Over the last decade, neural transplantation has progressed from being an experimental technique for studying regeneration and plasticity in the brain to clinical trials of reconstructive surgery in human neurodegenerative disease. Whereas clear evidence is only available at present for the viability of this technique in Parkinson’s disease, applications to several other diseases, including Huntington’s disease, multiple sclerosis, spinal cord injury, and chronic pain are currently under active consideration. It is clear that the techniques of transplantation can be functionally viable under certain well-defined biological circumstances, but significant problems remain in the availability of suitable donor tissues and defining the optimal conditions for reliable survival of the implanted cells. If we are to obtain improved reliability of the present techniques or identify suitable alternatives, we need a better understanding of the conditions for the survival and integration of grafts into the host brain, and the mechanisms by which they influence host function. In this review I consider the nature of the structural reconstruction required to achieve repair in animal models of Parkinson’s and Huntington’s diseases, contrasting the replacement of deficient neurochemicals within the striatum in the former case, and the need for reconstruction of input and output connections of the striatal circuitry in the latter.