Intrinsic functional connectivity of blue and red brains: neurobiological evidence of different stress resilience between political attitudes

Conservatives are more sensitive to threatening/anxious situations in perceptual and cognitive levels, experiencing emotional responses and stress, while liberals are more responsive to but tolerant of ambiguous and uncertain information. Interestingly, conservatives have greater psychological well-being and are more satisfied with their lives than liberals despite their psychological vulnerability to stress caused by threat and anxiety sensitivities. We investigated whether conservatives have greater resilience and self-regulation capacity, which are suggested to be psychological buffers that enhance psychological well-being, than liberals and moderates. We also explored associations between intrinsic functional brain organization and these psychological resources to expand our neurobiological understanding of self-regulatory processes in neuropolitics. We found that conservatives, compared to liberals and moderates, had greater psychological resilience and self-regulation capacity that were attributable to greater impulse control and causal reasoning. Stronger intrinsic connectivities between the orbitofrontal cortex (OFC) and precuneus and between the insula and frontal pole/OFC in conservatives were correlated with greater resilience and self-regulation capacity. These results suggest the neural underpinnings that may allow conservatives to manage the psychological stress and achieve greater life satisfaction. This study provides neuroscientific evidence for the different responses of liberals and conservatives to politically relevant social issues.

When exposed to politically relevant experiences, individuals make judgments based on the subjective evaluation of the social issues and form a political attitude that motivates their political behaviors 1 . This process is called political socialization and is the process by which individuals shape their own political perspectives that conform to the norms and structures of political and economic institutions where they live 2 . Participating in political activities with one's own political beliefs and even engaging in daily conversations about political issues often inevitably leads to psychological stress, such as worry and irritability, about threatening or conflicting social issues 3 . Political psychology research has found that psychological responses to stressful events are different in people with left and right political attitudes because the way in which individuals perceive and manage threats or cognitive conflicts influences the formation of their political perspectives [4][5][6] .
Liberals and conservatives hold opposite political belief systems and have marked differences in psychological responses, especially in the self-regulatory processes dealing with stressful issues 7 . Compared with conservatives, liberals are more responsive to ambiguous and uncertain information, but they are also more tolerant of those stresses 8 . Neuroimaging studies have identified the neurobiological mechanisms underlying the tolerance

Results
Sociodemographic and psychological characteristics. We first demonstrated that homoscedasticity of sociodemographic and psychological variables, except the RQT sociability subtotal score (Fligner-Killeen chi-squared = 3.97, p = 0.046). The group comparison results of sociodemographic and psychological are presented in Table 1. Participants were not different in terms of their sociodemographic characteristics, except their political attitude scores. The multivariate analysis of variance (MANOVA) results of the Resilience Quotient Test (RQT) totals showed that there were statistically significant differences in psychological resilience scale scores based on political attitudes (Pillai's trace V = 0.14, F(6, 196) = 2.40, p = 0.030). We found significant group effects for the RQT total and self-regulation & sociability subtotal scores (Welch's F total (2,50.88) = 7.19, p = 0.002; Welch's F self-regulation (2, 44.10) = 6.38, p = 0.004; Welch's F sociability (2, 52. The additional MANOVA analysis of the RQT self-regulation subscales showed significant differences depending on political attitudes (Pillai's trace V = 0.14, F(6, 196) = 2.43, p = 0.027). We observed group effects for causal reasoning and impulse control subscales of the RQT self-regulation scale (Welch's F causal reasoning (2, 45. Seed-based resting-state functional connectivity. The group comparison results showed that conservatives had stronger anticorrelations between the ACC and the left superior frontal gyrus (SFG)/dorsolateral prefrontal cortex (dlPFC), between the bilateral orbitofrontal cortex (OFC) and the precuneus, and of the right insula with the superior lateral occipital cortex (LOC) and the frontal pole/OFC in the left hemisphere than liberals. In addition, conservatives had a stronger positive correlation of the left insula with the left occipital Scientific RepoRtS | (2020) 10:15877 | https://doi.org/10.1038/s41598-020-72980-x www.nature.com/scientificreports/ pole than liberals. There was no difference in resting-state functional connectivities (rsFCs) from the bilateral amygdalae, and all the functional connectivity measures were homoscedastic. We have summarized the detailed results of the seed-based functional connectivity in Table 2 & Fig. 1. The findings of functional connectivities associated with political attitudes were mostly replicated in the regression analysis using one hundred two participants (Table S1).
