Dear DCM experts.
We have DCM model with five regions: A,B,C,D,E and two modulatory effects
X and Y.
Only regions A,B,C show greater activation for X > Y in the conventional
analysis; however, DCM shows differential modulatory effects (X > Y) only
for connections going in and out of region D (which is the input region).
My questions are:
1) How can we interpret the differential modulation of connections
with region D, given that this region does not show differential
activation in these conditions? If both the input and output from region D
are greater for X>Y, why is there no difference in the level of activation
there?
2) I was worried that these findings could be an artifact of
including an extra variable (Z) as a direct input into region D. Z
includes the onsets of both X and Y (as well as one other condition not
modeled in DCM) and was included in order to disambiguate the modulatory
effects from the direct input. However, when Z is included in the model,
the 1st order kernels of both X and Y are zero for all regions. Is that a
problem for DCM? Could that account for the unintuitive findings?
Thanks a lot
Tali Bitan
Northwestern University
|