Excessive avoidance and diminished approach behavior are both prominent features of anxiety, trauma and stress related disorders. Despite this, little is known about the neuronal mechanisms supporting gating of human approach-avoidance behavior. Here, we used functional magnetic resonance imaging (fMRI) to track dorsal anterior cingulate and medial prefrontal (dACC/dmPFC) activation along an approach-avoidance continuum to assess sensitivity to competing appetitive and aversive contingencies and correspondence with behavior change. Behavioral and fMRI experiments were conducted using a novel approach-avoidance task where a monetary reward appeared in the presence of a conditioned stimulus (CS), or threat, that signaled increasing probability of unconditioned stimulus (US) delivery. Approach produced the reward or probabilistic US, while avoidance prevented US delivery, and across trials, reward remained fixed while the CS threat level varied unpredictably. Increasing the CS threat level (i.e., US probability) produced the desired approach-avoidance transition and inverted U-shaped changes in decision times, electrodermal activity and activation in pregenual ACC, dACC/dmPFC, striatum, anterior insula and inferior frontal regions. Conversely, U-shaped changes in activation were observed in dorsolateral and ventromedial prefrontal cortex and bimodal changes in the orbitofrontal and ventral hippocampus. These new results show parallel dorsal-ventral frontal circuits support gating of human approach-avoidance behavior where dACC/dmPFC signals inversely correlate with value differences between approach and avoidance contingencies while ventral frontal signals correlate with the value of predictable outcomes. Our findings provide an important bridge between basic research on brain mechanisms of value-guided decision-making and value-focused clinical theories of anxiety and related interventions.

No abstract

a b s t r a c tSymbolic generalization of avoidance may underlie the aetiology and maintenance of anxiety disorders. The aim of the present study was to demonstrate inferred threat-avoidance and safety (non-avoidance) behaviours that occur in the presence of stimuli indirectly related to learned threat and safety cues. A laboratory experiment was conducted involving two symbolic stimulus equivalence relations consisting of three physically dissimilar stimuli (avoidance cues: AV1eAV2eAV3 and neutral cues: N1eN2eN3). During avoidance learning involving aversive images and sounds, a key-press avoidance response was trained for one member of one of the relations (AV2) and non-avoidance for another (N2). Inferred threat and safety behaviour and ratings of the likelihood of aversive events were tested with presentations of all remaining stimuli. Findings showed a significantly high percentage of avoidance to both the learned and inferred threat cues and less avoidance to both the learned and inferred safety cues. Ratings in the absence of avoidance were high during training and testing to threat cues and low to safety cues and were generally lower in the presence of avoidance. Implications for associative and behavioural accounts of avoidance, and modern therapies for anxiety disorders are discussed.