During active avoidance (AA), subjects learn to emit instrumental actions that escape conditioned threats and avoid pain. Little is known about the neural mechanisms mediating avoidance responses (ARs), because AA research stalled in the 1970s amid heated theoretical debates, inconsistent results with crude brain manipulations, and suggestions that Pavlovian processes were sufficient to explain ARs. This is unfortunate, as AA is the prototypical paradigm for studying aversively-motivated instrumental behavior, a class of defensive learning that almost certainly contributes to both adaptive and maladaptive active coping strategies. AA has several advantages as an adaptive coping/resilience mechanism: 1) it blunts fear and anxiety reactions, 2) it is less context-dependent and more persistent than fear extinction, and, unlike passive avoidance, 3) it allows subjects to remain engaged in environments that include danger. Conversely, maladaptive ARs may be particularly problematic because AA is extremely resistant to extinction and occasionally, paradoxically, enhanced by punishment (?vicious circle behavior?); processes that may be operating in human disorders ranging from anxiety (i.e. OCD) to addiction. Since AA likely reflects instrumental learning layered over previously acquired fear conditioning, the experiments in this proposal will leverage knowledge about Pavlovian fear and appetitive instrumental brain circuits to determine training conditions and neural mechanisms mediating cognitive vs. reflexive avoidance. Shuttlebox avoidance and a novel outcome-devaluation procedure will be used to examine the development of AA habits in rats. Experiments will focus on resolving conflicting predictions of our ?amygdala disengagement? model of habitual AA, derived from AA studies, and the ?parallel pathways? model of habit formation, derived from studies of positive reinforcement with drugs or food. Based on converging lines of preliminary data, we hypothesize that asymptotic ARs transition from goal-directed to habitual with time, as basolateral amygdala (BLA) disengages and competition between parallel circuits in dorsal striatum shifts to favor stimulus-response (S-R) over response-outcome (R-O) control. Explicit feedback cues will be included during AA training, and outcome-devaluation will be achieved by counterconditioning (pairing response-produced safety signals with shock). Our objectives are to determine 1) whether training- or time-dependent processes lead to habitual AA, 2) whether BLA projections to dorsomedial striatum mediate outcome-dependent AA, and 3) whether dorsolateral striatum (DLS) mediates habitual ARs independent of central amygdala. Chemogenetic suppression of neural activity in specific amygdalostriatal pathways will be used with devaluation to probe avoidance circuits. If successful, these studies will demonstrate that avoidance habits depend on a time- dependent disengagement of BLA and a shift towards DLS control of habitual ARs. Disrupting this S-R circuit in DLS while undermining safety signals could facilitate treatments aiming to break AA habits.
Unusually persistent or maladaptive active avoidance responses likely contribute to a range of human disorders including anxiety and addiction. We will use shock-avoidance and outcome-devaluation tests (counterconditioning of response-produced safety signals) in rats to identify and manipulate amygdalostriatal circuits that make unique contributions to goal-directed vs. habitual avoidance.
