Sign in

Manjari Narayan | @Neurostats

@neurostats.org
7.4K followers 861 following 1.6K posts

AI in Bio & Health & Therapeutic Development Bio: linktr.ee/mnarayan Substack: blog.neurostats.org Peek into my brain: notes.manjarinarayan.org Previously @dynotx @StanfordMed PhD@RiceU_ECE | BS@ECEILLINOIS 🧪🧮⚕️🧬🧠🖥🤖📈✍️🩺👩‍📈📉

PostsRepliesMedia
Reposted by Manjari Narayan | @Neurostats
Peter Tennant @pwgtennant.bsky.social · 17/07/2025
I think the problem here is that most people who do epidemiological research - including many who might call themselves 'epidemiologists' - have not actually had any epidemiology training. I wish it was a protected title. With an exam that would require proving you understand residual confounding!
2121
Manjari Narayan | @Neurostats @neurostats.org · 17/07/2025
Otherwise, I'm more considered about pre-validation issues — correcting for chronological age without nested cross-validation in biomarker prediction models. The nonlinear thing becomes an issue if you want to make a claim that residuals are independent of age.
030
Reposted by Manjari Narayan | @Neurostats
Thomas Dietterich @tdietterich.bsky.social · 26/06/2025
For example, current evidence is that LLMs rely too much on interpolation rather than discovering underlying invariant causal abstractions.
3101
Manjari Narayan | @Neurostats @neurostats.org · 27/06/2025
📌
010
Manjari Narayan | @Neurostats @neurostats.org · 27/06/2025
Same goes for structure determines function concepts
000
Manjari Narayan | @Neurostats @neurostats.org · 27/06/2025
I think Cox is earnest. It is scary to listen to him because he is right that the AI presentation of "facts" is largely based on what is popular, but you can also steer it to selectively give you evidence that confirms your own worldview.
020
Manjari Narayan | @Neurostats @neurostats.org · 27/06/2025
This appears to be a very problematic weaponization of casual reasoning that feels 'rigorous' because it involves a selective presentation of evidence. Some very excellent researchers have looked into the problems from air pollution. #CausalSky #MetaSky www.theguardian.com/technology/n...
theguardian.com
Inside a plan to use AI to amplify doubts about the dangers of pollutants
Risk analyst Tony Cox’s work has been backed by the chemical lobby, and some health experts are alarmed
121
Manjari Narayan | @Neurostats @neurostats.org · 25/06/2025
📌
000
Manjari Narayan | @Neurostats @neurostats.org · 22/06/2025
What year was this picture taken in?
000
Manjari Narayan | @Neurostats @neurostats.org · 22/06/2025
Exactly > The people who care about a topic enough to research it will often have ideological motivations. The ideological influence is there. That doesn’t mean it’s bad science.
010
Manjari Narayan | @Neurostats @neurostats.org · 22/06/2025
Every few years there is a new article like this, but not much will change unless scientists stop relying on GraphPad Prism that doesn't offer any of this and have something they can use more easily. LLM powered workflows might change this going forward.
060
Manjari Narayan | @Neurostats @neurostats.org · 22/06/2025
📌
000
Manjari Narayan | @Neurostats @neurostats.org · 21/06/2025
The grass is not greener, I'm afraid.
020
Manjari Narayan | @Neurostats @neurostats.org · 21/06/2025
It's a niche for sure. But it exists and it's underutilized and could be further developed
010
Manjari Narayan | @Neurostats @neurostats.org · 21/06/2025
📌
000
Manjari Narayan | @Neurostats @neurostats.org · 21/06/2025
Unless you are using inferential clustering methods that are designed to severely test any hypotheses against the baseline that "how often would you find clusters if the data was generated by these high dimensional null patterns or if data were corrupted by structured measurement error"
172
Manjari Narayan | @Neurostats @neurostats.org · 21/06/2025
Can you demonstrate the extent and degree of different kinds of sequential selection biases from a #CausalSky #StatSky #MedSky perspective in public clinical trial datasets? DM me if you are interested. For instance in TrialBench? github.com/ML2Health/ML...
github.com
ML2ClinicalTrials/Trialbench at main · ML2Health/ML2ClinicalTrials
Contribute to ML2Health/ML2ClinicalTrials development by creating an account on GitHub.
