Justine
Well that's a blast from the past.
I think since that nascent posting there have been many multicenter
studies so obviously many different scanners were used, and I'm sure
thinking has evolved. I suggest searching medline for these studies.
Also the BIRN website has a page on combining studies from different
sites: See https://xwiki.nbirn.org:8443/bin/view/Function-BIRN/FBIRN_Best_Practices
One important point is to make sure that your groups aren't confounded
with different scanners- i.e., if you have 3 groups they weren't
scanned each on a separate scanner, which would confound separating
group differences from scanner differences.
Darren
On Wed, Jan 12, 2011 at 3:53 PM, justine dupont
<[log in to unmask]> wrote:
> Dear SPMers
> In 2003 Darren Gitelman gave a great summary of the issues surrounding
> combining scans from different scanners
> https://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind03&L=SPM&P=R94534&I=-3&d=No+Match;Match;Matches
> If possible, I want to combine scans from 3 scanners. Darren's article
> suggests this is not generally advisable, but may be acceptable if you have
> equal number of subjects from each group for each scanner, and the results
> seem to be going in the same direction from each sample.
> This is well and good, however, I am wondering if people's thoughts on this
> had evolved at all in the 8 years since that post?
> Thank you
> Justine
>
>
--
Darren Gitelman, MD
Northwestern University
710 N. Lake Shore Dr., 1122
Chicago, IL 60611
Ph: (312) 908-8614
Fax: (312) 908-5073
|