This one hit me in the feels today
29.06.2025 20:29 โ ๐ 11957 ๐ 1510 ๐ฌ 304 ๐ 185@mfs.ms.bsky.social
Neuroscientist, curious about math, fMRI, psychiatry, moods, and individual differences.
This one hit me in the feels today
29.06.2025 20:29 โ ๐ 11957 ๐ 1510 ๐ฌ 304 ๐ 185It's clippy again!
30.06.2025 15:20 โ ๐ 1 ๐ 0 ๐ฌ 1 ๐ 0"Machine vision systemsโwhich have neither bodily states nor conscious thoughtsโcan predict with remarkable accuracy how humans will feel about the things they look at." ๐๐ฅ
The perceptual primacy of feeling (PNAS Jan 2025) www.pnas.org/doi/10.1073/...
#neuroskyence #vision #affectsci
What's it like at some agencies right now? I'm fresh off a conference where I got to speak to a bunch of people still at some important agencies. What's the climate like? Why are you staying? Generally the responses felt like this, and we are all Maggie:
12.05.2025 20:14 โ ๐ 6 ๐ 2 ๐ฌ 0 ๐ 0All of our neural dynamics tools, software, protocols, datasets, and even hardware plans are now available in one convenient place!
Check it out and let us know what you'd like to see next: www.allenneuraldynamics.org/tools
๐ง ๐
There's also doi.org/10.1038/s442..., where ketamine vs placebo was masked by anesthesia during surgery, so there shouldn't be any placebo effect. But it was "surprisingly high" anyway. They've reanalyzed the data to look at the opioid system as a cause in doi.org/10.1016/j.ja....
06.05.2025 13:59 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0All four hardcover volumes of Fundamental Papers in Complexity Science
So very excited [these](www.foundationalpapersincomplexityscience.org) arrived today!! I doubt I'll ever get through them all, but I always get inspiration and ambition from the complex systems perspective.
03.05.2025 23:05 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0I found joy today,
not cold, but slightly chilled,
emerging from the shade into the warm sunshine.
Harvard has set an example for other higher-ed institutions - rejecting an unlawful and ham-handed attempt to stifle academic freedom, while taking steps to make sure students can benefit from an environment of intellectual inquiry, rigorous debate and mutual respect. Letโs hope others follow suit.
15.04.2025 03:52 โ ๐ 90822 ๐ 18532 ๐ฌ 1616 ๐ 757Today I lost an NIH grant training clinicians to do research about intimate partner violence with perinatal women.
More women are murdered in the perinatal period than die of obstetric complications.
I didn't just lose a grant. We lost an entire cohort of people building careers to prevent that.
From 2010 to 2016 (latest data I have ), NIH research contributed to EVERY drug approved by the FDA
22.03.2025 10:44 โ ๐ 32434 ๐ 8610 ๐ฌ 730 ๐ 298In a new Science study, cryoโelectron tomography captures the in-cell architecture of the mitochondrial respiratory chain, illuminating how the coordinated action of molecular machines drives lifeโs fundamental energy conversion.
Learn more in this week's issue: scim.ag/3FA3Ygq
I think we have enough evidence to now argue that AI is not getting exponentially better, or even linearly better. It is asymptotically converging on a level of inadequacy, and even that is only happening because the data thieves are still stealing everything that isnโt nailed down.
21.03.2025 12:32 โ ๐ 8 ๐ 4 ๐ฌ 1 ๐ 0Last year, I was overjoyed to receive an NIH NRSA fellowship to study toddler brains and caregiving effects on memory at Columbia. Last night, my grant was terminated.
11.03.2025 15:17 โ ๐ 1021 ๐ 378 ๐ฌ 85 ๐ 18Have not been able to tear myself away from exploring these just received spatial transcriptomics data from the septum and surrounding nuclei in human ๐ง . Fascinating to explore since there are no comprehensive spatial gene expression data across these regions yet available in human brain.
01.03.2025 15:01 โ ๐ 29 ๐ 3 ๐ฌ 1 ๐ 0I currently have two fully-funded PhD opportunities in my lab - both providing a chance to work on the incredible LEAF data set that we have been collecting the last couple of years (220 children followed longitudinally from 10-30 months on measures of attention and early EF). Please share!
