Introduction The subthalamic nucleus (STN), a well known target for deep brain stimulation (DBS) in Parkinson’s disease (PD), is involved in motor, limbic, and cognitive processes ( Temel et al., 2005) . To achieve a successful outcome of DBS surgery the electrode should be placed in its motor part. In current clinical practice, targeting is performed with anatomical 1.5 T or 3T MR images, that display the STN as a homogenous structure. Consistent and patient specific identification of the motor, associative, and limbic zones of the STN with 3T MRI has been demonstrated to be challenging ( Lambert et al., 2012) . Objectives In this study, ultra-high field (7T) MRI was used to map the motor and non-motor areas of the STN in individual patients with PD. Patients and methods The study was approved by the Institutional Review Board at the University of Minnesota and the Local Medical Ethics Committee at Maastricht University Medical Center. Seventeen PD patients (five female, mean age = 62 years), were scanned on a 7T MR scanner (Magnetom 7T Siemens). Three scans were performed: a T1-weighted, a T2-weighted, and a diffusion-weighted scan performed twice with opposite phase encoding directions, along 50 directions with a b-value of 1500 s/mm 2 and 4 b0-volumes (see Table 1 ). Five patients were scanned with slight deviations to this protocol. The motor and non-motor territories of each patient’s STN were identified by tracing their structural connections to the limbic, associative, motor and remaining cortical areas. To this end, the cortex of a reference brain ( Grabner et al., 2006) , was manually divided into these four areas ( Fig. 1 A,B) and registered to each patient’s T1-weighted image. The STNs were manually delineated on each patient’s T2-weighted image and registered to the T1-space as well. Finally probabilistic fiber tracking with each STN as a seed region was performed in T1-space with the FMRIB Software Library (FSL) ( Jenkinson et al., 2012) based on a three-fiber model ( Behrens, 2007) . Results The STNs, with volumes ranging from 75.8 mm 3 to 183.9 mm 3 , were clearly visible on the T2-weighted images( Fig. 1 C) and a clear segregation of the territories was observed ( Fig. 1 D). In all 34 STNs, the superior posterolateral area was connected to the motor cortical area. The volume of the motor area varied from 17.2% to 78.7% of the total STN volume, with an average of 55.3%. In all 34 STNs, the overlapping region in the center was connected to the associative cortical area. Connections to the limbic cortical areas were found in 30 STNs, all from the inferior anteromedial tip of the STN, but with varying sizes. Finally connections to the remaining cortical areas were only found sporadically. Conclusions These results show that ultra-high field MRI can be used to identify the motor and non-motor areas of the STN for each patient individually. The great variability between patients, in volumes of the STN and the motor area also indicate that this mapping should be performed on an individual patient basis. Further studies are currently underway to infer if these maps indeed improve clinical outcome.