Hi, Marko,
Thanks for your reply. Dan Kimberg also pointed out that there will be
more misregistration errors if using template just from control group. So it
is a balance between loss of signficance and decreas of misregistration
errors. Is there anyway to control this balance? I would stop here if this
has been already discussed before.
-----Original Message-----
From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] On
Behalf Of Marko Wilke
Sent: Friday, April 13, 2007 3:22 AM
To: [log in to unmask]
Subject: Re: [SPM] VBM, local template just from controls or from both
patients and controls?
Hi Ze Wang,
> Would you think it is better to generate local template just from
> the control group for VBM? Suppose patients have more grey matter (or
> white matter) deficiency than controls, if we use a template from both
> groups, would this reduce the salience of VBM? Since extra deformation
> will be introduced for the control group during the underling
> normalization, which however maybe much less if we use a local template
> from the control group.
This question has been addressed several times by now, with the
punchline being that it makes more sense to include ALL subjects under
study if you want or need (!) to create a custom template in the first
place (see previous discussions on the necessity to do this in spm5).
The reason is that otherwise segmentation/normalization will prefer the
group that contributed to the template which will introduce a bias into
the processing stream and hence, the results.
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
=====================================================================
|