Dear DCM experts
An old message by Karl Friston to the list says:
"You can certainly use DCM for ER designs. The only caveat is that you
cannot disambiguate between a direct effect of the stimulus-bound context
(e.g. valence) and a bilinear effect on a connection. This is because the
context is not present during the absence of the stimulus. This is not
a severe problem but does oblige you to restrict the contextual effect
to a direct effect OR a bilinear effect."
I am not sure what exactly this means.
If I have a model with regions X and Y, and an input condition i:
I guess this means that I can't specify a model in which X receives a
direct input (i) while the connection from Y to X is modulated by i. (is
my understanding correct?)
1)can I specify a model in which region X receives a direct input (i) and
the connection from X to Y is modulated by i?
2) Are there any other implications for using DCM in an ER design?
thanks in advance
Tali Bitan
Northwestern University
|