I encourage everyone to read the paper and rather than relying on social media impressions. It’s an excellent paper speaking about a specific category of work in NeuroAI that is still hard to generalize to the whole field.

Our paper is out in @natneuro.nature.com! www.nature.com/articles/s41... We develop a geometric theory of how neural populations support generalization across many tasks. @zuckermanbrain.bsky.social @flatironinstitute.org @kempnerinstitute.bsky.social 1/14

Very excited about this work looking at the emergent mechanisms that vision language models use to perform structured visual processing, mirroring a computational strategy (visual indexing) proposed in cognitive science, but here learned by VLMs. Check out the paper/thread for more details!

The visual world is composed of objects, and those objects are composed of features. But do VLMs exploit this compositional structure when processing multi-object scenes? In our 🆒🆕 #ICLR2026 paper, we find they do – via emergent symbolic mechanisms for visual binding. 🧵👇

Google DeepMind's Mining Generalizable Activation Functions They argue that evolutionary search is a powerful framework for discovering new activation functions, showing that LLM-driven pipelines like AlphaEvolve replace manually designed search spaces with flexible,

How do you know how visual stimuli are represented internally for decision making? This is perhaps the central question in perceptual decision making. In a new paper, we show that one can use artificial neural networks to crack this problem. #NeuroAi #VisionScience direct.mit.edu/opmi/article...

They develop TinyLoRA, a new ft method. with TinyLoRA + RL, models learn well with dozens or hundreds of params. For example, they use only 13 parameters to train 7B Qwen model from 76 to 91% on GSM8K. "Learning to Reason in 13 Parameters" Paper: arxiv.org/abs/2602.04118

New @pnas.org paper out 🎉 “Representational drift reflects ongoing balancing of stochastic changes by Hebbian learning” 👉 doi.org/10.1073/pnas... What drives representational drift in neural populations? Here’s the short version. 👇 🧪🧠 1/5

Why don’t neural networks learn all at once, but instead progress from simple to complex solutions? And what does “simple” even mean across different neural network architectures? Sharing our new paper @iclr_conf led by Yedi Zhang with Peter Latham arxiv.org/abs/2512.20607

This paper was an awesome collaborative effort of a @fitngin.bsky.social working group. It provides a detailed review of how DNNs can be used to support dev neuro research @lauriebayet.bsky.social and I wrote the network modeling section about how DNNs can be used to test developmental theories 🧵

Fascinating neuropsych article from 1919 of a patient with spatial disorientation, with a loss of mental imagery for routes: watermark02.silverchair.com/archneurpsyc...

The hippocampal map has its own attentional control signal! Our new study reveals that theta #sweeps can be instantly biased towards behaviourally relevant locations. See 📹 in post 4/6 and preprint here 👉 www.biorxiv.org/content/10.6... 🧵(1/6)

Love this. “Spontaneous” has always really meant “we don’t know why X is happening”, a perspective tied to the dominance of structured task paradigms for studying cognition, behavior, and brain activity.

I know this isn't the content you followed me for, but I watched a father desperately searching through piles of bodies, crying out, "Sepehr, Baba, where are you?" A 12-fu…-minute video searching through the bodies. It is a level of pain I can’t put into words. …

New in Nature Neuroscience: We developed a flexible model that reveals how animals learn tasks—uncovering stages, sudden insights, and gradual improvements unique to each animal. Learning isn't monotonic, and our model captures that complexity 🐭📊 www.nature.com/articles/s41...

Excited to share our NEWEST PREPRINT led by @rochellekaper.bsky.social!! osf.io/preprints/ps... We ask: How do people learn multiple layers of environmental structure – w/o feedback – & how well do they *know* they’ve learned? Turns out, stimulus familiarity matters more than we thought! 🧵👇

Our paper on the "Oneirogen hypothesis" is now up in its revised form on eLife! This is the hypothesis that psychedelics induce a dream-like state, which we show via modelling could explain a variety of perceptual and learning effects from such drugs. elifesciences.org/reviewed-pre... 🧠📈 🧪

Thank you to @aaas.org for the opportunity to lead a new editorial pod at Science Advances on Systems, Computational, Neuro-AI & Neurotechnology (SCANN). Top opportunity to highlight outstanding work at the intersection of neuroscience, computation, and emerging technologies. Send your papers now!

We're already in the phase that each of us (Iranians) by now know someone killed/injured in our circle of friends/family. And this is despite the continued internet blackout, when millions haven't still managed to hear from their family/friends since Thursday, January 8th. This tells a lot ... 😥😥

I’m very happy to share the latest from my lab published in @Nature Hippocampal neurons that initially encode reward shift their tuning over the course of days to precede or predict reward. Full text here: rdcu.be/eY5nh

The depth of pride I feel for my brave compatriots in Iran — braver in a single minute than I will ever be in my entire life — is overwhelming.

When and why do modular representations emerge in neural networks? @stefanofusi.bsky.social and I posted a preprint answering this question last year, and now it has been extensively revised, refocused, and generalized. Read more here: doi.org/10.1101/2024... (1/7)

With some trepidation, I'm putting this out into the world: gershmanlab.com/textbook.html It's a textbook called Computational Foundations of Cognitive Neuroscience, which I wrote for my class. My hope is that this will be a living document, continuously improved as I get feedback.

