Repetitive transcranial magnetic stimulation (rTMS) of the prefrontal cortex has been thought to have great potential to treat refractory depression since the first studies published ten years ago. However, one of the potential limitations of rTMS is the poor definition of the localization of the prefrontal cortical target, which is based on a rather simplistic anatomical approach, i.e., 5cm anterior to the primary motor cortical representation of the hand. This "standard procedure" does not take into consideration interindividual variations in brain morphology. We report the case of a 40-year-old woman who underwent two weeks of 10Hz-rTMS for the treatment of a major, drug-resistant depressive episode. The rTMS target was determined with the "standard procedure" for the first week and with a dedicated navigation system as the left Brodmann area 46 for the second week. The clinical assessment of antidepressant effects was performed before and after each week of stimulation. Following the first week of stimulation, the patient improved, in particular regarding speech production. Using the navigation system, the location of the target stimulated during the first week was found to correspond to Broca's area, and not to the prefrontal area as intended. Antidepressant effects were more marked after the second week of navigated rTMS. In the present case, the prefrontal target was situated 8.3cm anterior to hand motor cortex. This illustrates that the "standard procedure" may inaccurately target the prefrontal cortex, although resulting in antidepressant-like effects. The use of navigation systems should limit the variability of the results reported so far in the treatment of depression by rTMS.