Bidirectional control of dorsomedial striatum on innate avoidance behavior

Avoidance behaviour, aimed at escaping dangerous stimuli and threatening situations, can become maladaptive when individuals avoid relatively safe situations, a hallmark of anxiety disorders. Innate avoidance, a natural aversion to ethologically relevant stimuli, involves a neural circuit in which the basolateral amygdala (BLA) and prefrontal cortex (PFC) play a central role. We have observed that both the PFC and the BLA send converging unidirectional excitatory inputs to the dorsomedial striatum (DMS), which thus may be ideally positioned to regulate the output of this circuit. Indeed, the DMS has been implicated in learned, and more recently in innate avoidance behaviours. To investigate the role of the PFC-DMS and BLA-DMS pathways in innate avoidance, we used a chemogenetic approach. CD1 male mice received a bilateral injection of AAVs expressing either the inhibitory DREADD hM4D(Gi) or the excitatory DREADD hM3D(Gq) in the PFC or BLA. Then, mice were focally injected with CNO or saline into the DMS and tested in the elevated plus maze (EPM) 30 minutes later. Inhibiting the PFC-DMS pathway did not affect the time spent in open arms of the EPM, but its activation significantly reduced anxiety. On the other hand, activation of the BLA-DMS pathway led to a reduction in the time spent in the open arms, while its inhibition had a strong anxiolytic effect. To understand why the same type of manipulation of two pathways converging in the same region generates opposite behavioral responses, we aimed to evaluate the contribution of striatal interneurons receiving direct projections from the PFC or the BLA. To this end, Ai14 mice were injected in the PFC or BLA with an anterograde virus (AAV1-hSyn-P2A-Cre-WPRE), capable of crossing the downstream synapse and infecting neurons that receive input from neurons at the injection site. Subsequently, brain sections were processed with a parvalbumin antibody to identify the percentage of double-positive striatal interneurons and verify whether a different percentage of these neurons received projections from the PFC or BLA. However, we did not observe significant differences in the percentage of parvalbumin-positive striatal interneurons receiving input from the PFC or BLA. As a final experiment, to verify how information from the PFC and BLA flows downstream of the basal ganglia, CD1 mice were injected with a viral vector expressing channelrhodopsin, while optical fibers were implanted in the GPi to specifically manipulate this projection during the EPM and OF tests. Activation of this pathway showed a reduction in time spent in the open arms of the EPM, suggesting an increase in avoidance behavior, without altering locomotion as measured in the OF. These results not only confirm the involvement of PFC-DMS and BLA-DMS projecting neurons in mice behaviour in the EPM, but also underscore that these pathways exert opposing bidirectional control over innate avoidance. Moreover, additional experimental subjects will be necessary to confirm the opposite effect of the projections does not involve a different contribution of parvalbumin-positive interneurons. Finally, this is the first evidence that information arriving in the DMS, which mediates avoidance behaviors, flows downstream to the GPi.