AI for deep brain stimulation still far from clinical deployment: systematic review
239 studies over 25 years show most AI systems remain at early translational stages.
A comprehensive systematic review published on arXiv by Souei et al. analyzed 239 peer-reviewed studies from 2000 to 2025 to assess the state of artificial intelligence in deep brain stimulation (DBS) for movement disorders. The research landscape is heavily skewed toward Parkinson’s disease and the subthalamic nucleus target, with minimal coverage of other disorders or brain targets. Most studies reported strong internal performance metrics, but external validation was extremely rare, evaluations were predominantly retrospective and single-centre, and over a quarter of the datasets were small-sample, high-dimensional — raising significant overfitting risks.
The authors performed a technology readiness assessment and found that most AI systems are still at early-to-intermediate translational stages. The primary bottleneck is not algorithmic inadequacy but rather limited validation and the inherent biological heterogeneity and dynamic complexity of DBS. Nonetheless, a small but growing number of external and prospective studies suggest the field is moving toward clinical maturity, with promising applications in targeting, programming, outcome prediction, and adaptive therapy delivery. The review concludes that concerted efforts in multi-centre validation and regulatory alignment are needed to bridge the gap to real-world deployment.
- 239 studies reviewed; 25+ years of research but only ∼5% had external validation.
- Parkinson’s disease and subthalamic nucleus targeting make up the vast majority of work.
- Over 25% of studies used small-sample, high-dimensional data, raising overfitting risk.
Why It Matters
AI-powered DBS could personalize treatment for Parkinson’s, but without rigorous validation, patient safety and efficacy remain uncertain.