Response inhibition deficits are positively associated with trait rumination, but attentional inhibition deficits are not: aggressive behaviors and interpersonal stressors as mediators

Rumination is repetitive and passive thinking about one’s depressive symptoms and possible causes and consequences of these symptoms (Nolen-Hoeksema, 1991; Watkins & Roberts, 2020). Previous studies suggested that rumination is a core vulnerability to depression. For example, dysphoric and depressed individuals who are induced to ruminate exacerbate their depressive mood (Donaldson & Lam, 2004; Nolen-Hoeksema & Morrow, 1993). In addition, trait rumination, which is assessed using the total score on the Ruminative Responses Scale (RRS; Nolen-Hoeksema & Morrow, 1991), predicts more severe depression (Butler & Nolen-Hoeksema, 1994; Nolan et al., 1998; Nolen-Hoeksema & Morrow, 1991), and the onset of major depression (Nolen-Hoeksema, 2000; Nolen-Hoeksema et al., 2007; Spasojević & Alloy, 2001). Previous studies have also indicated that rumination leads to deleterious consequences such as increases in associated negative thinking, ineffective social problem solving, interference with active instrumental behaviors, and reduced sensitivity to changing contingencies and contexts (for reviews, Watkins & Roberts, 2020).

During the past 2 decades, researchers have examined the possible role of executive function deficits in increasing rumination. Executive function is a family of top–down control processes that are needed when going automatic, or relying on instinct or intuition would be ill-advised, insufficient, or impossible (Diamond, 2013). Previous meta-analyses have suggested that deficits in executive functions were positively associated with trait rumination (Yang et al., 2017; Zetsche et al., 2018). However, no definite conclusions have been drawn to date on the subcomponents of executive function that could be related to increased rumination. Notably, the performance in tasks assessing inhibition has shown inconsistent associations with trait rumination, as described below.

Executive function tasks used in psychopathology research have involved emotional and non-emotional stimuli. The current study focused on executive function tasks using non-emotional stimuli because performance in emotional tasks might reflect not only impairments in executive function but also differences in emotional processing (Snyder et al., 2015).

Inhibition is the ability to deliberately inhibit dominant, automatic, or prepotent responses when necessary (Miyake et al., 2000). Several studies have examined the association between trait rumination and performance in emotionally neutral inhibition tasks. A previous study by Hasegawa et al. (2021) assessed the inhibition of prepotent behaviors and reported a significant positive association between trait rumination and inhibition deficits. The authors constructed a latent variable from the performance in the Go/No-Go Task (GNG), the Stop Signal Task (SST), and the Conners’ Continuous Performance Test 3rd Edition (CCPT) and examined the association between this latent variable and trait rumination in university students. Their results indicated that the latent variable reflecting deficits in inhibition of prepotent behaviors had a positive indirect association with rumination via increased stressors. These findings suggest that the impaired ability to inhibit prepotent behaviors could lead to problematic behaviors, and resulting stressors might, in turn, increase rumination.

In contrast, a contradictory finding was obtained in an initial study that examined associations between trait rumination and performance in the Modified Stroop Task (MST), which is a popular measure of inhibition. Altamirano et al. (2010) indicated that ruminative undergraduate students perform better in the MST, as indicated by fewer errors in incongruent trials, and this association was significant even after controlling for the participants’ depression level. They interpreted this result as reflecting higher active maintenance of a single task goal by trait ruminators in the face of distractions. The findings by Altamirano et al. (2010) contradict the assumption that high ruminators have inhibition impairments, although caution is needed because a subsequent study found no significant correlation between trait rumination and MST performance (Nishimura et al., 2020).

