Differential encoding of social identity, valence and unfamiliarity in the amygdala and piriform cortex

Abstract

To appropriately respond to social situations, animals must rapidly identify conspecifics as individuals, gauge their familiarity and recall any good or bad associations. In previous work (Mazuski & OKeefe, 2022), we characterized populations of neurons in the basolateral amygdala complex (BLA) that were strongly tuned to social conspecifics. In this current study, we asked whether these or other populations in the BLA encode properties necessary for social interaction: individual recognition, familiarity detection and reward association. In addition to cells in the rat lateral amygdala (LA) and basolateral amygdala (BLA), we recorded from cells in the anatomically-related piriform cortex (PIR) to see how much of the amygdala response was inherited from this input. Here, we show that these three interrelated areas encode distinct features of social conspecifics at the population and single neuron level. We recorded from large populations of neurons using 4-shank Neuropixels probes while rats learned a social discrimination task, where one male conspecific was paired with a food reward (S+) and the other its absence (S-). For comparison, two unfamiliar animals were presented on occasional probe trials. After successful discrimination learning, we looked for transfer effects to our previously studied task where implanted rats freely interacted with all four conspecifics in an open-field environment. The LA dynamically encoded social reward learning. Initially, LA neuronal population activity could not distinguish between the S+ and S-conspecifics, but after learning distinct representations emerged. In contrast, the PIR showed no effect of social reward learning. Single PIR units encoded social identity with strong tuning to individual conspecifics both before and after learning. The BLA encoded both reward learning and social identity suggesting this is where the PIR and LA information streams converge. Unexpectedly, the BLA and LA strongly coded for the unfamiliarity of the rat - single neurons responded to the probe trial animals with firing rates 2-5x higher than to the task animals. This cross-region encoding of identity, reward and unfamiliarity was reactivated during social interaction in the new open-field environment, demonstrating that the learning was not context specific. The results throw light on how distinct neural circuits contribute to social recognition and memory and how they interact with each other.