Self-regulated emergence of heavy-tailed weight distributions in evolving complex network architectures
Synaptic plasticity typically produces heavy-tailed distributions of synaptic strengths, consisting of a few strong connections among many weaker ones. Meanwhile, structural plasticity relies on distinct signaling cascades to reshape network topology. We propose a model in which both types of plasticity adhere to the Hebbian principle while operating within homeostatically regulated activity. Synaptic plasticity alone generates heavy-tailed weight distributions, but only when any activity spreading beyond neighboring units is discarded. However, when combined with Hebbian structural plasticity, i.e., adaptive rewiring, heavy-tailed weight distributions also arise with more extensive activity flow. Furthermore, adaptive rewiring provides complex network structures with convergent-divergent circuits similar to those that facilitate signal transmission throughout the nervous system. Having adaptive weight adjustment and rewiring driven by the same homeostatic dynamics gives our model a parsimonious and robust framework that simultaneously produces heavy-tailed weight distributions and convergent-divergent units under a wide range of dynamical regimes. Consequently, the model accounts for key connectivity structures in both C. elegans and the mouse, suggesting that its principles are shared across species of different complexities.