These inconsistent results could be explained by a recent, empirically supported model of inhibition. Tiego et al. (2018) proposed that inhibition could be divided into two subcomponents: response inhibition representing the ability to suppress a prepotent motor response, and attentional inhibition representing the ability to resist interference from distracting stimuli. They conducted a latent variable analysis using three tasks that were assumed to assess response inhibition and three tasks assessing attentional inhibition and reported data supporting their model, in which each subcomponent of inhibition is distinguished from the other. According to this classification, the GNG, the SST, and the CCPT were clarified as tasks assessing response inhibition, and the MST was clarified as a task assessing attentional inhibition. The inconsistent association between performance in these tasks and trait rumination in previous studies suggest that these two subcomponents of inhibition have different effects on trait rumination.

However, previous studies' findings are limited because they did not compare the relationships of response inhibition and attentional inhibition with trait rumination in the same study and possibly reporting contradictory effects of response inhibition and attentional inhibition on rumination due to differences in sample characteristics. In addition, previous studies did not adequately consider the pathways from each subcomponent of inhibition to rumination. For example, although Hasegawa et al. (2021) showed that a latent variable that reflected response inhibition deficits was indirectly associated with increased trait rumination via increased stressors, they did not examine the behaviors that mediated the association between response inhibition and stressors. Moreover, previous studies have not examined the possible mediators in the association between attentional inhibition and rumination (Altamirano et al., 2010; Nishimura et al., 2020).

The present study’s primary purposes were to compare the concurrent association of response inhibition and attentional inhibition with trait rumination and to examine possible pathways from these subcomponents of inhibition to rumination in university students. Following Hasegawa et al. (2021), we assumed that response inhibition deficits were positively associated with rumination via increased stressors. We also examined whether aggressive behaviors mediated the association between response inhibition and interpersonal stressors because previous studies have shown that aggressive behaviors were one behavioral correlate of impaired response inhibition (Qiao et al., 2016; Raaijmakers et al., 2008). On the other hand, following Altamirano et al. (2010), we assumed that attentional inhibition deficits were negatively associated with rumination. We used the GNG as an index of response inhibition and the MST as an index of attentional inhibition because previous studies have demonstrated that performance in these tasks is significantly related to trait rumination (Altamirano et al., 2010; Hasegawa et al., 2021).

Previous studies have examined whether depression confounded the relationship between rumination and other factors related to rumination, such as executive functions (see Zetsche et al., 2018). Therefore, we also investigated whether the relationships among each inhibition subcomponent, aggression, stressors, and rumination remained significant, even after controlling for the influence of depressive symptoms.

We also examined models that replace rumination assessed with the total RRS scores with brooding and reflection subcomponents of rumination. Brooding is “a passive comparison of one’s current situation with some unachieved standard,” and reflection is “a purposeful turning inward to engage in cognitive problem solving to alleviate one’s depressive symptoms” (Treynor et al., 2003, p. 256). Previous studies have suggested that brooding and reflection were distinct subcomponents because brooding was more strongly associated with concurrent and future depression levels than reflection (Pearson et al., 2010; Schoofs et al., 2010; Treynor et al., 2003). Hasegawa et al. (2021) demonstrated that the relationship between brooding and stressors was stronger than that between reflection and stressors. Therefore, it is plausible that the indirect effects of response inhibition on brooding via negative achievement and interpersonal events would be stronger than on reflection.

Finally, this study examined the relationships between trait worry and the two subcomponents of inhibition. Worry is defined as “a chain of thoughts and images, negatively affect-laden and relatively uncontrollable. The worry process represents an attempt to engage in mental problem solving on an issue whose outcome is uncertain but contains the possibility of one or more negative outcomes” (Borkovec et al., 1983, p. 10). Worry is a core feature of Generalized Anxiety Disorder (American Psychiatric Association, 2013). Rumination and worry share many characteristics, such as negative content and a perseverative style. However, rumination and worry are different because the former concerns the past, whereas the latter concerns the future (Watkins et al., 2005). Although previous studies have compared the associations of rumination and worry with executive functions, further investigations are necessary because subcomponents of executive functions that are specifically related to rumination or worry have not been identified (Zetsche et al., 2018 for meta-analysis). Therefore, the present study compared the associations of rumination and worry with response inhibition and attentional inhibition.

