丘脑底核
脑深部刺激
概率逻辑
刺激
体素
医学
计算机科学
帕金森病
人工智能
病理
内科学
疾病
作者
Ta Huynh Duy Nguyen,Andreas Nowacki,Ines Debove,Katrin Petermann,Gerd Tinkhauser,Roland Wiest,Michael Schüpbach,Paul Krack,Claudio Pollo
标识
DOI:10.1016/j.brs.2019.05.001
摘要
Directional deep brain stimulation (dDBS) of the subthalamic nucleus for Parkinson's disease (PD) increases the therapeutic window. However, empirical programming of the neurostimulator becomes more complex given the increasing number of stimulation parameters. A better understanding of dDBS is needed to improve therapy and help guide postoperative programming.To determine whether clinical effects of dDBS can be predicted in individual patients based on lead location and volume of tissue activated (VTA) modelling.We analysed a prospective series of 28 PD patients. Imaging analysis and systematic clinical testing performed 4-6 months postoperatively yielded location, clinical efficacy and corresponding therapeutic windows for 272 directional contacts. We calculated the corresponding VTAs to build a probabilistic stimulation map using voxel-wise statistical analysis.We found a positive and statistically significant correlation between the overlap ratio of a patient's individual stimulation volume and the probabilistic map's sweet spot -defined as the 10% voxels with the highest clinical efficacy values (average Spearman's rho = 0.43, average p ≤ 0.036). Patients who had a larger therapeutic window with directional compared to omnidirectional stimulation had a larger distance between the electrode and the sweet spot centroid (average distances 2.3 vs. 1.5 mm, p = 0.0019).Our analysis provides new insights into how the definition of a probabilistic sweet spot based on directional stimulation data and individual VTA modelling can be applied to predict clinically effective directional stimulation and help guide clinicians with the intricate postoperative DBS programming.
科研通智能强力驱动
Strongly Powered by AbleSci AI