Rumination and executive functions: Understanding cognitive vulnerability for psychopathology

Abstract

Introduction: Both rumination, a pattern of repetitive, self-focused thought in response to distress, and deficits in executive functions (EFs), a set of cognitive abilities that facilitate higher-order thinking, have transdiagnostic associations with psychopathology. Although empirical studies suggest associations between EFs and rumination, this literature has not examined subtypes of rumination and different components of EFs. It also has not examined whether rumination and EFs explain overlapping variance in psychopathology, which is relevant to theoretical models suggesting that rumination might mediate the EF-psychopathology association.

Methods: We used structural equation modeling to examine the association between latent factors for two types of rumination (anger and depressive) and three components of EF (a Common EF factor, and factors specific to updating working memory and shifting mental sets) and whether they independently relate to internalizing and externalizing psychopathology in a population sample of 764 young adults (mean age 23 years) from the Colorado Longitudinal Twin Study.

Results: Depressive and Anger Rumination showed small correlations with a Common EF factor (rs = -.09 to -.11). Anger Rumination and Common EF ability were associated with independent variance in externalizing psychopathology, whereas Depressive Rumination, but not Common EF, was associated with internalizing psychopathology.

Limitations: Examination of cross-sectional relations in a population sample led to low symptom endorsement for psychopathology and necessitated examination of lifetime, rather than past-year, psychopathology.

Conclusions: Inconsistent with mediation hypotheses, Common EF abilities and rumination are correlated yet largely independent constructs that both predict psychopathology.

Keywords: Brooding; Cognitive control; Executive control; Inhibition; Reflection; Task switching.

Figure 1

Measurement models of (A) rumination and (B) executive functions (EFs) and (C) psychopathology. Panel A and C adapted from du Pont et al. (2018). Ellipses represent latent variables; rectangles represent manifest variables or individual scales. Standardized parameters for men are depicted above those for women, standard errors are in brackets. Note that standardized loadings differ slightly across sex even when unstandardized parameters are sex-invariant, because variances differ. For model A, the factor variance of Anger Rumination and the covariance between Anger and Depressive Rumination were constrained across sex but all other parameters not used to identify the model varied across sex. For model B, all unstandardized loadings and the factor variances for Common EF and Shifting-Specific were constrained to equality across sex, but the factor variance for Updating-Specific, as well as task intercepts and residual variances, varied across sex. For model C, the factor variance of Internalizing and the Internalizing-Externalizing covariance were constrained across sex, and all other parameters not used to identify the model varied across sex. AA= angry afterthoughts; TR= thoughts of revenge; AM=angry memories; UC=understanding causes; RRQ-RU=rumination; RRS-B=brooding; RRS-R= reflection. Anti=antisaccade; Stop=stop-signal; keep=keep track; Letter=letter memory; Sback=spatial n-back; Num=number-letter; Col=color-shape; Cat=category-switch. GAD=generalized anxiety disorder; MDD=major depressive disorder; ASPD=antisocial personality disorder; ALC=alcohol use disorder; CAN=cannabis use disorder; TOB=tobacco use disorder; ILL=substance use disorder involving illicit drug use. *p<.05.

Figure 2

Regression model of rumination, executive functions (EFs), and psychopathology. Ellipses indicate latent variables. For simplicity, the manifest variables that load on the latent variables are not depicted. Standardized parameters are depicted for men above the standardized parameters for women. Standard errors are in brackets. All unstandardized covariances and regression paths were constrained to be equal across sex. For clarity, bold lines were used to depict significant associations, solid lines depict marginally significant associations, and dotted lines were used to depict nonsignificant associations. Internalizing Psychopathology r² = .59 and .40 in men and women, respectively, and Externalizing Psychopathology r² = .15 in both men and women. *p<.05, ⁺p<.10.