Research paper
The interplay between cognitive risk and resilience factors in remitted depression: A network analysis

https://doi.org/10.1016/j.jad.2016.02.001Get rights and content

Highlights

  • We present data-driven network models for remission from depression.

  • Resilience played a central role in all three networks.

  • Resilience linked cognitive complaints and emotion regulation to residual symptoms.

  • The contribution of a behavioral measure for cognitive control was negligible.

Abstract

Background

Individuals in remission from depression are at increased risk for developing future depressive episodes. Several cognitive risk- and resilience factors have been suggested to account for this vulnerability. In the current study we explored how risk- and protective factors such as cognitive control, adaptive and maladaptive emotion regulation, residual symptomatology, and resilience relate to one another in a remitted depressed (RMD) sample.

Methods

We examined the relationships between these constructs in a cross-sectional dataset of 69 RMD patients using network analyses in order to obtain a comprehensive, data-driven view on the interplay between these constructs. We subsequently present an association network, a concentration network, and a relative importance network.

Results

In all three networks resilience formed the central hub, connecting perceived cognitive control (i.e., working memory complaints), emotion regulation, and residual symptomatology. The contribution of the behavioral measure for cognitive control in the network was negligible. Moreover, the directed relative importance network indicates bidirectional influences between these constructs, with all indicators of centrality suggesting a key role of resilience in remission from depression.

Limitations

The presented findings are cross-sectional and networks are limited to a fixed set of key constructs in the literature pertaining cognitive vulnerability for depression.

Conclusions

These findings indicate the importance of resilience to successfully cope with stressors following remission from depression. Further in-depth studies will be essential to identify the specific underlying resilience mechanisms that may be key to successful remission from depression.

Introduction

Depression is a highly prevalent, severe mental illness that is related to substantial individual suffering (e.g., Cuijpers et al., 2004, Lima and Fleck, 2007). In terms of disability, estimations suggest that major depressive disorder (MDD) is among the leading causes of burden of diseases worldwide (e.g., Demyttenaere et al., 2004). Current treatment options (psychological, pharmacological, and neurostimulation interventions) are moderately successful in achieving initial symptom reduction but long-term effects are less encouraging, with research showing that recurrence of MDD (i.e., experiencing a depressive episode after having exhibited full and/or partial remission from a previous depressive episode) is high in the general population (35% after 15 years), and even higher in those treated at specialized mental health centers (60% after 5 years and 85% after 15 years; Hardeveld et al., 2010). This has led to the realization that studying individuals remitted from depression (RMD) is crucial in understanding who remains well after initial remission and who is at-risk for new depressive episodes (e.g., De Raedt and Koster, 2010, Marchetti et al., 2012).

Current research has successfully identified a number of interindividual variables that seem to play a key role in risk as well as resilience in RMD. At the level of information-processing, previous depressive episodes have a negative impact on cognitive control processes (Vanderhasselt and De Raedt, 2009), which are crucial for goal-directed behavior. Importantly, cognitive control has been found to play a major role in emotion regulation, the process of influencing which emotions one has, including when and how these emotions are experienced (Gross, 1998). For instance, cognitive control impairments have been associated with maladaptive emotion regulation strategies such as rumination, self-blame, and catastrophizing (e.g., Hoorelbeke et al., 2016, Joormann and Gotlib, 2008, Whitmer and Banich, 2007), known to have detrimental effects on mental well-being (Aldao and Nolen-Hoeksema, 2010, Garnefski and Kraaij, 2006). Moreover, cognitive control moderates the effects of maladaptive emotion regulation on mood in daily life, with for instance low levels of cognitive control predicting a stronger increase in negative affect following rumination (Pe et al., 2013). Furthermore, in the context of remission from depression, impaired cognitive control has shown to predict rumination, linking cognitive control impairments to recurrent depressive symptoms in a RMD sample (Demeyer et al., 2012). Importantly, cognitive control impairments may also disrupt adaptive emotion regulation processes (Cohen et al., 2014, Joormann and D'Avanzato, 2010, Joormann and Vanderlind, 2014), which are key to resilience and mental well-being (Gross and John, 2003, Hu et al., 2014, Kalisch et al., 2015). Despite increasing research linking RMD to information-processing factors that are involved in emotion regulation strategies, which subsequently influence resilience or alternatively increase depressive symptoms, there are limitations to the current available research. Most importantly, research has often tested simple, unidirectional relationships between these constructs, which ignores the notion that many of the constructs involved can have reciprocal relationships. For instance, there is empirical evidence showing that levels of cognitive control can influence ruminative tendencies (Cohen et al., 2015) as well as evidence that levels of rumination influence cognitive control (Philippot and Brutoux, 2008). Currently, there is very little work integrating risk- and protective factors in RMD.

