Depression is a common disorder with an annual prevalence that varies from 1% to 6% in the community. It is associated with a high rate of recurrence; up to 50-60% of depressed individuals may not respond to conventional pharmacotherapy. This reinforces the need to find alternative treatments to improve the outcome of such patients. Transcranial magnetic stimulation (TMS) allows for the non-invasive stimulation of superficial cortex in patients, both for investigative and therapeutic purposes. It is usually applied over the prefrontal cortex at frequencies of 1-20 Hz and at an intensity around motor threshold. A meta-analysis is presented, including 31 studies evaluating the antidepressant effect of TMS. Repetitive (r)TMS was more effective in the treatment of depression than sham rTMS, with a medium-sized effect. Studies that have examined rTMS efficacy in the treatment of depression are highly heterogeneous in terms of both sample characteristics and treatment parameters. Moreover, most of them have employed a relatively small number of participants. Strict double-blinding cannot be guaranteed due to flawed sham conditions. All of these factors have the potential to lead to errors in the results of the individual studies and may have affected the result of the present meta-analysis. Unfortunately, there is no compelling evidence regarding the most effective combination of rTMS parameters. The literature suggests that future trials should employ a greater number of rTMS sessions, adequate concealment allocation and an individualized approach to locating the DLPFC using neuroimaging. Until the optimal treatment parameters are delineated, rTMS remains an experimental approach for the treatment of depression. More knowledge regarding the characteristics of patients who benefit from this treatment and the size and persistence of clinical effects is needed. © 2006 Elsevier Ltd. All rights reserved.