Repair of CNS injury or degeneration by cell replacement may lead to significant functional recovery only through faithful reconstruction of the original anatomical architecture. This is particularly relevant for point-to-point systems, where precisely patterned connections have to be re-established to regain adaptive function. Despite the major interest recently drawn on cell therapies, little is known about the mechanisms and the potentialities for specific integration of new neurons in the mature CNS. Major findings and concepts about this issue will be reviewed here, with special focus on work dealing with the Purkinje cell transplantation in the rodent cerebellum. These studies show that the adult CNS may provide some efficient information to direct cell engraftment and process outgrowth. On their side, immature cells may be able to induce adaptive changes in their adult partners to facilitate their incorporation in the recipient network. Despite the rather high degree of specific integration achieved in several different CNS regions, these processes are usually defective and long-distance connections are not rewired. Thus, although some potentialities for cell replacement exist in the mature CNS, full incorporation of new neurons in adult circuits is rarely observed. Indeed, intrinsic mechanisms for growth control as well as injury-induced changes in the properties and architecture of the nervous tissue contribute to hamper repair processes. As a consequence, crucial to obtain successful cell replacement and integration in the mature CNS is a deep understanding of the basic biological mechanisms that regulate the interactions between newly added elements and the recipient environment.