While motor cortical areas are the main targets of the integrative activity of basal ganglia, their main output consists of the corticospinal system. Transcranial magnetic stimulation (TMS), a relatively new method to investigate corticospinal physiology, has been widely used to assess possible changes secondary to Parkinson's disease (PD). The use of single- and paired-pulse TMS, two varieties of the original technique, disclosed multiple functional alterations of the corticospinal pathway. For instance, when the latter was tested at 'rest', or in response to somesthetic afferents, it showed excess excitability or reduced inhibition. In turn, during production of a voluntary output, its activation was defective, or inadequately modulated. One major mechanism may be a dysfunction of the interneurons mediating the level of excitation within cortical area 4. For instance, there is a shortening of the so-termed 'central silent period', which is a complex, TMS-induced, inhibitory phenomenon possibly mediated by activation of GABA(B) receptors. The so-called 'short-interval intracortical inhibition', which is possibly mediated by GABA(A) receptors, is also diminished. Levodopa restores these and other TMS alterations, thus demonstrating that cortical area 4 is sensitive to dopamine modulation. Overall, TMS has provided substantial new pathophysiological insights, which point to a central role of the primary motor cortex in the movement disorder typical of PD. Repetitive (r-)TMS, another form of TMS, has been studied as a treatment for PD motor signs. Although some reports are favorable, others are not, and have raised the problem of appropriate control experiments. Although extremely interesting, the potential therapeutic role of r-TMS in PD needs further evaluation.