Neuroimaging experimental studies on brain plasticity in recovery from stroke.

Topographical cortical organization of sensorimotor area has been shown to be highly plastic, altering his configuration in response to training in different tasks in healthy controls and neurological patients. The term ''brain plasticity'' encompasses all possible mechanisms of neuronal reorganization: recruitment of pathways that are functionally homologous to, but anatomically distinct from, the damaged ones (eg, non-pyramidal corticospinal pathways), synaptogenesis, dendritic arborisation and reinforcement of existing but functionally silent synaptic connections (particularly at the periphery of core lesion). The study of neuroplasticity has clearly shown the ability of the developing brain--and of the adult and ageing brain--to be shaped by environmental inputs both under normal conditions (ie, learning) and after a lesion. Neuronal aggregates adjacent, or distant to a lesion in the sensorimotor area can progressively adopt the function of the injured area. Imaging studies indicate that recovery of motor function after a lesion (i.e. stroke) is associated with a progressive change of activation patterns in specific brain structures. Transcranial magnetic stimulation (TMS) and magnetoencephalography (MEG) can detect reshaping of sensorimotor areas; they have a high temporal resolution but have several limitations. TMS can only provide bidimensional scalp maps and MEG depicts three-dimensional spatial characteristics of virtual neural generators obtained by use of a mathematical model of the head and brain. However, the use of objective methods that assess brain reactivity to a physical stimulus (i.e., TMS) or to a sensory input (ie, electrical stimulation to hand and fingers) can integrate information from self-paced motor tasks, because the resolution of abnormal activation over time could be secondary to recovery. Functional MRI (fMRI) and positron emission tomography (PET), on their own, have insufficient time resolution to follow the hierarchical activation of relays within a neural network; however, because of their excellent spatial resolution, they can integrate the findings of TMS and MEG. An integrated approach constitutes, at present, the best way to assess the brain plasticity both under normal conditions and after a lesion.