Dear Paul, dear All,

> We could create a pediatric reference template by warping the adult one to a
> pediatric space - would this be helpful?

I am not sure this would help since there really is no pediatric space 
as yet (the CCHMC-data is still in template space of spm2, which of 
course is only an approximation to a more appropriate solution). 
However, the main difference between the kid's and the adult's data is 
in regional and global shape and regional and global tissue composition, 
very much less in size, so I think it is reasonable to proceed with the 
common space.

> The adult EPI template could be warped into pediatric space using the
> spatial transformation parameters from warping the adult T1 to the available
> CCHMC2 children's T1 template (or study-specific T1 anatomicals).

This would allow normalizing pediatric EPI directly, but it will not be 
any more accurate in terms of localization, I fear, as tissue 
compositionand regional shape differences will still not be taken into 
account. Another reason why I never made a pediatric EPI template is 
that scanner-specific influences will be much more evident in EPI data, 
and there is the issue of field strength, and does one use an undistort 
routine, and.... In my studies, I have therefore taken the approach of 
coregistering the anatomical and the EPI and normalize based on the T1, 
dragging the EPI along.

> I'm unclear whether a pediatric template will improve spatial normalization
> or whether it will simply localize the results more accurately.

It depends on what you mean by accurate. It will improve normalization 
in so far as it will lead to a much lesser deformation of images, as, 
e.g. the pediatric parietal lobe (shedding lots of GM in the process of 
maturation) is not forced to align with an adult parietal lobe.

I believe that the "accuracy of localization" may actually decrease when 
using a pediatric template, if one defines accurate as "conforming to a 
standard space". This is because the pediatric brain certainly conforms 
less to the available adult reference brains to begin with; MNI, the 
version within SPM is based on about 150 about 127-year-olds, but 
certainly even more discrepancy must be expected between a pediatric 
brain and a single, about 60-year-old lady (i.e., Talairach-space). I 
therefore try to refrain from using Talairach coordinates when reporting 
pediatric data (based on pediatric reference data). I agree it makes 
comparing results more difficult; then again, comparing apples with 
oranges (or pears) is difficult, too, so I rather prefer to be blunt in 
this regard.
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
=====================================================================