Learn faster and retain more with Recall. Use my code artem25 for 25% off a subscription, that is valid until 1 April 2026 - Recall: https://www.getrecall.ai/?t=artem

Take my challenge on this video: https://app.getrecall.ai/c/b1eaba93-d755-5710-b075-373b6a26d02d

=====
My name is Artem, I'm a neuroscience PhD student at Harvard University. 
🌎 Website and Social links: https://kirsanov.ai/
📥 "Receptive Field" neuro-newsletter: https://artemkirsanov.substack.com/
✨ Support me on Patreon to get access to Discord community:  https://patreon.com/artemkirsanov
=====

How does the brain learn new tasks so quickly, without starting from scratch each time? In this video, we explore a recent paper from Tim Buschman’s lab at Princeton that provides compelling evidence for compositionality in the prefrontal cortex. By recording neural population activity while monkeys flexibly switch between tasks, the researchers uncover stable “sensory” and “motor” subspaces that are reused across contexts.
We’ll examine how these reusable neural modules enable rapid generalization, how task-irrelevant information is actively suppressed, and how an internal “task belief” signal dynamically reconfigures cortical geometry to route information to the appropriate output.
Together, these findings suggest that flexible intelligence may emerge from the brain’s ability to compose, amplify, and route reusable neural components, rather than rebuilding circuitry for every new problem.

🕒 OUTLINE:
00:00 Introduction
01:20 Sponsor: Recall
2:32 Task Setup
4:14 Detecting Neural Subspaces
5:54 Task Interference & Shared Subspaces
7:34 Neural Switchboard
8:50 Task Belief signal
10:50 Conclusion


📚 FURTHER READING & REFERENCES:
Link to the paper (open access): https://www.nature.com/articles/s41586-025-09805-2

=====
Music by Artlist
=====

**Disclaimer:** This channel is my personal project. The views and content expressed here are my own and are separate from my research role at Harvard University.

Artem Kirsanov

A Princeton study published in Nature investigates how biological brains reuse modular components when learning new tasks, rather than starting from scratch. Using monkeys performing visual discrimination tasks, researchers recorded prefrontal cortex neuron activity and found that the brain maintains stable, reusable subspaces for features like color, shape, and motor commands. Cross-task decoding experiments confirmed these modules are physically reused across different tasks. A central 'task belief' signal acts as a gain control mechanism, amplifying relevant sensory subspaces and suppressing irrelevant ones, while also routing information to the correct motor output. This compositional architecture — analogous to Lego bricks — explains the brain's remarkable ability to rapidly adapt to novel situations.

Why the Brain Doesn’t Start From Scratch