Wiring cost in the organization of a biological network

Abstract

To find out the role of the wiring cost in the organization of the neural network of the nematode \textit{Caenorhapditis elegans} (\textit{C. elegans}), we build the neuronal map of \textit{C. elegans} based on geometrical positions of neurons and define the cost as inter-neuronal Euclidean distance \textit{d}. We show that the wiring probability decays exponentially as a function of \textit{d}. Using the edge exchanging method and the component placement optimization scheme, we show that positions of neurons are not randomly distributed but organized to reduce the total wiring cost. Furthermore, we numerically study the trade-off between the wiring cost and the performance of the Hopfield model on the neural network.