Model-Based Comparison of Deep Brain Stimulation Array Functionality with Varying Number of Radial Electrodes and...

Area of Science:

• Neurosurgery
• Biomedical Engineering
• Computational Neuroscience

Background:

• Deep brain stimulation (DBS) uses leads with electrodes to modulate neural activity.
• Radially distributed electrodes offer potential for more precise targeting in DBS.
• Increasing electrode density presents design and programming challenges.

Purpose of the Study:

• To investigate the impact of radial electrode count on steering neural activation regions (RoA).
• To identify optimal RoA features for machine learning-based programming prediction in DBS.
• To establish design constraints for future high-density DBS arrays.

Main Methods:

• Computational modeling using finite element tissue and multi-compartment biophysical axon models.
• Simulation of 27 lead designs with 1-9 radially distributed electrodes.
• Calculation of 2D threshold-dependent RoAs and analysis of geometric features.

Main Results:

• More radial electrodes allow finer RoA steering but are limited by charge injection constraints beyond four electrodes.
• Uniform multi-cathode stimulation improves RoA shifting with more electrodes; non-uniform stimulation offers similar or better results without increased electrodes.
• Machine learning robustly classified 15 stimulation configurations using three RoA geometric features.

Conclusions:

• For clinical DBS leads, >4 radial electrodes provide marginal benefits in steering, shifting, and sculpting axonal activation.
• RoA center of mass and orientation are sufficient for robust machine learning classification of stimulation settings.
• Study findings inform design limitations for developing advanced high-density DBS electrode arrays.