Thread by @KateWassum

***New paper out from our lab***

@MelissaMalvaez discovered a critical brain circuit that regulates reward learning & decision making

Here is what she found. Thread 👇

Free full text: rdcu.be/bvn4u

nature.com/articles/s4159…

When making a decision, like what to have for lunch, you have consider your options 🌮🍔🍕

Chief among these considerations is how good each item would be right now: their incentive value.

This value is state dependent.
A 🌮 is super valuable when you’re hungry, less so when you’re full.

An item’s incentive value is also learned.
If you’ve never had a 🌮 when hungry, then it won’t be a go-to. But if you’re famished and you eat a 🌮, then you’ll learn tacos are really valuable when you’re hungry.

Live view of Cookie Monster learning the incentive value of 🌮s:

So the next time you’re hungry you’re more likely to be like:

We thought the basolateral amygdala might be involved in at least learning this value bc the BLA is a brain hub for emotional learning. (You’re probably familar with it’s greatest hit: Fear Learning)

@MelissaMalvaez found that the basolateral amygdala (specifically glutamate signaling in the BLA) is critical for BOTH encoding AND retrieving the value of an anticipated rewarding event.

Thus the basolateral amygdala is doing more than just emotional learning, it’s actually contributing to reward pursuit decisions.

@MelissaMalvaez was then curious what brain regions might support the BLA in its value encoding and retrieval functions.

She thought the orbitofrontal cortex might collaborate with the basolateral amygdala.

The OFC and BLA go way back:

But there isn’t just one OFC

@A_Izquierdo1 wrote a nice piece on this:
jneurosci.org/content/37/44/…

@MelissaMalvaez found that both the lateral and medial subdivisions of the OFC project to the BLA.

Turns out, lateral OFC→BLA projections are critical for learning a reward’s value— their activity is both necessary & sufficient for value encoding— medial OFC→BLA not so much.

We were actually able to cause the subjects to think a food was better than it actually was and thus later pursue it more vigorously than they should, just by stimulating lateral OFC→BLA projections.

Here’s where it gets surprising.

You might think (we did) if lateral OFC→BLA projections help learn a reward’s value, that they must come back online to retrieve that information...

...but no! Lateral OFC→BLA projections mediate value encoding but NOT retrieval.

Instead medial OFC→BLA projections drive the state-dependent retrieval of a reward’s value: medial OFC→BLA activity is both necessary & sufficient for retrieving a reward’s value from memory.

So you don’t need medial OFC→BLA projections to learn a reward’s value, but you do need them to retrieve that memory under the right circumstances.

Thus the BLA is critical for you to learn about the value of a rewarding event and to retrieve this information to make good reward pursuit decisions, and its supported in this function by doubly dissociable contributions from the lateral and medial OFC.

(Yes, I stole my daughter’s crayons to draw that picture)

We’re very curious whether this architecture or organizing principle applies to other memory systems.

But more broadly, we think these data, showing functional & neuroanatomical dissociability of reward value encoding and retrieval, reveal a clear vulnerability in the brain for poor decision making.

Big THX to the rockstar team who worked on this. @MelissaMalvaez had a lot of help: @ChristineShieh @TheeMikeMurphy & Venuz Greenfied