OBJECT: The authors describe a new method of localizing electrodes on magnetic resonance (MR) images and focus on the positions of both the most efficient contact and the electrode related to the MR imaging target. METHODS: Thirty-one patients who had undergone bilateral subthalamic nucleus (STN) deep brain stimulation (DBS) were included in this study. Target coordinates were calculated in the anterior commissure-posterior commissure referential. A study of the correlation between the artifact and the related contact allowed one to deduce the contact position from the identification of the distal artifact on MR imaging. The best stimulation point corresponded with the contact resulting in the best Unified Parkinson's Disease Rating Scale (UPDRS) motor score improvement. It was compared (Student t-test) with the dorsal margin of the STN (DM STN), which was determined electrophysiologically. The distance between the target and the electrode was calculated individually in each axis. The best stimulation point was located at anteroposterior -2.34 +/- 1.63 mm, lateral 12.04 +/- 1.62 mm, and vertical -2.57 +/- 1.68 mm. This point was not significantly different from the DM STN (p < 0.05). The postoperative UPDRS motor score was 28.07 +/- 12.16, as opposed to the preoperative score of 46.27 +/- 13.89. The distance between the expected and actual target in the x- and y-axes was 1.34 +/- 1.02 and 1.03 +/- 0.76 mm, respectively. In the z-axis, 39.7% of the distal contacts were located proximal to the target. CONCLUSIONS: This approach proposed for the localization of the electrodes on MR imaging shows that DBS is most effective in the dorsal and lateral part of the STN and indicates that the DBS electrode can be located more proximally than originally expected because of the caudal brain shift that may occur during the implantation procedure.