| Author |
 Topic  |
|
|
cantlon
Starting Member
4 Posts |
 Posted - Oct 12 2008 : 2:03:39 PM
|
hi. I'm getting some very bad normalization outcomes in SPM2 and I'm not sure why. The normalization stretching and warping is pulling visual cortex on the subjects' brains into the cerebellum on the normalized brain.
The subjects' original anatomy images look like they have a little bit of cerebellum cut off and I was wondering if this is why or if there is something else about SPM2 normalization that is causing the problem...? If so, is there an easy way to fix this?
jessica
Here is a screen shot of the original anatomy for one subject:
And here is the normalized anatomy from that subject (top) and the normalized canonical brain (bottom):
|
|
|
dvsmith
Advanced Member
USA
218 Posts |
Posted - Oct 12 2008 : 3:07:04 PM
|
Hi Jessica,
I'm not really an SPM user, but your normalization is probably being thrown off by the missing cerebellum. I think this situation is analogous to having a lesion somewhere. Since the intensities in and around the lesion are so different from the template, the normalization functions do not work very well. You basically just need to mask those areas out so they don't distort the normalization.
For details, take a look at the following:
http://imaging.mrc-cbu.cam.ac.uk/imaging/MaskedEpiNormalization (maybe this could be generalized into a fix for you?)
Brett et al (2001). Spatial Normalization of Brain Images with Focal Lesions Using Cost Function Masking. NeuroImage.
Rorden & Brett (2001). Stereotaxic display of brain lesions. Behavioural Neurology.
There has to be an easy way to implement these fixes in SPM. Maybe someone else can chime in on that?
Also, and this is a stretch, but you might also try looking at the origin of your image to make sure it's actually on the AC. Moving the origin to the AC should help if it's not already there (for whatever reason).
Hope this helps,
David |
 |
|
|
cantlon
Starting Member
4 Posts |
Posted - Oct 13 2008 : 07:07:38 AM
|
thanks david...I saw the lesion thing but, like you alluded to, it seems like potentially a lot of work since all 40 of my subjects have a chopped cerebellum.
am i the only on with anatomy like this? if not then are other people able to deal with this better with a different normalization routine or in a package other than SPM2?
or, is it possible that the problem isn't due to the anaomy but some weird normalization parameter?
thanks again! |
|
|
|
josh.bizzell
BIAC Staff
USA
118 Posts |
Posted - Oct 13 2008 : 09:31:24 AM
|
Jessica,
Like David said, I'm pretty sure it's due to the missing cerebellum in the subject's anatomical. Any chance you have a full-brain, higher-resolution T1 for each subject (I'm assuming the T1 you showed is coplanar with the functionals of your study)? If so, you could do a rigid registration of the coplanar T1 to the full-brain T1, normalize the full-brain T1, and apply those normalization parameters to the coplanar.
If not, you might be able to use a "template weighting image" or just a different template image (one without a cerebellum).
-Josh |
|
|
| |
Topic |
|
bad normalization
