Comprehensive Summary
This study investigates how to improve DBS programming for Parkinson's and how local field potential (LFP) recordings from implanted deep brain stimulation (DBS) electrodes can help identify the best stimulation contact in Parkinson’s. The research was a systematic review of human studies and the authors searched 9 databases up to November 2023, screening hundreds of papers, including 25 relevant ones. They found that high frequency oscillations occasionally showed beneficial effects. While other single frequency features, such as theta, alpha, gamma, often were weaker when compared to beta band power which consistently predicted optimal DBS contacts, outperforming random selection in most studies. By combining multiple different LFP features they were able to be more accurate and reliable than with just a single marker. They found that the type of electrode had some influence on the results but the timing of recording or state of medication mattered less and concluded that multiple features and machine learning methods are the best path going forward in this field.
Outcomes and Implications
Programming DBS is extremely time consuming when trying to find the best contact, but through automated LFP methods, the time could be significantly reduced. Not only will LFP based methods improve the time it takes to find optimal contact regions but could also find more accurate regions than those identified by standard programming which would help motor outcomes and reduce side effects significantly. LFP programming already has shown comparable clinical benefits to standard clinical methods in several studies. Multi-feature and machine learning methods allow for fast, objective contact selection during follow up visits and in early in the treatment process. This could eventually lead to widespread usage of this sort of technology but further clinical trials still need to occur before any steps are taken.