The Role of Emotion Regulation for General Self-Efficacy in Adolescents Assessed Through Both Neurophysiological and Self-Reported Measures

Introduction

An individual’s beliefs in their own abilities and capacity to organize and execute the actions required to achieve a desired result constitute the concept of self-efficacy (SE).¹ SE represents a positive youth developmental construct that has been found associated with reduced risk for adolescent psychopathology and higher academic achievement in several domains.^2–4 If people believe they have a personal sense of control, they are prone to a more proactive approach in their life.⁵ This is reflected in individuals with a higher sense of SE, as they seem to utilize metacognitive strategies to a greater extent, which further enhances self-regulatory capabilities.^2,6 Self-efficacious students appear to work harder, persist longer in challenging situations, and are less anxious.¹ A great body of research suggests that adolescent emotion regulation (ER) represents an important component in the relationship between SE and developmental outcomes.^7–11 However, the neurophysiological mechanisms underlying the relationship between ER and SE, and the development of adolescents’ SE beliefs in general, are still poorly understood.¹²

Studies on the relationship between SE and physiological measures has been conducted in conjunction with neuroanatomical studies on regional gray matter and mean diffusivity, and behavioral studies on phobias and fear conditioning.^13–18 In the behavioral studies, it was found that individuals who believe they can exert control over aversive events show less autonomic arousal than those who believe that they have less control. While these studies indicate how SE may influence emotional responses to challenge in line with the concept of ER-SE, they do not explain the neurophysiological mechanisms facilitating the effective formation of general-domain SE beliefs.¹⁹ Given the relationship between SE and ER, it is of interest to determine if SE beliefs are influenced by neurophysiological processes involved in ER.^12,15–17

SE beliefs are traditionally considered to be influenced in four different ways.¹⁵ The first three processes; mastery of experiences, role-modeling, and persuasion, are well researched and understood. The fourth process is less understood and concerns one’s own judgment about the experienced emotional and somatic states of arousal occurring in response to challenge exposure.^1,20 Thus, being able to regulate one’s own emotions may prevent an individual from feeling overwhelmed by challenging tasks. ER may also allow for processing task-relevant features needed to ensure desired outcomes, thus ensuring the individual a sense of mastery.^21,22 ER ability is influenced by factors such as ER strategies, ER-SE, and neurophysiological ER capacity, among others.^19,23–26 While ER strategies and ER-SE are subject to change as a result of experience, the neurocognitive processes underlying ER capacity such as executive function and modulation of arousal are largely influenced by genetics.^{9,22,24,27–30} Thus, ER capacity represents a determinant of an individual’s absolute ER threshold, essentially serving as a psychophysiological constant that cuts across situations, potentially having a more chronic influence on the development of general SE beliefs.

Both ER and prospective and retrospective self-referential judgements of performance abilities are dependent on an overlapping set of prefrontal cortical structures, including the dorsolateral and the medial prefrontal cortex.^22,31–33 Together, these prefrontal structures are intimately involved in brain networks coordinating internally and externally guided behaviors to reach desired outcomes.^22,34,35 Vagally mediated heart rate variability (vmHRV), which refers to the difference in time intervals between heart beats resulting from vagal input, is considered an index for vagal tone and a proxy for the prefrontal activity that reflects ER capacity.^24,25,36 A growing body of evidence suggests that vagal tone is influenced by dorsolateral prefrontal cortex activity.^37–42 In a recent study, resting vagal tone was associated with functional connectivity between dorsolateral prefrontal- and medial prefrontal-related brain networks responsible for solving tasks and directing attention towards internal states, but this association diminished when participants were exposed to stress.³⁹ Thus, ER capacity, indicated by vagal tone at rest, may be an indicator of the threshold for when coordination between internally and externally guided cognitions is no longer possible, serving as a possible modulator of processes influencing SE.