Really interesting work, introducing Epiplexity (an observer-dependent measure of information under computational constraints) Feels like something theoretical neuroscience could use right away. 🧠📈

🚨📜+🧵🚨 Very excited about this work showing that people with no hand function following a spinal cord injury can control the activity of motor units from those muscles to perform 1D, 2D and 3D tasks, play video games, or navigate a virtual wheelchair By a wonderful team co-mentored w Dario Farina

Just saw this paper today (even though it was out yesterday😱) and it is super interesting and powerful that so much clinical data with predictive power can be extracted from one night of sleep data www.nature.com/articles/s41...

As this ramps up and we are asked to cheerlead for it, I remain genuinely shocked that Canada, once again, under the banner of a **nation-building**, has created a funding program that excludes Canadian students and researchers based on presence in Canada. /1 universityaffairs.ca/news/feds-la...

Wondering if the "slow death of scaling" is the hidden force behind this trend. papers.ssrn.com/sol3/papers.... Very nice write up by @sarahooker.bsky.social Rethinking architectures and training curricula can definitely open up a new space for #NeuroAI

Reminder that this job is open! If you have any questions, feel free to get in touch. Deadline is February 20th.

Is it just my timeline, or are we seeing a spike in loose neuro-based AI ideation again? I’m seeing way more references to specific mechanisms by AI folks (hippocampus, oscillations, etc.). It's often handwavy, but interesting to see the pendulum swing back after years of fading interest. 🧠🤖

What can cognitive science learn from AI? In infinitefaculty.substack.com/p/what-cogni... I outline how AI has found that scale and richness of learning experiences fundamentally change learning & generalization — and how I believe we should rethink cognitive experiments & theories in response.

This paper had a pretty shocking headline result (40% of voxels!), so I dug into it, and I think it is wrong. Essentially: they compare two noisy measures and find that about 40% of voxels have different sign between the two. I think this is just noise!

New preprint. We show that in addition to reward prediction errors (RPEs), dorsal striatal dopamine signals encode sensory prediction errors (SPEs), the difference between sensory prior & observed stimulus. www.biorxiv.org/content/10.6...

The last year was very hard as my father was dying. It became that much harder when I discovered that AI played a role in amplifying his physical pain, and perhaps hastening the end of his life. It wasn’t easy to write about what happened, but I’ve tried. open.substack.com/pub/buildcog...

If this result generalises, then the field's great dendritic computation dream is over

Apple’s real 'edge' might be Edge AI. On-device continual learning is still a tough nut to crack (catastrophic forgetting, energy consumption, etc.), but Apple may have the right pieces in place to get there first.

Great piece as always by @markdhumphries.bsky.social. I’d bet a lot of the pushback foundation models in neuroscience receive is simply because of the name. Call them anything else and they’d be perfectly fine :)

Just published my review of neuroscience in 2025, on The Spike. The 10th of these, would you believe? This year we have foundation models, breakthroughs in using light to understand the brain, a gene therapy, and more Enjoy! medium.com/the-spike/20...

Most likely captured on a phone, this image perfectly captures the reality of Iran today.

James makes a crucial point in this thread about AI in data analysis. It reminded me of this recent paper: osf.io/preprints/ps.... While I don’t agree with every argument made, I found the particular segment below especially useful.

Andrew’s latest blog series has the potential to be a full curriculum. It’s made me think more about the value of non-interdisciplinary foundations. There’s a strong case for ensuring students are expert cognitive (neuro)scientists before they transition into interdisciplinary spaces eg neuroAI 1/3

If you're (1) into nhp neurophysiology and/or biological motor control, (2) a foreign big shot and (3) interested in moving here, please reach out -- this new program looks extremely attractive (min. $8M over 8y)! www.uwo.ca/research/can...

Our paper on data constrained RNN that generalize to optogenetic perturbations now citable on eLife: doi.org/10.7554/eLif...

Generative AI systems are being built primarily for entertainment, design and communication, but their potential for neuroscience is vast. @shahabbakht.bsky.social explores how this technology could help capture an animal’s ecological experience. #neuroskyence www.thetransmitter.org/artificial-i...

Excited to announce a new book telling the story of mathematical approaches to studying the mind, from the origins of cognitive science to modern AI! The Laws of Thought will be published in February and is available for pre-order now.

On why AlphaFold may not revolutionize drug discovery. Great piece by @philipcball.bsky.social

Why isn’t modern AI built around principles from cognitive science or neuroscience? Starting a substack (infinitefaculty.substack.com/p/why-isnt-m...) by writing down my thoughts on that question: as part of a first series of posts giving my current thoughts on the relation between these fields. 1/3

“Why AGI Will Not Happen” by Tim Dettmers. timdettmers.com/2025/12/10/w... This essay is worth reading. Discusses diminishing returns (and risks) of scaling. The contrast between West and East: “Winner takes all” approach of building the biggest thing vs a long-term focus on practicality.

Always such a vibrant NeuroAI community in Montreal !

Yup I realized that when one of his New Yorker articles discussed his great idea that the brain might have a special region for face recognition, all presented as his idea long after this had been widely published.