- Email: [email protected]
An Investigation of the Relationship Between Cognitive Reactivity and Rumination
Available online at www.sciencedirect.com
Behavior Therapy 39 (2008) 65 – 71
www.elsevier.com/locate/bt
An Investigation of the Relationship Between Cognitive Reactivity and Rumination Michelle L. Moulds, Eva Kandris, Alishia D. Williams, Tamara Lang, Carol Yap, Karolin Hoffmeister The University of New South Wales, Sydney
Teasdale’s (Teasdale, J.D. (1988). Cognitive vulnerability to persistent depression. Cognition and Emotion, 2, 247–274) differential activation hypothesis refers to the ease with which maladaptive cognitive processes are triggered by mild dysphoria as cognitive reactivity. Supporting this model is evidence of a differential association between sad mood and dysfunctional cognitions in formerly depressed and neverdepressed individuals and the finding that cognitive reactivity predicts depression recurrence in remitted depressives. The Leiden Index of Depression Sensitivity–Revised (LEIDS-R; Van der Does, A.J.W., Williams, J.M.G. (2003). Leiden Index of Depression Sensitivity–Revised (LEIDS-R). Retrieved September 4, 2007, from http://www.dousa.nl/ publications_depression.htm#LEIDS) is a recently developed self-report measure that provides clinicians and researchers with a time-efficient means by which to assess cognitive reactivity. This study investigated the relationship between cognitive reactivity (indexed by the LEIDS-R) and depressive rumination in a nonclinical sample (N = 324). As predicted, partial correlations between the LEIDS-R (subscale and Total scores) and the Ruminative Response Scale (RRS; Nolen-Hoeksema, S., and Morrow, J. (1991). A prospective study of depression and posttraumatic stress symptoms after a natural disaster: The 1989 Loma Prieta earthquake. Journal of Personality and Social Psychology, 61, 115–121) were significant after controlling for current depressive symptoms. A subsample of participants (n= 130)
This research was supported by a grant from the Australian Research Council awarded to the first author. We thank Dr. Willem Van der Does for kindly providing the Leiden Index of Depression Sensitivity and for his helpful correspondence regarding this measure. Address correspondence to Michelle L. Moulds, School of Psychology, The University of New South Wales, NSW 2052 Australia; e-mail: [email protected]. 0005-7894/07/0065–0071$1.00/0 © 2007 Association for Behavioral and Cognitive Therapies. Published by Elsevier Ltd. All rights reserved.
was administered a structured interview to determine current and past depression diagnostic status. Recovered depressed individuals scored more highly on the LEIDS-R Total and LEIDS-R Rumination subscale; however, the groups did not differ on the remaining subscales. Regression analyses indicated that (across all participants) LEIDS-R Total made a unique contribution to the prediction of depression over and above trait level of depressive rumination. Overall, the LEIDS-R is a time-efficient self-report index of cognitive reactivity that demonstrates promise in distinguishing recovered and never-depressed individuals.
BIOLOGICAL , BEHAVIORAL , AND COGNITIVE MODELS have been advanced to account for the processes and features that underpin vulnerability to the onset and recurrence of depression. Teasdale’s (1988) differential activation hypothesis offers a cognitive account of depression vulnerability. The central premise of this model is that the cognitive processes and content that are activated when an individual experiences sad mood are the critical determinants of whether the mood disturbance is transient or, alternatively, whether it persists and develops into an episode of clinical depression. The degree to which negative cognitive content and maladaptive patterns of cognitive processing are triggered by mild dysphoric states is referred to as cognitive reactivity. Teasdale (1988) posits that mild or transient dysphoria will persist and become clinical depression if “a vicious cycle based on the reciprocal relationship between the depressed state and negative thinking” (Teasdale, 1988, p. 254) has been established. Preliminary support for this model comes from the finding that, whereas initial depressive episodes are typically precipitated by negative life events, the relationship between environmental stressors and
66
moulds et al.
