Spontaneous structure formation in a network of chaotic units with variable connection strengths

As a model of temporally evolving networks, we consider a globally coupled logistic map with variable connection weights. The model exhibits self-organization of network structure, reflected by the collective behavior of units. Structural order emerges even without any interunit synchronization of dynamics. Within this structure, units spontaneously separate into two groups whose distinguishing feature is that the first group possesses many outwardly directed connections to the second group, while the second group possesses only a few outwardly directed connections to the first. The relevance of the results to structure formation in neural networks is briefly discussed.