Post

For a neural circuit to be "specialized", does it need specialized neurons? In our new work - now out in @NatureNeuro - we address this question in the BLA, using representational geometry to propose a new population-centric way of defining circuit specialization. 🧵👇 1/

2/ The BLA is known to have specialized circuits for valence (+/-), but what about other features of emotional experiences, such as behavior? In a virtual burrow assay, we observed that mice both tremble (~freeze) and ingress (~flee) in response to aversive (shock-paired) odors.

3/ Both trembling and freezing are more frequent in response to aversive odors (CS+), but far from exclusive. Mice typically tremble and then ingress when CS+ is presented, in accordance with the predator-imminence theory (freeze -> flee).

4/ What about the neural code? All variables were highly decodable, but representing the world is not enough to act on it. What matters is "how" this information is organized!

5/ For example, to make quick, precise decisions when faced with threats, the brain needs to read out information precisely (low interference) and flexibly (high generalization) - this is a "specialized" circuit. How can we design one?

6/ A circuit made of specialized sub-populations, such as "valence neurons" and "tremble neurons", clearly satisfies these requirements: sub-populations encode variables independently, and each trivially generalizes across contexts because it cares about only one thing.

7/ But we can also characterize these properties (high generalization, low interference) on a population level. For generalization, we can use cross-condition generalization performance (CCGP). For interference, we developed a new measure: cross-variable interference (CVI):

8/ The key is that specialized neurons are not necessary to satisfy these requirements! We can have the same properties (high CCGP, low CVI) as long as the representational geometry of the encoded variables is disentangled / compositional (left panel):

10/ Crucially, it does so without specialized populations! Selectivity is largely mixed across all variables:

11/ The geometry then changes after ingressing, and the mouse no longer represents valence. Instead, only the exposed, aversive condition stands out from the bulk. This is compatible with the coding of a new variable, "safety", which has the highest decoding and CCGP.

12/ In short, specialized readouts can be supported by a neural population even with mixed selectivity, as long as the geometry is right. We find that BLA simultaneously encodes multiple variables within an emotional experience, with the right geometry for specialized readouts.