Hi all,
I have two questions about dual regression analysis of resting state data. I have a two-group repeated measures design (each group is scanned at two timepoints) and I am interested in the group x time interaction.
First, those groups have unequal sample sizes (the second group has 4 more participants than the first). Would it make sense to randomly select four people out of the second group for elimination from analysis (in my mind, making the sample sizes for each group equal would help avoid biasing the group ICA result towards one group over another)? If not, is there some way I should deal with this or is it not an issue?
Second, when I conduct the group ICA, I should conduct it on all of the runs that will be entered into the dual regression analysis (so I should enter in a run for each individual at each timepoint), right?
Thanks!
Cindy
|