We’ll be the last ones falling.

"The tragedy of the em dash"

The title of a book I’ll be writing soon.

Acute requirement for the hippocampus in putatively conscious vision revealed by a mouse model of blindsight The phenomenon of blindsight provides a unique opportunity to uncover brain areas important for conscious vision. Patients with blindsight lose the co…

www.sciencedirect.com/science/arti...

very nice paper on a mice model of blindsight

i'm not sure if i'm fully convinced by their saliency control, but maybe i'm just picking hair

let me explain: (a thread to follow)

New preprint! 🧠
How do RNNs learn abstract rules from sequences, independent of specific stimuli?

By Vezha Boboeva, with Alberto Pezzotta & George Dimitriadis

"From sequences to schemas: low-rank recurrent dynamics underlie abstract relational representations"
www.biorxiv.org/content/10.6...

New preprint from my lab! We study how reinforcement learning & selective attention interact. To do so, we built a set of models describing different ways that value & reward prediction error can modulate top-down attention. We compare model outcomes to monkey data from a color value learning task

Check the lecture notes I contributed to: a lecture note from Analytical Connectionism 2024

“A Computational Basis of Natural Intelligence” (PMLR, Analytical Connectionism 2024), based on lectures by Jonathan D. Cohen.

Brain-inspired warm-up training with random noise for uncertainty calibration - Nature Machine Intelligence Cheon and Paik show that overconfidence in deep neural networks arises from standard initialization practices, and that brief warm-up training with random noise improves uncertainty calibration and meta-cognitive recognition of unknown inputs.

A neurodevelopment-inspired warm-up strategy to address uncertainty calibration: networks are briefly trained on random noise and labels before exposure to real data, leading to well-calibrated confidence and strong detection of unknown inputs.

Cool results!

#NeuroAI
www.nature.com/articles/s42...

To be fair, I wasn't exactly hitting the deadline perfectly, but I was keeping them in the loop the whole time.That’s what really annoyed me.
And I actually cared about that paper and was putting in the work, not just ghosting.

My first foray into explicitly trying to bridge Marr’s levels, with @bealebrains.bsky.social. Inspired by Hahn and Wei’s models (pubmed.ncbi.nlm.nih.gov/38360947/), we wondered how the brain could instantiate sensory inference with efficient /decoding/ properties.

1/

Nope, got an email this morning directly from the editor.

Bitter reviewer rant alert! 🚨

Nothing kills the motivation to do review like being "uninvited" when you’re only two days late; especially after keeping the editor updated.

I’m nearly finished, I take the work seriously, and I’m doing it on top of a million other things 🙄

Wonderful work by Amin!

Want a dataset to test ideas on neural basis of decision making or how areas interact as we make choices? Check out our data published today @rudebecklab.bsky.social. >16,000 single neurons from 22 anatomically confirmed areas in macaques performing a decision task. www.nature.com/articles/s41...

What are the systems in neuroscience that we really have something that we can call “explanation” at all relevant levels, other than reflexive feed-forward like circuits.

Here are a few that I would argue are getting there. Obviously not complete explanations but genuinely satisfying.

A perfect antidote to the Eon story.

The RL component is exactly what Eon’s press release tried to sweep under the rug.

🧵 New preprint led by @bingbrunton.bsky.social, @elliottabe.bsky.social, @lawrencehu.bsky.social

We gave a worm brain control of a fly body and it walked

What did we learn? Nothing, other than deep reinforcement learning is effective

We call it the digital sphinx

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

Inhibitory normalization of error signals improves learning in neural circuits Normalization is a critical operation in neural circuits. In the brain, there is evidence that normalization is implemented via inhibitory interneurons and allows neural populations to adjust to chang...

How do neural circuits in the brain implement normalization? 🧠

In our new paper, we show that just normalizing sensory input isn't enough. Crucially, we must also normalize the error signals! 🧵👇

Paper: arxiv.org/abs/2603.17676

Here's a lovely #blueprint on a new study from our lab led by @royeyono.bsky.social.

tl;dr: it implies that there may be interneurons whose role is to normalize credit assignment signals during learning.

#neuroscience 🧪

Excellent summary of the issues surrounding the 'uploaded' fruit fly.

As I noted in my response to @theroberthart.bsky.social, given the flashy claims, it’s our responsibility to set a high bar for evidence. It definitely needs to be more than just video demos.

JEDI: Jointly Embedded Inference of Neural Dynamics Animal brains flexibly and efficiently achieve many behavioral tasks with a single neural network. A core goal in modern neuroscience is to map the mechanisms of the brain's flexibility onto the dynam...

New paper! We introduce JEDI, Jointly Embedded Dynamics Inference for neural dynamics.
arxiv.org/abs/2603.10489. JEDI flexibly infers dynamical principles (across behaviors/contexts) from neural population data through RNNs constrained at single-neuron resolution to reproduce that data.

Representation Biases: Variance Is Not Always a Good Proxy for Importance A central approach in neuroscience is to analyze neural representations as a means to understand a system's function, through the use of methods like principal component analysis, regression, and repr...

Pleased to share that our paper "Representation Biases: Variance is Not Always a Good Proxy for Importance" is now out as Theory/New Concepts paper in eNeuro!
www.eneuro.org/content/13/3... 1/

This is not a fly uploaded to a computer C’est ne pas une fly.

Viral claim on X: This is a "real uploaded animal" with “91% behavior accuracy”

@theroberthart.bsky.social with @shahabbakht.bsky.social @asbates.bsky.social @anayebi.bsky.social @birchlse.bsky.social @tom-mcclelland.bsky.social: Is it though?

That is great. I’m very curious to learn more about your paradigms.
Also, eager to see where you’ll go with your Option C solution. To me, integrating the two systems at the neural level sounds like a serious but exciting challenge.

That is it … the experimenter should ideally live with the subject for a relatively long period of time for developing a reliable measure.

Idk, hard to measure :)

These are also the hardest factors to control in human experiments.

💯

To my colleagues: as you review the next round of applications, please take into account the unique challenges these candidates have faced.

I’m personally witnessing how hard they strive to stay connected despite the brutality of the regime and a devastating war.

Hopeless in Tehran Essays on neuroscience

Today I received a note from a grad student who lives in Tehran. Her note gives you firsthand experience of what it’s like to live in a city that is being bombed, and what it’s like to be young and feel despair about your future.

rezashadmehr.blogspot.com/2026/03/hope...

This is super cool!

The neuroscience implications are also interesting. Putting developmental stage aside, does the adult brain even need complicated and expensive credit assignment to learn?

Perhaps in situations where the animal is pushed outside its ecological niche (eg experiment tasks)?