In neurorehabilitation, transcranial magnetic stimulation (TMS) offers information regarding prognosis and pathophysiology and could also be useful for therapeutic purposes. Numerous studies have indicated that, after stroke, the absence of motor evoked potentials is associated with a poor motor recovery. In contrast, MEPs obtained in the paretic muscle with low stimulus intensities suggest a good restitution of motor function. TMS studies have shown that the location of a brain lesion determines motor cortex excitability changes: Patients with central somatosensory lesions show a disinhibition in the ipsilesional motor cortex. Lesions in the territory of the superior cerebellar artery are associated with a loss of motor cortex excitability. Stroke patients participating in a Constraint-induced movement therapy show an enlargement of the motor output area in the affected hemisphere after therapy. This enhancement of motor excitability is associated with an improvement of motor function. Some evidence is emerging that the application of low frequency repetitive TMS over the non-lesioned hemisphere improves neglect phenomena by down-regulation of the excitability of the non-lesioned hemisphere.