Athena Akrami

Hello @cosynemeeting.bsky.social, is the talk schedule out yet for #cosyne2026?

After several years of work, my lab is starting to put out our first papers on learning in a unicellular organism (Stentor coeruleus).

Here we show evidence for a form of associative learning in Stentor:
www.biorxiv.org/content/10.6...

A new preprint, co-authored with @johnwkrakauer.bsky.social:

The Deliberation Taboo

Cognitive science is, nominally, the science of thinking. We argue that the field has no theory of what thinking is and, even worse, that the topic has largely dropped out of focus. 1/

osf.io/preprints/ps...

Vectorized instructive signals in cortical dendrites - Nature Mice learning a neurofeedback brain–computer interface task show neuron-specific teaching signals in cortical dendrites, consistent with a vectorized solution for credit assignment in the brain.

This paper on how the brain may do gradient descent is very cool: www.nature.com/articles/s41...

This is great from the @athenaakrami.bsky.social lab — hippocampus is required for unsupervised statistical learning, and dCA1 populations split into subspaces for sensory features vs abstract rules. Elegant optogenetic + recording approach.

https://doi.org/10.64898/2026.02.14.705916

Thank you, Caswell! Very glad to hear you found it interesting :)

🚨🚨New Preprint Alert!🚨🚨

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

Animal learning is painfully slow (at least initially). Yet, well trained animals can learn very fast, sometimes displaying few-shot inference. How does this transition occur?

Neuronal spiking in the mammalian forebrain is dominated by a heterogeneous ground state Neuronal firing patterns have significant spatiotemporal variability with no agreed-upon theoretical framework. Using a combined experimental and mode…

With this one in print, I think I finally earned that PhD... 😅
Presented for the first time at the cosyne when the world ended (March 2020). I'll bring over a summary thread from twitter when it was still twitter...

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

Thank you, Anna! Your works have been a big source of inspiration for this to happen!

New preprint from @athenaakrami.bsky.social Lab at SWC shows how mice can learn abstract rules and statistical structure without reward or instruction. 📢

The hippocampus plays a key mechanistic role.

Blog: www.sainsburywellcome.org/web/blog/fin...

Preprint ⤵️

Thanks, Cedric, for the shout out! :)

Thank you, Dan!

I won't disagree!
We're now looking into the classical reward system (aka dopamine) in the same paradigm ... stay tuned! :)

A big shout to Dammy who's been working on this project for ~7 years and the co-authors, Lida Pentousi, Ryan Shen & Vezha Boboeva.

And going after hippocampus in statistical learning was all inspired by @annaschapiro.bsky.social (after a fun chat at sfn 2019) & Nick Turk-Browne's works in humans!

We believe this provides the first causal, circuit-level account of unsupervised abstract structure learning in mice.

And the hippocampus acts as a general-purpose model-builder.

Read the preprint here for all the details, and we'd love to hear your feedback!

7/7

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

At the neural level, all sounds, relevant or irrelevant to the cover task, are represented in dCA1 & we can decode all statistical contexts from dCA1.

dCA1 factorises sensory features and abstract structure into orthogonal population subspaces, enabling generalisation across sensory contexts.

6/7

Silencing dCA1 abolishes these learning signatures, without affecting global arousal, attention, or performance in cover task.

Crucially, we show that dCA1 is 'required' to update the internal models: when silenced, updating was disrupted & mice got stuck with the old knowledge!

5/7

We used a cover task to keep subjects engaged, then manipulated irrelevant sound statistics in various ways:

mice (like humans) spontaneously learned event probabilities, sequence identity, and abstract rules (eg "rising-in-pitch tones" vs "rising-then-falling") very quickly (within minutes).
4/7

We developed a reinforcement-free paradigm to study statistical learning in mice and humans, inspired by infant works as well as @mariachait.bsky.social's work in adult humans.

Learning is tracked implicitly via pupil-linked surprise — no explicit training, no reward for the structure.

3/7

Statistical Learning by 8-Month-Old Infants Learners rely on a combination of experience-independent and experience-dependent mechanisms to extract information from the environment. Language acquisition involves both types of mechanisms, but mo...

Brains learn structure without reward, instruction, or feedback — a process called statistical learning.

Human infants do this effortlessly (famous work by Saffran et al 1996), but how does the brain implement it?

2/7
www.science.org/doi/10.1126/...

Thrilled to finally share this work! 🧠🔊

Using a new reinforcement-free task we show mice (like humans) extract abstract structure from sound (unsupervised) & dCA1 is causally required by building factorised, orthogonal subspaces of abstract rules.

Led by Dammy Onih!
www.biorxiv.org/content/10.6...

Excited to launch Principia, a nonprofit research organisation at the intersection of deep learning theory and AI safety.

Our goal is to develop theory for modern machine learning systems that can help us understand complex network behaviors, including those critical for AI safety and alignment.

1

Are you doing ephys in humans? Come do neuroscience with us!

Building compositional tasks with shared neural subspaces Nature - The brain can flexibly perform multiple tasks by compositionally combining task-relevant neural representations.

Thrilled that my paper is out in the @nature.com. We explored how the brain builds complex tasks by compositionally combining simpler sub-task representations. The brain flexibly performs multiple tasks by dynamically reusing neural subspaces for sensory inputs and motor actions

rdcu.be/eRVUk

New paper alert! 🚨

We found that the brain's compass is remarkably stable at two scales

1️⃣ the system maintains its internal organization for weeks
2️⃣ It "remembers" its orientation for weeks, even after a single visit

This may be key to how the brain aligns its other maps.

Paper: rdcu.be/e3waP

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

From Chand: “of potential interest”. Our lab’s first dataset! Tian and friends used Neuropixels to record from 7,500 units across DLPFC and PMd to show how contextual decision making (specifically, XOR computation) occurs and how this differs by location in DLPFC and between areas.

Apes Share Human Ability to Imagine YouTube video by Johns Hopkins University

Imagination in bonobos!

I am thrilled to share a new paper w/ Amalia Bastos, out now in @science.org

We provide the first experimental evidence that a nonhuman animal can follow along a pretend scenario & track imaginary objects. Work w/ Kanzi, the bonobo, at Ape Initiative

youtu.be/NUSHcQQz2Ko

Working with rats, & not happy w MRI based brain atlases from adult males? We've got you! High resolution (2photon), 3D atlases (female rats) are now available in @brainglobe.info. Plz start using these atlases & let us know your feedback for the official release!
brainglobe.info/blog/swc-fem...

🐭🧠

Postdoc position in Paris: come help develop new generation human brain computer interfaces ⚡🧠💻

Interested? Contact me if you have experience with machine learning (e.g. simulation-based inference, RL, generative/diffusion models) or dynamical systems.

See below for + details and retweet 🙏