Hello all,
Is it possible to perform a repeated measures analysis in randomise?
For example, with 2 cohorts and a variable number of return visits per participant (anywhere from 0 to 4).
If not, then what's the next best way to perform such an analysis?
Let's say there are 10 participants with bipolar disorder and 15 matched controls. Each participant has a variable number of followup visits. We'd like to know the effect of [visit] on voxelwise perfusion. One possible solution could be to add a [months since baseline] covariate to the Randomise design matrix and just ignore the specific subject effect. But, this seems like a poor solution.
The idea here is that it may be possible to implement some kind of GLM to detect the effect of time.
Thanks!
-Ilia
|