Threats reliably evoke shifts in high-level ideological investment, but little work to date has explored the neural mechanisms underlying these dynamics. This paper describes how continuous theta burst transcranial magnetic stimulation may be employed to test the contribution of the posterior medial frontal cortex (and/or other regions) to threat-related ideological shifts.
Decades of behavioral science research have documented functional shifts in attitudes and ideological adherence in response to various challenges, but little work to date has illuminated the neural mechanisms underlying these dynamics. This paper describes how continuous theta burst transcranial magnetic stimulation may be employed to experimentally assess the causal contribution of cortical regions to threat-related ideological shifts. In the example protocol provided here, participants are exposed to a threat prime-an explicit reminder of their own inevitable death and bodily decomposition-following a downregulation of the posterior medial frontal cortex (pMFC) or a sham stimulation. Next, disguised within a series of distracter tasks, participants’ relative degree of ideological adherence is assessed-in the present example, with regard to coalitional prejudice and religious belief. Participants for whom the pMFC has been downregulated exhibit less coalitionally biased responses to an immigrant critical of the participants’ national in-group, and less conviction in positive afterlife beliefs (i.e., God, angels, and heaven), despite having recently been reminded of death. These results complement prior findings that continuous theta burst stimulation of the pMFC influences social conformity and sharing and illustrate the feasibility of investigating the neural basis of high-level social cognitive shifts using transcranial magnetic stimulation.
This paper presents a recently developed method for experimentally neuromodulating ideological threat-reactions, with a particular focus on nationalistic bias and religiosity1. Importantly, however, the procedure presented in what follows should be taken as an illustrative token of a promising general approach to the study of the neural substrate of high-level social and ideological cognition (e.g., with regard to normative judgments, political attitudes) using transcranial magnetic stimulation (TMS). To theoretically situate this 'proof-of-concept' example, prior work on links between threat-detection and ideological investment, including plausible neural correlates of these effects, is briefly reviewed.
Threat and Ethnocentrism
People live in, and sometimes die for, social groups2. By assorting into coalitions, individuals benefit from shared access to both knowledge and material resources. Because sharing valued material or informational resources renders people vulnerable, individuals are incentivized to calculate whether others are likely to reciprocate or abuse their generosity3. Categorizing another person as sharing an investment in an in-group is thought to privilege coordination between fellow in-group members by enhancing mutual care and trust. This group-favoring ethnocentrism can also lead to a negative valuation of those perceived to be aligned with out-groups-and hence to be unreliable if not overtly antagonistic and, therefore, undeserving of in-group resources4,5. Under contexts of conflict, group prejudice appears not only to discourage cooperation but to motivate aggression against individuals perceived to belong to or sympathize with enemy coalitions6. If, over deep time, in-group favoritism advances reproductive fitness7,8, then the mind may have been shaped by selection to support ethnocentrism9,10, particularly under contexts of threat11,12. Consistent with this functional interpretation of ethnocentrism, the extent to which individuals express commitment to group ideologies has been observed to increase following primes of threat13,14,15. Although social scientists have studied the influence of threats on ideological commitment for decades, only recently has attention turned to the brain mechanisms at work16,17,18,19,20. In the present protocol, a brain region previously associated with low-level problem solving (e.g., updating motor responses to obtain a reward) is demonstrated to contribute to facultative shifts in ideological beliefs.
Threat, Ideological Commitment, and the Posterior Medial Frontal Cortex
The pMFC includes the dorsal anterior cingulate cortex (dACC) and the dorsomedial prefrontal area anterior to the supplementary motor cortex (dmPFC). The pMFC has been implicated in an array of reactions to negative stimuli21,22,23. The pMFC contributes to the detection of discrepancies between current and preferred conditions, as well as to subsequent decision-making adjustment of behavior to reduce such discrepancies24,25,26. For example, the dACC has been implicated in relatively low-level cognitive control functions such as those measured in the Stroop, Flanker, divided attention, or Go/No-Go tasks25. Similarly, at a higher level of abstraction, the dACC component of the pMFC is hypothesized to evoke intensified expressions of ideological commitment to moral or cultural values following exposure to threats (e.g., reminders of uncertainty or death)17,18. Cues of the concept of death have been found to similarly trigger activity in the dmPFC27,28 and to intensify ideological expressions (e.g., national identification, punishment of norm-violators)14. Cues of social isolation have likewise been observed to heighten ethnocentrism12 and to activate the dACC29.
