Hello,
I'm using diffusion-weighted data in which two sets of scans were acquired with opposite phase-encoding (right-left, and left-right). After pre-processing (topup and eddy correction) I've noticed that the brain is a slightly different shape in the b-zero weighted volumes compared to the diffusion weighted volumes. The brain appears to be wider in the b-zero volumes. This is not simply because the CSF lights up in b-zero images ; it is because the brain tissue itself is stretched out.
Is this expected?
My current solution is to find the average over all diffusion-weighted volumes, and to align the result with the high-resolution structural image, rather than using the b-zero average. This is important because brain masks on the structural image must align accurately to the diffusion-weighted data when reverse-transformed – and there would be a small misalignment if I were to use the b-zero image.
Is this a good work around? Is there anything else that I should be aware of?
Kind regards,
Jackson
|