010
Reposted by Manjari Narayan | @Neurostats
Evgeni V Pavlov @evgenivpavlov.bsky.social · 20/06/2025
She wrote this in May…
open.substack.com
Can’t take it with you
I am in hospice care and reflecting a lot on what a good life is.
110417
Reposted by Manjari Narayan | @Neurostats
Marco A. Maximo Prado @maximoprado.bsky.social · 20/06/2025
the Mesoscopic Integrated Neuroimaging Data (MIND) platform will soon be opening also for outside researchers leveraging our new 15.2 T 🐁 MRI, SHIELD clearing and fully upgraded lightsheet microscope plus computational resources developed by @neuroak.bsky.social to handle large datasets.
0152
Manjari Narayan | @Neurostats @neurostats.org · 20/06/2025
📌
000
Manjari Narayan | @Neurostats @neurostats.org · 20/06/2025
📌
000
Manjari Narayan | @Neurostats @neurostats.org · 19/06/2025
I also think most of the authors were not solid neuroimaging researchers. It was like a fun project. Not particularly good science.
061
Manjari Narayan | @Neurostats @neurostats.org · 19/06/2025
Neurobullshit is alive and well in education research. Reminds me of the good old days of @neuroskeptic.bsky.social on this 10+ years ago
2294
Manjari Narayan | @Neurostats @neurostats.org · 18/06/2025
substack.com/@manjarinara...
substack.com
Manjari Narayan on Substack
The part Box said that is often left out. “The practical question is how wrong do (models) have to be to not be useful.”
1112
Manjari Narayan | @Neurostats @neurostats.org · 18/06/2025
📌
000
Manjari Narayan | @Neurostats @neurostats.org · 18/06/2025
Also, longitudinal studies don't fix causality problems. Fortunately, they help you discover entirely new kinds of confounding.
0111
Manjari Narayan | @Neurostats @neurostats.org · 18/06/2025
📌
000
Reposted by Manjari Narayan | @Neurostats
Ted Underwood @tedunderwood.com · 18/06/2025
If you use a model dialogically, it’s effectively guiding you through the free-writing/revision cycle composition teachers recommend. I get why this mode of use isn’t very visible. All the rhetoric around AI has been that it’s “generative” and does the work for you. Things like the “Dear Sydney” +
4453
Manjari Narayan | @Neurostats @neurostats.org · 18/06/2025
I feel vindicated
010
Manjari Narayan | @Neurostats @neurostats.org · 17/06/2025
cc: @jackiesullivan.bsky.social @kordinglab.bsky.social nglab.bsky.social and Paul Cisek.
120
Manjari Narayan | @Neurostats @neurostats.org · 17/06/2025
5 years later, it is finally on youtube. Thanks @ohbmofficial.bsky.social youtu.be/X5syQN9ZkQE?... #neuroskyence #CausalSky
180
Manjari Narayan | @Neurostats @neurostats.org · 17/06/2025
I wish PDFs were on arxiv rather than academia.edu. It's not accessible this way.
000
Manjari Narayan | @Neurostats @neurostats.org · 17/06/2025
👀 A performative prediction paper
010
Manjari Narayan | @Neurostats @neurostats.org · 17/06/2025
Are there things that ought to be in the DAGs that people don't know to put in the DAGs are the kind of thing I care about here.
120
Reposted by Manjari Narayan | @Neurostats
A. Brad Schwartz @abradschwartz.bsky.social · 17/06/2025
What’s common knowledge in your field but shocks outsiders? The vast majority of archival documents are neither digitized nor searchable online. 🗃️
412423
Manjari Narayan | @Neurostats @neurostats.org · 17/06/2025
Reductionism never falls out of fashion because one has to embrace theory to question it.
