Resting-state connectivity predicts patient-specific effects of deep brain stimulation for Parkinson’s disease

Neural circuit-based guidance for optimizing patient screening, target selection and parameter tuning for deep brain stimulation (DBS) remains limited. To this end, we propose a functional brain connectome-based modeling approach that simulates network-spreading effects of stimulating different brain regions and quantifies rectification of abnormal network topology in silico. We validate these analyses by predicting nuclei in basal-ganglia circuits as top-ranked targets for 43 local patients with Parkinsons disease and 90 patients from public database. However, individual connectome-based predictions demonstrate that globus pallidus and subthalamic nucleus (STN) constituted as the best choice for 21.1% and 19.5% of patients, respectively. Notably, the priority rank of STN significantly correlated with motor symptom severity in the local cohort. By introducing whole-brain network diffusion dynamics, these findings unfold a new dimension of brain connectomics and underscore the importance of neural network modeling for personalized DBS therapy, which warrants experimental investigation to validate its clinical utility.