Hi Sophie,
> good morning, I was hoping to please get some help with these questions,
it's almost evening here but your kind words shall not go without an
answer :)
> attached is an example of our results.
Uh-oh... I hate to be harsh but I do not really believe that these are
real effects. They are scattered and small and would not survive a
decent extent threshold (remember that, at 1x1x1mm resolution, you need
1000 voxels to yield a cluster of 1 ccm). Having said that, a lot of
people argue that an extent threshold is not valid in VBM. Why? Because
of the highly non-stationary smoothness, which is also the answer to
> 1) why are my P corrected values listed as NaN?
I am not wholly sure, to be honest, why they show up as NaN, but the
point is, cluster-level inference is not valid in VBM anyway, as stated
by Ashburner and Friston in their legendary "VBM the methods" in 2001
(or so). Also, people tend to move away from uncorrected thresholds at
the voxel level.
> 2) why am I getting a cluster so far outside the brain and how do I fix
> this?
As to why, I think it is probably chance. As to how to get rid of it,
use an explicit mask that only contains your gray matter (allow for some
smoothing). Also, you may want to specify an absolute threshold of 0.05
or so. Everything below that in a segmented image is probably junk anyway.
> We don't stop playing because we grow old, we grow old because we stop
> playing.
Working in pediatrics, I naturally like that quote ;)
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
=====================================================================
|