In order to obtain a more comprehensive view on the interaction between information-processing and emotion regulation strategies in relation to risk and resilience we conducted a network analysis on these constructs in a RMD sample. Based on graph theory, network modeling represents an important innovation to examine the interplay between different constructs in a largely data-driven manner (Borsboom and Cramer, 2013). Within a network model each variable is represented by a node, while the edge between two nodes shows the relationship between them. Typically, studies have relied on this type of analysis to explore how observable behaviors (i.e., symptoms) relate to one another, aiming to overcome the use of unobservable, latent variables (i.e., depression) (e.g., Borsboom et al., 2011, Cramer et al., 2010, De Schryver et al., 2015, Fried, 2015, McNally et al., 2014). However, network modeling can also be employed to decipher the interrelationship between constructs (i.e., structural network analysis) and, in turn, explore the nomological universe in which the different constructs are placed (Costantini et al., 2015b). To do so, relying on weighted and directed networks represents a great advancement, in that it is possible to obtain a fine-grained representation of the centrality (i.e., the extent to which a construct plays a central role in the network) of all the constructs considered and the possible directionality among them (Borsboom and Cramer, 2013, Costantini et al., 2015a).

In order to gain further insight in the mechanisms underlying remission from depression, we propose the use of this latter approach to examine how key constructs in the context of vulnerability for depression and resilience are related in a RMD sample. For this purpose, based on the literature, we selected four key risk factors (cognitive control impairments, working memory complaints, maladaptive emotion regulation, and residual depressive symptomatology) and two protective factors (adaptive emotion regulation and resilience) for the network analyses: (1) At the level of information-processing we obtained information about cognitive control measured with a well-validated performance based task, the Paced Auditory Serial Addition Task (PASAT; Gronwall, 1977; for a review see Tombaugh, 2006), and (2) an indicator of experienced working memory complaints, the Working Memory scale of the Behavior Rating Inventory of Executive Function (BRIEF-WM; Roth et al., 2013). Previous studies with MDD and other clinical samples indicate that self-reported cognitive functioning in daily life and performance on cognitive tasks may capture different aspects of cognitive control, as they are not necessarily associated with each other and may differ in their predictive value for well-being and symptomatology (Chan et al., 2008, Middleton et al., 2006, Mowla et al., 2008, Svendsen et al., 2012). Furthermore, the Cognitive Emotion Regulation Questionnaire (CERQ; Garnefski et al., 2001) was used to assess a broad range of emotion regulation strategies, which allows calculation of compound scores for (3) adaptive and (4) maladaptive emotion regulation processes. (5) The Remission from Depression Questionnaire (RDQ; Zimmerman et al., 2013) was used as an indicator of residual symptoms following (partial) remission from depression given that previous work indicates that residual symptomatology increases the chance of recurrence of depressive episodes (e.g., Solomon et al., 2000). This questionnaire provides a more nuanced assessment of remission than standard measures of depressive symptomatology as it combines assessment of residual depressive- and related symptoms with indicators of functioning such as sense of well-being. (6) Finally, given the importance of resilience to mental health (e.g., Griffiths et al., 2015), resilience was selected as a protective factor for the network analysis. Resilience – connoting “inner strength, competence, optimism, flexibility, and the ability to cope effectively when faced with adversity” (Wagnild, 2009, p. 105) – was assessed using the Resilience Scale (RS; Portzky et al., 2010, Wagnild and Young, 1993). This self-report measure is based on five characteristics assumed to be central to resilience: perseverance, equanimity, meaningfulness, being self-reliant, and the realization that each person is unique (for a review, see Wagnild, 2009).

