Dear fellow SPMers,
I think it is useful from time-to-time to remind the list of the
issues with 'within session' scaling, or proportional scaling (that pesky
'scale' button) in SPM2. I continue to feel that this issue is the single
biggest problem in image analyses today. I continue to get papers and grants
to review in which the users blindly use this 'feature'.
I encourage those who wish to do use scaling to read the below
papers (better now then when your grant is being reviewed).
Note that this is different than the between session or the grand
mean scaling in SPM2 - which is not prone to the issues outlined in the
papers below. Best regards, Kent
Aguirre, G. K., Zarahn, E., & D'Esposito, M. (1998). The inferential impact
of global signal covariates in functional neuroimaging analyses. Neuroimage,
8(3), 302-306.
Andersson, J. L. (1997). How to estimate global activity independent of
changes in local activity. Neuroimage, 6(4), 237-244.
Andersson, J. L., Ashburner, J., & Friston, K. (2001). A global estimator
unbiased by local changes. Neuroimage, 13(6 Pt 1), 1193-1206.
Desjardins, A. E., Kiehl, K. A., & Liddle, P. F. (2001). Removal of
confounding effects of global signal in functional magnetic resonance
imaging analyses. Neuroimage, 13, 751-758.
------------------------------------------------------
Kent A. Kiehl, Ph.D., Director
Clinical Cognitive Neuroscience Laboratory
Olin Neuropsychiatry Research Center, Institute of Living
Assistant Clinical Professor
Department of Psychiatry
Yale University School of Medicine
Mailing address and contact information:
Institute of Living
200 Retreat Avenue
Olin Neuropsychiatry Research Center
Hartford, CT USA 06106
Tel: 860-545-7385; Fax: 860-545-7797
Email [log in to unmask]
http://www.nrc-iol.org/ccnlab
-----Original Message-----
From: Stephen J. Fromm [mailto:[log in to unmask]]
Sent: Sunday, March 06, 2005 7:12 AM
Subject: Re: Global scaling and grand mean
On Sat, 5 Mar 2005 17:06:19 +0000, Joe Maisog <[log in to unmask]> wrote:
>> A few years ago it was common for people to do global scaling
>> in SPM fMRI analyses. Nowadays the advice is that, in most
>> situations, it shouldn't be done.
>
>Does anybody out there still routinely perform global scaling on their
fMRI
>data? I am curious to see what the consensus is.
>
>Thanks very much!
>
>Joe
I don't know what individual groups do, but the consensus among SPM folk
seems to be not to do it.
One notable exception is Keith Worsley.
For people I know, the debate assumes that global scaling shouldn't be
done, and the only question is whether to use grand mean scaling only, or
to do "voxelwise scaling" so that effect sizes are in percent signal
change.
|