Two of the most studied emotion regulation strategies are cognitive reappraisal and expressive suppression. Cognitive reappraisal is an antecedent- and goal-focused, flexible strategy aiming to modify the emotional impact of a situation by means of cognitive change, whereas expressive suppression is a response-focused strategy characterized by modulating current emotions through inhibiting the expression of them.^23,43,44 When measured at rest, higher vagal tone is associated with various ER outcomes.^{21,26,45–47} This includes individual differences in daily reported affect in both challenging and everyday settings and more frequently using cognitive reappraisal as ER strategy in response to negative stimuli.^21,46–48

The aim of this study is to assess the role of self-reported and neurophysiological indicators of ER on general SE in adolescents, thus covering a gap in the SE literature.¹² SE beliefs are considered to be influenced by one’s own judgment about the experienced emotional and somatic states of arousal occurring in response to challenge exposure.^1,20 A previous study found that vagal tone moderated the relationship between psychosocial stressors and psychopathology in adolescents.⁴⁹ Because ER strategies arguably represent the steps to achieve a desired outcome, while ER capacity represents an individual’s neurophysiological threshold for when the strategies can no longer be successfully implemented, we hypothesize that vagal tone, as an indicator of prefrontal ER capacity, moderates the relationship between self-reported ER strategies and general SE beliefs (hypothesis: H).^23,24,36,50

Materials and Methods

Ethics

The study was approved by the Norwegian Centre for Research Data. The study followed institutional guidelines when applying to the Norwegian Social Science Data Services (NSD) ethical guidelines for experimental studies. Approval was obtained by filling out the initial NSD online form. In accordance with NSD guidelines, informed consent from parents was not needed when all participants were 16 years or older. Only non-identifiable data was used, and all participants provided written informed consent and were debriefed about the purpose of the study after data collection. The present study complies with the Declaration of Helsinki, International Ethical Guidelines for Biomedical Research Involving Human Subjects and the International Guidelines for Ethical Review of Epidemiological Studies.

Sample Description

Forty-five healthy high school students (69% female, n = 31), 17–18 years of age at the time of data collection, were selected to participate in the present study. Four participants were excluded from data analysis due to invalid and missing results. Participants were recruited from high school classes, and everyone had chosen psychology as an optional subject.

Assessment, Design and Procedure

All self-report questionnaires were administered in paper format prior to recording of inter-beat intervals. Items on the questionnaires used to measure SE and ER were presented as statements and the participants used a pen to select the statements according to what fit them. There were no explicit options to leave items unanswered.

General Self-Efficacy Scale

SE beliefs are linked to distinct life domains of functioning and must be tailored to the specific situation of interest. However, some researchers have proposed a generalized sense of SE which is less context specific and explains a broader range of coping and performance.^51,52 The present study was conducted utilizing a Norwegian translated version of the General Self-Efficacy Scale (GSE).³ The scale consists of ten Likert-scale items from 1 to 4, where higher scores indicate higher SE. GSE construct scores were generated from averaging all the scores of individual items. Construct validity has been shown in numerous domains and across cultures.^53,54 Besides, according to Scholz et al, several research projects have yielded internal consistency (α = 0.75–0.91).³

Emotion Regulation Questionnaire

ER strategies were assessed through a Norwegian translated version of the Emotion Regulation Questionnaire (ERQ).⁵⁴ ERQ consists of 10 Likert-scale items from 1 to 7 capturing habitual usage of cognitive reappraisal and expressive suppression. Higher scores indicate higher usage of the two emotion regulation strategies. ER construct scores were generated by averaging all the scores in each subscale. Psychometric evaluations of ERQ indicate high construct validity and high internal consistency for both reappraisal (Cronbach’s α = 0.91) and suppression (Cronbach’s α = 0.80).^55–57

Days of Exercising

The participants were also asked to report how many days per week they engaged in physical exercising activities (days of exercising; DOE), as well as their usage of tobacco, caffeine, beta blockers, or other drugs that can cause disruptions in heart rate.

Vagally Mediated Heart Rate Variability

VmHRV was quantified from inter-beat intervals that were recorded for seven minutes with photoplethysmography by using the Alive Software (Alive^TM by Somatic Vision, Inc.). Alive is used for recording inter-beat intervals in both clinical and non-clinical samples.^58,59 The protocol was based on established recommendations for short-term recordings of inter-beat intervals for vmHRV quantification which suggests that five minutes of recording conducted at rest is required for quantifying reliable vmHRV indices.⁶⁰ Thus, five minutes of the seven-minute recordings were intended for vmHRV analysis. Recordings were carried out after the participants had completed the self-report questionnaires, during school hours over four days in sound-attenuated rooms, with only one participant being recorded per room at any given time. Participants were asked to refrain from caffeine two hours before the assessment. Upon arrival, participants were seated, and some time was spent establishing rapport with each participant and confirming that they had abstained from coffee as agreed. Participants were then informed about the recording procedure and instructed to sit back in a comfortable position they could relax in. Three IOM finger sensors were then placed on the participant’s non-dominant hand. This was followed by instructions to relax by themselves for some minutes while breathing consistently at a natural pace, and to avoid bodily movements for the duration of the recording until the experimenter came back. Any participants that brought with them their phones to the site of recording were instructed to put their phone away after setting it to silent with vibration turned off. The participants were then left alone in the room for seven–eight minutes. Five minutes of highest quality, between the first 150 and the last 50 seconds of the inter-beat interval recordings were used for HRV quantification.

