Andreas Horn

Please support Lead-DBS by buying our Lego Kit. Its also fun!

Andreas Horn breaks down recent exchanges on the lesion network mapping method.

Just a reminder that Trump didn't just bumble us into an unnecessary war -- he deliberately created the conditions that he's now screaming about

Cool initiative by @micahgallen.com to restore what Elon destroyed: bsky.app/profile/did:...

This week's cover @thelancet.com

Yes exactly. Totally agree w you!

So maybe taking these as null models isnt too sensible either, no?

Yes, I know. As always, your insights were spot on and I found the Edelman insights super helpful, as well. I do wonder, however, whether the properties of the connectome are truly meaningless, biologically. I dont think so. They may say something about "default" modes of the brain.. – insightful!

a logo that says " spread the word " on a white background ALT: a logo that says " spread the word " on a white background

Since @natneuro.nature.com did not give us the opportunity to comment on the article when it was released, we currently stand preprint against high-impact article in terms of visibility.

We would therefore very much appreciate if you could spread the word about our reply – Thank you so much!

a man is standing in front of a blackboard and says i am the smartest man alive . ALT: a man is standing in front of a blackboard and says i am the smartest man alive .

…how can one ever be so sure to truly be the smartest person in the room?

The general debate is healthy and welcome – but it may be good advice to stay humble and constructive when criticizing a method from the outside without real hands-on experience.

I must say that it does puzzle me a bit that some folks believe that ~200 smart people across the world used a method for ~10 years across ~200 papers and simply did not see its fundamental flaws – that can then be quickly revealed with some basic math and permutation testing..

Methodological flaw may upend network mapping tool The lesion network mapping method, used to identify disease-specific brain networks for clinical stimulation, produces a nearly identical network map for any given condition, according to a new study.

To me, these reports stand in stark contrast to quotes e.g. by Luca Cocchi that “the method itself is probably not salvageable” www.thetransmitter.org/brain-imagin...

While these suggestions (by Petersen & Edelman) are certainly great tools in our belt to investigate results rigorously, our own analysis shows that the method – as currently implemented at least in the majority of papers – is sound.

Dimensionality reduction establishes specificity in lesion network mapping Background Lesion network mapping (LNM) has emerged as a powerful tool for linking focal brain lesions to distributed functional networks. However, the biological specificity of these networks has bee...

Edelman et al. also suggested constructive control analyses that could be helpful: www.medrxiv.org/content/10.6...

Brain stimulation and brain lesions converge on common causal circuits in neuropsychiatric disease - Nature Human Behaviour Which brain circuits are causally involved in depression? Using the human connectome as a wiring diagram, Siddiqi et al. combine data from lesions, deep brain stimulation and transcranial magnetic sti...

…This can be seen as a generalization of the methods already taken in the LNM field (e.g. @shansiddiqi.bsky.social's permutation based network comparisons introduced in www.nature.com/articles/s41...

@petersenm.bsky.social et al. have suggested to create better null models re the properties of the human connectome (maybe a good thing for almost any fMRI result)?
www.medrxiv.org/content/10.6...
…

Others have also commented on the article. For example, Meng et al. see the original paper potentially even more critical than us: www.biorxiv.org/content/10.6...

Nonetheless, we are generally excited about attempts to critically evaluate the LNM technique and the article has motivated a lot of control analyses across labs across the world, which is a healthy undertaking and debate for any scientific method.

ii) the only metric implied by the paper was whole-brain spatial correlations of unthresholded maps. For almost any fMRI derived map (task-fMRI, rs-fMRI, etc) this will always be comparably low-dimensional. Significance thresholding of results is a standard measure in the field of fMRI.

Without adding a specificity analysis step, this compares to doing a task-fMRI experiment without a control condition. Without a baseline contrast, the analysis will favor connections often hit by stroke / dominant modes of the connectome.

There are numerous issues but the main two are:
i) the paper only analyzed sensitivity analysis of LNM. As the name implies, this first step is meant to be inclusive and comparing networks of this step will lead to higher similarity across findings.

A recent @natneuro.nature.com paper analyzed lesion network mapping and raised concerns about the validity of the method.
See below 👇 for our response.

www.biorxiv.org/content/10.6...

That is a very good question!

EVEN MORE D4TA - Clear Vinyl Clear special limited edition vinyl. Only 400 being made worldwide! How much is too much? For Moderat, a group that have never been shy about going big—or opening up their production vault—the idea of...

Why stop at More Data when Even More Data exists?

moderat.rosecityworks.com/products/eve...

Could extend iEEG to DBS/iEEG :)

I recently came into possession of a new item for my office, which will undoubtedly become a critical all-purpose fix for our scientific projects.

If, in the future, folks at @netstim.org should ever require more data, they can just reach behind the sofa and get it!

(in case you shouldn't know moderat, they are certainly worth checking out)

Fantastic talk by @barbarahollunder.bsky.social on weaving together findings across studies in the field of connectomic DBS at the Würzburg expert summit!

Netstim on route to the DBS summit in Würzburg! Looking forward to seeing everybody soon!

For the first time, we characterized the Parkinson's Disease Response Network across space and time, simultaneously. Fantastic work led by @bahnebahners.bsky.social out in BRAIN now 👇👇