In line with previous literature (Costantini et al., 2015a, McNally et al., 2014), we relied on different types of network models to obtain a more comprehensive representation of factors related to remission from depression. First, we examined simple correlational patterns (i.e, association network). Second, the underlying structure of the network was examined by means of a concentration network, where the correlations between every pair of variables were controlled for all the other variables of the network. Third, we examined a relative importance network to index predictive directionality within cross-sectional data, although this does not necessarily imply causality (McNally et al., 2014). Based on the literature we expected to find a model depicting reciprocal relationships between cognitive control and emotion regulation. Maladaptive emotion regulation strategies would link cognitive control impairments to increased residual symptomatology, whereas adaptive emotion regulation strategies would link cognitive control to resilience, which should show the opposite relation to residual symptomatology.

Section snippets

Participants

The sample consisted of 69 RMD patients that were recruited for a cognitive control training study registered as NCT02407652 at ClinicalTrials.Gov. The protocol of this training study was published online (Hoorelbeke et al., 2015). For our network analyses, baseline measures were used from the 68 participants of the training study plus one additional participant that was only willing to contribute to the correlational study. To be eligible for participation in this study, participants should be

Results

Descriptive statistics of the variables of interest are reported in Table 2 (added as supplemental material). The association network (Fig. 1) highlights that all the nodes were related to one another, with resilience, residual depressive symptoms (RDQ), and self-reported working memory complaints (BRIEF-WM) showing the strongest connectivity and being positioned at the center of the network. In general, resilience showed the strongest correlations compared to RDQ and BRIEF-WM, therefore

Discussion

Provided that individuals who remit from depression have a larger chance to develop new depressive episodes, we aimed to obtain a comprehensive view on how risk- and protective factors relate in this population. Based on previous work we identified cognitive control, adaptive and maladaptive emotion regulation as well as resilience and residual depressive symptoms as key constructs. The relationships between these constructs were examined using network analyses in order to obtain a

Acknowledgements

This research was supported by a Special Research Fund (BOF) of Ghent University awarded to Kristof Hoorelbeke (B/13808/01) and by Grant BOF10/GOA/014 for a Concerted Research Action of Ghent University awarded to Ernst Koster. Igor Marchetti is a postdoctoral research fellow of the Research Foundation Flanders (FWO; FWO14/PDO/115). Maarten De Schryver was supported by Grant BOF09/01M00209 awarded to Jan De Houwer.

References (74)

  • N. Garnefski et al.

    Negative life events, cognitive emotion regulation and emotional problems

    Personal. Individ. Differ.

    (2001)
  • S.L. Keng et al.

    Effects of mindfulness on psychological health: a review of empirical studies

    Clin. Psychol. Rev.

    (2011)
  • L.S. Middleton et al.

    The relationship between perceived and objective cognitive functioning in multiple sclerosis

    Arch. Clin. Neuropsychol.

    (2006)
  • P. Philippot et al.

    Induced rumination dampens executive processes in dysphoric young adults

    J. Behav. Ther. Exp. Psychiatry

    (2008)
  • T.N. Tombaugh

    A comprehensive review of the Paced Auditory Serial Addition Test (PASAT)

    Arch. Clin. Neuropsychol.

    (2006)
  • M.-A. Vanderhasselt et al.

    Impairments in cognitive control persist during remission from depression and are related to the number of past episodes: an event related potentials study

    Biol. Psychol.

    (2009)
  • M.-A. Vanderhasselt et al.

    Adaptive cognitive emotion regulation moderates the relationship between dysfunctional attitudes and depressive symptoms during a stressful life period: a prospective study

    J. Behav. Ther. Exp. Psychiatry

    (2014)
  • C.E. Waugh et al.

    A resilience framework for promoting stable remission from depression

    Clin. Psychol. Rev.

    (2015)
  • M. Zimmerman et al.

