New results for a new year! “Linking neural population formatting to function” describes our modern take on an old question: how can we understand the contribution of a brain area to behavior? www.biorxiv.org/content/10.1... 🧠👩🏻‍🔬🧪🧵 #neuroskyence 1/ - ThreadSky

marlenecohen.bsky.social • 131 days ago

New results for a new year! “Linking neural population formatting to function” describes our modern take on an old question: how can we understand the contribution of a brain area to behavior?
https://www.biorxiv.org/content/10.1101/2025.01.03.631242v1
🧠👩🏻‍🔬🧪🧵
#neuroskyence
1/

Comments

coldasskatie.bsky.social•129 days ago

This is super interesting work! Do you have any sense of whether there is flexibility within these modules? For instance, do you think a given module might format information differently based on environmental, social, etc. context?

marlenecohen.bsky.social•127 days ago

Yes! That's a really interesting question and one that is guiding some of our current/future work. We see some evidence that task uncertainty changes formatting (mushing otherwise orthogonal variables together: https://www.biorxiv.org/content/10.1101/2024.03.04.583375v1). 1/2

marlenecohen.bsky.social•127 days ago

There is reason to believe that attention orthogonalizes the attended information & we have preliminary evidence that disorders that affect cognition (e.g. Alzheimer's disease) might change formatting as well. As you suggest, it will be really interesting to see how this depends on other processes.

marlenecohen.bsky.social•131 days ago

We harnessed a variety of new approaches to resolve an emerging paradox. On one hand, animals with more cognitively complex, flexible behavior tend to have more distinct brain areas, and models that perform many tasks organize into modules.
(illustrations from the awesome https://bioart.niaid.nih.gov/)
2/

marlenecohen.bsky.social•131 days ago

On the other hand, you all keep publishing evidence that everything an animal perceives, knows, or does can be decoded from essentially any brain area. If everything is everywhere, why have distinct brain areas?
3/

marlenecohen.bsky.social•131 days ago

A clue: the way different information (e.g. the size of a strawberry and how I predict it will taste) is formatted in neural populations affects how it can be used for computation. If size and taste prediction are encoded independently, I can flexibly use either or both to guide a decision.
4/

marlenecohen.bsky.social•131 days ago

That idea that neural population formatting implies a neural population’s role in decision-making came from terrific theoretical work and correlative observations but had not to our knowledge been tested causally. That’s what we set out to do.
5/

marlenecohen.bsky.social•131 days ago

Superstar staff scientist @douglasruff.bsky.social trained subjects to flexibly make decisions based on a combination of vision (motion direction) and rewards expected from different choices, and he recorded groups of neurons in visual area MT and decision area dlPFC.
6/

marlenecohen.bsky.social•131 days ago

Consistent with ‘everything is everywhere’, Doug could decode both motion and reward information in both areas. But it was formatted differently: MT kept that information separate while dlPFC mushed them together in ways that reflected the decision strategy.
7/

biologicalart.bsky.social•129 days ago

Thank you for the shout out! We’re really proud of bioart and will keep updating it when we can.

marlenecohen.bsky.social•127 days ago

It's such a wonderful service to our community - thank you!

biologicalart.bsky.social•127 days ago

You’re welcome! I’m glad to see how well received it’s been

Comments

Posting Rules

Reply