Table 1.

Parameters of the neuronal model (see also Fig. 1).

	Description	Parameterisation	Prior
${\kappa }_i$	Postsynaptic rate constant of the $i^{th}$ neuronal population in each of N regions	$\begin{array}{l}\exp \left({\theta }_{\kappa }\right)\cdot {\kappa }_i\\ \kappa =\left[\mathrm{256,128,16,32}\right]\end{array}$	$p\left({\theta }_{\kappa }\right)=N\left(\mathrm{0,0}\right)$
$a_{i\to k}$	Intrinsic connectivity between populations i and k in each region.	$\exp \left({\theta }_a\right)\cdot a$ $a=\left[2111\right]\ast 512$	$p\left({\theta }_a\right)=N\left(\mathrm{0,0}\right)$
$B_{b,f}$	Condition-specific matrices. Elements are zero unless forward, backward or intrinsic connections are allowed to change in different conditions.	${\theta }_{b,f}$	$p\left({\theta }_b\right)=N(0,\frac{1}{8})$
$A_{f,b}$	Forward and backward extrinsic connectivity matrices. If there is any forward (backward) connection between from region $j$ to $i$, the corresponding element $(i,j)$ in $A_f$ (respectively $A_b)$ is set to one.	$\exp \left({\theta }_A\right).A_{f,b}$	$p\left({\theta }_A\right)=N(0,\frac{1}{8})$
C	Scalar matrix to driving input	${\theta }_c$	$p\left({\theta }_C\right)=N(0,\frac{1}{32})$