Hi Rogier & Wim-
I am no expert, but I have seen this behavior before. With some
software packages, negative values can be truncated at 0. If you are using
a sinc type of interpolation, there will be ringing artifacts which will
be especially noticeable near the edge of the brain, since large values
from the brain get carried into low-valued extra-brain regions. The
sinc function introduces negative values, especially near sharp
boundaries. If these are truncated by your software, then you will notice
a checkerboard pattern.
Try using a trilinear interpolation and see if the problem persists. You
can also test this by reslicing a 3D volume containing a single point in
the center and zeros elsewhere.
Terry
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Terrence R. Oakes, Ph.D.
Director of PET Research and Image Analysis
Waisman Laboratory for Brain Imaging and Behavior
Rm T-133, Waisman Center
1500 Highland Ave.
Madison, WI 53705 U.S.A.
(608)-265-6610
[log in to unmask]
http://tezpur.keck.waisman.wisc.edu/~oakes/
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
|