Super excited to share our first preprint of work from the lab! We monitor and manipulate the CA1 representation of mice exploring a multicompartment environment to contrast geometric vs. predictive theories of cognitive mapping. TLDR: All signs point to a predictive theory! tinyurl.com/yeyv5p3x - ThreadSky

atkeinath.bsky.social • 96 days ago

Super excited to share our first preprint of work from the lab! We monitor and manipulate the CA1 representation of mice exploring a multicompartment environment to contrast geometric vs. predictive theories of cognitive mapping.

TLDR: All signs point to a predictive theory!

https://tinyurl.com/yeyv5p3x

Comments

apeyrache.bsky.social•96 days ago

congrats! Sounds really interesting, looking forward to reading this in details.

atkeinath.bsky.social•96 days ago

Thanks Adrien! It's been exciting to see it develop from (what we thought would be) a simple story about orientation into a much more interesting tale of individual differences and behavioral determinants. Feedback is definitely welcome!

hugospiers.bsky.social•96 days ago

Cool. Opposite of what we got in 2021 Cur Biol. I look forward to reading this Alexandra!

atkeinath.bsky.social•96 days ago

Yes indeed! I think there’s a few reasons why, happy to chat about it. :)

hugospiers.bsky.social•96 days ago

#1 I'm guessing is that our goal-directed task was rather weak providing little modulation. Rats certainly learned the transition structure, but I think very self motivated by exploration, rather than reward prediction.

atkeinath.bsky.social•96 days ago

Definitely one reason I think. Also, the scale matters a lot in our data: similarity of movement within compartment on a few centimeter- and 1 sec scale matched CA1 but not transitions between compartments. Also, behavior during early experience, not after familiarization, was strongest.

hugospiers.bsky.social•96 days ago

Very interesting. So glad you've been exploring this. :)

atkeinath.bsky.social•96 days ago

Thanks! And thanks for establishing the paradigm! Happy that the reality turned out more interesting than we had anticipated. :)

steveread.bsky.social•96 days ago

📌

neuroetho.bsky.social•96 days ago

📌

dahery.bsky.social•96 days ago

Congratulations to you and Sergio!

amkashef.bsky.social•96 days ago

Very interesting: Haven't read it yet, but quick question: Can the animal's movement patterns influence their geometric representations of that space? The environment appears the same to us, but mice could potentially utilize different landmarks based on their exploration patterns.

atkeinath.bsky.social•96 days ago

Great Q (@dahery.bsky.social brought this up in lab meeting)! We can't rule this out during the free navigation experiment, but the finding that changing navigation (while keeping geo constant) changes CA1 in a directional and transient way is harder to account for from a purely geo theory.

atkeinath.bsky.social•96 days ago

These results demonstrate that idiosyncratic navigation determines CA1 representational structure (behavior matters!), consistent with predictive cognitive mapping. Big thanks to my first graduate student @sergiopecirno.bsky.social who led this, and Mice 4 and 5 the real MVPs! Feedback welcome!

atkeinath.bsky.social•96 days ago

(3) Finally, we limit the mouse’s navigational options to change their navigational patterns. This induces a change to CA1 remapping structure to match the new pattern of navigation. Afterwards both behavior and CA1 revert to their OG structures. (Yes, that is 43 sessions of data for each mouse)

atkeinath.bsky.social•96 days ago

(2) Within mouse, compartments that have similar navigational patterns (i.e. similar successor matrices) on a second-long timescale during early experience are represented more similarly in CA1. Between-compartment transitions did not match CA1. (Spatiotemporal scale looks suspiciously like BTSP?)

atkeinath.bsky.social•96 days ago

Three primary findings. (1) Different mice show different patterns of remapping across identical unrewarded compartments. Doesn’t matter if we use rate map-based measures or location-agnostic measures. Same environment, different maps for different mice. Already a strike against geometric theories.

Comments

Posting Rules

Reply