Conservatism and the neural circuitry of threat

Conservatism and the neural circuitry of threat: economic conservatism predicts greater amygdala–BNST connectivity during periods of threat vs safety
academic.oup.com/scan/article/13/1/43/4596542
Walker S Pedersen, L Tugan Muftuler, Christine L Larson
Social Cognitive and Affective Neuroscience, Volume 13, Issue 1, January 2018, Pages 43–51, doi.org/10.1093/scan/nsx133
Published: 06 November 2017

Abstract
Political conservatism is associated with an increased negativity bias, including increased attention and reactivity toward negative and threatening stimuli. Although the human amygdala has been implicated in the response to threatening stimuli, no studies to date have investigated whether conservatism is associated with altered amygdala function toward threat. Furthermore, although an influential theory posits that connectivity between the amygdala and bed nucleus of the stria terminalis (BNST) is important in initiating the response to sustained or uncertain threat, whether individual differences in conservatism modulate this connectivity is unknown. To test whether conservatism is associated with increased reactivity in neural threat circuitry, we measured participants’ self-reported social and economic conservatism and asked them to complete high-resolution fMRI scans while under threat of an unpredictable shock and while safe. We found that economic conservatism predicted greater connectivity between the BNST and a cluster of voxels in the left amygdala during threat vs safety. These results suggest that increased amygdala–BNST connectivity during threat may be a key neural correlate of the enhanced negativity bias found in conservatism.

conservatism, bed nucleus of the stria terminalis, BNST, amygdala, 7T fMRI, threat bias

Introduction
It has been argued that an enhanced negativity bias—the tendency to give greater attention and react more strongly toward negative vs positive stimuli (Norris et al., 2010)—predisposes individuals to gravitate toward conservative ideology, because conservatism prioritizes stability and the minimization of tangible threats, while liberal ideology prioritizes social change and egalitarianism (Jost et al., 2009; Hibbing et al., 2014). This line of reasoning is based on research demonstrating an increased bias toward negative and threatening stimuli among those who endorse conservative views (Jost et al., 2003; Hibbing et al., 2014; Lilienfeld and Latzman, 2014). Conservatism is associated with a greater attentional bias toward negative words, images and angry faces (Carraro et al., 2011; Dodd et al., 2012; McLean et al., 2014), increased interpretation of ambiguous facial expressions as threatening (Vigil, 2010) and increased physiological reactivity to negative stimuli (Oxley et al., 2008; Smith et al., 2011; Dodd et al., 2012). Moreover, the link between conservative views and sensitivity to threat has been confirmed in multiple meta-analyses (Jost et al., 2003, 2017; Burke et al., 2013).

Despite this evidence, the link between conservatism and threat has not been universally accepted (Jost et al., 2017). For example, some have argued that threat sensitivity is associated with extremism on both sides of the political spectrum (Greenberg and Jonas, 2003; van Prooijen et al., 2015). Other theories predict that threat sensitivity is positively associated with greater social conservatism but also with greater economic liberalism (Duckitt and Sibley, 2010), particularly among those who are low on political engagement (Malka and Soto, 2015; Crawford, 2017).

Despite the wealth of research on conservatism and the negativity bias, little is known about the neural mechanisms underlying this association. Based on research demonstrating that the amygdala is involved in the detection of and response toward threat (Davis and Whalen, 2001; Sander et al., 2003; Öhman, 2005; Davis et al., 2010), studies have investigated how amygdala structure and function is related to individual differences in conservatism (Kanai et al., 2011; Schreiber et al., 2013). Kanai et al. (2011) found that conservatism was associated with greater gray matter volume in the amygdala, and suggested that this finding may be associated with the emotional and cognitive differences across political orientation, particularly those associated with ‘managing fear and uncertainty’ (p. 678). Schreiber et al. (2013) examined amygdala activation in Democrats and Republicans during a risk-taking task in which participants had the option of receiving a small, but guaranteed monetary reward, or taking a gamble that would sometimes result in a larger reward, and sometimes result in a commensurate monetary loss. They found that Republicans exhibited a larger amygdala response for trials in which they took the risky gamble and won a large reward vs trials in which they took the safe option and received a small reward.