Descriptive statistics of each measure are shown in Table 1, and correlations between each measure are displayed in Table 2. Response inhibition deficits showed significant positive correlations with all study variables except worry, whereas attentional inhibition deficits had significant positive correlations with depression and response inhibition deficits. Aggressive behaviors showed significant positive correlations with brooding, total RRS scores, and negative interpersonal events.⁵ The Z test for comparing two dependent correlations showed that response inhibition was more strongly correlated with rumination than attentional inhibition (z = 2.475, q = 0.208, p = 0.014). Similarly, Z tests compared correlations of rumination and worry with the two inhibition measures. Results indicated that neither the correlation with response inhibition deficits nor the correlation with attentional inhibition deficits differed significantly between rumination and worry (z = 1.740, q = 0.116, p = 0.082; z = 0.635, q = 0.042, p = 0.526).

We conducted path analysis with FIML to examine whether and how response inhibition and attentional inhibition deficits were related to rumination assessed with the total RRS scores. The model provided an excellent fit to the data (χ² (2) = 1.530, p = 0.466; CFI = 1.000; RMSEA = 0.000, 90% CI: [0.000, 0.125]). The residuals of the correlations in this analysis are displayed in Table 3. Moreover, the standardized estimates and R² of this model are illustrated in Fig. 1. In addition, the standardized and unstandardized estimates are shown in Table 4.

Response inhibition deficits were significantly and positively associated with aggressive behaviors, negative interpersonal events, negative achievement events, and rumination. On the other hand, attentional inhibition deficits were not significantly associated with aggressive behaviors, both negative events, and rumination. Aggressive behaviors were positively associated with negative interpersonal events, and negative interpersonal events were positively associated with rumination.

A bias-corrected bootstrap test was conducted to determine the significance of the hypothesized indirect effects. The results indicated significant indirect effects from response inhibition deficits to aggressive behaviors to negative interpersonal events to rumination (b = 0.010, SE = 0.007, β = 0.009, 95% CI: [0.001, 0.029]).

We also conducted a path analysis with depression as an exogenous variable influencing aggressive behaviors, stressors, and rumination. This model had a good fit for the data (χ² (2) = 1.751, p = 0.417; CFI = 1.000; RMSEA = 0.000, 90% CI: [0.000, 0.130]). The residuals of the correlations in this analysis ranged from − 0.006 to 0.072. Results showed that deficits in response inhibition had a significant positive association with aggressive behaviors (b = 0.120, SE = 0.055, β = 0.187, p = 0.020, 95% CI: [0.023, 0.340]) but not with negative interpersonal events (b = 0.059, SE = 0.041, β = 0.101, p = 0.158, 95% CI: [− 0.045, 0.236]) and achievement events (b = 0.053, SE = 0.049, β = 0.093, p = 0.275, 95% CI: [− 0.074, 0.261]). In addition, negative interpersonal events were positively associated with rumination (b = 0.312, SE = 0.129, β = 0.165, p = 0.017, 95% CI: [0.023, 0.292]), whereas response inhibition deficits, attentional inhibition deficits, and negative achievement events did not show significant associations with rumination (− 0.131 ≤ bs ≤ 0.135, − 0.090 ≤ βs ≤ 0.122, ps ≥ 0.054). Each variable explained 3.9% of the variances in aggressive behaviors, 23.6% of the variances in negative interpersonal events, 25.2% of the variances in negative achievement events, and 31.6% of the variances in rumination. Bootstrap tests indicated the 95% CI of the indirect effect from response inhibition deficits to aggressive behaviors to negative interpersonal events to rumination included zero (b = 0.005, SE = 0.004, β = 0.005, 95% CI: [0.000, 0.017]).

