so I tried to identify the origin of the problem. I went back to the raw data, where I found voxels (coinciding with those with high FA) where the intensity of the b0 volume was higher than the intensity of the b1000, which I found quite strange.
-
I discussed this ‘artifact’ in the b0 with the scanner technician, who raised the hypothesis that it could be a scanner reconstruction problem, with no clear resolution possible.
Have you ever encountered an error like this? If so, is there any possible solution/correction?
Thank you very much for all your assistance.
Best regards,
Nelson
