Background: The threshold current for inducing muscle contractions by stimulation of pyramidal tract fibres adjacent to the globus pallidus internus (GPi) is, besides microelectrode recordings for the determination of nuclear boundaries, currently the only neurophysiological marker for intraoperative refinement of the anatomically planned target point for pallidal deep brain stimulation (GPi-DBS) in dystonia. Objectives: To determine the relationship between intraoperative thresholds for muscle contractions under general anaesthesia and postoperative thresholds in GPi-DBS. Methods: Intraoperatively, current amplitude thresholds (120 µs, 130 Hz) were determined in 6 dystonic patients under general anaesthesia (through the uninsulated tip of the microelectrode guide tube). Postoperative localization of chronic stimulation electrodes by MRI and image fusion with the stereotactic planning determined the stimulation contact for comparing thresholds with intraoperative values. Results: Current thresholds were 3.3 ± 0.8 mA intraoperatively (follow-up 0, FU0; n = 12), 2.9 ± 1.2 mA within 1 week after surgery (FU1; n = 12), and 3.5 ± 1.6 mA after 6-17 months (FU2; n = 8). FU0 and FU1 differed by trend, and FU1 and FU2 were significantly different (Friedman test, p = 0.0048; post hoc Dunn multiple comparison test, p < 0.05). FU0 and FU2 were not different. Discussion: The threshold amplitude to induce tonic muscular contractions may constitute a valid approach of functionally refining the anatomically guided electrode placement in GPi-DBS for dystonia, because intraoperative values are predictive for postoperative thresholds with the chronically implanted neurostimulation system.