Figure 7. .

Relationship between the interparcellation variations of the empirical graph-theoretical metrics and the goodness-of-fit. (A) Cross-correlations among the inverted granularities, the graph-theoretical measures of the empirical connectomes (network propertties depicted in Figure 2 and Figure 3), the structure-function relationship, and the goodness-of-fit of the models to the empirical data. The corelation was calculated across parcellations between the median values over all subjects. Significant correlations are highlighted by colors (p < 0.05, two-sided, Bonferroni corrected). (B) Loadings of the first (PC1) and the second (PC2) principal components of the group-averaged graph-theoretical metrics, that is, the contributions of the original empirical data variables to PC1 and PC2. (C) Regressions of the PC1 scores with the medians of the goodness-of-fit between empirical (eFC) and simulated (sFC) functional connectivity. The medians were calculated across subjects for each considered parcellation for the phase oscillator (red) and the neural mass model (blue) as indicated in the legend together with the fraction of the explained variance. The symbols stand for the individual parcellations from Table 1. (D) Cumulative amount of explained variance in the group-averaged graph-theoretical measures as a function of the number of included PCs. (E) Fraction of the interparcellation variance of the goodness-of-fit being explained by the (multivariate) linear regression model as a function of the number of PCs included in the model. Other abbreviation: a.u. = arbitrary unit, cumul. = cumulative, expl. = explained, var. = variance.