Related Experiment Video
Updated: May 4, 2026
14:14
Targeting Neuronal Fiber Tracts for Deep Brain Stimulation Therapy Using Interactive, Patient-Specific Models
Published on: August 12, 2018
11.2K
Robustness of Lead Reconstruction for Deep Brain Stimulation Modeling and Probabilistic Mapping
Sabry L Barlatey1, Alexis P R Terrapon2, Gerd Tinkhauser3
1Department of Neurosurgery, Bern University Hospital and University of Bern, Bern, Switzerland, sabry.barlatey@insel.ch.
Stereotactic and Functional Neurosurgery
|May 2, 2026
Summary
Lead reconstruction for deep brain stimulation (DBS) is robust across different CT scans, even with patient movement. This ensures reliable automated programming for Parkinson's Disease and Essential Tremor.
Area of Science:
- Neurosurgery
- Medical Imaging
- Computational Neuroscience
Background:
- Deep brain stimulation (DBS) is an effective therapy for various neurological conditions.
- Advancements in directional leads increase programming complexity, necessitating automated solutions.
- Probabilistic mapping and computational lead reconstruction are key for predicting DBS parameters.
Purpose of the Study:
- To systematically assess the robustness of computational lead reconstruction across distinct postoperative CT scans in DBS patients.
- To evaluate the impact of image set variability on lead tip localization and volume of tissue activation (VTA) prediction.
- To determine the reliability of lead reconstruction for automated programming algorithms in Parkinson's Disease (PD) and Essential Tremor.
Main Methods:
- Retrospective analysis of 34 DBS patients (PD/Essential Tremor) with two postoperative CT scans.
- Independent processing of each CT scan using the Lead-DBS toolbox for lead reconstruction.
- Comparison of lead tip coordinates and VTA between distinct image sets, including assessment of group-level probabilistic maps for PD patients.
Main Results:
- Mean lead tip translation between CT scans was 0.79mm, with no significant impact from pneumocephalus.
- Individual-level VTA comparison showed a mean Dice coefficient of 0.73, decreasing with lower stimulation amplitudes.
- Group-level analysis demonstrated robust N-images (Dice 0.88) and clinical improvement maps (Dice 0.90), indicating reliability for probabilistic sweet spot identification.
Conclusions:
- Current computational lead reconstruction routines are sufficiently robust for probabilistic sweet spot identification in DBS programming.
- While individual VTA variability exists, it diminishes at the group level, supporting the use of automated programming.
- Further investigation into sources of individual-level variability, such as CT-to-MRI co-registration accuracy, may refine future programming strategies.
