Stratifying Parkinson's Patients With STN-DBS Into High-Frequency or 60 Hz-Frequency Modulation Using a Computational...

Area of Science:

• Neurology
• Computational Neuroscience
• Biomedical Engineering

Background:

• Subthalamic nucleus deep brain stimulation (STN-DBS) is effective for Parkinson's disease (PD) motor symptoms.
• Recent research suggests 60 Hz stimulation may benefit PD patients with gait disorders.
• Personalizing STN-DBS frequency settings is crucial for optimizing treatment outcomes.

Purpose of the Study:

• To develop a computational model for stratifying PD patients into distinct STN-DBS frequency settings (60 Hz vs. high-frequency stimulation [HFS]).
• To identify preoperative clinical indicators that predict optimal stimulation frequency for individual PD patients.

Main Methods:

• Retrospective analysis of preoperative MDS-UPDRS III scores from 20 PD patients undergoing STN-DBS.
• Random Forest classification algorithm used to build predictive models associating patient characteristics with stimulation frequency (60 Hz or HFS [130-185 Hz]).
• Models validated using a leave-one-out cross-validation approach.

Main Results:

• Computational models achieved 95% accuracy in stratifying patients into 60 Hz or HFS groups.
• The most effective models utilized two to three preoperative predictors.
• Gait disturbance and right hand rest tremor were the most significant predictors for classification.

Conclusions:

• Preoperative clinical indicators can be used to develop accurate computational models for STN-DBS frequency selection in PD.
• These models enable a priori stratification of PD patients into 60 Hz or HFS groups.
• This approach has the potential to enhance STN-DBS therapy utilization by tailoring it to specific PD clinical subtypes.