For a long time now, researchers have worked ondefining different metrics able to characterize the importanceof nodes in networks. Among them, centrality measures haveproved to be pertinent as they relate the position of a node in thestructure to its ability to diffuse an information efficiently. Thecase of dynamic networks, in which nodes and links appear anddisappear over time, led the community to propose extensions ofthose classical measures. Yet, they do not investigate the fact thatthe network structure evolves and that node importance mayevolve accordingly. In the present paper, we propose temporalextensions of notions of centrality, which take into account thepaths existing at any given time, in order to study the timeevolution of nodes' importance in dynamic networks. We applythis to two datasets and show that the importance of nodes doesindeed vary greatly with time. We also show that in some cases itmight be meaningless to try to identify nodes that are consistentlyimportant over time, thus strengthening the interest of temporalextensions of centrality measures.