Ultimately we need to move beyond thinking about individual neural circuits in isolation and towards understanding how neural circuits function in relation to each other (14/14)

This work hi-lites the often overlooked role of redundancy in neural circuit control of behavior. Redundant encoding allows for simple integration of multiple circuits. Specificity in how each circuit is engaged supports granular encoding (amplifying impact of unrewarded outcomes) (13/14)

While each input is specialized to execute this under distinct conditions, once engaged, they redundantly modulate behavior pointing to complementary roles in control of reward seeking (12/14)

Putting this altogether- we show that mPFC-NAc & vHip-NAc dynamically track outcome information to modulate engagement (11/14)

We showed this is causal- opto stim after the outcome increased latency to press on the next trial - and also we show that it is the total glutamatergic input that tunes engagement, independent of input identity. The NAc doesn’t seem to care where the glutamate comes from (10/14)

Which is exactly what we found! As mPFC-NAc & vHip-NAc activity increased, so did the latency to respond on the following trial (9/14)

So what exactly is the behavioral relevance of this neural integration of reward? We found that removing response requirements disrupted encoding in both circuits, leading us to hypothesize that outcome-associated neural activity in mPFC & vHip-NAc might be modulating task engagement… (8/14)

But when we explicitly tested redundancy between these circuits, we found there were differences. While input from the mPFC invariantly encodes reward, encoding in vHip input is uniquely anchored to unrewarded outcomes (7/14)

This suggested the two circuits might be redundantly integrating reward... (6/14)

Unexpectedly, we found that mPFC-NAc & vHip-NAc seemed to both encode outcomes in the same way. When a choice was rewarded, activity in both inputs was suppressed, and with each consecutive unrewarded choice, activity gradually increased in both inputs (5/14)

We asked if each contributes unique information or if there is redundancy in how these circuits encode reward. We simultaneously recorded from mPFC-NAc and vHip-NAc neurons using dual-site fiber photometry in mice engaged in a two-armed bandit task (4/14)

There’s a lot known about mPFC-NAc and reward. Given that inputs from the vHip converge with mPFC inputs in NAc, we wanted to understand how these two circuits might be working together to support reward processing (3/14)

Here we unpack the apparent redundancy in how mPFC and vHip inputs to the NAc shell integrate reward. We identify a novel motif of reward integration that is common to both inputs but used differently…(2/14)