Social hierarchies exert a powerful influence on behavior, but the neurobiological mechanisms that detect and regulate hierarchical interactions are not well understood, especially at the level of neural circuits. Here, we use fiber photometry and chemogenetic tools to record and manipulate the activity of nucleus accumbens-projecting cells in the ventromedial prefrontal cortex (vmPFC-NAcSh) during tube test social competitions. We show that vmPFC-NAcSh projections signal learned hierarchical relationships, and are selectively recruited by subordinate mice when they initiate effortful social dominance behavior during encounters with a dominant competitor from an established hierarchy. After repeated bouts of social defeat stress, this circuit is preferentially activated during social interactions initiated by stress resilient individuals, and plays a necessary role in supporting social approach behavior in subordinated mice. These results define a necessary role for vmPFC-NAcSh cells in the adaptive regulation of social interaction behavior based on prior hierarchical interactions.

Impairments in social behavior are observed in a range of neuropsychiatric disorders and several lines of evidence have demonstrated that dysfunction of the prefrontal cortex (PFC) plays a central role in social deficits. We have previously shown that loss of neuropsychiatric risk genevCacna1cvthat codes for the Cav1.2 isoform of L-type calcium channels (LTCCs) in the PFC result in impaired sociability as tested using the three-chamber social approach test. In this study we aimed to further characterize the nature of the social deficit associated with a reduction in PFC Cav1.2 channels (Cav1.2PFCKOmice) by testing male mice in a range of social and non-social tests while examining PFC neural activity using in vivo GCaMP6s fiber photometry. We found that during the first investigation of the social and non-social stimulus in the three-chamber test, both Cav1.2PFCKO male mice and Cav1.2PFCGFPGFP controls spent significantly more time with the social stimulus compared to a non-social object. In contrast, during repeat investigations while Cav1.2PFCWTmice continued to spend more time with the social stimulus, Cav1.2PFCKOmice spent equal amount of time with both social and non-social stimuli. Neural activity recordings paralleled social behavior with increase in PFC population activity in Cav1.2PFCWTmice during first and repeat investigations, which was predictive of social preference behavior. In Cav1.2PFCKOmice, there was an increase in PFC activity during first social investigation but not during repeat investigations. These behavioral and neural differences were not observed during a reciprocal social interaction test nor during a forced alternation novelty test. To evaluate a potential deficit in reward-related processes, we tested mice in a three-chamber test wherein the social stimulus was replaced by food. Behavioral testing revealed that both Cav1.2PFCWTand Cav1.2PFCKO mice showed a preference for food over object with significantly greater preference during repeat investigation. Interestingly, there was no increase in PFC activity when Cav1.2PFCWTor Cav1.2PFCKOfirst investigated the food however activity significantly increased in Cav1.2PFCWTmice during repeat investigations of the food. This was not observed in Cav1.2PFCKOmice. In summary, a reduction in Cav1.2 channels in the PFC suppresses the development of a sustained social preference in mice that is associated with lack of PFC neuronal population activity that may be related to deficits in social reward.