Spinal cord injury is characterized by a complex set of events, which include the disruption of connectivity between the brain and the periphery with little or no spontaneous regeneration, resulting in motor, sensory and autonomic deficits. Transplantation of neural stem cells has the potential to provide the cellular components for repair of spinal cord injury (SCI), including oligodendrocytes, astrocytes, and neurons. The ability to generate graft-derived neurons can be used to restore connectivity by formation of functional relays. The critical requirements for building a relay are to achieve long-term survival of graft-derived neurons and promote axon growth into and out of the transplant. Recent studies have demonstrated that mixed populations of glial and neuronal progenitors provide a permissive microenvironment for survival and differentiation of early-stage neurons, but inclusion of growth factors with the transplant or cues for directional axon growth outside the transplant may also be needed. Other important considerations include the timing of the transplantation and the selection of a population of neurons that maximizes the effective transmission of signals. In some experiments, the essential neuronal relay formation has been developed in both sensory and motor systems related to locomotion, respiration, and autonomic functions. Despite impressive advances, the poor regenerative capacity of adult CNS combined with the inhibitory environment of the injury remain a challenge for achieving functional connectivity via supraspinal tracts, but it is possible that recruitment of local propriospinal neurons may facilitate the formation of relays. Furthermore, it is clear that the new connections will not be identical to the original innervation, and therefore there needs to be a mechanism for translating the resulting connectivity into useful function. A promising strategy is to mimic the process of neural development by exploiting the remarkable plasticity associated with activity and exercise to prune and strengthen synaptic connections. In the meantime, the sources of neural cells for transplantation are rapidly expanding beyond the use of fetal CNS tissue and now include pluripotent ES and iPS cells as well as cells obtained by direct reprogramming. These new options can provide considerable advantages with respect to preparation of cell stocks and the use of autologous grafting, but they present challenges of complex differentiation protocols and risks of tumor formation. It is important to note that although neural stem cell transplantation into the injured spinal cord is primarily designed to provide preclinical data for the potential treatment of patients with SCI, it can also be used to develop analogous protocols for repair of neuronal circuits in other regions of the CNS damaged by injury or neurodegeneration. The advantages of the spinal cord system include well-defined structures and a large array of quantitative functional tests. Therefore, studying the formation of a functional relay will address the fundamental aspects of neuronal cell replacement without the additional complexities associated with brain circuits.