The dmPFC component of the pMFC is a particularly plausible inhibitor of prosociality toward out-group members, as the dmPFC is considered important to modulating the treatment of self versus other during social decision-making30,31. A growing literature suggests that human prosocial inclinations-most pronounced toward fellow in-group members, all else being equal-may partially derive from a tendency for feelings and behaviors to activate similar neural systems, whether originating in the self or in another32. The extent of this self-other neural resonance has been found to predict prosocial behavior33,34,35. Consistent with the role of coalitional bias in moderating prosociality, neural resonance and related prosocial behaviors are influenced by factors such as group identity36,32. The modulation of prosociality contingent on group identification may owe to mechanisms of inhibitory prefrontal cognitive control involving the dmPFC, as the dmPFC is important for the tonic control of spontaneous imitation37,30, as well as for shifting between the perspectives of self and other38. Most compellingly, downregulating the dmPFC causes greater financial sharing behavior39, directly implicating the dmPFC as inhibiting prosociality, plausibly including the suppression of prosociality on the basis of out-group affiliation. These relatively high-level social functions of the dmPFC may be understood as manifestations of a larger role of the dmPFC in various cognitive control functions40. For example, experimentally upregulating the dmPFC was recently shown to enhance impulse control in an asocial delay-discounting paradigm in which participants who postpone immediate reward receive greater future reward41.
The pMFC complex appears responsive to cues of the presence of various sorts of social discrepancies, and pMFC activity predicts behavioral shifts designed to reduce said discrepancies42. For instance, pMFC activity correlates with preference changes in a manner that appears to reduce cognitive dissonance43,44,45 or to heighten social conformity following evidence that a person's opinions deviate from the group consensus46,47. In a demonstration of the causal role of the pMFC in enabling such dynamics, the experimental downregulation of pMFC activity via TMS has been shown to decrease social conformity despite cues of disagreement with the group48. In sum, upon detection of relatively low-level problems such as receiving a negative outcome in a motor task24, or relatively high-level problems such as a deviation between a person's stated views and those of his/her peers, the pMFC appears involved in the activation of networks that coordinate problem-relevant responses23,47,49.
Considered together, the overall pattern of results implicates the pMFC as part of neurobiological architecture evolved to manage challenges spanning various low-level and high-level domains. Accordingly, when participants are presented with an out-group member's critique of their group, the pMFC would hypothetically be involved in detecting this conflict and with coordinating a typical response: derogation of that out-group critic and their ideas. By the same logic, the pMFC is predicted to help enable individuals confronted with their own mortality to amplify their belief in a pleasant afterlife. If so, then participants for whom the pMFC complex has been downregulated may be expected to evince less derogation of critical out-group members and less religious belief following reminders of the inevitability of death.
Downregulating Targeted Cortical Areas Via Continuous Theta Burst Stimulation
Theta burst stimulation (TBS) is a form of patterned TMS. TMS stimulates the brain non-invasively by producing a rapidly varying magnetic field over the stimulated subject's scalp. This rapidly varying magnetic field induces electric currents in the brain, which in turn lead brain cells to fire50,51,52. In this manner, the manipulation of targeted brain regions via TMS allows research to move past the correlational discoveries of traditional brain mapping methods employing neuroimaging. By stimulating a given brain region and, thereby, decreasing or increasing its activity, causal inferences about the relevance of that region on a variety of behavioral tasks may be inferred.
TBS protocols have been modeled from repetitive electrical stimulation protocols that induced long-term potentiation (LTP) or long-term depression (LTD) in animal studies53. Continuous theta burst stimulation (cTBS), which consists of 50 Hz triplets of pulses delivered at 5 Hz for 40 s, for a total of 600 pulses, has an effect similar to LTD, reducing the activity in the stimulated area for an estimated period of at least 1 hour. Intermittent TBS (iTBS) consists of the same pattern of bursts at the same frequency of cTBS. However, in iTBS, the subject is stimulated for 2 s at a time, which is repeated every 10 s for 190 s (totaling 600 pulses, as in cTBS). iTBS has an effect similar to LTP, enhancing the activity in the stimulated area for a period of time comparable to cTBS. Whereas the cTBS method highlighted here can reduce ideological threat-responses, theoretically, iTBS may heighten ideological threat-responses.
The protocol that follows details methods recently used to experimentally downregulate group bias and religious belief with cTBS1, in the hope that researchers interested in alternate modes of ideological threat-response might replicate these effects and/or modify this general approach for their own purposes (e.g., substituting alternate threat primes and/or judgment outcomes, or by adding a control stimulation site).
All methods described in what follows have been approved by the Office of the Human Research Protection Program (OHRPP) of the University of California, Los Angeles.
1. Pre-experiment Steps
2. cTBS Procedure
3. Survey Tasks
In the example study, the final sample consisted of 38 participants (58% female, Mage = 20.9 years, SD = 2.67). About 36.8% of the participants identified as White, 36.8% as East Asian, 13.2% as South Asian, 7.9% as Middle Eastern, and 5.3% as Other. As intended, the sample was politically moderate (M = 4.68, SD = 1.51; 1 = "Extremely Liberal"; 5 = "Moderate"; 9 = "Extremely Conservative").
