Figure 2
From: Homeostatic plasticity and emergence of functional networks in a whole-brain model at criticality
(a) Pearson correlation (solid lines) between simulated and empirical functional connectivity matrices, ρ(Fm, Fe), as a function of T/Tc (black color for W and red color for \(\tilde{W}\)). Chi-squared distance (dots) between the corresponding (normalized) probability distribution functions. The normalization of the excitatory input (\(\tilde{W}\)) enhances the match between empirical and simulated data by a factor of about 1.5. The best match (ρ ≈ 0.6) occurs at T corresponding to the peak of σ(A), while the smallest distance, χ2 ≈ 0.4, occurs at T corresponding to the peak of 〈S2〉. (b) Overall match between empirical resting state networks (templates obtained from75) and simulated RSNs using sICA. We use the Cohen’s Kappa index κ as a measure of similarity. (c) Probability distribution functions at the corresponding minimum of χ2, that is, T/Tc = 0.8 and T/Tc = 1 for the non normalized and normalized networks, respectively. The green line represents the empirical distribution. (d–f) Empirical and simulated functional connectivity matrices for the same parameters used in (c). The functional matrices are organized in blocks with RH (right hemisphere) and LH (left hemisphere).