Although these studies suggest that differences in amygdala structure and function are associated with political orientation, neither of these studies examined whether conservatism is associated with an increase in amygdala reactivity toward threat or toward negative stimuli more generally. Although Kanai et al. (2011) found greater gray matter volume in the amygdala of conservatives, they did not attempt to determine whether this increase in amygdala volume predicted a greater response toward threat. Similarly, Schreiber et al.’s (2013) finding that Republicans exhibit a greater amygdala response on winning risky trials vs winning safe trials could imply either that conservatism is associated with greater amygdala reactivity to the uncertainty accompanying risk taking or that conservatism is associated with greater amygdala reactivity to receiving a large (vs small) reward. As noted by Tritt et al. (2014), given that the amygdala has been implicated in the response to reward (Baxter and Murray, 2002; Murray, 2007; Cunningham and Brosch, 2012), neither of these studies definitively demonstrate that conservatism is associated with increased amygdala reactivity toward threat.

In addition to the amygdala, the bed nucleus of the stria terminalis (BNST) is another neural structure that is relevant for individual differences in threat reactivity. The BNST is a region of the basal forebrain, which is heavily connected, both structurally and functionally, with the amygdala. An influential model of the neural threat response states that connectivity between the amygdala and BNST is critical in the initiation of the response to sustained or uncertain threat (Davis et al., 2010). In support of this theory, rodent research has implicated amygdala–BNST connectivity in the response to sustained threat (Lee and Davis, 1997; Keen-Rhinehart et al., 2009; Cai et al., 2012; Flandreau et al., 2012; Asok et al., 2016). Although there is little human research examining amygdala–BNST connectivity, a growing body of literature indicates that the BNST responds to a variety of threat-related stimuli (Mobbs et al., 2010; Somerville et al., 2010; Choi et al., 2012; Klumpers et al., 2015; Pedersen et al., 2016). The BNST mediates this response via numerous pathways, including outputs which affect the functioning of the hypothalamic–pituitary-adrenal axis and the autonomic nervous system (Davis et al., 2010). Although past findings highlight altered amygdala–BNST connectivity as a likely candidate as a neural correlate for the increased threat bias associated with conservatism, no studies have investigated this possibility. In fact, the function and connectivity of the BNST is understudied more generally, in part because the small size of the BNST makes it difficult to study with standard neuroimaging techniques such as 3-Tesla fMRI scanning. Thus, high-resolution imaging is needed to investigate whether changes in amygdala–BNST connectivity are associated with individual differences in conservatism.

Additionally, while past research has investigated amygdala differences associated with conservatism, no studies have investigated the separable influences of social and economic conservatism on amygdala structure or activity. Social and economic conservatism are distinct, although correlated, constructs (Treier and Hillygus, 2009; Carmines et al., 2012; Feldman and Johnston, 2014; Ksiazkiewicz et al., 2016) and may differentially predict some personality and cognitive traits. Arguments that sensitivity to threat is associated with social conservatism and economic liberalism (Malka and Soto, 2015; Crawford, 2017) highlight the need to measure social and economic political attitudes separately, particularly when studying the threat bias.

To test whether social and economic conservatism are associated with increased reactivity in neural threat circuitry, we monitored participants’ resting brain activity via fMRI while under threat of an unpredictable shock and while safe. We used 7-Tesla magnetic resonance imaging (MRI) to enable the collection of high-resolution functional images (0.86 × 0.86 × 1 mm) to adequately detect amygdala–BNST connectivity. We predicted that conservatism would be associated with greater changes in resting-state connectivity between the BNST and the amygdala during periods of threat vs safety.

Discussion
Based on past research demonstrating that conservatism is associated with an increased bias toward negative or threatening stimuli (Jost et al., 2003; Oxley et al., 2008; Vigil, 2010; Carraro et al., 2011; Smith et al., 2011; Dodd et al., 2012; Hibbing et al., 2014; Lilienfeld and Latzman, 2014; McLean et al., 2014), and that conservatism is related to altered amygdala structure (Kanai et al., 2011) and function (Schreiber et al., 2013), we sought to test whether conservatism is associated with an increased neural threat response. Based on an influential model (Davis et al., 2010) stating that connectivity between the amygdala and BNST is a critical component of the response to prolonged or uncertain threats, we predicted that economic and social conservatism would be associated with greater changes in connectivity between these regions across conditions of potential threat and safety. We tested this by having participants complete task-free scans, one while under threat of shock and one while safe. We assessed connectivity between the amygdala and BNST by calculating a correlation between the blood-oxygen-level dependent (BOLD) response of the BNST and that of the amygdala during these respective scans. This gave us a measure of the degree to which the activity in these regions was coupled, suggesting connectivity.

