Hi Bernhard,
the differences between the analyses arise from the different DFs and the
different non-sphericity estimations for the one-sample t-test and the
withing-subject ANOVA.
hth,
Jan
Bernhard Staresina wrote:
> Dear SPM experts,
>
> I have a quick question on the difference between using con images as opposed to beta images for
> the second-level analysis in SPM2. Let’s say I have 4 conditions A, B, C and D, and I’m interested in
> the effect AB > CD. Using the con images, I would create the linear contrast [1 1 -1 -1] on the
> subject-specific first level, and would then take the resulting con images from all my subjects to the
> second level, calculating a 1-sample T test. Alternatively, using the beta images that result for each
> condition from the GLM (bA, bB, bC, bD), I would set up a 2nd level repeated-measures ANOVA with
> the 4-condition beta images as my repeated measures. To get at AB > CD, I would then set up the
> contrast [1 1 -1 -1]. Having done the analysis both ways, I get very similar results, but they are not
> identical. It would be great if you could help me understand what the exact difference is between
> these two approaches.
>
> Any help would be appreciated.
>
> Cheers,
> Bernhard
>
--
Jan Gläscher, Ph.D. Caltech Brain Imaging Center
+1 (626) 395-4976 (office) Caltech, Broad Center, M/C 114-96
+1 (626) 395-2000 (fax) 1200 California Blvd
[log in to unmask] Pasadena, CA 91125
|