Skip to main content
. 2017 Jul 12;37(28):6751–6760. doi: 10.1523/JNEUROSCI.0291-17.2017

Table 3.

Optimized DCM connectivity parameters

Parameter name Prior mean Prior (log) variance Posterior mean Posterior variance Description
A{1}(2,1) 0 1/16 −1.0919 0.0283 Ascending connection from AC to IF (SP to SS)
A{2}(2,1) 0 1/16 −0.0901 0.0553 Ascending connection from AC to IF (SP to DP)
A{3}(1,2) 0 1/16 −0.0877 0.0341 Descending connection from IF to AC (DP to SP)
A{4}(1,2) 0 1/16 −0.3759 0.0527 Descending connection from IF to AC (DP to II)
G(1,:) −1.39; −2.02 1/8; 1/8 −1.3455; 0.2480 0.0568; 0.0157 Intrinsic connections in AC (SP gain; II gain)
G(2,:) −1.39; −2.02 1/8; 1/8 −1.1852; −2.2009 0.0498; 0.0575 Intrinsic connections in IF (SP gain; II gain)
B{1}(1,1) 0 1/8 −0.1324 0.0013 Modulation of SP gain in AC by regularity
B{1}(2,2) 0 1/8 0.0624 0.0001 Modulation of SP gain in IF by regularity
B{2}(1,1) 0 1/8 0.0001 0 Modulation of SP gain in AC by alphabet size
B{2}(2,2) 0 1/8 −0.0437 0.0001 Modulation of SP gain in IF by alphabet size

The posterior means were used as priors while inverting the initial time window in the Bayesian belief updating analysis.