Thank you for answering my message. I run these analyses in an event-
related design with four conditions, time and space derivatives and
movement regressors (14 subjects). Regarding Matt’s question, I do not
have an explicit baseline regressor since I included null events in the
model but I did not model them explicitly. I will perform the random
analysis effects to make population inferences but I did the multiple
subject fixed effects model to plot the hemodynamic response for all
subjects.
I do have different activations, T and p values between the single and the
multiple-subjects models for the same subject. I defined the contrast for
all subjects as suggested, and the display is now correct but the results
are different than those for the same contrast when is done automatically.
So it seems there are some problems.
I will send my spm.mat files of both models and my batch files to Matt and
Volkmar, Thank you for your interest and help,
Selene
|