BACKGROUND: There is evidence that both high and low frequency rTMS may have therapeutic effects on motor performance of Parkinson's disease. OBJECTIVE: The aim of the study was to conduct the first direct comparison of the two approaches. METHODS: 52 PD patients were randomly classified into two groups. The first group received 20Hz and the 2 nd group received 1Hz rTMS with a total of 2000 pulses over M1of each hemisphere for ten days. Effects were assessed with the Unified Parkinson's Disease Rating Scale part III (UPDRS), Instrumental Activity of Daily Living (IADL), and a self-assessment score (SA) before, after the last session, and one month later. Cortical excitability was measured before and after the end of sessions. RESULTS: There was a significant improvement on all rating scales after either 1 Hz or 20 Hz rTMS, although but the effect was greaterpersisted for longer at after 20 Hz (treatment X time interaction for UPDRS and IADL (P = 0.075 and 0.04 respectively). Neither treatment affected motor thresholds, but 20 Hz rTMS increased MEP amplitude and the duration of transcallosal inhibition. In an exploratory analysis, Eeach group was subdivided into akinetic-rigid and tremor dominant subgroups and the effects of 1 Hz and 20 Hz treatment recalculated. There is was weak evidence that patients with an akinetic-rigid presentation may respond better than those with predominant tremor. CONCLUSION: Both 20 Hz and 1 Hz rTMS improve motor function in PD, but 20 Hz rTMS is more effective and patients with an akinetic-rigid respond better than predominant tremor.