Sign-specific stimulation "hot" and "cold" spots in Parkinson's disease validated with machine learning
Creators
- 1. Joint Department of Medical Imaging, University of Toronto, ON, Canada; University Health Network, Toronto, ON, Canada
- 2. University Health Network, Toronto, ON, Canada
- 3. Department of Neurology, Charité – Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität zu Berlin, and Berlin Institute of Health, Germany
- 4. Department of Neurology, Charité – Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität zu Berlin, and Berlin Institute of Health, Germany; Berlin School of Mind and Brain, Humboldt-Universität zu Berlin, Germany; Deutsches Zentrum für Neurodegenerative Erkrankungen, Berlin, Germany; Neurocure Cluster of Excellence, Charité – Universitätsmedizin Berlin, Berlin, Germany
- 5. Edmond J. Safra Program in Parkinson's Disease, Morton and Gloria Shulman Movement Disorders Clinic, Toronto Western Hospital, UHN, Division of Neurology, University of Toronto, Toronto, Ontario, Canada
- 6. University Health Network, Toronto, ON, Canada; Department of Neurosurgery, University of Toronto, Toronto, Ontario, Canada; Center for Advancing Neurotechnological Innovation to Application (CRANIA), Toronto, ON, Canada
- 7. University Health Network, Toronto, ON, Canada; Department of Neurosurgery, University of Toronto, Toronto, Ontario, Canada
- 8. Edmond J. Safra Program in Parkinson's Disease, Morton and Gloria Shulman Movement Disorders Clinic, Toronto Western Hospital, UHN, Division of Neurology, University of Toronto, Toronto, Ontario, Canada; Krembil Brain Institute, Toronto, Ontario, Canada; Center for Advancing Neurotechnological Innovation to Application (CRANIA), Toronto, ON, Canada
Description
Deep brain stimulation (DBS) of the subthalamic nucleus (STN) has become a standard therapy for Parkinson’s disease (PD). Despite extensive experience, however, the precise target of optimal stimulation and the relationship between site of stimulation and alleviation of individual signs remains unclear. We examined whether machine learning could predict the benefits in specific parkinsonian signs when informed by precise locations of stimulation.
We studied 275 PD patients who underwent STN-DBS between 2003 and 2018. We selected pre-DBS and best available post-DBS scores from motor items of the Unified Parkinson's Disease Rating Scale (UPDRS-III) to discern sign-specific changes attributable to DBS. Volumes of tissue activated (VTAs) were computed and weighted by i) tremor, ii) rigidity, iii) bradykinesia, and iv) axial signs changes. Then, sign-specific sites of optimal (“hot spots”) and suboptimal efficacy (“cold spots”) were defined. These areas were subsequently validated using machine learning prediction of sign-specific outcomes with in-sample and out-of-sample data (n=51 STN-DBS patients from another institution).
Tremor and rigidity hot spots were largely located outside and dorsolateral to STN whereas hot spots for bradykinesia and axial signs had larger overlap with STN. Using VTA overlap with sign-specific hot and cold spots, support vector machine (SVM) classified patients into quartiles of efficacy with ≥92% accuracy. The accuracy remained high (68-98%) when only considering VTA overlap with hot spots but was markedly lower (41-72%) when only using cold spots. The model also performed poorly (44-48%) when using only stimulation voltage, irrespective of stimulation location. Out-of-sample validation accuracy was ≥96% when using VTA overlap with the sign-specific hot and cold spots.
In two independent datasets, distinct brain areas could predict sign-specific clinical changes in PD patients with STN-DBS. With future prospective validation, these findings could individualize stimulation delivery to optimize quality of life improvement.
Hot and cold spots for each sign are publicly available as binary labels in NIfTI format.
Files
Files
(2.2 GB)
| Name | Size | Download all |
|---|---|---|
|
md5:8b6da31455e72a5174166d8dac6d2ea9
|
277.6 MB | Download |
|
md5:934aa4e689bc39808465b13233c4e49d
|
277.6 MB | Download |
|
md5:008b376a8b83d0e1dc13d93162eb5c73
|
277.6 MB | Download |
|
md5:0bad15d34fcca6696eb7417669ea1e55
|
277.6 MB | Download |
|
md5:e3d4f3535cdc1612f86391cc616bbd14
|
277.6 MB | Download |
|
md5:2f08907834615b5d005cdc874d2b7703
|
277.6 MB | Download |
|
md5:d898bbaefb55d52cbaa76c3db29b771f
|
277.6 MB | Download |
|
md5:72ac309ca599792e3e8be87430b47f53
|
277.6 MB | Download |