 |
 |
Donald, that's fantastic. Thanks. (Except for the part about including one EV per subject. I have 22 subjects; that's just cruel.) As far as the time course thing: I'm constructing the design matrix in FEAT (or really in Glm), but I'm entering it into dual_regression, which operates on time course data (using the spatial maps from the group ICA to extract temporal components from the individual time course data, then using the individual temporal components to construct individual-level spatial maps, then combining the individual spatial maps using FEAT -- I think). I could probably generate the relevant difference images if I broke into the dual_regression script, but I'm just as happy to do it this way, since I'm not good at FSL yet and the straight Group effect isn't that interesting to me.