|
This one was a tough one – but Ningfei just pushed v2.5 to our servers! |
| |
Unfortunately, a bug first
there are tons of updates – but before we start, we must report there has been a bug in Lead-DBS that applied for reconstruction of segmented electrode orientations of segmented Abbott electrodes in native space. Currently, to our knowledge, such analyses were rare among users. No single publication we are aware of should be affected by this bug. But if your study involves reconstruction of segmented leads in native space, please update to the new version and reconstruct orientations of these directional leads, again. Feel free to contact us in case of questions. The bug does not apply to analyses carried out in template / MNI space.
Big thanks go to Till & Ningfei for identifying and fixing the bug! |
| |
|
This story should remind us that Lead-DBS is
- pure research software
- created by a small team of enthusiasts
- and we need the help of the whole community to detect bugs & improve utility of the software
As with any type of research software, users should not apply these methods thought- or carelessly but be encouraged to look under the hood, show, inspect & visualize data as much as they can and to stay in contact with developers about questions.
But now, let's see what's new! |
| |
Three pillars to map DBS effects
Within Lead-DBS v.2.5, we introduce three novel analysis tools to
- Map localized "sweet" and "sour" spots to DBS targets based on group level data
- Map improvement variables to fiber tracts
- Map improvement variables to brain networks
|
| |
|
Three new GUIs to analyze relationships between stimulation site and clinical or behavioral changes following DBS on a group level |
| |
|
We have added a rather long walkthrough tutorial that may get you started to explore the new tools in case they seem interesting.
We plan to add extensive documentation for all three tools to the Lead-DBS manual, in the near future. |
| |
Sweetspot Explorer
One of the three tools is the Sweetspot explorer and it is important to highlight the strong contributions of Till Dembek in this project. While Ningfei and myself programmed most of the GUI, Till did the actual work that is based on his current preprint comparing existing approaches to map Sweetspots in the DBS literature.
Building upon his work, we included main concepts that have been used in the literature into the new tool: |
| |
|
The aim to harmonize fundamental choices made in existing publications into a single tool |
| |
|
In this context, it's crucial to note that some of the mentioned published approaches used different concepts (such as completely different VTA models or even absence of VTA models as in the paper by Eisenstein 2014) as well as post-hoc processing (such as permutation statistics or cluster-level statistics) that are not implemented in the tool. Our aim was to guide some methodological choices by referencing published examples. "Inspiring" your analysis from one of the published papers will not mean that it will exactly reproduce them. |
| |
New developers
Enrico Opri (Emory), Johannes Achtzehn and Nanditha Rajamani (Charite) have issued first pull requests with code enhancements to the Lead-DBS project. Congratulations & thank you so much for joining our team!
Especially the contributions by Enrico should be highlighted since he made the whole pipeline of Lead-DBS fit to support unilateral leads. (He also has an amazing recent paper on closed-loop DBS with the UF Fixel team which can be found here).
The purpose of Lead-DBS is to be a tool developed by the community for the community & git technology is optimal to make that happen. Anyone, anywhere in the world should feel welcome to submit pull requests to improve our code, fix glitches and/or add new features, as many have done, in the past. See a (likely non-exhaustive) list of contributors here. If your name should be missing, please do let us know! |
| |
|
We're really excited about pull requests from new people that help to make Lead-DBS better! |
| |
TOR-PSM Atlas (Elias 2020)
The team at Toronto Western Hospital (led by Andreas Lozano, paper spearheaded by Gavin Elias) has calculated probabilistic sweetspots from a total of N = 482 patients with 5 diseases from 15 years of experience at a single center. The team has graciously shared the results as an atlas showing optimal stimulation sites for Parkinson's Disease, Dystonia, Essential Tremor & Psychiatric Diseases within Lead-DBS. |
| |
|
The TOR-PSM Atlas summarizing the experience of 15 years of one of the leading DBS centers, worldwide |
| |
TOR-signPD (Boutet 2021)
…as if the above would not be enough, there is a second paper from the same group coming soon which mapped symptom-specific improvements (bradykinesia, rigidity, tremor & axial symptoms) in PD. The endeavor was spearheaded by Alexandre Boutet & the maps are also available within Lead-DBS 2.5 and on Zenodo. |
| |
|
The TOR-signPD atlas showing probabilistic sweet- and sourspots for STN-DBS in PD |
| |
STN Sweetspots (Dembek 2019)
In a similar endeavor, the Cologne group (led by Lars Timmermann, spearheaded by Till Dembek & Jan Roediger) have analyzed symptom-specific sweetspots that were predictive in out-of-sample data. As mentioned above, Till may be one of the people that has thought most thoroughly about how to map sweetspots to the brain. Furthermore, the atlas builds upon unilateral monopolar review data and hence shows the most direct clinical relationship between stimulation site and clinical effects. |
| |
|
Probabilistic sweetspots for rigidity, akinesia & side-effects |
| |
Essential Tremor Lesion Network Atlas (Joutsa 2018)
Can brain lesions be used to identify DBS targets? Yes! Juho Joutsa, Ludy Shih and Michael Fox used lesion network mapping of stroke lesions that resulted in tremor relief to define their common network. From all voxels in the brain, this network map precisely peaked at the VIM – hence indirectly confirms the widely used DBS tremor target.
The resulting map is now part of Lead-DBS in form of an atlas that can be thresholded deliberately (see animation) and could help to better understand tremor neuromodulation effects in the brain. |
| |
|
Lesion Network Analysis of beneficial brain lesions that had improved pre-existing tremor – honing down to the VIM in the thalamus |
| |
Subcortical White Matter Pathways Atlas (Avecillas 2019)
Josue Avecillas-Chasin has done beautiful work to identify relationships between clinical improvements and stimulation sites and has aggregated tracts relevant to DBS in this novel atlas. The plan is to let the atlas grow in subsequent versions to include further bundles, but we believe it is already very helpful in its current form.
