hello DCMers-
The B matrix values are an 'over and above the A matrix' kind of value due to the basic DCM bilinear equation. However, is the B matrix from bayesian model averaging (i.e. BMS.DCM.rfx.bma.mEp.B) interpreted the same way or should we instead test for significant modulation of a parameter by performing a two-tailed, two-sample t-test between the A and B matrix BMA distributions, i.e. a test of a significant distributional 'shift'? thanks!
-Sam
|