Preliminary ANOVA tests detected a significant effect of the order of essay presentation on ratings of the "anti-US" immigrant, F(1, 36) = 5.30, p = 0.027, η2 p = 0.13, 95% confidence interval (CI) [-2.07, -.13], with no order effect observed for ratings of the "pro-US" immigrant, p = 0.74. Accordingly, the essay order was included as a covariate in subsequent analyses. (Follow-up tests confirmed that controlling for order does not alter the overall pattern of results.) Such order effects may occur using this measure of group bias and should be routinely checked for as potential sources of noise.
As predicted, cTBS of the pMFC increased positive evaluations of the critical immigrant author, who was rated 28.5% more positively in the TMS condition (M = 4.10, SD = 1.66) than in the control condition (M = 2.93, SD = 1.22), F(1, 35) = 7.01, p = 0.012, η2 p = 0.17, 95% CI [-2.06, -0.27]. Also consistent with the hypothesis that the pMFC reacts to ideological conflict, ratings of the pro-US author (who posed no ideological threat) were an average of 8.2% higher in the cTBS condition (M = 5.90, SD = .87) than in the sham condition (M = 5.42, SD = 1.17), a difference that was not statistically significant, F(1, 36) = 2.09, p = 0.157, η2 p = 0.06, 95% CI [-1.16, 0.20]. Remarkably, exploratory tests revealed that the effects of cTBS were equivalent with respect to both the ratings of the immigrants' personal qualities and the endorsement of their harshly critical arguments, attesting to an effect of the cTBS manipulation on the intensity of the participants' ideological investment in nationalistic values, rather than an effect limited to interpersonal amity (see Figure 1). This finding suggests that applications of this cTBS method addressing forms of ideological adherence relatively unrelated to social affiliation may be attenuated.
Also consistent with predictions, participants who received cTBS reported an average of 32.8% less positive religious conviction (M = 3.05, SD = 1.92) relative to the sham participants (M = 4.54, SD = 2.26), F(1, 36) = 4.80, p = 0.035, η2 p = 0.12, 95% CI [0.11, 2.87]. Participants in the cTBS condition also reported a less negative religious conviction (M = 2.84, SD = 1.89) relative to the sham participants (M = 3.98, SD = 2.50), but this difference did not reach statistical significance, F(1, 36) = 2.52, p = 0.122, η2 p = 0.07, 95% CI [0.32, 2.60]. This result underscores the apparent relevance of the pMFC to recruiting ideological stances (e.g., positive religious beliefs) pertinent to particular problems (e.g., mortality), rather than to ideological adherence in general, as heaven presents a better solution than hell to the problem of death.
There were no apparent effects of condition on any of the 11 PANAS-X affect subscales (ps 0.09–0.92). Participants in both conditions reported moderately positive affect (cTBS: M = 2.44, SD = 0.61; Sham: M = 2.37, SD = 0.75) and low levels of negative affect (cTBS: M = 1.22, SD = 0.27; Sham: M = 1.43, SD = 0.51). The null effects of cTBS on self-reported emotion suggest that the findings with regard to group bias or religiosity were not driven by changes in emotional reactivity to the reminder of death.
Figure 1: Effects of cTBS on US citizens' endorsement of the personal appeal and arguments of pro-US and anti-US immigrants. Please click here to view a larger version of this figure.
Downregulating the pMFC via cTBS significantly reduced bias toward an out-group member critical of the US and avowed religious belief in the aftermath of a vivid reminder of death, consistent with both the hypothesized role of the pMFC in ideological threat-reactions and with the promise of TMS as a method for experimentally manipulating ideological cognition. While encouraging and consonant with a substantive background literature, the results of any single study should be considered to be highly preliminary. In addition, the cTBS methods employed here should be understood as subject to significant improvement.
Although TMS has a relatively good spatial accuracy (a few mm), there are limitations that impede accurate focal stimulation, such as between-subject brain variability. The site of stimulation is selected based on the average coordinates of a particular brain region, and the actual location of any region of interest will vary from individual to individual. Accuracy can be improved by using individual subjects' magnetic resonance imaging (MRI) images, as these images can be used in concert with a neuronavigation system to target particular regions on the imaged brain as a best practice. Unfortunately, because MRI scans are often prohibitively costly, most TMS studies (including the example given in this paper) have not had access to neuronavigation.
The differential recruitment of cortical neurons may also contribute to a variability in the effects of cTBS, resulting in potentially opposite effects of both cTBS and iTBS in some individuals. Hamada and colleagues64 report finding no overall difference in MEP modulation among 52 participants receiving cTBS versus iTBS to the primary motor cortex. In their study, some participants exhibited a facilitatory response to cTBS and an inhibitory response to iTBS, while others displayed the reverse pattern, and still others showed uniformly inhibitory or excitatory responses to either form of stimulation. Although many other studies have been able to induce the expected response in their overall sample, Hamada and colleagues' results militate for assessing individual differences in responses to cTBS (or iTBS if applicable) in future experimental designs; Hamada et al. provide a detailed description of how MEP latencies may be used to predict a cTBS/iTBS response within subjects64.
