OBJECTIVE: Cerebral ventriculography (Vg) and magnetic resonance imaging (MRI) scanning are routine procedures to determine the implanted electrode placement into the subthalamic nucleus (STN) and are used in several centers that provide deep brain stimulation for Parkinson's disease patients. However, because of image distortion, MRI scan accuracy in determining electrode placement is still matter of debate. The objectives of this study were to verify the expected localization of the electrode contacts within the STN and to compare the stereotactic coordinates of these contacts determined intraoperatively by Vg with those calculated postoperatively by MRI scans. To our knowledge, this is the first study attempting to compare the "gold standard" of stereotactic accuracy (Vg) with the anatomic resolution provided by MRI scans. METHODS: Images from 18 patients with Parkinson's disease who underwent bilateral operation were used in this study. Among the 36 chronically stimulated contacts, 28 contacts (78%) were localized in the dorsolateral part of the STN. The remaining eight contacts (22%) were located more dorsally in the zona incerta, close to the upper border of the STN. RESULTS: Significant differences were found between Vg and MRI scans regarding the mediolateral x coordinate of the contacts for both left and right electrodes and regarding the right-sided anteroposterior y coordinate. No statistical difference was found for the left-sided y coordinate and the dorsoventral z coordinate for both sides. CONCLUSION: If we assume that Vg is an imaging gold standard, our results suggest that postoperative MRI scanning may induce a slight image translation compared with Vg. However, MRI scans allowed localization of most of the contacts within the STN.