Avoidance of harm and anxiety: A role for the nucleus accumbens

Abstract

Active and passive avoidance behaviors involve either emitting or omitting a response to avoid potential harm. Both are key components in the etiology and maintenance of anxiety disorders, yet the neural circuitry that mediates avoidance is underexplored. Using functional magnetic resonance imaging greater hemodynamic activation of the nucleus accumbens was found during active avoidance, whereas greater deactivation of the nucleus accumbens was observed during passive avoidance. These findings extend the role of the NAcc from purely reward-based action-contingencies, to one that also involves either emitting or withholding a response to avoid harm. Critically, the degree of activation and deactivation of the NAcc during avoidance was associated with individual levels of anxiety, which supports the idea that the NAcc may play a key role in the etiology and maintenance of aberrant avoidance behaviors in disorders of anxiety.

Introduction

Examining the neural basis of avoidance is central to gaining a better understanding of anxiety disorders, which are characterized by high levels of both active and passive avoidance behaviors (Dymond and Roche, 2009, Risbrough, 2010). Indeed, models of avoidance behavior emphasize the role not only of classical aversive conditioning, as initially proposed by the two-factor theory (Mowrer, 1951), but also of anxiety in the acquisition and persistence of avoidance behavior (Deakin and Graeff, 1991, Lovibond et al., 2008). However, to date research in the field of human anxiety has tended to focus mainly on brain regions mediating classical aversive conditioning, rather than the neural circuitry mediating instrumental avoidance behavior (Dymond and Roche, 2009). To redress this imbalance, here we investigate the neural basis of avoidance behaviors and associated anxiety using functional magnetic resonance imaging (fMRI), with an emphasis on the role of the nucleus accumbens (NAcc); a region increasingly shown to be involved in modulating behavioral responses to both rewarding and aversive events (e.g., Carelli, 2002, Jensen et al., 2007, Levita et al., 2009, Reynolds and Berridge, 2002, Wadenberg, 2010).

There are a number of brain imaging studies in humans that support a role for the NAcc in active avoidance behavior. For example, Mobbs et al. (2007) found that NAcc activity was associated with a decrease in dread and an increase in confidence of avoiding a virtual predator, while Jensen et al. (2003), Delgado et al. (2009) and Niznikiewicz and Delgado (2011) all found significant activation in the NAcc in response to a stimulus that signaled an aversive event that could be actively avoided. Additional support for a role for the NAcc in avoidance behavior is provided by the efficacy of deep-brain stimulation of the NAcc in otherwise treatment resistant obsessive–compulsive disorder patients who show debilitating avoidance behaviors (Sturm et al., 2003). However, no study to date has explicitly tested whether the NAcc is also engaged when humans are required to withhold a specific response to avoid punishment (passive avoidance). Furthermore, the role of the NAcc in human avoidance behavior has not always been supported by functional brain imaging, as some studies using monetary loss paradigms have failed to find NAcc activation during active avoidance (Schlund and Cataldo, 2010, Schlund et al., 2010, Schlund et al., 2011, but see Niznikiewicz and Delgado, 2011).

The ambiguity regarding the exact role of the NAcc in avoidance behavior in humans contrasts with studies in rodents that consistently demonstrate that the NAcc does play a key role in both passive and active avoidance behavior (Burhans and Gabriel, 2007, David et al., 2001, Manago et al., 2009, Reynolds and Berridge, 2002, Schoenbaum and Setlow, 2003), and that NAcc-dopamine plays a major role in acquisition and performance of these behaviors (Bracs et al., 1984, McCullough et al., 1993, Valenti and Grace, 2010, Wadenberg, 2010, Wadenberg et al., 1990). The role of dopamine in avoidance behavior is consistent with the responsivity of dopaminergic neurons to both fearful and rewarding stimuli (Brischoux et al., 2009, Matsumoto and Hikosaka, 2009), and the anxiogenic effect of cocaine, a dopamine reuptake inhibitor, in the NAcc (David et al., 2001). Moreover, work in non-human primates and rodents suggest that the NAcc may be responsible for the etiology and persistence of pathological avoidance behavior in anxiety disorders (Kalin et al., 2005, Zarrindast et al., 2008).

To address the uncertainty regarding NAcc involvement in anxiety-dependent avoidance behavior in humans we developed a novel task designed to test the role of the NAcc in both active and passive avoidance behavior. In addition, we also tested whether individual anxiety levels would correlate with both task performance and the degree of NAcc engagement as measured by BOLD response. Given negative findings relating to NAcc activation to the warning stimulus in some active avoidance paradigms using secondary reinforcers such as monetary loss (Schlund and Cataldo, 2010, Schlund et al., 2010, Schlund et al., 2011), in this study we used a primary reinforcer (aversive images that produce disgust) due to the possibility that the inconsistent findings regarding secondary reinforcers may be due to such reinforcers not being deemed as aversive or punishing as primary reinforcers, as the participants know that there is a potential for a rewarding outcome in other trials (but see, Niznikiewicz and Delgado, 2011). Likewise it is also possible that avoidance behavior motivated by preventing reward loss may recruit distinct neural networks to those engaged during avoidance of an aversive/noxious stimulus, as demonstrated during classical aversive conditioning (Delgado et al., 2011), or that the affective value of money may vary significantly between individuals to a much greater extent than a primary reinforcer would, thus explaining the inconsistent results when such reinforcers are used.

This study represents the first time, to our knowledge, that a direct comparison has been made between the neural circuits that mediate passive and active avoidance behaviors in humans. We demonstrate that the NAcc shows activation during active avoidance, but deactivation during passive avoidance. Moreover, a largely overlapping neural circuitry involved in both forms of avoidance was found, which included the caudate nucleus and anterior insular cortex bilaterally and the right middle (Brodmann area 6) and right medial frontal gyrus (dorsal anterior cingulate: area 32). Significantly, we also show that individual anxiety levels were associated with the degree of engagement of the NAcc during both forms of avoidance. Thus, high levels of anxiety were found to be associated with increased activation of the NAcc during active avoidance, but with greater deactivation of the region during passive avoidance, suggesting that the NAcc does play a key role in the expression of anxiety in humans, acting to influence the development and persistence of avoidance behaviors.