Understanding the factors underlying invasive processes, such as establishment and spread, is critical to predict the occurrence and magnitude of biological invasions. In general, those factors do not only depend on the invader itself, but also on the characteristics of the invaded area. In the metapopulation theoretical framework, landscapes are approximated as networks of discrete habitat patches. These patches are nodes of a network connected by edges that represent dispersal pathways between them. Many properties can be quantified to accurately characterize spatial heterogeneity in the connectivity between patches in a network. Node centrality for instance, measures the contribution of an individual habitat patch to landscape connectivity; highly central nodes are called “hubs”. By facilitating individual dispersal, those hubs can promote spread, but also decrease local establishment because of mechanisms like demographic stochasticity and Allee effects. By combining modeling and experimental approaches, we investigated how invasion dynamics are affected by the presence of hubs in a landscape. We first simulated invasions using a stochastic discrete-time metapopulation model. Then we challenged model predictions by performing experimental introductions of minute parasitoid wasps in artificial laboratory landscapes with patches characterized by different levels of centrality.