Although we were unable to include social conservatism as a variable in our analysis due to lack of variation in our sample, our results were congruent with our hypothesis—economic conservatism predicted greater connectivity between a cluster in the left amygdala and the BNST for a threat minus safe contrast. This suggests that greater reactivity in threat-related neural circuitry is associated with economic conservatism. Critically, this is the first study to demonstrate that conservatism is associated with changes in connectivity between the amygdala and BNST. This is particularly important, given that no past studies have tied conservatism to altered amygdala function in the context of threat.

Follow-up tests revealed that the increased amygdala–BNST connectivity for threat vs safe conditions in economically conservative participants was driven by both increased connectivity during threat and reduced connectivity during the safe condition. The former finding suggests that economic conservatism is associated with increased sensitivity in neural circuitry associated with the threat response. However, the latter finding, that economic conservatism is associated with a decrease in amygdala–BNST connectivity during conditions of safety was unexpected. This finding suggests that individuals who exhibit more reactivity to threat also exhibit a reduced baseline in amygdala–BNST connectivity, possibly due to some negative feedback mechanism.

A growing body of literature links conservatism to an increased threat bias (Hibbing et al., 2014). Participants who are high in conservatism exhibit increased capture of attention by task-irrelevant aversive stimuli (Carraro et al., 2011; McLean et al., 2014), have better memory for negative vs positive scenes (Mills et al., 2016), and when shown a variety of images, spend more time viewing negative ones (Dodd et al., 2012). High conservatism is also related to greater skin conductance responses to aversive stimuli (Oxley et al., 2008; Smith et al., 2011; Dodd et al., 2012), and increased levels of self-reported phobic-fears (Hatemi et al., 2013). Furthermore, participants who are high in conservatism anticipate experiencing more negative emotional reactions when imagining a future negative event, and report experiencing a more negative emotional reaction toward receiving a lower than expected exam grade (Joel et al., 2014). This study found that conservatism is associated with a greater increase in connectivity between the amygdala and BNST during conditions of unpredictable threat compared with safety. As this connectivity is thought to be important in initiating the response to sustained threat (Davis et al., 2010), changes in this connectivity may be an important neural mechanism underlying the increased threat bias accompanying high conservatism.

Greenberg and Jonas (2003) have argued that sensitivity to threat is associated with extreme views on either side of the political spectrum. Our results run contrary to this argument, demonstrating a linear relationship between economic conservatism and threat sensitivity in an important component of neural threat circuitry. Van Prooijen et al. (2015) suggest that there may be both an overall linear relationship between conservatism and sensitivity to threat, and an underlying quadratic trend, with individuals at both political extremes exhibiting greater threat sensitivity than their more moderate counterparts. Future research with larger, more politically diverse samples is needed to test whether sensitivity in neural threat circuitry follows this pattern.

Some have proposed a model of the personality differences underlying political ideology that predicts that sensitivity to threat is associated with social conservatism but also economic liberalism (Duckitt and Sibley, 2010). Our findings, however, suggests that economic conservatism is associated with greater reactivity in an important component of neural threat circuitry. Malka and Soto (2015) have proposed that individuals with traits like threat sensitivity may be heavily predisposed toward social conservatism and more modestly toward economic liberalism but that this pattern may only manifest in participants who are low in political engagement. In this view, those who are sensitive to threat are drawn toward social conservatism. However, those who are high in political engagement may be motivated to adopt views that are consistently conservative across both social and economic domains. This may cause individuals who are high in threat sensitivity to adopt economic conservatism to maintain consistency with their socially conservative views. Because of this, economic conservatism may predict sensitivity to threat among the highly politically engaged but not because of a direct relationship between the two. One would expect that among our socially liberal sample, those who endorsed economically conservative views were relatively unaffected by motivations to maintain consistency across social and economic domains. As such, our findings seem inconsistent with Malka and Soto’s (2015) model. However, future research should more directly test whether the relationship between economic conservatism and sensitivity in neural threat circuitry is dependent on political engagement.

