For a large number of vertebrate species it is now indisputable that spinal networks have the capability of generating the basic locomotor rhythm. However, because of technical difficulties, the rate of progress in defining the intrinsic properties of mammalian locomotor rhythm generators has been slow in comparison to that made in the study of such networks in lower vertebrates. Investigations on afferent and descending control of locomotor activity in mammals have demonstrated that many of these pathways interact with the rhythm generator. In this review we discuss how these interactions (resetting) can be used for outlining relevant spinal circuits as a basis for a future identification of individual neurons of the spinal locomotor networks. In this overview we have given particular emphasis to selected afferent systems to illustrate the possibilities and problems with this approach.