Fig. 6: Qualitative comparison between empirical and simulated FC properties.

Top two rows display network properties of the empirical rsFC, compared against the observations from simulated FCs (bottom two rows), which are obtained for the optimal parameter choices. Number of test subjects are: n = 23 young and n = 22 older adults for CamCAN data, and n = 23 young and n = 17 older adults for NKI data. The variability is coming solely from individual’s SC and is mainly assessed in the test cohort. The MDMF model inversion reproduces almost similar alterations between two age groups occurring for the test cohorts, where, the modeling parameters are computed from training cohorts. Except, alterations in betweenness centrality is not fully predicted from the test subjects of NKI data. Statistical differences are assessed by Mann–Whitney U test, where p < 0.05 implies statistically significance, otherwise not significant (NS). A solid black line connecting two mean values of younger and elderly groups, is plotted to visually tracking alterations. The errorbars show standard deviations.