Greetings,
We would like to perform a DCM analysis on an event-related dataset. I have
searched the archives of the SPM listserv and there are a number of postings
that allude the possibility of performing such an analysis with an
experimental design such as the one we employed, however it does not seem
that anyone has had the issue that we are having.
The estimated model seems not to explain (hardly) any of the variance of our
data (for some of our regions it is simply a zero vector). Our original
model included 8 ROIs. We obtained similar results when it was reduced to
3. We entered 0 for the event durations during the GLM portion of our
analysis. Is this correct for performing DCM on event-related data, or
should a duration be specified here?
Thank you in advance for any help that you may offer
Kevin
|