jcfr
commented
4 years ago Thanks for the report :pray:
To help the broader community to know about this issue and provide guidance, I suggest you summarize it on the slicer forum: https://discourse.slicer.org/
For diffusion-weighted-imaging protocols that collect DWIs with opposite phase-encoding directions (e.g. AP/PA), B0 volumes may have b-values that are slightly greater than 0 (e.g. b-value = 5 ). It appears that dtiestim is not able to recognize these volumes as B0s (since their b-values are not equal to 0), and it defaults to using the first gradient in the DWI as the B0 it users for tensor estimation and other calculations. Assuming that the first gradient in a DWI is a B0 may be incorrect since quality-control protocols could have excluded the first gradient. (The second gradient, which may also be a B0, could also have been excluded.) This issue also affects IDWI calculation and baseline averaging. It would be great if a new version of dtiestim had a flag that allowed the user to specify a threshold under which it would consider a gradient to be a B0.