Dual-attribute ring-star neural network

Nowadays, the research of brain science is becoming more and more important. The study of the complex dynamic behavior of neural networks plays an irreplaceable role in the in-depth understanding of neurological diseases and brain behavior. However, research on the heterogeneity of neuronal connections is still insufficient. Based on this background, a Dual-Attribute Ring-Star Neural Network (DRSNN) based on Small-World-like Coupling Strength Changing Scheme (SCSCS) is proposed in this paper. Dual-Attribute in DRSNN means that the network proposed in this paper adopts ring-star network in topology and the connections between neurons are heterogeneous using the SCSCS to simulate the complexity of real neural network connections. Through simulation of the DRSNN, it is found that when the initial strengths are set appropriately, changing the star coupling strengths can significantly increase the probability of lag synchronization behavior; the probability of global synchronization increases significantly when the ring coupling strengths are changed. When all the coupling strengths change, the probability of synchronization behavior, especially lag synchronization behavior, will increase greatly. The influence of changing two important parameters of the SCSCS (the number of iterations M and the changing probability P₂) on the dynamic behavior of DRSNN is studied: increasing M and P₂ can improve the probability of synchronization of the DRSNN, especially lag synchronization. The DRSNN proposed in this paper is helpful to simulate the connectivity and heterogeneity of real neural networks.