Figure 2. Brain network modeling.

(A) Group-averaged human and chimpanzee networks visualized on the same brain template. Top 20% of connections by strength are shown. (B) Schematic diagram of the model. Each brain region is recurrently connected with strength $w$ and driven by an excitatory input $I_0$ and white noise with standard deviation $D$. The connection between regions $i$ and $j$ is weighted by $A_{ij}$ based on the connectomic data. The regional neural dynamics are represented by the synaptic response variable $S$; high $S$ translates to high neural activity. (C) Method for calculating the dynamic range of each brain region from its mean synaptic response $\overline{S}$ versus global recurrent strength $w$ curve. Note that ${\displaystyle {\overline{S}}_x={\overline{S}}_{\mathrm{m}\mathrm{i}\mathrm{n}}+\left(x/100\right)\ast \left({\overline{S}}_{\mathrm{m}\mathrm{a}\mathrm{x}}-{\overline{S}}_{\mathrm{m}\mathrm{i}\mathrm{n}}\right)}$, with $w_x$ being the corresponding global recurrent strength at ${\overline{S}}_x$ and $x=\{10,90\}$. (D) Example time series of regions with different (top panel) and similar (bottom panel) dynamic ranges at $w$ = 0.6 and 0.8. The time series in the top panel have correlation values (Pearson’s r) of 0.06 and 0.08 at $w$ = 0.6 and $w$ = 0.8, respectively. The time series in the bottom panel have correlations of 0.40 and 0.14 at $w$ = 0.6 and $w$ = 0.8, respectively.

Figure 2—figure supplement 1. Validation of simulated dynamics on empirical functional neuroimaging data.

(A) From the connectome, neural activity is simulated using the model presented in Figure 2B. This activity is fed into a hemodynamic model to obtain a simulated fMRI signal for each brain region. Finally, functional connectivity ($FC$) is calculated by taking pairwise Pearson correlations of the simulated fMRI signals across all regions. (B) Association between simulated and empirical FC for humans and macaques. The simulated FC is calculated using model parameters that optimize model-data fitting. Each dot represents the pairwise $FC$. The solid line represents a linear fit with Pearson’s correlation coefficient (r) and p value (p).