The understanding of the physiological basis of basic functions of brains requires detailed information about the functional structures of neuronal networks. Feedback loops are crucial dynamic motifs playing a pivotal role in the regulation and control of many important physiological and biochemical processes such as gene transcription, signal transduction, and metabolism (intracellular processes), and neuronal coding and decoding (intercellular processes). In the study, we focus on investigating a prerequisite for generating the synchronizations in spike neural networks and also a crucial structure for inducing the most robust synchronized network behaviors. Based on simulations for synthetic spiking neuronal network models, it is shown that only the networks with positive feedback loop (PFL) motifs can arrive at synchronizations under a variety of initial states. Importantly, coupled direct and indirect PFL motifs induce most robust synchronized bursting behaviors, which are one of the most remarkable characteristics of biological neuronal networks. This result may infer a crucial structural module for designing both in vitro and in vivo neuronal networks.