Data Reduction

Four participants were excluded due to missing participant code between self-report measure and recording of inter-beat intervals. There were no missing values in the remaining dataset. Five minutes of high-quality recording was selected from the middle parts of the seven minute recordings for vmHRV analysis, in line with established recommendations.⁶⁰ R-peak extraction was used to detect inter-beat intervals and vmHRV was analyzed in the ARTiiFACT software.⁶¹ The detection of artifacts was done through the software, as well as by confirmatory visual inspection. Cubic spline interpolation was used to replace inaccurate IBIs.

The root mean square of successive differences (RMSSD) and the high-frequency component (HFms²; Fast-Fourier-Transformation, bandwidth 0.15 to 0.40 Hz), are the most commonly used indicators of vagal tone.⁶⁰ We were mainly interested in RMSSD. HFms² and RMSSD are usually highly correlated; thus, we included HFms² in our initial analyses to test for this correlation.⁶² Mean HR and the percentage of successive NN intervals exceeding 50 ms (pNN50) were also quantified from the inter-beat interval recordings. Mean HR and pNN50 do not primarily indicate vagal mediation and were included for the possibility of comparison and replication with other studies. Normative data of vagal tone in healthy adolescents (non-athletes) with normal BMI indicate that RMSSD (ms) measures for girls are minimum 12.80 and maximum 256.30, and for boys, minimum 9.10 and maximum 350.40.⁶³ Exclusion criteria followed the normative range as cut-off limits and was also compared to other studies.²⁶ All participants were within the normal range.

Statistical Analysis

The means, standard deviations (SD), range, minimum, and maximum values were quantified for each variable and presented in Table 1. The relationships between GSE, ERQ, and vmHRV were initially explored through partial correlations controlling for DOE and sex. Results from all correlations are presented in Table 2.

Table 1 Descriptive Statistics (N = 41)

Table 2 Partial Correlations Controlling for DOE and Sex (N = 41)

Hierarchical multiple linear regression analyses controlling for the effect of sex and DOE were conducted using RMSSD, HFms², and ERQ subscales as predictors of GSE. First, sex and DOE were entered as predictors of GSE in the first model (model 1), reappraisal was included as predictor in the second model (model 2), and RMSSD was included as predictor in the third model (model 3). Results of the regression analysis are presented in Table 3.

Table 3 Hierarchical Linear Regressions Predicting GSE (N = 41)

To test the hypothesis that ER capacity moderates the relationship between self-reported ER strategies and general SE beliefs, moderation regression analyses were conducted using ERQ scores as independent variable, RMSSD as moderator variable, and GSE as dependent variable to assess if RMSSD influenced change in the relationship between ER and SE. Moderation analysis was conducted with the PROCESS macro-version 3.5 for SPSS.⁶⁴

How an individual judges their ability to perform the actions needed to reach desired outcomes is influenced by metacognition, where favorable estimations can be due to overconfidence rather than metacognitive accuracy.^21,65 The accuracy to which one judges one’s own performance abilities and vagal tone are influenced by the same prefrontal structures.^{32,33,37,39,42} Vagal tone is negatively associated with overconfidence and positively associated with metacognitive accuracy.^21,66 As we use self-report measures of ER strategies and GSE, it is possible that metacognition is a confounding variable in the present analysis. In a follow-up analysis, and because we did not measure metacognition, this possibility was explored by plotting the quadratic relationship between the z-transformed RMSSD variable (predictor) and the z-transformed GSE and Reappraisal variables (dependent variables). The plots are presented in Figure 1.

Figure 1 Scatter plots for relationships between RMSSD and GSE, and RMSSD and reappraisal with quadratic regression line. (A) GSE and RMSSD. There is a trend in the quadratic regression towards individuals with both lower and higher values of RMSSD to report high GSE. (B) Reappraisal and RMSSD. There is a slight trend in the quadratic regression towards individuals with both lower and higher values of RMSSD to report more use of reappraisal as emotion regulation strategy. Abbreviations: RMSSD, Root mean square of successive N-N differences; GSE, General Self-Efficacy.