Multiple papers went into creating the atlas, that are best summarized on this page. |
| |
|
The new Subcortical White Matter Pathway Atlas |
| |
Cluster Headache Atlas (Nowacki 2020)
A team between Bern and Oxford spearheaded by Andreas Nowacki analyzed DBS effects to treat cluster headache across published cases from multiple groups and now graciously contribute results as a probabilistic atlas within Lead-DBS v2.5!
They report an overall response rate in 75% of 40 patients – potentially an optimal target map could help to revive this field? |
| |
|
Finding a sweetspot to treat Cluster Headache with DBS |
| |
Atlas of the Human Hypothalamus (Neudorfer & Germann 2020)
To map the effects of DBS effects in the hypothalamus region (as the ones described above) further, yet another beautiful atlas by the Toronto center spearheaded by Clemens Neudorfer and Jürgen Germann could become extremely helpful, in the future.
In their work, a minimum deformation averaging pipeline was employed to produce a normalized, high-resolution template from multimodal magnetic resonance imaging (MRI) datasets. This template was used to delineate hypothalamic (n = 13) and extrahypothalamic (n = 12) gray and white matter structures. |
| |
|
A high-resolution in vivo anatomical atlas of the human hypothalamic region. The animation also showcases some novel rendering possibilities introduced in Lead-DBS v2.5. The BigBrain dataset is shown in the brackground. |
| |
Pallidal Dyskinesia Atlas (Tsuboi 2020)
A team between the Fixel Institute at U Florida and the Mayo Clinic in Jacksonville spearheaded by Takashi Tsuboi with Michael Okun and Erik Middlebrooks have analyzed stimulation-induced dyskinesia in pallidal DBS for Parkinson's Disease. They were able to successfully identify regions associated with stimulation-induced dyskinesia that they graciously contributed in form of an atlas to Lead-DBS. |
| |
|
Contacts resulting in stimulation-induced dyskinesias are shown as larger red markers on the right & the estimated VTA-based region in red on the left. The 7T Ex Vivo 100 um Template is shown as background. |
| |
The Golden Electrode Award 2020
Over 85 publications have been empowered by Lead-DBS in 2020 and we are continuingly amazed about the growing creativity & genius ways the software is applied by teams world-wide. |
| |
|
From our perspective, it is fabulous to see how different labs validate each other by means of open science and sometimes just by the availability of datasets such as the atlases shown above.
How collaborations are fostered and how – for instance on our Slack channel – we can see a growing interest in these topics with people helping each other out and aggregating results across the globe.
By no means do we feel entitled to give out an award (see this as a joke) – but simply as a reading list of personal highlights – we felt that two papers were outstandingly creative and/or used Lead-DBS in truly integrative ways, last year.
The first is a study by Weaver et al.
from Washington, which showed that STN DBS reduces the current needed to elicit a motor-evoked potential using focal direct cortical stimulation, supporting the idea that STN DBS in PD may work in part through modulation of the hyperdirect pathway.
What makes the study special is that it used imaging (i.e. diffusion tractography and electrode localizations), electrophysiological recordings and direct cortical stimulation in humans to confirm a specific hypothesis. |
| |
|
In 7 of 8 patients, DBS significantly reduced the M1 direct current stimulation intensity required to elicit motor‐evoked potentials. |
| |
|
The second is a study by Kehnemouyi et al.
from the Bronte-Stewart lab at Stanford, which showed (among many things) that the overlap between the VTA model and sensorimotor STN accounted for changes in beta burst duration (which again accounted for changes in movement velocity).
What makes the study special is that it integrated effects of VTA modeling (on the specific target identified by image processing), with both behavioral changes and changes in electrophysiological biomarker expression. |
| |
|
Lead-DBS based electrode localizations & VTA models accounted for changes in beta burst duration in STN-DBS for PD |
| |
Using the Medtronic Percept? Take a look at Perceive
The new implantable sensing IPG will bring invasive neurophysiology to many clinical centers. To facilitate open research on basal ganglia and PD, Julian Neumann's group at Charité has built an MNE Python based toolbox that we think could be interesting to many of you. |
| |
|
With the perceive toolbox, you can anonymize and convert your data to BIDSstandard .edf format using #mne_bids with icn.perceive.convert_to_bids and can use all MNE python functions for processing, stats and visualization from there. |
| |
Advanced modeling of DBS effects
As mentioned in the last newsletter, we have been hard at work to create an interface between an advanced modeling pipeline, OSS-DBS, and Lead-DBS. Namely, Konstantin Butenko (from Ursula van Rienen's group in Rostock) and Ningfei Li have put in countless hours to make this happen and we are happy to say that the process is now complete. This interface is currently not public and we will need more time to properly test, document and evaluate the integration.
A similar endeavor taking place between the same labs (van Rienen & Horn) is to create a version of Lead-DBS for rodent research. |
| |
|
First tract activation models calculated by Konstantin Butenko using OSS-DBS via the interface with Lead-DBS. Fibers from the Basal Ganglia Pathway Atlas (Petersen et al. 2019) are shown, activated fibers in red, damaged fibers in purple, non-activated fibers in white. |
| |
|
As always, feel free to connect with us via Twitter, ResearchGate or Slack.
…or fork/meet us on github to contribute directly!
Best regards on behalf of the Lead-DBS team,
Andy |
| |
|
