819 White Matter Activation Pathways in STN and GPi DBS for Parkinson’s Disease

Abstract

INTRODUCTION: The subthalamic nucleus (STN) and globus pallidus interna (GPi) are common deep brain stimulation (DBS) targets for Parkinson’s disease (PD). DBS is thought to function by modulating downstream circuits, yet targeting is based on anatomy rather than structural connectivity. While tractography using normative connectomes has been used to study the impact of the hyperdirect pathway, cortical connectivity patterns only account for 20% of the variance in DBS response. Less is known about subcortical and cerebellar connectivity, and patient-specific tractography in DBS for PD. METHODS: Preoperative T1-weighted and diffusion-weighted MRI, postoperative CT, and clinical data were acquired for 49 patients (24 STN, 25 GPi) who underwent DBS for PD. After initial programming, the Lead DBS toolbox was used to model VAT based on the individual subject electrode and stimulation parameters. Whole-brain probabilistic tractography was performed from VAT using FSL (FMRIB's Software Library) to estimate activation pathways. Average activation pathways were generated for each target by weighting each subject’s pathways by unilateral MDS-UPDRS-III response to stimulation at initial programming. Comparison tractography was performed using 85-subject PD normative connectome. RESULTS: Analysis of weighted activation pathways reveals that STN-DBS strongly activates pathways to the caudate nucleus and bilateral cerebellum. In contrast, GPi-DBS activates pathways to the putamen and the ansa lenticularis. Furthermore, normative connectome tractography shows that the subcortical and cerebellar structural connectivity is not as well appreciated. Lastly,cortical pathways show that GPi DBS has increased premotor connectivity while STN DBS has robust motor connectivity, and confining the seeds to the STN and GPi does not diminish cortical connectivity. CONCLUSIONS: Activation pathways from STN and GPi therapeutic contacts have unique subcortical and cerebellar connectivity. White matter connections within the basal ganglia and cerebellum, including ansa lenticularis and regions associated with the DRTT, are connected to the therapeutic contacts, and this connectivity is not as well appreciated using normative connectome data. Activation pathways via the internal capsule alone cannot explain cortical connectivity. Further research to test the utility of defined connections using patient-specific tractography may help better predict DBS response or prospectively inform targeting.