Topographic organization of metabolic networks associated with cognitive deficits in PD patients. a The previously documented PD-cognitive deficit-related metabolic pattern (PDCP) (Huang et al. 2007b). b Eigenvector Centrality (EC) values (i.e., relative nodal importance in information flow) analyzed by graph theory correlated with corresponding region weights of the PDCP. c The EC (radius of sphere) and EC derivatives within PDCP-relevant hubs (DEC; color of sphere) were visualized. DEC estimates the “sensitivity” of information flow to discrete perturbation against nodal centrality. The EC is correlated with the disease-related metabolic patterns’ region weights, and that the nodal sensitivity toward the metabolic network configuration can be estimated using a novel perturbation (simulation) method (Ko et al. 2018), which computes the DEC. d Bi-plot of the EC vs. DEC. The right middle frontal gyrus (rMFG, BA 8/9) and superior frontal gyrus (rSFG, BA 6/8) were identified with relatively high “importance” and “sensitivity” (red arrow).