Ventromedial Prefrontal Cortex and Basolateral Amygdala Projections to Ventromedial Striatum Encode Active Avoidance Behavior

Active avoidance is a fundamental defensive behavior, defined as performing a voluntary action to escape a noxious stimulus. Active avoidance serves as an adaptive response to threat, ensuring safety. In individuals with anxiety disorders and post-traumatic stress disorder, avoidance behavior can become maladaptive when even safe situations are avoided, at great psychosocial cost. Corticostriatal and amygdalostriatal pathways have been implicated in active avoidance, though their role in learning and expression of this behavior is not fully understood. Projections from the ventromedial prefrontal cortex (vmPFC) to the ventromedial striatum (VMS) have been shown to regulate a variety of defensive behaviors; however, their role in active avoidance has not previously been examined. Additionally, basolateral amygdala (BLA) projections to the VMS have been shown to promote active avoidance in studies utilizing pharmacological inactivation and optogenetic inhibition, though in vivo real time neural activity in this pathway has not been recorded during active avoidance learning. Here we utilized fiber photometry for in vivo recordings of neural activity in vmPFC-VMS and BLA-VMS projections during active avoidance learning and expression. We implemented a two-way signaled active avoidance paradigm in which a light cue served as the conditioned stimulus (CS) signaling an impending foot shock. We examined changes in neural activity in these two projections during CS presentation, onset of active avoidance, and conditioned freezing. We found that vmPFC-VMS projections develop learning-related increases in activity at CS onset across training, while BLA-VMS projections do not show learning-related encoding of the CS. Additionally, we found that both vmPFC-VMS and BLA-VMS projections develop an increase in activity at avoidance onset. No changes in neural activity were observed during cued freezing in either vmPFC-VMS or BLA-VMS projections. Together these results indicate that vmPFC-VMS projections encode both CS and avoidance, while BLA-VMS projections may simply encode avoidance. Finally, we utilized optogenetic inhibition of vmPFC-VMS projections during the CS to investigate the necessity of this pathway for expression of active avoidance behavior. We found that inhibition of vmPFC-VMS projections attenuates learned active avoidance behavior, indicating that activity in this pathway is required for proper active avoidance. In summary, our results demonstrate task-relevant encoding of active avoidance behavior in vmPFC-VMS and BLA-VMS projections.