👀 Check out @marcelhussing.bsky.social and @liv-daliberti.bsky.social's amazing new work on behavioral consistency! ❓The challenge? Retraining an RL agent might give you a completely different policy than before! This makes everything harder, as we never quite know whether we simply got unlucky 🤔

1/ New preprint + VSS poster! Attention has long been thought to enable efficient vision. But does it? First demonstration that attention—consuming just 4-5% of the energy budget—can cut the energetic costs of vision in half. #VSS2026

In case you missed it: 🧪 #Neuroskyence #PsychSkySci #AcademicSky #BioSky

Apple's Scaling Laws for Mixture Pretraining Under Data Constraints They find a fundamental trade-off: too little target data in the mixture underexposes the model to the target domain, while too much target data repeats the same examples excessively,

People will be like ‘Bluesky is all academics & historians’ & I’ll be like, ‘You mean the people who enthusiastically tell me the coolest shit I ever heard?????’

New paper! Multistep inference across the human lifespan can be improved with individualized memory interventions, with co-first @kwcooper.bsky.social @lovecrabmeat.bsky.social @dalezhou.bsky.social, Craig Stark, and @aaronbornstein.bsky.social (🧵): academic.oup.com/psychsocgero...

Do brain circuits really sit in a 'soup' of neuromodulators? Are neurons like soggy bits of bread soaked in signaling factors? Is the whole brain just squelching around in one big pool of warm modulators?

Where, exactly, does learning happen in the brain? Out today in @nature.com, we identify a synaptic locus of birdsong learning and show that the circuit can be tuned to make birds learn faster - but at a cost. Read on👇 #neuroskyence 🧪 #prattle 💬 #bioacoustics Shareable link: rdcu.be/fiyrS

Now out in Nature Machine Intelligence @NatMachIntell “Adopting a human developmental visual diet yields robust and shape-based AI vision”: doi.org/10.1038/s422.... A wonderful case where brain inspiration improved AI. With @martisamuser.bsky.social, Radek Cichy and @timkietzmann.bsky.social .

Many of us were taught experiments are for testing hypotheses. In Ch 1 of Experimentology, our free, open methods textbook, my coauthors and I argue differently: experiments are for estimating the magnitude of causal effects. This reframing has important consequences. 🧵 experimentology.io

A series of lovely recent papers have framed *in vivo* ACh modulation of DA release as unconfirmed. *Last year* @jiaxuanqi.bsky.social showed clear in vivo cholinergic modulation of DA during birdsong learning, including evidence that ACh is permissive for learning while DA is instructive. Read on👇

🚨 New paper alert! 🎉🎉🎉 The largest meta-analysis to date on physical activity & mood in daily life is out in Nature Human Behaviour. An amazing effort led by Johanna Rehder & Markus Reichert et al. (2026)!

A web page just told me everything it learned about me without asking. It knew my city, my device, my GPU, and how long I hesitated before scrolling. sinceyouarrived.world/taken

Do RL agents really need to replan from scratch for every new goal? Why do rodents "preplay" goals they haven't pursued? We think these questions share a candidate answer.

Our new paper with @maxtaylordavies.bsky.social introduces a resource-rational model of Theory of Mind. The model can explain many of the successes and failures of mindreading in human adults and children, and non-human primates. 🧵 royalsocietypublishing.org/rspb/article...

Posting for Vibhu Sahni: Due to unforeseen circumstances the Burke Neurological Institute will have to terminate its research operations 5/22. There are international postdocs on visas who are looking for new homes. Please reach out to Vibhu if you can help

bump for visibility!

Do you think that the canonical cortical circuit is complicated? It's so much worse than that! Overlayed on top of the wired connectome, there are *18* different signal-carrying peptides in cortex! Their range is ~similar to the size of a cortical column elifesciences.org/articles/47889 1/

Why are oxytocin study results so inconsistent? Methods get most of the blame, but the problem is also theoretical as existing accounts are too vague to falsify. In a new preprint, I formalise the Allostatic Theory of Oxytocin as a computational model osf.io/preprints/ps... 🧵

New preprint: "Linear equivalence of nonlinear recurrent neural networks." For large nonlinear (potentially chaotic) RNNs with random connectivity, the full N×N covariance matrix takes the same form as that of a ~linear~ network with the same couplings, driven by independent noise.

Two of my favourite things in one paper — replay and SR. Wittkuhn et al find that during brief task pauses, human visual cortex replays task states, and the structure of those replays tracks implicit learning of a successor representation. Nice. https://doi.org/10.1073/pnas.2507516122

It’s published… the first manuscript from the Rich Lab is officially typeset and published!!! What an absolutely surreal moment… that I’m going to try my best to enjoy while awaiting the results of a study section from last week 😅😅😅 journals.plos.org/ploscompbiol...

New work w/ Zach Kelso and @madeleinecsnyder.bsky.social www.biorxiv.org/content/10.6... Our negative results on classical conditioning in planarian flatworms. This was surprising, given the long history of work (including sensational findings of memory transfer and retention through decapitation).

