Dear Craig,
Don't do any swapping of outputs as a quick fix.
This sounds quite serious and is probably indicative of a more fundamental
problem.
Are the labels correct on all of your original images? That is, do
they correctly describe the anatomy that you can see in those images? And
please check all images, structurals and functionals.
Also, have you checked the registrations carefully?
I suspect that one of the above is the true cause of the problems and needs
to be fixed before trying anything else. It isn't possible that you have correctly
labelled input images and have a correct registration but only get a swap of
A-P labels. That just cannot happen, so please check the above and let us
know.
All the best,
Mark
On 10 Aug 2011, at 19:28, Craig Moodie wrote:
> Good afternoon experts,
>
> I'm having problems with a feat GLM Block Design analysis that I'm trying to run and I'd appreciate some advice. The higher level analyses are showing activation in areas outside the brain and, in general, don't show activation in regions that would make sense based on what was seen in the lower level copes. After looking at the results and wondering why they don't make sense, I realized that the orientations are flipped in the higher level gfeat thresh_zstat images as compared to the lower level nifti images. It seems that the A-P axes in the nifti files that I used are opposite to that of the MNI 2mm brain.
>
> Hence, my question is, do I simply go through and reverse the A-P orientation of all of the original nifti files and this will solve all my problems, or is there some other systemic error that could be occurring between the lower level registration to the MNI brain and the group level normalization? Also, is FSL Maths the best/most efficient way to ensure my alignments will be radiologically accurate?
>
> Please let me know what you think.
>
> Best,
> Craig.
>
> --
> Craig Moodie,
> PhD Student,
> Department of Neuroscience,
> University of Minnesota Medical School.
>
|