Recovery of the function after stroke is a consequence of many factors including resolution of oedema and survival of the ischaemic penumbra. In addition there is a growing interest in the role of central nervous system (CNS) reorganization. Much of the evidence supporting this comes from animal models of focal brain injury, but non-invasive techniques such as functional magnetic resonance imaging, transcranial magnetic stimulation, electroencephalography and magnetoencephalography now allow the study of the working human brain. Using these techniques it is apparent that the motor system of the brain adapts to damage in a way that attempts to preserve motor function. This has been demonstrated after stroke, as part of the ageing process, and even after disruption of normal motor cortex with repetitive transcranial magnetic stimulation. The result of this reorganization is a new functional architecture, one which will vary from patient to patient depending on the anatomy of the damage, the biological age of the patient and lastly the chronicity of the lesion. The success of any given therapeutic intervention will depend on how well it interacts with this new functional architecture. Thus it is crucial that the study of novel therapeutic strategies for treating motor impairment after stroke take account of this. This review maps out the attempts to describe functionally relevant adaptive changes in the human brain following focal damage. A greater understanding of how these changes are related to the recovery process will allow not only the development of novel therapeutic techniques that are based on neurobiological principles and designed to minimize impairment in patients suffering from stroke, but also to target these therapies at the appropriate patients.