All analyses were conducted using the SPSS software package (version 26).⁶⁷

Results

Descriptive Statistics

Table 1 provides descriptive statistics for mean HR, vagal tone indices RMSSD and HFms², pNN50, scores on ERQ subscales, GSE, and DOE.

Gender differences in RMSSD measures were non-significant (t(39) = −0.404, p = 0.688). This must, however, be considered in light of the fact that the sample size for the male group is small (n_male = 12).

Results from Correlation Analysis

Results from the partial correlations controlling for DOE and sex are shown in Table 2. HFms² and RMSSD were highly correlated (p < 0.001) indicating good quality of indices.⁶²

When controlling for sex and DOE, RMSSD was not significantly correlated with reappraisal (p = 0.751), suppression (p = 0.632), or GSE (p = 0.539). HFms² was not significantly correlated with reappraisal (p = 0.632), suppression (p = 0.255), or GSE (p = 0.725). Reappraisal was not significantly correlated with suppression (p = 0.205). Reappraisal was significantly correlated with GSE (p = 0.004). Suppression was not significantly correlated with GSE (p = 0.996). Due to RMSSD and HFms² being highly correlated, only RMSSD was used in subsequent analyses.

Table 3 presents the result from the hierarchical multiple regression analysis control for DOE and sex.

As indicated from the initial correlation analyses, RMSSD (p = 0.404) was not a significant predictor of GSE when controlling for sex and DOE. Reappraisal (p = 0.005) significantly predicted GSE when controlling for sex and DOE.

H: VmHRV moderates the relationship between self-reported ER strategies and general SE beliefs

Due to the non-significant relationship between suppression and GSE, only reappraisal was included in the moderation analysis. The relationship between reappraisal and general SE beliefs was not moderated by vagal tone (reappraisal x RMSSD: B = 0.005, p < 0.279).

Post-Hoc Exploratory Quadratic Regression Analysis

Figure 1 shows the quadratic relationship between RMSSD and GSE, and RMSSD and Reappraisal. As shown in Figure 1, results indicated a bidirectional relationship between RMSSD and GSE and reappraisal, where both higher and lower RMSSD resulted in higher scores on GSE (p > 0.05) and reappraisal (p > 0.05), but this was not statistically significant.

Discussion

The aim of the present study was to investigate ER capacity, indicated by vagal tone, as a potential moderator in the relationship between self-reported ER strategies and general SE in a sample of high school students. ER is an important component in the relationship between SE and developmental outcomes but the neurophysiology underlying this relationship is poorly understood.¹² Our hypothesis was that vagal tone as an indicator of ER capacity would moderate the relationship between self-reported ER strategies and general SE beliefs. In our initial analyses, we found that cognitive reappraisal, but not expressive suppression, was significantly and positively associated with SE. The significant association between cognitive reappraisal and general self-efficacy was expected, given that cognitive reappraisal allows an individual to find ways of contextualizing situational outcomes in ways that are conducive for motivation towards desired outcomes.^68–70 As opposed to cognitive reappraisal, expressive suppression has traditionally been considered a maladaptive strategy.^23,43,44,71 This is arguably a far too static view on ER, and the more recent view is that expressive suppression increases resilience if applied flexibly and appropriately.^72–74 For instance, a recent study found that both the flexible suppression and enhancement of emotion expression were positively associated with general self-efficacy beliefs.⁷⁴ Because expressive suppression is also associated with psychopathology, which is negatively associated with SE, the flexibility to which individuals apply expressive suppression may have been a bidirectional confounder in the relationship between expressive suppression and SE in our results. Nevertheless, expressive suppression was excluded from subsequent analyses.^75–78

Vagal tone was not significantly associated with either self-reported SE beliefs nor any of the self-reported ER measures. A growing body of evidence suggests that the relationship between vagal tone and measures of ER depends on whether ER strategies are inferred from measures conducted at rest (eg, using the ERQ as in the present study) or from measuring daily emotional states or responses to emotional stimuli.^21,46–48 Follow-up regression analyses confirmed that vagal tone did not predict cognitive reappraisal.

