Hello again,
> The template I used is CCHMC2 medium for age group 8-12, since my
> subjects are in the same age group.
OK (see below for comments).
> For normalisation, i used the T1
> weighted pediatric whole brain template image.
As previously discussed on the list, it is now recommended to use
unified segmentation for spatial normalization as this uses much more
parameters.
> I mentioned the bounding
> box only to indicate i have not used a different one than what was
> mentioned for the pediatric template (-90:91 -126:91 -72:109).
You should not usually have to change that option when you write
normalized images.
> I use SPM5 for analysis and it is
> an fMRI study.After normalization the images were just fine.But once the
> spm.mat is created and the results are displayed,overlaid on the
> subject's T1 it shows activations as falling outside.
The results in your SPM.mat only reflect what is in your images, so the
problem is likely to be there before estimation. And again, the issue of
smoothing predictably leads to "activation" outside of the strict
boundaries. You may also want to have a look at the mask.img created
during analysis which tells you which voxels were analyzed in the first
place. If this is far too liberal, you could play with the masking
options during estimation.
> Also when rendered
> with the .mat file which came along the template, the activations dont
> coincide with the brain image.
Rendering will only display activations falling within a couple of mm in
front and behind the surface of the render-file. If I remember correctly
I once fiddled with this setting in order to get more "realistic"
renderings, but would have to look this up. So some discrepancy between
the glass brain and the rendering must also be expected. As I said, the
best impression is usually gained by overlaying the results on the
template used for spatial normalization.
> However, I have not tried using
> the structural for normalisation.I also donot have enough subjects right
> now so as to make a template with TOM.
For the former, I suggest you try that. For the latter, that is the
beauty about TOM: you do not need enough subjects as the parameters for
generating a template (based on 400+ kids from the NIH study) are
already included (well, you need to download them separately but they
are on the same page :) So I would suggest trying these options (use the
TOM priors as priors in unified segmentation) and then check again.
Best,
Marko
--
=====================================================================
Marko Wilke (Dr.med./M.D.)
[log in to unmask]
Universitäts-Kinderklinik University Children's Hospital
Abt. III (Neuropädiatrie) Dept. III (Pediatric neurology)
Hoppe-Seyler-Str. 1, D - 72076 Tübingen
Tel.: (+49) 07071 29-83416 Fax: (+49) 07071 29-5473
=====================================================================
|