Next, we analyzed the brooding subscale score instead of the total RRS score as the dependent variable. The model provided an excellent fit to the data (χ² (2) = 1.434, p = 0.489; CFI = 1.000; RMSEA = 0.000, 90% CI: [0.000, 0.123]). The residuals of the correlations in this analysis ranged from 0.000 to 0.075. Negative interpersonal events showed significant positive associations with brooding (b = 0.184, SE = 0.045, β = 0.318, p < 0.001, 95% CI: [0.166, 0.455]), whereas other variables including response inhibition deficits did not show significant associations with brooding (− 0.022 ≤ bs ≤ 0.051, − 0.050 ≤ βs ≤ 0.108, ps ≥ 0.143). Each variable explained 16.6% of the variances in brooding. Bootstrap tests indicated significant positive indirect effect from response inhibition deficits to aggressive behaviors to negative interpersonal events to brooding (b = 0.004, SE = 0.002, β = 0.011, 95% CI: [0.002, 0.031]).

We repeated the identical analysis using the reflection subscale score instead of the brooding subscale score. The model provided an excellent fit to the data (χ² (2) = 1.357, p = 0.508; CFI = 1.000; RMSEA = 0.000, 90% CI: [0.000, 0.121]). The residuals of the correlations in this analysis ranged from 0.000 to 0.075. Results showed that neither variable was significantly related to reflection (− 0.022 ≤ bs ≤ 0.075, − 0.060 ≤ βs ≤ 0.158, ps ≥ 0.069). Each variable explained 5.4% of the variances in reflection. Bootstrap tests showed the 95% CI of the indirect effect from response inhibition deficits to aggressive behaviors to negative interpersonal events to reflection included zero (b = 0.001, SE = 0.001, β = 0.005, 95% CI: [0.000, 0.022]).

This study indicated that response inhibition deficits had an indirect positive association with rumination via increases in aggressive behaviors and negative interpersonal events. The significant positive relationship between response inhibition deficits and aggressive behaviors is consistent with previous findings (Qiao et al., 2016; Raaijmakers et al., 2008), and suggest that individuals with response inhibition impairments have a reduced ability to control aggressive behaviors. Results also suggested that negative interpersonal events increased as a result of aggressive behaviors leading to rumination, which was not the case with negative achievement events, although caution is needed in making this interpretation because differences between simple correlation coefficients of negative interpersonal events with rumination and negative achievement events with rumination were small. Nevertheless, the finding that negative interpersonal events are primarily related to rumination is consistent with the psychobiological theory of depression proposed by Slavich et al. (2010). They suggested that because humans, who are social animals, have a fundamental drive to maintain positive social status, social values, and social regard, social rejections that threaten these needs might elicit negative self-referential cognitions concerning social worth and esteem (i.e., rumination). This could be the reason why only negative interpersonal events, mostly consisting of rejection events, were positively related to rumination.

Associations between each of the variables described above remained significant even after controlling for the influence of depression level, although caution is needed because the indirect effect from response inhibition deficits to rumination via increases in aggressive behaviors and negative interpersonal events was not significant. These results indicated that depression might not be a confounding factor in the relationship among these variables. On the other hand, the significant direct associations between response inhibition deficits and the two types of negative events disappeared after controlling for depression. These findings suggest that response inhibition deficits might be positively related to rumination through interpersonal processes.

Inconsistent with the study by Altamirano et al. (2010), the current study indicated that attentional inhibition deficits were unrelated to trait rumination. A nonsignificant association between attentional inhibition deficits and trait rumination was also reported in Nishimura et al. (2020). Therefore, it is implausible that high ruminators could better resist interference from distracting stimuli. On the other hand, the incongruent-trial error rates in the MST were correlated with increased depressive symptoms, which is consistent with the finding by Altamirano et al. (2010), although it should be noted that the significant correlation disappeared when the data of participants with multivariate outliers were excluded (see “Statistical analysis”). These results might reflect symptoms of impaired concentration in dysphoric and depressed individuals (American Psychiatric Association, 2013; Beck et al., 1996).

