Emotional stress and social learning -- Annie Bryant

The distinction between consequences of physical stress versus those of emotional stress is relevant in studying PTSD, in which afflicted individuals may experience intense fear of trauma-related stimuli even if they only witnessed a traumatic event. In Sial 2015, they emphasize that although witness models of stress have been published (e.g. van den Berg et al. 1998), these paradigms lack ethological validity. Instead, they propose a modified chronic social defeat stress (CSDS) protocol that incorporates a witness component, such that the witness mouse experiences “vicarious” emotional stress (VSDS).

In this paradigm, male C57BL/6 mice were exposed to physical stress (PS) with a CD-1 aggressor, emotional stress (ES) via observing PS mouse defeat, or control stress (CON). I was surprised that placing two males together in the CON group wouldn’t result in even mild aggression, but the authors note they used a particularly non-aggressive C57 strain. Sial et al. obtained physiological readouts (serum corticosterone [CORT] levels, body weight) and behavioral readouts (social interaction test [SIT], elevated plus maze test [EPMT]) to measure susceptibility to stress. Despite the very different natures of the stress in the PS and ES mice, mice in both groups displayed reduced weight gain; increased social avoidance (up to one month after last chronic stress session); increased anxiety-like behavior in the EPMT; and increased CORT levels following acute and chronic social defeat stress. To me, the most intriguing finding was that VSDS induced a smaller degree of social avoidance in ES mice at 24 hours compared to CSDS in PS mice, but that ES mice exhibited greater social avoidance (comparable to that of PS mice) one month later. This suggests that physical trauma elicits an immediate response, but emotional stress must incubate before it produces maladaptive social behavior. Do emotional stress memories go through neuronal circuitry that physical stress memories bypass? Does this different circuitry increase the propensity for emotional stress memory engrams to be inappropriately activated in humans in response to trauma-related stimuli?

Sial et al. was a methods paper that extended upon Warren et al. 2013; as such, they don’t make any broad-sweeping conclusions or speculate as to brain regions involved in encoding these emotional stress memories. Cue (pun intended) Allsop et al., who explored the biological mechanisms by which mice use emotional stress derived from social cues to learn which stimuli to avoid. They focused on two brain structures that have been previously linked to observational learning: the anterior cingulate cortex (ACC) and basolateral amygdala (BLA). Although there are reciprocal connections between these structures, they focused predominantly on ACC à BLA projections. Using an observational fear learning paradigm with optogenetic manipulation, this group demonstrated that projections from ACC à BLA mediate socially-derived learning.

Prior studies have reported conflicting findings as to whether or not secondhand observation of aversive stimuli is sufficient for observational learning. Allsop et al. exposed one group of “observer” mice to a foot shock before watching the “demonstrator” mice get shocked with simultaneous tone and light cues, while another group of “naïve” mice was not shocked first. While both groups froze 24 hours later upon tone and light cue, the experienced observers exhibited a greater magnitude of a response, highlighting the importance of prior experience in observational learning associations. I thought it was fascinating that the naïve mice froze at all, since they had to rely solely on interpreting the response of the demonstrator mouse without knowing what exactly they were going through. The authors don’t get into how a mouse senses distress in another mouse, but perhaps this is through a release of pheromones? It makes sense from an evolutionary perspective that an animal wouldn’t need to physically experience a threat themselves to learn that it’s harmful, since many aversive stimuli in the wild are fatal. I wonder to what extent this type of observational learning could be encoded in epigenetics and passed from one generation to the next.

Allsop et al. measured in vivo recordings of observer ACC and BLA neurons during “habituation trials,” during which cues were delivered but no shocks were given to the demonstrator, as well as trials in which shocks and cues were delivered in a temporally paired or unpaired manner. They found that the magnitude and direction of change in neuronal activity across neuronal populations strongly corresponded to predictive value of cues in paired and unpaired groups. Optogenetic inhibition of ACC à BLA neurons via NpHR during cue-shock pairing reduced BLA activity during freezing; these mice also froze less upon cue delivery 24 hours later. However, inhibition of these neurons after cue-shock pairing (during the recall phase) or during classical fear conditioning didn’t affect freezing response. Allsop et al. concluded that ACC à BLA transmission is necessary for observational learning – specifically, social learning – but not for the subsequent expression of fear memory. This temporal sensitivity reminded me of last week’s papers, in which the precise timing of neuronal activity in the LA was crucial for biasing the memory trace to incorporate a given neuron.

While the foot shock model is certainly valid for investigating consequences of ES and PS, it is severe and isn’t necessarily ethologically relevant, particularly to humans. Peer bullying, on the other hand, is a common stressor in children and adolescents, and can contribute to development of psychopathology like PTSD. Allsop et al. recognized the need to translate their findings in a more ethologically relevant setting, so they investigated whether ACC à BLA neurons may transmit socially-derived information in a social defeat setting. Optogenetic inhibition of ACC à BLA neurons in observer mice while the demonstrator was placed with either an aggressive CD-1 mouse or juvenile intruder mouse subsequently decreased the observers’ social avoidance in both contexts.

Taken together, these findings indicate that ACC neurons encode socially-derived observation information and transmit it to the BLA, which then consolidates the memory to encode the aversive value of the cue, driving behavioral outcomes without firsthand experience of cue-shock pairing.