AUTHOR=Rentzeperis Ilias , van Leeuwen Cees 

TITLE=Adaptive Rewiring in Weighted Networks Shows Specificity, Robustness, and Flexibility

JOURNAL=Frontiers in Systems Neuroscience

VOLUME=Volume 15 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/systems-neuroscience/articles/10.3389/fnsys.2021.580569

DOI=10.3389/fnsys.2021.580569

ISSN=1662-5137

ABSTRACT=The brain adaptively rewires its network connections in response to neural activity. Adaptive rewiring can be described as optimizing the network structure for the efficiency of neural signal diffusion. Thus, adaptive rewiring amounts to creating shortcut connections in regions with high diffusion and pruning where diffusion is low.  In evolving model networks, adaptive rewiring leads to topologies akin to brain anatomy: small worlds with rich club and modular or centralized structures. We continue our model simulations of this mechanism, focusing on three desiderata for the rewiring mechanism: a bias to evolve towards specific architectures (specificity), the ability to dynamically maintain an initial architecture (robustness), and the tendency to stochastically deviate from the rules of specificity or robustness (flexibility). We find that adaptive rewiring models show both specificity and robustness, controlled by a single parameter, the rewiring rate. Furthermore, models show greater flexibility for a skewed, lognormal weight distribution of their connections than for a normally distributed one. The results show that adaptive rewiring offers robustness and specificity of the evolving network architectures, and that small control parameter shifts allow flexibly switching between them. These results contribute to the notion of adaptive rewiring as a basic mechanism in establishing the variety of brain functional architecture.