Subthalamic and pallidal deep brain stimulation: are we modulating the same network?

The subthalamic nucleus and internal pallidum are main target sites for deep brain stimulation in Parkinson's disease. Multiple trials that investigated subthalamic versus pallidal stimulation were unable to settle on a definitive optimal target between the two. One reason could be that the effect is mediated via a common functional network. To test this hypothesis, we calculated connectivity profiles seeding from deep brain stimulation electrodes in 94 patients that underwent subthalamic and 28 patients with pallidal treatment based on a normative connectome atlas calculated from 1000 healthy subjects. In each cohort, we calculated connectivity profiles that were associated with optimal clinical improvements. The two maps showed striking similarity and were able to cross-predict outcomes in the respective other cohort (R = 0.37 at P < 0.001; R = 0.34 at P = 0.032). Next, we calculated an agreement map, which retained regions common to both target sites. Crucially, this map was able to explain an additional amount of variance in clinical improvements of either cohort when compared to the maps calculated on each cohort alone. Finally, we tested profiles and predictive utility of connectivity maps calculated from different motor symptom subscores with a specific focus on bradykinesia and rigidity. While our study is based on retrospective data and indirect connectivity metrics, it may deliver empirical data to support the hypothesis of a largely overlapping network associated with effective deep brain stimulation in Parkinson's disease irrespective of the specific target.

Automated summary

This study investigated the connectivity profiles of deep brain stimulation electrodes in patients with Parkinson's disease and found a common functional network between subthalamic and pallidal stimulation that is associated with optimal clinical improvements. Furthermore, the study showed that connectivity maps based on different motor symptom subscores can predict treatment outcomes.