Contrary to our hypothesis, we did not find evidence in support of ER capacity moderating the relationship between ER strategies and general SE beliefs. Several factors could explain these results. As ER strategies and SE beliefs are considered to be modified by experience, while vagal tone is largely determined by genetics, it could be that genetic factors play a lesser role in the development of SE.^9,27,29,30 This interpretation, however, is in stark contrast to evidence from genetic research suggesting that adolescent SE is mostly determined by genetics.⁷⁸ Thus, another and perhaps more plausible interpretation of our results, may be that the relationship between ER capacity and SE beliefs is influenced by the degree to which participants are able to use experience to develop an accurate understanding of their abilities to master various tasks. Another possible moderator of this relationship can be interoception. Previous studies suggest interoception as a central mechanism in neurological and self-reported processing of emotions, emotion regulation, and neural activity and performance in emotional and counter-intuitive decision-making tasks.^{26,48,79–85} In counter-intuitive decision-making tasks, interoception was found to moderate the relationship between SE and performance.⁸² Vagal tone has also been found related to emotion regulation during decision-making tasks.⁸³ Thus, future research should test the influence of interoception on the relationship between ER and SE.

Whether self-reported ER strategies correspond to actual ER strategies, and whether a high score on self-reported SE represents overconfidence or an accurate assessment of an individual’s ability to master the environment, will arguably depend on the metacognitive accuracy of the respondent. Metacognitive accuracy is dependent on the dorsolateral prefrontal cortex.^32,33 A growing body of evidence suggests that vagal tone is influenced by dorsolateral prefrontal cortex activity.^37–42 Vagal tone, in turn, is positively associated with prospective metacognitive accuracy for performance estimations, and negatively associated with overconfidence.^21,66 Thus, individuals with low vagal tone may have high SE due to overconfidence resulting from lower dorsolateral prefrontal cortex function, while individuals with high vagal tone may have high SE due to increased dorsolateral prefrontal capacity for 1) metacognitive judgements of their abilities, 2) metacognitive ER, and 3) applying metacognitive learning strategies.^{2,21,31–33,47,48,66} To explore this possible confounder, and because we did not include a measure of metacognition in our study, we conducted quadratic regression analyses on the relationship between vagal tone and SE, and vagal tone and ER in a follow-up exploratory analysis. The findings (Figure 1) revealed a u-shape indicative of the suspected interference of metacognition. The relationship was more pronounced for SE than for reappraisal, although none of the relationships were significant, which could be due to the modest sample size.

Limitations and Future Directions

The results should be interpreted in light of several limitations. Our sample size is relatively modest; a post hoc power analysis showed that with a sample size of 41 achieved test power was 1-β = 0.77. We needed a sample size of N = 44 to achieve a test power above the normative level of 80%, thus we cannot exclude the possibility of type two error. The sample population consisted of healthy high school students, which means that we already include participants who have experienced some sense of mastery, and that are on track for academic achievement. Furthermore, almost 70% of our sample consisted of females, while research indicates that adolescent males tend to have higher SE than adolescent females.^86–88 Although we controlled for the effect of sex in the present study, generalizing our findings to a broader context should be done with caution.

While both RMSSD and HFms² are accepted indicators of vagal tone, and both indicators are associated with prefrontal cortex activity, research indicates that RMSSD is also influenced by sympathetic activity.^{37–42,60,89,90} Thus, it can be argued that the choice of index could have slightly impacted results, especially considering our modest sample size.

All limitations considered, future studies should repeat measurements in a larger sample of adolescents, with a more balanced distribution of males and females, preferably with a greater age distribution, and include interoceptive and metacognitive measurements. Future studies should also consider including measurements of daily affect or an emotion regulation task, and both RMSSD and HFms² in analyses.^21,46 Utilizing other methods (eg, electroencephalogram, functional near-infrared spectroscopy, or functional magnetic resonance imaging) for measuring neurophysiological correlates of ER capacity, such as dorsolateral prefrontal cortex activity and connectivity with other prefrontal structures and amygdala during ER tasks, may serve as more objective indicators of ER capacity thus more appropriate predictors of self-efficacy than vagal tone in isolation.^31,91

Conclusion

This is the first study to assess whether emotion regulation capacity, indicated by vagal tone, moderates the relationship between emotion regulation and general self-efficacy beliefs in adolescents. Using both psychological and neurophysiological measures is a first step towards filling a neuroscience gap in the research on self-efficacy. While we did find an association between emotion regulation and self-efficacy in the present study, we did not find evidence for neurophysiological emotion regulation capacity moderating the influence of emotion regulation on self-efficacy beliefs. While we controlled for the influence of both sex and exercise in our analyses, we cannot account for the possible role that metacognition or interoception may have had on our results. Thus, the study should be repeated while including those variables in a larger and more representative sample before findings can be generalized.