gms | German Medical Science

Objective: There is scarce knowledge available about the position of electrodes for deep brain stimulation (DBS) of the posteroventrolateral (PVL) internal pallidum (GPI).

Methods: 18 patients (3 male; mean age 42) with dystonia (4 cervical, 5 segmental, 1 axial, 3 generalized, 3 myoclonus), spinocerebellar ataxia (1), and tardive dyskinesia (1) were included. Using open-source software developed by us (J.A.K.) and others, an algorithm was generated to segment and reconstruct the pallidum three-dimensionally from a preoperative proton-dense triple-echo MRI. The borders of the pallida were correlated with intraoperative microrecordings and the stereotactic electrode position was derived from postoperative CT scans.

Results: A complete set of data could be obtained for nine patients. The average volume of the pallidum was 2.2 ± 0.2 cm³. The absolute distance of the center of gravity of the segmented pallida from the midcommissural point (MCP) was 20.2 ± 1.5 mm. Individual variation was highest in the PVL pallida at the level of MCP, in particular in the lateral direction. The distance of the center of gravity of the pallida from MCP was positively correlated with the width (r = 0.76, p = 0.017) and length of the third ventricle (r = 0.804, p = 0.009). The borders of the pallidum as determined by microrecordings differed from MRI-based segmentations on average by 2.3 ± 1.2 mm (entrance) and 1.2 ± 0.9 mm (exit). In most instances the electrophysiological borders were found more ventrally. The active contacts were found within the ventral pallidum (9 of 18 electrodes) or below the pallidal base. The absolute distance between the center of electrical current and the ventral base of the pallidum was 1.8 ± 1.3 mm. Most active contacts were located within or below the external segment of the pallidum as interpolated from a digitized version of the Morel atlas co-registered with the segmented pallida. The average location of the active contacts of all patients (n = 18) relative to MCP (in mm) was: x = 21.6 ± 1.2; y = 4.1 ± 4.1; z = -3.7 ± 1.5. The location of active contacts did not correlate with improvement of dystonia.

Conclusions: Individual variation in the PVL pallidum is high. Although our data are in accordance with the anti-dystonic effects of the ventral pallidal region, it is unclear whether the external pallidum proper mediates therapeutic effects or if it merely was targeted to assure an adequate distance from the corticospinal tract. Stimulation outside the targeted nuclei within adjacent fiber tracts has also been described for other targets.