Efficient Physical Embedding of Topologically Complex Information Processing Networks in Brains and Computer Circuits
Wiring costs of nervous and computational networks are lower than expected in a random network, but can be reduced by a rewiring algorithm designed to minimize connection distance between elements (A); see Table 1. However, the fractal dimension of network topology, while lower than expected in a randomly rewired network, is also reduced by rewiring for cost minimization (B). Rentian scaling is either destroyed or disrupted by minimally rewiring the C. elegans connectome (C) or the human MRI network (D). The inset panels in (C,D) show the hierarchical modularity of each minimally wired system, which has been represented by a co-classification matrix as in Figure 2.