Although our results demonstrate an association between neural function and economic conservatism, these results do not address causality. The structure and function of the brain can be shaped by experience (Mechelli et al., 2004; Ceccarelli et al., 2009; Fu and Zuo, 2011; Woollett and Maguire, 2011; Klimecki et al., 2014). As such, the functional differences in neural threat reactivity associated with conservatism that we have observed could either be a heritable trait that predisposes individuals toward economic conservatism or a neural change that has developed because of the adoption of conservative economic views. In practice, political ideology and neural structure and function likely influence one another in a dynamic process that unfolds over time (Jost et al., 2014).

It is important to note that while these results add to current research suggesting an increased negativity bias associated with conservatism (Hibbing et al., 2014), this line of research should not be construed as implying that one ideology is superior to another. As past researchers have noted, the finding that conservatives exhibit an enhanced negativity bias has sometimes been used to paint conservatives in a negative light (Dodd et al., 2012; Motyl and Iyer, 2014). Although stronger in conservatives on average, the negativity bias—the tendency to attend and react more strongly to negative vs positive stimuli—is a general characteristic of human psychology (Norris et al., 2010). Furthermore, the negativity bias is thought to be an evolutionarily adaptive trait that helps individuals avoid danger, and individual variation in this trait is not associated with decreased life satisfaction (Norris et al., 2011). Surely attending to threat is necessary not only on a personal level but also on a societal level as well. The specifics of what should be considered a threat, the amount of attention each threat should be given and the best way to mitigate threats are matters that warrant careful consideration along with thoughtful public debate.

Although the finding that economic conservatism is accompanied by increased neural reactivity toward threat is a valuable starting point for future research, readers should also consider the limitations of this study. To avoid anticipation effects during the safe condition—in which participants may exhibit heightened anxiety during the safe scan if they know that a threat scan will follow shortly—we did not counterbalance the order of the threat and safe scans. As such, it is possible that the neural differences associated with conservatism observed were the result of an altered response to the order of the blocks, rather than an altered response to threat. Additionally, this study involved a small sample size (n = 24), and a relatively large number of participants whose data were excluded due to both artifacts in the data and participant attrition. Future research should seek to replicate these results while counterbalancing the order of conditions in a larger sample.

In addition, conservatives were underrepresented in our sample. The most common response to our political alignment question was ‘Somewhat Liberal’ with no participants endorsing the ‘Very Conservative’ option. This issue is not unique to this study. For example, the past studies investigating changes in the amygdala associated with conservatism also involved left-leaning samples. Of the 90 participants who took part in Kanai et al.’s (2011) studies, none identified as ‘Very Conservative’, while 60 Democrats and 22 Republicans took part in Schreiber et al.’s (2013) study. Thus, while existing research demonstrates that variations in conservatism is accompanied by altered amygdala structure and function in somewhat liberal samples, future research is needed to confirm that a similar pattern exists within samples that are distributed more evenly along the liberal-conservative spectrum.

Our results suggest that increased reactivity to potential threat in the amygdala and BNST may be an important neural correlate of the increased reactivity to threat that accompanies conservatism (Hibbing et al., 2014). However, although we did find increased neural reactivity to threat associated with conservatism, conservatives did not report greater changes in state anxiety from threat to safety. Given the high face validity of our threat of shock manipulation, it is very likely that participants were aware that the purpose of the manipulation was to induce state anxiety. As such, reports of state anxiety may have been colored by expectation or good-participant effects that may have washed out individual differences in state anxiety associated with conservatism. Alternatively, our small sample size may not have enabled adequate power to detect a relationship between conservatism and self-reported anxiety. Future research should seek to tie alterations in the function of the amygdala and BNST to behavioral and self-reported indicators of the enhanced negativity bias in conservatism.

Although future research is needed to further examine the neural mechanisms underlying the increased negativity bias in conservatism, this study adds to existing literature suggesting that conservatism is associated with altered amygdala function (Schreiber et al., 2013). This is the first study to show altered amygdala function in conservatism during threat by employing high-resolution 7T fMRI to demonstrate that conservatism is associated with increased amygdala–BNST connectivity during the anticipation of threat vs safety. This is critical, because connectivity between the amygdala and BNST is thought to be an important component of the neural circuitry that coordinates the response to sustained and uncertain threat (Davis et al., 2010). As such, increased amygdala–BNST connectivity during threat may be a key neural correlate of the enhanced negativity bias found in conservatism.