In this work we focus on the temporal interactions within a network. We focus on information transmission delays, both in terms of cognitive systems and complex systems. Based on a model of neuron network which core is a set of interconnected neurons, with no predefined organization, we try to understand the relationship between the dynamics that take place within the network and the network topology. To what extent does the topology determine the learning ability ? We are specifically interested in information transmission delays in this neuron network. Indeed, the existence of various transmission delays within a network can introduce the concept of temporal pattern of activity, or polychronous group. The adaptation of the topology by unsupervised learning, based on the activity, determines the selection of preferential patterns. We seek to highlight and exploit the existence of stimulus-specific patterns. Such results would allow us to propose new tools to study complex networks, as well as new models of classification.