040
Manjari Narayan | @Neurostats @neurostats.org · 17/06/2025
CEP ~ 1 - Var(Effect on Outcome | Effect on Proxy) / Var(Effect on Outcome)
110
Manjari Narayan | @Neurostats @neurostats.org · 17/06/2025
Cool will have to read more carefully. Hypothetical question: if someone were using adverse events data as a way of verifying the causal effect predictiveness of non-clinical / preclinical methods of forecasting of toxicity or adverse effects, do you see any problems?
110
Manjari Narayan | @Neurostats @neurostats.org · 17/06/2025
FDR metrics only offer error-control of the false discovery proportion under expectation. It is already a liberal standard of evidence when you realize that. Most don't. If one wants to defend a paper in nature, you might find AI reasoning agents quite useful to check your work. #StatSky 🖥️ 🧬 📈💻
010
Manjari Narayan | @Neurostats @neurostats.org · 16/06/2025
📌 very interesting. Is there a way to do this that changes what actions scientists might take in doing follow-up studies?
110
Reposted by Manjari Narayan | @Neurostats
Philip Ball @philipcball.bsky.social · 24/05/2025
What was interesting to me was that, despite my best efforts, SCB was unable to understand that this was a framing of the observations, and not simply a description of them. The response was that to be curious about this framing revealed total ignorance of the field, despite the following... /3
1461
Manjari Narayan | @Neurostats @neurostats.org · 16/06/2025
Awesome answers. Can we have a follow-up meta science workshop?
031
Manjari Narayan | @Neurostats @neurostats.org · 16/06/2025
The generous variation is that the assumption is mechanisms are separable, can be studied independently and then combined together in a way that will converge to the truth.
120
Reposted by Manjari Narayan | @Neurostats
Cian O'Donnell @cianodonnell.bsky.social · 07/06/2025
Apart from at a few over-plucked trees, for many subtopics it is not clear how to make a good theoretical contribution to the area. Involves inventing new plucking tools, deciding which fruit are both within grasp and edible for your experimental colleagues
1161
Manjari Narayan | @Neurostats @neurostats.org · 16/06/2025
@stuartbuck.bsky.social @ruxandrabio.bsky.social know anyone who would care about this?
000
Manjari Narayan | @Neurostats @neurostats.org · 16/06/2025
Instead I think the reframe needs to be a change in the interpretation of regulating drugs that only work in combination with a non-pharmacological therapy under circumstances that need to be discovered as part of therapeutic development. #MedSky #CasualSky #PharmaSky @fda.gov
100
Manjari Narayan | @Neurostats @neurostats.org · 16/06/2025
This is very interesting. It is indeed problematic that psychotherapy efficaciousness is not being evaluated. But I'm not sure about the anesthetic analogy. The problem is that the two together represent a kind of non-additive therapeutic interaction in a way that isn't necessary with anesthesia.
101
Reposted by Manjari Narayan | @Neurostats
Maria Glymour @mariaglymour.bsky.social · 15/06/2025
I wish some of the folks with terminated grants- especially Harvard, Columbia folks - would post the summary statements from NIH study sections about their grants. The public should know what independent scientists thought about the projects that are being terminated. #PublicHealth #EpiSky #EndAlz
110034
Reposted by Manjari Narayan | @Neurostats
Zeta Of 1 @zetaof1.bsky.social · 15/06/2025
One reason I hate the term "Polynomial Regression" so much is because linear regression means "linear in parameters", not "linear in variables". I'm not calling it logarithmic regression either when logging some variables. Still linear regression. Stop making squaring a variable sound like magic.
2423
Manjari Narayan | @Neurostats @neurostats.org · 15/06/2025
Capacity to notice implicit assumptions would be a good start. bsky.app/profile/laur...
130