Emergent Behavior in Spiking Neural Networks: A Study on Complex Network Structures

Spiking neural networks (SNNs), often referred to as the "third-generation" artificial neural networks, closely emulate the bio-inspired characteristics of global weakly supervised learning and local unsupervised bio-optimization. The neurons in SNNs more accurately mirror biological neurons. Increasing evidence suggests that biological neural networks exhibit complex network structures. To gain a deeper understanding of spiking neural networks, it is essential to focus on the organization and topology of their networks. In this paper, we construct several spiking neural network models with complex network structures and apply continuous external inputs to the target networks. By observing the emergent behavior of these models under different structural configurations, we analyze the influence of external inputs on networks with varying topologies. Our findings aim to provide a foundation for uncovering and harnessing the intelligent behaviors inherent in spiking neural networks.