Table A1.

Detailed overview of exponential family parameter settings for the simulation of healthy- and chronic pain inference.

Parameter	Healthy inference	Pathological inference
Prior on 1st node	$\lambda =p(\text{pain})=0.222$,	$\lambda =p(\text{pain})=0.777$,
	$\nu =100$	$\nu =100$
Top-down prior	$\lambda =p({\text{noxious}}_t\mid {\text{pain}}_t)=0.8$,	$\lambda =p({\text{noxious}}_t\mid {\text{pain}}_t)=0.6$,
	$\lambda =p({\text{harmless}}_t\mid {\overline{\text{pain}}}_t)=0.9$,	$\lambda =p({\text{harmless}}_t\mid {\overline{\text{pain}}}_t)=0.4$,
	$\nu =100$	$\nu =20$
Transition prior	$\lambda =p(H_t=\text{pain}\mid H_{t-1}=\overline{\text{pain}})=0.2$,	$\lambda =p(H_t=\text{pain}\mid H_{t-1}=\overline{\text{pain}})=0.7$,
	$\lambda =p(H_t=\overline{\text{pain}}\mid H_{t-1}=\text{pain})=0.7$,	$\lambda =p(H_t=\overline{\text{pain}}\mid H_{t-1}=\text{pain})=0.2$,
	$\nu =100$	$\nu =100$
Observations	$S_1-S_5$= noxious input, $S_{12}-S_{18}$ = harmless input for both cases

Further, the first observation scheme to produce Figure 2 is shown below. Probabilities and their respective counter-probabilities sum up to 1 (83).