Dear Spm'ers,
I'm an intern and I'm quite new in the field of fMRI and SPM.
At the moment I'm studying language activity using a letter fluency task in
a blocked design in both patients and controls. My regressors consist of 1
regressor for language and 6 regressors for confounds due to motion. I'm
interested in studying effective or functional connectivity using this
language task.
At the moment I'm looking into PPI's. I've read that if you want to use
PPI's you need to start with a factorial design. My factors are: 1 Verbal
fluency ( language vs rest) and group (patients vs healthy controls).
However as I understood you need to have two task factors to do a PPI
instead of one task factor(language) and one factor consisting of group.
As I see it two task factors could for example be constructed as 1 =
language vs rest and 2= control condition vs rest.
So in order to use the PPI routine I need to add another task (for example
control consition vs rest). Please correct me if I'm wrong.
Is there perhaps a way of circumventing adding another task or should I (in
this case) stick to simple correlations?
And would it in the case of a different method for determing connectivity
for example structural equation modelling, were you use the variance-
covariance matrix, also be necessary?
I hope my questions are not too dull!
Thanks in advance.
Kelly Diederen
|