Global network metrics of functional brain. Although we observed that liberals (M = 0.77, SD = 0.08) had higher normalized modularity (Q) of the brain network than conservatives (M = 0.73, SD = 0.07) in the trend-level of uncorrected results (p = 0.079), there was no significant group difference in the network metrics when after correction for multiple comparisons (Fig. 2).

Associations between intrinsic functional organization and RQt measures.
A stronger anticorrelation between the left OFC and the precuneus was inversely correlated with the greater RQT total (ρ = − 0.50, p = 0.050) and self-regulation subscale scores (ρ = − 0.59, p = 0.017) in conservatives. Among the RQT self-regulation subscales, the causal reasoning and impulse control scores were also correlated with the connectivity between the left OFC and the precuneus in conservatives (ρ = − 0.60, p = 0.015; ρ = − 0.51, p = 0.045). At the trend level, there was a correlation between the RQT self-regulation subscale score and right insular connectivity with the left frontal pole/OFC (ρ = − 0.46, p = 0.073). None of the intrinsic functional organization measures were associated with RQT scores in liberals.

Discussion
To expand the neurobiological understanding of different psychological responses to stress in liberals and conservatives, we investigated associations between intrinsic brain organization and psychological resources, such as resilience and self-regulation capacity. We found greater psychological resilience and self-regulation capacity in conservatives than liberals and moderates. We also found that greater impulse control and causal reasoning www.nature.com/scientificreports/ contributed to the regulation capacity in conservatives. Conservatives had stronger intrinsic connectivities between the OFC and the precuneus and between the insula and the frontal pole/OFC than liberals, and these brain organization characteristics are associated with the buffering roles of psychological resources. These results identified the neural underpinnings that may help conservatives achieve greater psychological well-being and life satisfaction despite the susceptibility to psychological stress caused by threats and anxiety. Political psychology studies have investigated liberals' and conservatives' psychological responses to politically relevant social events and the accompanying stress because political perspectives are interactively influenced by threats and cognitive conflict perceived from social environments 4,7 . Regardless of political attitudes, people respond sensitively to stressful events that occur in a politically relevant social environment 3 . However, liberals and conservatives differ in the ways in which they recognize and respond to various political and social issues. Political psychologists have explained these psychological differences on the political spectrum from the perspectives of morality. Though relationships between ideology and morality is not simply explained in terms of a causal model, liberals are more likely to prioritize some moral values like care/harm, liberty/oppression, and fairness/cheating, whereas conservatives are more sensitive to a broader range of moral issues [31][32][33] . In addition to the different moral intuitions, they are thought to have different psychological resources and neurocognitive strategies. Liberals use a neurocognitive strategy that facilitates complementary cognitive control when facing conflicting information 10 , and this resource may allow them to accept more politically liberal views. However, it is still unknown which psychological resources and neurocognitive strategies allow conservatives to cope with the stress resulting from the sensitive perception of threatening and anxious situations while still having www.nature.com/scientificreports/ no physiological reactions to those stimuli and achieving enhanced satisfaction with their lives [15][16][17] . Compared to liberals, conservatives are more likely to conform to social rules and control their internal and external environments 19,22 . Along with a greater self-regulation capacity in conservatives, we found that they are more resilient to psychological stress. As psychological resilience and self-regulation lessen stress and enhance psychological well-being 21,34 , we suggest that these psychological resources of conservatives help them cope with the stress in a self-protective manner from threatening and anxious social circumstances and achieve greater life satisfaction. Among various factors affecting psychological resilience 35 , we identified self-regulation capacity as the psychological resource that might be used to manage stress and improve psychological well-being in conservatives, while sociability (empathy and reaching out) and positivity (optimism and self-efficacy) were not as influential as the regulation capacity. We then showed that the greater self-regulation capacity was a consequence of a higher degree of impulse control and causal reasoning in conservatives. These psychological traits may be related to their personalities. The most evident personality trait of conservatives compared to liberals is conscientiousness 36 . Conscientiousness consists of low-order structures, such as orderliness and impulse control 37 ; thus, we assume that conservatives, based on this personality trait, are more likely to follow orders and control the impulsiveness that possibly causes problematic outcomes. Having the tendency towards impulse control is also linked to affective well-being and life satisfaction 34 . Likewise, identifying