Gang,
I've been thinking about the questions I want to ask of the data.
One hypothesis is that the HRF estimates will be different between the two groups (a patient group and a control group). A second hypothesis is that aspects of the HRF may correlate with the covariate (response time to a task). I don't know for sure that the two groups have the same covariate effect. I'm interested in looking at the overall mean of the covariate across groups, rather than in regard to each group's mean. There is a slight difference in the average covariate score between the two groups, although the standard deviation of the second group is much larger than that of the first group and the total range of covariate values in the second group includes the full range of covariate values in the first group.
Given the statements above, would it be better to set up two 3dLME analyses, one looking at each hypothesis (one analysis omitting group membership and one analysis omitting the covariate), or to set up a single analysis to try to estimate the effects of group membership and covariate value at the same time?
Thanks,
-John