Huge thanks to bernardo for all the helps and ideas and encouragement through the years, all lab members for the discussions, and to all the talented and wonderful people that made this work possible: Wengang, Grace, Ellliot, Paolo, Catherine, and Ellie!! 🙏

(end)

For more experiments, details of analysis/code, and everything else, check out our preprint:

www.biorxiv.org/content/10.1...

(12/13)

In addition, we also showed that EP>LHb synaptic sign tightly correlates with DA updates, such that:

1. EP>LHb are more inhibitory when DA is positively updated
2. EP>LHb are more excitatory when DA is negatively updated
3. Such correlation exists within ~20 min window

(11/13)

We do this by opto stimulating EP Sst fibers while measuring monosynaptic EPSC or IPSC in LHb.

We show that:
1. EP>LHb synapses are more inhibitory after punish-to-reward pairing reversal
2. EP>LHb synapses are more excitatory after reward-to-punish pairing reversal

(10/13)

Using an optogenetic cue to induce plasticity has an important benefit:

We can induce plasticity in vivo, then examine these exact synapses that underwent plasticity ex vivo in acute brain slices!

This allow us to measure synaptic signs of EP>LHb after learning.

(9/13)

We also performed a series of ctrl & loss of function experiments, showing that such bi-directional modulation of DA by EP Sst neurons is:

1. specific to opto itself, not opto-associated sensory cues
2. driven by postsynaptic LHb signaling cascades
(8/13)

We found that EP opto stim did NOT modulate DA before any pairing (i.e. I = E).

However, DA & anticipatory licking increases upon EP opto stim after pairing with reward.

Moreover, such DA & anticipatory licking increase reverses when EP opto was paired with punishment.

(7/13)

We tested this hypothesis by exactly creating such link. To do so, we created a task where licking upon opto activation of EP>LHb axons leads to water reward or air puff punishment.

Simultaneously, we also record DA release in NAc as the final output of the circuit.

(6/13)

Since LHb inhibits dopamine (DA), we thus hypothesize a sign-switching plasticity rule for EP>LHb synapses:

1. If EP Sst activation is linked to good outcomes, make EP>LHb more inhibitory
2. If EP Sst activation is linked to bad outcomes, make EP>LHb more excitatory

(5/13)

Previous work from our lab by @WallaceLab1 and @SeulAhKim2 found that Glu & GABA are co-packaged inside individual synaptic vesicles.

This raises the intriguing possibility that EP>LHb synapses can be a graded synapses that can be net excitatory or net inhibitory🚦

(4/13)

However, our brain does contain synapses that theoretically are able to implement sign-switching plasticity 🔃

We think the synapses between Sst neurons in the entopeduncular nucleus (EP) and their targets in lateral habenula (LHb) are intriguing candidates ☯️

(3/13)

In our brain, neurotransmitter glutamate and GABA are normally released by different neurons. Therefore, the sign of a synapse is typically either excitatory or inhibitory.

Brain plasticity rules are thus thought to modify synaptic weights but not signs, unlike ANNs.

(2/13)