problems and understanding causal relationships are strategies for managing psychological stress, and the causal reasoning may further improve psychological well-being 35,38 . Our findings provide evidence that greater impulse control and causal reasoning contribute to the psychological buffering roles of both resilience and self-regulation capacity in conservatives. Neuropolitics research has examined the self-regulatory processes of single brain regions 10,14,39,40 ; however, these limited regions are not sufficient to explain the complex structures of the psychological responses that manage stress 24 . Thus, we adopted intrinsic brain organization analyses to investigate multiple regions communicating with each other and to expand the neurobiological understanding of neuropolitics 25 . Conservatives had stronger rsFCs from the ACC, OFC, and insula that function cognitive and emotional processing to multiple www.nature.com/scientificreports/ brain regions of the default mode network (DMN), frontoparietal network (FPN), and dorsal attention network (DAN) when compared with liberals. If the OFC and precuneus became more strongly connected, conservatives were more likely to be resilient to psychological stress and have greater self-regulation capacity, specifically impulse control and causal reasoning. The OFC is one of key regions involved in self-regulation and psychological resilience, while the precuneus processes emotion regulation as coactivated with the amygdala 24,41,42 . The connectivity between these regions is related to the self-processing of negative emotional distractors and mental representation 29 . Indeed, a disruption of the functional coupling between these regions explains dysfunctional emotional processing and the negative self-view of patients with major depression 29,43 . Together with the above evidence, we suggest that the stronger intrinsic connectivity between the OFC and precuneus is a neurobiological resource that makes conservatives resilient to psychological stress. The greater self-regulation capacity of conservatives was also related to the stronger connectivity of the insula with the frontal pole/OFC, which was identified as a subregion of the DMN 44 . The insula, together with the dorsal ACC, forms a salience network (SN) that functions to detect the saliency of external stimuli and facilitates task-related processing by switching between the DMN and the central executive network 45,46 . Having an interaction with the DMN, the insula in the SN engages in an attentionally demanding task and cognitive control, especially attention shifting 47,48 . The attentional switching away from emotional stimuli is a primary strategy to regulate negative emotions and might explain the approach-avoidance motivational responses of conservatives 49,50 . Indeed, conservatives use an avoidance-based inhibitory strategy to prevent negative feelings and outcomes in the face of stressful events 50 . The association of insular connectivity with self-regulation capacity in conservatives is supported by a recent finding showing weaker insular connectivity with the frontal pole/ OFC as a brain dysfunction in disruptive behavior disorders that are characterized by cognitive and emotional dysregulations 30 . Thus, we propose that the stronger connectivity between the insula and frontal pole/OFC in conservatives is the other neurobiological resource that enhances self-regulation and lessens psychological stress.
Even through our hypothesis-driven tests assess the neural associations with psychological resilience and selfregulation, we missed to provide psychological explanations for the rest of the intrinsic brain organization differences; however, we think that these neurobiological differences might reflect some other cognitive psychological characteristics of liberals and conservatives. Conservatives had stronger connectivity between the ACC and the SFG/dlPFC and between the insula and the occipital regions than liberals. The ACC plays a substantial role in conflict monitoring and interacts with the dlPFC in compensatory processes of top-down cognitive control 51 . Thus, the different connectivity between the ACC and the SFG/dlPFC may be related to the greater sensitivity to and simultaneously greater tolerance of conflicting information in liberals. On the other hand, we propose that the different insular connectivity with the occipital regions might reflect different personality traits related to novelty seeking behavior between liberals and conservatives 52 . The insula, especially the posterior subdivision, is associated with risk prediction during decision-making and the novelty-seeking personality trait 53 . In addition, an insular network with posterior brain regions, including the occipital cortex, is involved in monitoring environmental information, which may be involved in interest in new experiences 54 . Based on the linkage between the insula and the novelty-seeking trait, our finding of insular connectivity differences raises the question of the underlying neurobiological basis of greater openness in liberals, which should be addressed in future studies. In addition to the seed-based intrinsic connectivities, the segregation and integration patterns of the intrinsic global brain network were comparable between political attitudes. This might indicate that liberals and conservatives have different intrinsic brain organization structures confined to specific connections between regions involved in self-regulatory processes and personality traits, such as conscientiousness and openness to new experiences.
