Hi Irina
> Trying to follow the suggested procedure (coregistration:estimate -> segmentation -> coregistration: reslice), I come across a segmentation problem. I.e. I get a weird GM and WM images from the segmentation. I would appreciate if you could have a look on my result:
> https://docs.google.com/fileview?id=0B212KwbJX4jvYTM0MmM5MTUtZmEzMC00Mjg1LTlhNGYtNTY3ZDMwOWE2NGY2&hl=en
> and could tell what went wrong there. I get these corrupted images only if the structural image has been previously coregistered with the functional mean. When I segment a raw structural image, the segmentation goes fine.
The fact that segmentation works ok without coregistering to the
functional suggests that it's something about the coregistering that
causes the problem; for example, if the mean functional is further
away from the template images, coregistering to the mean would result
in poorer starting estimates for segmentation.
I would try the following:
1) Coregister the structural to the mean functional (so now structural
and all functionals are lined up);
2) Use "display" to line up the structural with template space;
3) When you select "reorient images", select both the structural and
all the functional images.
When you're done, your structural and all functional images should (a)
still be aligned with each other and (b) be roughly aligned with the
MNI-space templates. (It's worth double-checking all of this using
CheckReg.)
Hopefully that will fix it—let me know if not (although in that case,
I'm at a bit of a loss...).
Best regards,
Jonathan
|