Hello,
I am running a DCM analysis of 3 regions. For one of the regions in this model in the pfc, there are two possible clusters we can use. We constructed two identical model spaces: one with regions A1, B and C and the other with regions A2, B, and C. The BMS analysis found the same model to be optimal for both, and that the driving input was to the A1 region in the first case and the A2 region in the second case. We are particularly interested in these A regions as driving inputs to this system. One major difference between the models is that the probability of the driving input (pC) is nonsignificant for the A1 model and significant for the A2 model. I know BMS was not designed to decide which regions should be included in the model, but does this difference in significance provide evidence that region A2 is more likely than A1 to be the region that sends the experimental "driving input" through the rest of the system?
Thank you,
Ian
|