The present study has several limitations that should be addressed in future studies. First, the number of conservative subjects was relatively small for high statistical power. We recruited a sufficient number of study participants at first, but most participants had moderately liberal or moderate attitudes. This might reflect the nature of political attitude distribution, which is skewed towards moderately liberal and moderate among the young adult population in the Republic of Korea 55 . The small sample size of conservatives may have increased risk of false discovery 56 ; however, the results of intrinsic connectivity differences were replicated in the regression analysis that employed the large enough samples. Thus, our findings suggest the association between intrinsic connectivities and political attitudes are likely to be true positives. Second, extreme partisan attitudes, regardless of political orientation, might have influenced our behavioral results because there were individuals with strong political attitudes (i.e., "very liberal" or "very conservative") only among the liberals. A recent study has shown the effects of partisan extremity on mental rigidity 57 . However, we argue that our participants with a very liberal orientation may not be so radical as to have extremely high mental rigidity because we observed comparable psychological resilience and self-regulation capacity between liberals and moderates. Third, the sample characteristics could be biased because the present study involved only young adult participants. However, our findings are in line with previous neuropolitics studies with populations with similar sociodemographic characteristics (e.g., college students or young adults, distribution of political attitudes, socioeconomic status) 10,13,40 . However, replicate studies with large samples across a broad age range are still needed. Finally, we assessed only one axis of political attitudes (liberalism vs. conservativism) even though other axes (economic freedom, government control, etc.) could exist in diverse socioeconomic and cultural situations in the real world 58 . However, the simple questionnaire we used to ask a degree of political attitude between liberal and conservative beliefs has been validated as a reliable measure of political attitudes in many other studies of political psychology and neuropolitics 10,13,32,52 .
In conclusion, we identified that conservatives, compared to liberals and moderates, have greater psychological resilience and self-regulation capacity that may buffer the stress caused by threatening and anxiety-inducing political information. These psychological resources are neurobiologically related to the stronger connectivities of the OFC and insula with the precuneus and frontal pole/OFC, respectively, in the intrinsic brain network of conservatives. This study may provide evidence explaining the different responses of liberals and conservatives Scientific RepoRtS | (2020) 10:15877 | https://doi.org/10.1038/s41598-020-72980-x www.nature.com/scientificreports/ to politically relevant social issues and expand the neurobiological understanding of neuropolitics by showing the interactions of brain regions that are involved in self-regulatory processes and resilience.

participants.
One hundred sixteen healthy volunteers were recruited from the student participant pool of the Transdisciplinary Research Center for Culture-Brain Dynamics at Seoul National University (SNU). We collected sociodemographic information, including political attitudes and socioeconomic status 59 . Political attitude was asked using a five-point Likert scale measuring the current political orientation ranging from very liberal (1) to somewhat liberal (2), moderate (3), somewhat conservative (4), and very conservative (5) 13 . We assigned participants to liberal (1, 2), moderate (3), and conservative groups (4,5). To ensure an absence of mental illness, we conducted an interview using the Structured Clinical Interview for DSM-IV Axis-I Disorders, Non-Patient edition 60 . Additionally, we administered the Beck Depression Inventory (BDI) and Beck Anxiety Inventory (BAI) to assess the severity of depression and anxiety symptoms 61,62 . We excluded participants who met the following criteria: a lifetime history of psychiatric/neurological disorders, severe depression/anxiety symptoms, or exposure to psychiatric medications. After this screening process, one hundred two (42 liberals, 44 moderates, and 16 conservatives) of the original 116 participants remained (Table 1). We assessed psychological resilience and self-regulation capacity using the Korean version of the RQT, which consists of a total score for resilience and three subscores for self-regulation, sociability, and positivity 35,63 . All participants were informed of a complete description about all experimental procedures, and we were provided with written informed consent obtained from the participants. All experimental protocols and methods in this study were approved by the Institutional Review Board at SNU and performed in accordance with ethical guidelines of the Declaration of Helsinki.