    A new type of scale for determining remission from depression: the remission from depression questionnaire

    J. Psychiatr. Res.

    (2013)
  • D. Borsboom et al.

    Network analysis: an integrative approach to the structure of psychopathology

    Annu. Rev. Clin. Psychol.

    (2013)
  • D. Borsboom et al.

    The small world of psychopathology

    PLoS One

    (2011)
  • N. Cohen et al.

    Inhibition of negative content – a shared process in rumination and reappraisal

    Front. Psychol.

    (2014)
  • N. Cohen et al.

    Linking executive control and emotional response: a training procedure to reduce rumination

    Clin. Psychol. Sci.

    (2015)
  • G. Costantini et al.

    Development of indirect measures of conscientiousness: combining a facets approach and network analysis

    Eur. J. Personal.

    (2015)
  • D.R. Cox et al.

    Linear dependencies represented by chain graphs

    Stat. Sci.

    (1993)
  • A.O.J. Cramer et al.

    Comorbidity: a network perspective

    Behav. Brain Sci.

    (2010)
  • R. De Raedt et al.

    Understanding vulnerability for depression from a cognitive neuroscience perspective: a reappraisal of attentional factors and a new conceptual framework

    Cogn. Affect. Behav. Neurosci.

    (2010)
  • K. Demyttenaere et al.

    Prevalence, severity, and unmet need for treatment of mental disorders in the World Health Organization World Mental Health Surveys

    J. Am. Med. Assoc.

    (2004)
  • M. De Schryver et al.

    Unpacking constructs: a network approach for studying war exposure, daily stressors and post-traumatic stress disorder

    Front. Psychol.

    (2015)
  • S. Epskamp et al.

    qgraph: network visualizations of relationships in psychometric data

    J. Stat. Softw.

    (2012)
  • B.L. Fredrickson

    The role of positive emotions in positive psychology-the broaden-and-build theory of positive emotions

    Am. Psychol.

    (2001)
  • B.L. Fredrickson et al.

    What good are positive emotions in crises? A prospective study of resilience and emotions following the terrorist attacks on the United States on September 11th, 2001

    J. Personal. Soc. Psychol.

    (2003)
  • E.I. Fried

    Problematic assumptions have slowed down depression research: why symptoms, not syndromes are the way forward

    Front. Psychol.

    (2015)
  • T.M.J. Fruchterman et al.

    Graph drawing by force-directed placement

    Softw.:Pract. Exp.

    (1991)
  • N. Geschwind et al.

    Mindfulness training increases momentary positive emotions and reward experience in adults vulnerable to depression: a randomized controlled trial

    J. Consult. Clin. Psychol.

    (2011)
  • F.E. Griffiths et al.

    The effect of strategies of personal resilience on depression recovery in an Australian cohort: a mixed methods study

    Health

    (2015)
  • U. Groemping

    Relative importance for linear regression in R: the package relaimpo

    J. Stat. Softw.

    (2006)
  • Cited by (70)

    • Individual differences associated with treatment adherence and transfer effects following gamified web-based cognitive control training for repetitive negative thinking

      2022, Internet Interventions
      Citation Excerpt :

      In the current data, compared to the subjective indicator of executive functioning (effortful control), the objective indicator of executive functioning (i.e., naPASAT task performance) was not indicative for at-risk status (i.e., this indicator did not differentiate well between the low-, moderate- and high-risk group based on User Profile). This finding is line with recent findings suggesting poor correspondence between subjective and objective indicators of cognitive functioning (Hoorelbeke et al., 2016b; Mohn and Rund, 2016; Potvin et al., 2016; Van den Bergh et al., 2021), which in the context of depression may be explained by factors such as rumination, self-efficacy (Hagen et al., in press), and level of symptomatology (Serra-Blasco et al., 2019). In contrast to the predictive role of User Profile for change in cognitive task performance and severity of internalizing symptomatology over time, and anxiety- and stress symptoms in particular (i.e., for depression only a non-significant trend was observed), User Profile did not predict change in RNT.

    View all citing articles on Scopus
    View full text