| Active avoidance (AA) learning is a two-step form of threat learning in which the learner acquires a behavioral strategy to avoid exposure to a threatening stimulus. About eighty percent of wild type rodents evince effective avoidance after adequate training; the other twenty percent continue to display conditioned fear responses and rarely avoid aversive stimuli. Scientists have exploited this natural variation in AA behavior both to illuminate the neurobiology of animal AA, and to model both anxiety and addiction in rodents. The experimental paradigms designed for use in humans, however, have failed to capture individual differences in AA learning or expression. It has been challenging, as a result, in humans both to identify the neural correlates of effective AA and to determine the nature of the relationship between AA and psychological constructs.;In order to meet these objectives, we developed a novel and translational task to investigate AA in humans. In the current manuscript we describe four studies in which we: pilot the task (Study 1); use functional magnetic resonance imaging (fMRI) to document the brain regions that mediate human AA (Study 2); and explore the role of noradrenergic neurotransmission in human AA (Studies 3 and 4). In Study 1, we demonstrate the capacity of our task to elicit variable patterns of AA behavior. We find evidence in Study 2 that the amygdala, striatum, and medial prefrontal cortex (mPFC) are key mediators of this process. Specifically, participants suppressed the right amygdala but activated the right caudate during AA acquisition, while the mPFC influenced avoidance by interacting with both of these subcortical structures. Though beta-adrenergic antagonist propranolol enabled the expression of AA in animals that exhibited robust conditioned fear responses in initial training sessions, we did not document an effect of the drug on human AA in either Study 3 or 4. We review how our data might inform future investigations of avoidance.;Finally we discuss the potential utility of our task as a tool to probe the pathophysiology of psychiatric disease. While we did not detect consistent correlations between AA performance and self-reports of either anxiety or impulsivity, we propose that such associations may only exist in clinical populations. We argue that, even if additional research does not demonstrate that AA behavior is an effective gauge of vulnerability to mental illness, our experimental protocol engages many of the neurocircuits implicated in disorders of emotion regulation and could still advance scientific understanding of mood, anxiety, and substance use disorders. |