One of the things that has always bothered me about TV. Everything is fundamentally filtered through how a TV writer sees the world. It's a very...*particular* sort of worldview — only a very particular sort of person lives in LA and gets work as a TV writer — and once you see it you can't unsee it.

When an axon of cell A is near enough to excite a cell B and repeatedly or persistently takes part in firing it, some growth process or metabolic change takes place in one or both cells such that A's efficiency, as one of the cells firing B, is increased.

Now out in PNAS with @jaeyoungson.bsky.social, Alice Xia, @apaxon.bsky.social & @orielf.bsky.social. Medial temporal lobe encodes predictive representations of people's real-world social networks which afford them key advantages in social navigation. www.pnas.org/doi/10.1073/... 🧵

Looking forward to presenting our work on the origin and flexibility of reward signals at the @arc-mpib.bsky.social next week!

🚨 Why can’t mammals regenerate limbs like frog tadpoles or salamanders? In our new paper in @science.org , we show that species-specific oxygen sensing acts as a gatekeeper for initiating limb regeneration 🐭🐸 🔗 www.science.org/doi/10.1126/... #EvoDevo

Amazing study and the coolest results (from my point of view!) is that LC norepinephrine neurons projecting to the frontal cortex encode reward prediction error

Do you want to use vibration stimuli in remote research studies? 👀 📳🤳🏼 Our latest paper in Behaviour Research Methods might be of interest to you! Coauthors include: @kalvinroberts.bsky.social @peircej.bsky.social @multisensorylab.bsky.social link.springer.com/article/10.3...

Delighted to share our discoveries about one of the brain's neurotransmitter systems: www.biorxiv.org/content/10.6... Together with colleagues at the @alleninstitute.org, we have learned a lot about a tiny cluster of neurons in the brainstem locus coeruleus (LC) that releases norepinephrine (NE). 1

what's your ai strategy? what's your quantum computing strategy? what's your ww3 strategy? what's your collapse of consensus reality strategy? what's your restructuring of global supply chains strategy? what's your pedophile billionaire cabal blackmail ring strategy?

"The cerebellar circuitry may be uniquely poised to learn the probability of events in the world and internalize these as prior knowledge." pubmed.ncbi.nlm.nih.gov/41946969/ The Bayesian little brain?

🚨 arxiv.org/abs/2604.06129 PoM: A Linear-Time Replacement for Attention with the Polynomial Mixer This paper is the result of doing a lab-wide hackathon on an idea I've had for some time. Probably the paper with the highest number of authors I've ever done. It's a CVPR Findings 26. Thread 🧵👇

This wonderful collaboration, led by @idoba.bsky.social and co-supervised by me and @shaharnitzan.bsky.social, is now published in Translational Psychiatry (www.nature.com/articles/s41...). We find that autistic individuals show more accurate and less biased credit assignment during an RL paradigm.

new proof of concept: ATProtoCall! video calling over atproto with just static files and no dedicated signaling or streaming infrastructure.

"The Brain, In Theory" is out today! A short excerpt in The Transmitter @thetransmitter.bsky.social www.thetransmitter.org/theoretical-...

I like how Anthropic is just obliquely releasing their work on recursive self-improvement.

If racism was a country: www.bbc.co.uk/news/article... “In addition, the return of more than 600,000 residents of southern Lebanon who evacuated northward will be completely prohibited south of the Litani until the safety and security of northern (Israeli) residents are ensured”

I liked this thread: bsky.app/profile/gui.... It highlights that the power of the platform is in the API access it exposes and controls. The great thing about ATproto is that most of the data is freely accessible: posts, likes, follows. 🧵

So what this describes is ethnic cleansing www.nytimes.com/2026/04/01/w...

Majestic, well done @rickcarlsson.bsky.social

Stoked this is finally out! We ask: how can we simulate the brain from the bottom up? It's not sufficient to grab the connectome and wire it up in silico! We need 1) ultrastructure 2) (causal) calibration data 3) functional data. Then we can build a simulation compiler. 1/

What is happening is that locally the species is actually becoming significantly _less_ genetically diverse, but the fragmentation disconnects parts of the species, which slowly diverge from each other. The species-wide diversity masks the plummeting local diversity. Again, this is BAD.

This is BAD. If we monitor the genetic diversity of a species, we might be convinced that it's doing okay, when in fact it may continue to lose genetic diversity for a long time even without losing any more of its habitat!

Delighted to share our latest preprint: Julie Zhu identifies a surprising reason why the genetic architecture of brain-related traits is so different than other traits. Check it out!

The hardest part of science is posing the right question, not answering it.

New work from our lab spearheaded by the amazing @kaixiangzhuang.bsky.social offering a fresh look at the big five personality traits.

Super excited to see the first paper from my PhD out! 🐭 Very satisfying to get to build an interpretable model of how the brain does a thing. A nice palette cleanser in the current sea of LLMs and deep ML.