Hi,
I'm trying to analyze an experiment whereby I have 30 subjects, each with 2
runs. I have 8 or so contrasts of interest at the first-level.
As per the FSL website
(http://www.fmrib.ox.ac.uk/fsl/feat5/detail.html#MultiSessionMultiSubject) I
have analyzed each run for each subject separately at the first level, then
fed the 2 resulting feat directories for each subject into one large
fixed-effects second-level analysis.
When it comes to the third-level, higher-level group analysis, the website
suggests: "Now we want the mean group effect, across subjects, achieved with
a third-level ME analysis. Select Inputs are lower-level FEAT directories
and select the 5 relevant directories created at second-level, named
something like subject_N.gfeat/cope1.feat." However, from the second-level
analysis, I have a feat directory for each of my first-level contrasts, and
within that I have cope images for each subject. So, when it comes to
third-level I have to run a separate analysis for each contrast, with the 30
cope images from that contrast as my input. If I do it this way, I'm not
sure how I could then perform ANOVA, where I'll need to combine contrasts at
the group level?
Alternatively, I'm wondering whether at the second level, I need to run an
analysis for each contrast, so then I assume I'll end up with 30 feat
directories for each contrast? (Which I'll then feed up to a third-level
analysis).
Thanks,
Sarah.
|