Dear SPM,
a question about FFX-, RFX analysis and multiple measurements of
subjects:
We have measured some subjects more than once, but not all. How do you
treat multiple measurements of subjects in a group-analysis?
Do you include all runs in a FFX-analysis? Do you use different
contrasts for those subjects you have measured more than once, compared
to these subjects you have just scanned once (e.g. 1/2 -1/2 1/2 -1/2 1
-1 for a hypothetical ffx-analysis of 2 subjects, the first measured two
times)?
What about a RFX-analysis? To make inferences about the population, does
it make sense to take 2 or 3 con-files from these subjects you have
measured 2 or 3 times, but just 1 con-file from the subject that were
scanned just once? I guess not.
Or do you just stick all runs of one subject together?
grateful for any suggestions
andreas
|