Hi Mark:
IMHO these are Fourier truncation (edge ringing) artifacts. These result from the relatively
small number of voxels used in your diffusion images, combined with the high SNR and
contrast visible in your low B-Value images. These are well understood and explained in a
variety of places. I have a rather old page at
http://airto.ccn.ucla.edu/BMCweb/SharedCode/MRArtifacts/MRArtifacts.html that shows
some of this. You can find lots of examples at other artifact collection web sites.
BTW: These have a very interesting history in MRI. If you are easily amused, I have an
article discussing some of this at: Cohen, MS, Baird, D. "Why Trade?"
Perspectives on Science - Volume 7, Number 2, 1999, pp. 231-254.
If you were dead set on getting rid of this, you could increase the matrix size. Just stay
away from non-square pixels in plane (see the Perspective on Science article for more
detail.)
|