I'm not sure who told you that you do not need to use the structural in registration
but it would not have been us. We *never* recommend registration directly from
functional to standard. Hence I would strongly advise you to repeat the registration
using the structural.
If your labels are all in the anatomically correct places in the image (as seen in
FSLView) before the registration then it sounds like you have a 180 rotation
with respect to the standard. It is possible to get this to work with a full search
but it might be easier to reorient them initially. You do not need to go back to
the dicoms to do this. Simply run fslreorient2std on the structurals and the
functionals and this should do the trick.
Note that if your labels are *not* in the anatomically correct places in the
original images then this needs to be fixed before you do anything else, as
that is a serious problem. I hope that this is not the case and the you simply
have the labels, and the brain, 180 degrees different from the standard
brain, which is not a problem and easily fixed by fslreorient2std.
All the best,
On 19 Aug 2011, at 04:28, Craig Moodie wrote:
> Hi Mark,
> I re-ran my lower level feats using full search and this did not seem to solve the problem. I just compared the example_func2standard.nii.gz files from the "normal search" and "full search" folders using fslview and fslhd and they seem to be identical. However, I did not use the registration to the structural but instead just registered the images straight to the standard MNI 2mm brain since I was told the registration to structural was not necessary for what I was doing. In addition, the structural nifties have the same A-P orientation as the functionals (which is opposite to the MNI A-P orientation).
> Could omitting the registration to the structural image be the problem or could it be some other problem with the labels? I've looked at the file names and the details using fslhd and I don't see anything that is obviously wrong. Do you have any ideas as to what/where the labels could be wrong and why this would affect the registration?
> Lastly, I've been wondering if I need to just reconvert the dicoms using dcm2nii or mricron and try to get the A-P orientation to match the MNI standard.
> Please let me know what you think.
> Thanks again for all the advice.