Objective: To investigate whether functional sweet spots of deep brain stimulation (DBS) in the subthalamic nucleus (STN) can predict motor improvement in Parkinson disease (PD) patients.Methods: Stimulation effects of 449 DBS settings in 21 PD patients were clinically and quantitatively assessed through standardized monopolar reviews and mapped into standard space. A sweet spot for best motor outcome was determined using voxelwise and nonparametric permutation statistics. Two independent cohorts were used to investigate whether stimulation overlap with the sweet spot could predict acute motor outcome (10 patients, 163 settings) and long‐term overall Unified Parkinson's Disease Rating Scale Part III (UPDRS‐III) improvement (63 patients).Results: Significant clusters for suppression of rigidity and akinesia, as well as for overall motor improvement, resided around the dorsolateral border of the STN. Overlap of the volume of tissue activated with the sweet spot for overall motor improvement explained R² = 37% of the variance in acute motor improvement, more than triple what was explained by overlap with the STN (R² = 9%) and its sensorimotor subpart (R² = 10%). In the second independent cohort, sweet spot overlap explained R² = 20% of the variance in long‐term UPDRS‐III improvement, which was equivalent to the variance explained by overlap with the STN (R² = 21%) and sensorimotor STN (R² = 19%).Interpretation: This study is the first to predict clinical improvement of parkinsonian motor symptoms across cohorts based on local DBS effects only. The new approach revealed a distinct sweet spot for STN DBS in PD. Stimulation overlap with the sweet spot can predict short‐ and long‐term motor outcome and may be used to guide DBS programming. ANN NEUROL 2019;86:527–538

Deep brain stimulation of the subthalamic nucleus improves non-motor symptoms in Parkinson’s disease, but with considerable inter-individual variability. Petry-Schmelzer et al. show that neurostimulation in specific subregions of the subthalamic nucleus has differential effects on mood/apathy, attention/memory and sleep-related outcomes. Neurostimulation could thus be tailored to patients’ individual non-motor profiles.