Achieving the correct coil orientation and positioning for each subject is another effective means of reducing potential variability in the effects of TBS. When available, stereotaxic neuronavigation software that guides the positioning and orientation can help to ensure accurate stimulation. In any event, running a large number of subjects in each experimental condition can also help to mitigate potential noise related to differential responsiveness to TBS.
An additional consideration when performing cTBS is how to implement an appropriate control condition. As was done here, cTBS may be contrasted with a sham condition (i.e., pointing the coil away to avoid actual brain stimulation, or lowering the intensity far below threshold). However, as it is conceivable that the cTBS of any region might induce changes in behavior, a common and arguably best practice is to introduce a condition wherein stimulation is delivered to a region thought to be unrelated to the task.
Stimulation of any one region can lead to a spreading activation to regions that are functionally downstream from or proximal to the area of interest, such that causal interpretations of the functional contribution of individual regions must be undertaken with great caution. Relatedly, augmenting cTBS of the pMFC with neuroimaging would importantly clarify the relative effect of the manipulation on the dmPFC versus dACC subcomponents of the pMFC, in addition to the potential collateral enervation of connected regions. At present, it remains unclear whether the observed effects are owed to downregulation of the dACC, dmPFC, or both. In addition to conventional neuroimaging, connectomic analyses might shed light on how pMFC mechanisms articulate with other regions throughout the brain to modulate ideological modes of cognition.
While cTBS of the pMFC to influence high-level social and ideological judgments remains a relatively new and little-tested method, it has been shown to reduce social conformity48 and monetary sharing39, in addition to the present reductions in group bias and religious belief1. While these findings are encouraging, the extent to which cTBS of the pMFC will influence such high-level judgments remains unclear, as is the replicability of previous effects. In the study highlighted here, prospective participants were carefully prescreened to exclude 'hardliners' with respect to both political orientation (strong liberals or conservatives) and religiosity (committed believers or atheists), leaving a sample of relatively moderate individuals. This effort was made to avoid individuals with hardened attitudes, who may produce reflexive responses via pathways distinct from those involved in the process of actively considering ideological positions. While sensible, these screening criteria may have inadvertently introduced a sample bias of some sort related to factors which deter 'middle-of-the-road' individuals from holding firm positions. Where resources allow for larger and more diverse samples, future researchers should include a broad spectrum of political and religious orientations in order to assess potential moderating effects of such individual differences.
In the present design, all participants were reminded of death to establish a context in which participants might be expected to draw on positive religious beliefs as an ideological solution. Although the findings were consistent with this interpretation, as well as with prior work linking religiosity with confidence in the face of physical danger65,66,67,68, exposing all participants to a death prime constitutes a serious limitation, as the data cannot reveal whether downregulation of the pMFC would produce comparable effects at baseline, in the absence of a recent threat cue. Further, it is not possible to determine whether the reduction in out-group bias toward the critical immigrant observed following cTBS of the pMFC reflects a muting in the impact of the death prime, a muting in the impact of the ideological conflict posed by the criticism of the US, or an interaction between the two. Thus, future work should incorporate a non-threat control condition to test whether, regardless of the presence of a background threat such as a reminder of death, downregulation of the pMFC reduces group bias or religious belief.
The present threat-induction was chosen because of the link between the prospect of dying and reassuring concepts of the afterlife, and secondarily because mortality-salience induction has been shown to heighten intergroup prejudice69. Alternate non-death threat inductions related to challenges in domains pertinent to other social judgments might be comparably employed. Importantly, non-death threat primes have been demonstrated to reliably modulate social judgments14,55,69.
Finally, one of the most challenging aspects of this work is that it requires expertise not only in TMS techniques but also in manipulating and measuring relatively abstract constructs such as ideological adherence. Thus, wherever possible, research teams should include investigators with relevant experience in disciplines such as social psychology, political psychology, or anthropology, as well as neuroscience.
The authors have nothing to disclose.
Colin Holbrook was supported by the U.S. Air Force Office of Scientific Research Award FA9550-115-1-0469.
Isopropyl alcohol | |||
Ear plugs | |||
Adult swim cap | Sprint Aquatics | 304 | |
MobiMini 2-channel recording system | TMSi | ||
EMG electrodes | TMSi | ||
Covidien Kendall disposable surface electrodes (24 mm) | Bio-medical | H124SG | |
Magstim Rapid2 TMS device | Magstim | ||
D70 figure-of-eight coil | Magstim | ||
Visor2 Neuronavigation software | ANT Neuro |