image acquisition and preprocessing. We acquired functional and structural brain images using a 3 T Trio magnetic resonance imaging (MRI) scanner (Siemens Magnetom Trio) with a 32-channel head coil at SNU. Functional imaging data were acquired using a gradient echo-planar imaging pulse sequence (echo time = 30 ms, repetition time = 2000 ms, flip angle = 80°, voxel size = 3.4 × 3.4 × 3.4 mm 3 , and 34 slices). We instructed participants to relax with their eyes open during the session. Subjects were reminded to stay awake before the acquisition and monitored through the use of an eye tracker to ensure that they did not fall asleep. High-resolution T1-weighted brain imaging data were also collected using a magnetization-prepared rapid gradient echo sequence with the following parameters: echo time = 2.19 ms, repetition time = 2400 ms, flip angle = 8°, voxel size = 0.85 × 0.85 × 0.85 mm 3 , 192 slices). We preprocessed the brain images using the CONN toolbox v17a 64 , implemented in the Statistical Parametric Mapping toolbox version 12 (SPM12; https ://www.fil.ion.ucl.ac.uk/spm/), with the following pipeline. After discarding the first four volumes for gradient field stabilization, the functional brain images were corrected for slice-timing discrepancies and realigned to the first scan via rigid-body alignment 65,66 . Subjects who exhibited head movement exceeding the criteria (translation > 1.5 mm and rotation > 1.5 degree in any direction) were excluded. After reorienting a center point of the images at the anterior commissure, we coregistered the functional to structural images. Using a nonlinear warping algorithm, both images were subsequently normalized to Montreal Neurological Institute (MNI) space and resampled to 2 mm isotropic voxel dimension 67 . The functional images were spatially smoothed with a full width at half maximum Gaussian kernel of 4 mm to increase the signal-to-noise ratio 68 . We cleaned noise signals by conducting nuisance regression using a component-based noise correction method (CompCor), linear detrending, and temporal bandpass filtering (0.008 < f < 0.09 Hz) 69,70 . functional connectivity analysis. To investigate seed-based functional connectivities, we selected the regions of interest (ROIs) according to previous neuroimaging studies reporting brain function in self-regulation and psychological resilience as follows: the ACC, amygdala, insula, and OFC 10,14,24,39 . The ROIs were defined using the Harvard-Oxford cortical and subcortical atlases. We calculated the connectivity strength using Pearson's bivariate correlation between the seed regions and voxels in the rest of the brain from the individuals' functional imaging data 64 . Subsequently, we transformed the correlation coefficients to a normal distribution by Fisher's z transformation. In addition to the functional analysis, we additionally conducted volumetric analyses using structural brain images for a replication study of previously reported neuroanatomical correlates in neuropolitics (Supplementary Methods & Results) 13 .
Graph theoretical analysis. We also conducted graph theory analysis to investigate a whole-brain topographical brain network complementing the functional connectivity analysis using the selected ROIs. A connectivity matrix for each participant was computed using bivariate correlation between each pair of 106 ROIs, followed by the r-to-z transformation. Single-subject connectivity matrices were binarized after thresholding by the proportion of the strongest connectivity strength 71 . To avoid a selection bias for one specific threshold, we applied a wide range of the sparsity threshold (0.03 ≤ S ≤ 0.35) and calculated the area under the curve (AUC) for each network metric 72 . Using the Brain Connectivity Toolbox, we assessed the segregation and integration of the global brain networks by clustering coefficients, modularity, and global efficiency 73 . All the network metrics were normalized by comparing them to 1000 random networks, which preserved the size, density, and degree distribution of the original networks 72 .
Statistical analyses. Because our samples are unequally sized, we checked the homogeneity of variance assumption before testing sociodemographic and psychological differences. The Fligner-Killeen test showed unequal variance of the RQT sociability variable; all the other variables were homoscedastic 74 . Sociodemographic variables were compared among liberals, moderates, and conservatives using the chi-squared test or Scientific RepoRtS | (2020) 10:15877 | https://doi.org/10.1038/s41598-020-72980-x www.nature.com/scientificreports/ one-way analysis of variance (ANOVA). To address the issue of unbalanced samples in terms of Type I error, we performed a MANOVA with Pillai's trace method when comparing the RQT total and subtotal scores between groups 75 . As self-regulation, in addition to resilience, is assumed to be the other psychological buffer, we conducted an additional MANOVA to compare the RQT self-regulation subscales (emotion regulation, causal reasoning, and impulse control) among groups. The above analyses were followed by the univariate Welch's ANOVA with Games-Howell post hoc comparisons, which are robust in case of unbalanced or heteroscedastic data, in SPSS version 23 (IBM, Armonk, N.Y.) 76,77 . We compared the connectivity strength between 42 liberals and 16 conservatives using two-tailed independent samples t-tests. Additionally, we conducted a linear regression analysis with the political orientation score as a predictor of resting-state brain connectivity in total one hundred two participants. This was to demonstrate the validity of the group comparison analysis with the small samples of conservatives, which was potentially vulnerable to false discovery rate. We defined regions of significant difference if the clusters survived the voxellevel height threshold with an uncorrected p < 0.001 and cluster-level extent threshold of the false-discovery rate (FDR) with a corrected p < 0.05 to address multiple comparison issues 78 . We examined the homoscedasticity of the clusters detected by using the Fligner-Killeen test. For group comparisons of the network metrics, we performed nonparametric permutation tests with 10,000 repetitions, followed by the Bonferroni correction for multiple comparisons 79 .
To explore the neural associations with resilience and self-regulation capacity, we conducted Spearman's correlation between intrinsic brain organization measures and the RQT scores within each group.