Fig. 8.

Parametric empirical Bayes analysis of simulated data. a) Estimated slow dynamics of parameters in the CMC mode over 21 segments of data with average of 2 second duration. b) PEB analysis with a design matrix based on the spectral envelope of the data. The original PEB posterior estimates show on top (full model) and the estimates following Bayesian model reduction are shown below (reduced). The reduced model reveals the underlying causes of seizures in Fig. 7 and clearly capture is the trends in parameters. ${\mathit{G}}_1$, ${\mathit{G}}_2$, ${\mathit{G}}_3$ and ${\mathit{G}}_4$ are self-inhibition of excitatory, superficial pyramidal, inhibitory, and deep pyramidal cells, respectively. As in previous figures, the effect sizes relate to unitless log scaling parameters.