Clinical Scorecard: Electrophysiological Markers Indicate the Optimal Therapeutic Range for Deep Brain Stimulation Electrode Contacts

At a Glance

Key Highlights

• Machine learning using electrophysiological markers can predict optimal DBS electrode contacts.
• STN power in fast frequency bands (>35 Hz) and STN-cortex coherence are key predictors of therapeutic window.
• Automated contact selection may reduce time and complexity of DBS programming compared to conventional monopolar review.

Guideline-Based Recommendations

Diagnosis

• Use electrophysiological recordings (MEG and LFP) from STN to assess neural oscillations related to Parkinson’s symptoms.

Management

• Apply machine learning models incorporating STN power and STN-cortex coherence to guide DBS contact selection.
• Perform monopolar review to clinically validate therapeutic windows and side effect thresholds.

Monitoring & Follow-up

• Monitor neural oscillatory activity in multiple frequency bands to optimize stimulation parameters.
• Assess symptom relief and side effects incrementally during programming.

Risks

• Suboptimal contact selection may activate non-target brain areas causing side effects and limiting therapeutic window.

Patient & Prescribing Data

Parkinson’s disease patients implanted with DBS electrodes targeting the STN

Electrophysiological features can predict therapeutic windows, enabling faster and potentially more precise DBS programming.

Clinical Best Practices

• Combine multiple electrophysiological markers rather than relying on a single signal such as STN beta power.
• Incorporate both subthalamic activity and subthalamo-cortical coherence for comprehensive assessment.
• Use machine learning tools to support and accelerate the monopolar review process.

References

• Roediger et al. - Automated DBS programming clinical trial
• Sarikhani et al. - DBS programming optimization via smartwatch accelerometry
• STN beta power as feedback signal for closed-loop DBS