The findings of this study suggest that response inhibition and attentional inhibition could be discriminated from each other based on their relationship with trait rumination because only response inhibition was significantly associated with rumination, and the magnitude of the correlation with rumination differed significantly between the two types of inhibition. A previous study using confirmatory factor analysis has also indicated that response inhibition and attentional inhibition could be discriminated (Tiego et al., 2018). Recent studies have emphasized that the latent variable named common executive function––composed of performance in tasks assessing inhibition, updating, and shifting––predicts trait rumination (du Pont et al., 2019; Gustavson et al., 2020; Snyder et al., 2019). However, we propose that subdivisions of inhibition might play a crucial role in the background to rumination.

It is possible that the significant relationship between response inhibition and trait rumination was caused by the shared variance in working memory capacity, which reflects the ability to actively maintain a goal in working memory. Tiego et al. (2018) used confirmatory factor analysis and indicated that working memory capacity is a distinct construct from response inhibition and attentional inhibition, and the two inhibition subcomponents are dependent on working memory capacity. We did not assess working memory capacity, and therefore, could not examine whether this ability might be a confounding factor in the association between response inhibition and rumination, although this is unlikely because a previous study has indicated that working memory capacity assessed by the Operation Span Task had a significant positive association with brooding (Onraedt & Koster, 2014). We suggest that future studies compare associations of rumination with response inhibition, attentional inhibition, and working memory capacity, after controlling for the influences of other variables.

In line with the total RRS score analysis, the indirect effect from response inhibition deficits to the brooding subscale scores via increases in aggressive behaviors and negative interpersonal events was significant. On the other hand, the indirect effect from response inhibition deficits to reflection was not significant and was weaker than brooding. The magnitudes of these indirect effects were due to different relationships between negative interpersonal events and each rumination subcomponent. The finding that brooding was more strongly related to interpersonal stressors than reflection is consistent with Hasegawa et al. (2021). These results also supported Slavich et al. (2010), who suggested that social rejection leads to negative self-referential thoughts concerning social worth and self-esteem, which largely overlap with brooding. The reduced negative valence of reflection might explain the weaker association between reflection and negative interpersonal events than the association between brooding and negative interpersonal events.

Results of the current study indicated that neither response inhibition nor attentional inhibition was significantly correlated with trait worry, which contrasted with the significant association between response inhibition and rumination. However, differences between rumination and worry in the magnitude of correlation coefficients with response inhibition were small and statistically nonsignificant, which supported the findings of a meta-analysis reporting that executive functions were not differentially related to rumination or worry (Zetsche et al., 2018). The nonsignificant difference between rumination and worry in the magnitude of the correlation with response inhibition might have resulted from low statistical power. In addition, Watson et al. (2017) indicated that the PSWQ is not a specific measure of anxiety-related constructs and that it is better considered a general measure of neuroticism. Therefore, we suggest that the current study be replicated in the future using a more specific measure of worry.

There are some limitations in this study in addition to the issue discussed above. First, the sample consisted only of university students, making it unclear whether the study’s findings are generalizable to other age groups or clinical samples. In addition, the post hoc power of the correlation analyses indicated that larger sample size was required for detecting the hypothesized association (see Footnote 5). Future studies are needed to replicate these findings in different and larger samples. Second, the cross-sectional design of the study precludes identifying causal relationships, which require a longitudinal design. Third, this study used single tasks to measure response inhibition and attentional inhibition rather than latent variables constructed from multiple task performance. Executive function tasks have a task-impurity problem and low test–retest reliability, and therefore, a latent variable approach is ideal for examining relationships between executive functions and other variables of interest, including psychopathological symptoms (Snyder et al., 2015). Future studies should adopt a latent variable approach for examining associations between each subcomponent of inhibition and trait rumination and worry. Finally, this study did not use a scale dedicated to assessing dependent stressors that are generated by individuals with maladaptive characteristics (Hammen, 1991; Liu & Alloy, 2010), because scales for separately assessing independent and dependent stressors are unavailable in Japan. It is suggested that future studies develop a measure for separately evaluating independent and dependent stressors and examine whether response inhibition increases rumination via an increase in dependent interpersonal stressors, regardless of dependent achievement stressors and independent stressors.