01.03.2025 13:02 โ ๐ 1396 ๐ 736 ๐ฌ 16 ๐ 16RFK Jr. Vows To Make Measles Deaths So Common They Wonโt Be Upsetting Anymore
27.02.2025 20:00 โ ๐ 23112 ๐ 4875 ๐ฌ 473 ๐ 393Now accepted at Nature Computational Science:
Nicole shows how #Bayesian #hierarchical #models can explicitly model major #source of #population #stratification in 10,000 ABCD subjects.
pmc.ncbi.nlm.nih.gov/articles/PMC...
This looks like amazing new work from my colleagues at Stanford demonstrating the opponent interactions of DA and 5HT in reinforcement learning.
02.01.2025 15:44 โ ๐ 63 ๐ 13 ๐ฌ 2 ๐ 0๐๐๐ป๐ฎ๐บ๐ถ๐ฐ๐ฎ๐น ๐๐ต๐ฒ๐ผ๐ฟ๐ ๐ผ๐ณ ๐ฐ๐ผ๐บ๐ฝ๐น๐ฒ๐
๐๐๐๐๐ฒ๐บ๐ ๐๐ถ๐๐ต ๐๐๐ผ-๐๐ฎ๐
๐บ๐ถ๐ฐ๐ฟ๐ผโ๐บ๐ฎ๐ฐ๐ฟ๐ผ ๐ฐ๐ฎ๐๐๐ฎ๐๐ถ๐ผ๐ป
Interesting paper but top-down causation seems to be way of capturing how micro-level entities respond to *aggregate state* of all other micro-level entities
#complexity
ht @evanthompson.bsky.social
www.pnas.org/doi/10.1073/...
๐ฏ
Hallucination is totally the wrong word, implying it is perceiving the world incorrectly.
But it's generating false, plausible sounding statements. Confabulation is literally the perfect word.
So, let's all please start referring to any junk that an LLM makes up as "confabulations".
I need this!! Now, how highly should I prioritize reading this paper, and which to-do list should I put it on?
10.12.2024 16:36 โ ๐ 2 ๐ 0 ๐ฌ 0 ๐ 0Try the sites listed at www.sci-hub.pub, but before you do, learn about the context at en.wikipedia.org/wiki/Sci-Hub to determine your comfort level with the legality of "stealing" publications your tax dollars funded.
10.12.2024 02:44 โ ๐ 6 ๐ 0 ๐ฌ 2 ๐ 0๐จNew WP๐จ
We examine news sharing on 7 platforms:
1)Right-leaning platforms=lower quality news
2)Echo-platforms: Right-leaning news gets more engagement on right-leaning platforms, vice-versa for left-leaning
3)But low-quality news gets more engagement EVERYWHERE, even BlueSky!
osf.io/preprints/ps...
I finally got a chance to watch that talk last night, and I agree it was excellent. We've been debating different regression schemes for training similar classifiers for a long time and this discussion was helpful. Thanks for sharing, Deon!
09.12.2024 17:08 โ ๐ 1 ๐ 0 ๐ฌ 1 ๐ 0Good news everyone! A new version of graph-tool is just out! @graph-tool.skewed.de
graph-tool.skewed.de
Graph-tool is a comprehensive and efficient Python library to work with networks, including structural, dynamical, and statistical algorithms, as well as visualization. 1/N
#networkscience
"Reproducibility is hard" - nice piece by @petebachant.me on trying reproduce one's previous work.
08.12.2024 02:23 โ ๐ 32 ๐ 5 ๐ฌ 1 ๐ 0A VERY PRETTY brain connectome diagram.
Took me two solid days of data munging and coding but I finally figured out how to create these gorgeous connectome/circle plot diagrams. This is the Yeo 7-network brain parcellation, with major connections in a group of healthy adults.
05.12.2024 08:11 โ ๐ 79 ๐ 6 ๐ฌ 9 ๐ 1Students should study the humanities not because it makes them better workers but because it rips you open and breaks your brain and changes everything you thought you knew. I *promise* you this. I *promise* it has this ability, no matter how smart you think you are. I have seen it countless times.
05.12.2024 22:15 โ ๐ 1744 ๐ 491 ๐ฌ 24 ๐ 42Figure 1 from the paper describes the methods.
@frauleindoktor.bsky.social shares the details and writes insightfully about this spontaneous engagement in doi.org/10.1016/j.bp....
06.12.2024 02:28 โ ๐ 2 ๐ 0 ๐ฌ 0 ๐ 1