Connectivity-based selection of optimal deep brain stimulation contacts: A feasibility study.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Neurology

Background:

• Deep brain stimulation (DBS) programming is complex and relies on experienced clinicians observing acute effects.
• Optimizing DBS parameters, such as contact location, is crucial for therapeutic success and minimizing side effects.

Purpose of the Study:

• To investigate if cortical connections of DBS contacts can predict stimulation efficacy.
• To develop a method for estimating optimal DBS stimulation locations based on brain connectivity.

Main Methods:

• Whole-brain probabilistic tractography was used to map cortical connections of DBS contacts in 38 Parkinson's disease patients.
• A support vector machine classifier was trained to categorize contacts based on their connections to efficacy-related vs. side effect-related cortical regions.
• Connectivity-based classifications were validated against actual stimulation outcomes using receiver-operating characteristic (ROC) curves.

Main Results:

• Distinct cortical clusters were identified for stimulation-induced efficacy and side effects.
• The classifier accurately predicted efficacious DBS contacts in a training dataset (42/47 contacts).
• This predictive accuracy was successfully replicated in an independent test cohort (Area Under ROC = 0.83).

Conclusions:

• Cortical connectivity patterns can reliably predict the efficacy of deep brain stimulation contacts.
• This connectivity-based approach shows potential for facilitating and optimizing DBS programming.
• Further prospective studies are warranted to confirm clinical utility, particularly with directional DBS electrodes.