depression progressively diminishes with recurrent episodes (Kendler, Thornton, and Gardner, 2000). Cognitive theorists have proposed that cognitive reactivity can account for the onset of depression in the absence of a precipitating event. That is, maladaptive cognitive processing becomes increasingly likely to be activated as more depressive episodes are experienced, and thus, over time, serves to trigger depressive relapse in the absence of external stressors (Segal, Williams, & Teasdale, 2002; Segal, Williams, Teasdale, & Gemar, 1996). More direct support for the differential activation hypothesis has been provided by studies that demonstrate that although recovered depressed patients and never-depressed controls do not differ on self-report of dysfunctional cognitions (e.g., “If I fail at my work, then I am a failure as a person”) in neutral mood, following a sad mood induction, recovered depressed patients report elevated levels of such cognitions compared to controls (Miranda & Persons, 1988). This finding supports a differential association between sad mood and dysfunctional cognitions for formerly depressed and neverdepressed groups; however, its correlational nature precludes interpretations about the causal role of this mood-cognition relationship in the maintenance of depression. An important validation of Teasdale’s (1988) model was the observation that the degree of activation of dysfunctional beliefs following a mood induction predicted depressive relapse by formerly depressed patients at 30-month follow-up (Segal, Gemar, & Williams, 1999). The index of cognitive reactivity employed by Segal et al. (1999) involved participants completing the Dysfunctional Attitudes Scale (DAS; Weissman & Beck, 1978) prior to and following a sad mood induction procedure. In this mood induction, participants listened to a piece of sad music while recalling a negative autobiographical memory. This methodology is recognized as the gold standard procedure for indexing cognitive reactivity and has been used in numerous studies. Nonetheless, the procedure is time-consuming, and moreover, it involves either repeated administration of the DAS in one testing session or administration of parallel forms of the DAS. The latter option is problematic, given that the degree to which the two forms are comparable has been questioned (Van der Does, 2002a). In order to address these challenges, Van der Does (2002a) developed the Leiden Index of Depression Sensitivity (LEIDS) as an alternative self-report and time-efficient tool with which clinicians and researchers can measure cognitive reactivity, independent of the standard mood induction procedure described above. Examples of
LEIDS items include: “When in a sad mood, I become more bothered by perfectionism”and“When I feel sad, I feel less able to cope with everyday tasks and interests.” The original LEIDS was a 52-item self-report measure. Initial investigation of its psychometric properties indicated that the LEIDS was best described by four factors comprised of 26 items (Negative Self-Evaluation, Acceptance/Coping, Indifference, and Risk Aversion) and that these factors possessed good psychometric properties (Van der Does, 2002a). Importantly, analyses confirmed that, as intended, the 26-item LEIDS predicted cognitive change (as measured by the DAS) in response to the standard mood induction procedure (see also Van der Does, 2002b). By comparison, baseline depression and cognitive dysfunction did not predict cognitive change. Van der Does therefore concluded that the LEIDS “is a promising measure of cognitive reactivity” (2002a, p.105). A more recent form of this measure, the LEIDS-R (Van der Does & Williams, 2003) contains 34 items on six subscales (Hopelessness/Suicidality, Acceptance/Coping, Aggression, Control/Perfectionism, Risk Aversion, and Rumination). This version of the LEIDS-R has now been used in a number of studies (e.g., Merens et al., 2005; Williams, Van der Does, Barnhofer, Crane, & Segal, in press), and Van der Does recommends use of this version (personal communication, May 20, 2005). Despite the advantages of employing a validated self-report instrument to index cognitive reactivity, to date the LEIDS-R has not been used widely. Moreover, it has not been extensively applied to compare formerly and never-depressed individuals. In part one of the initial validation study, participants were asked to indicate whether they had experienced a period of sad mood and/or anhedonia, and to clarify whether it lasted 2 or more weeks. This question provided a probable index of depression history. Under this classification, the formerly depressed and never-depressed groups differed as expected on all of the LEIDS subscales (Van der Does, 2002a). In the second part of this study a structured clinical interview was administered to establish the presence of a history of depression. Eight of the 48 participants in this part of the study reported a previous depressive episode (Van der Does, 2002a). Accordingly, Van der Does highlighted the need for future research to employ the LEIDS to examine cognitive reactivity in these two groups. To our knowledge, only two additional studies in the published literature (Merens et al., 2005; Van der Does, 2005) have compared recovered and never-depressed participants on the LEIDS. Van der Does (2005) reported higher LEIDS
cognitive reactivity and rumination scores for recovered depressed compared to neverdepressed participants. In addition, cognitive reactivity as indexed by the LEIDS was positively correlated with self-reported thought suppression and an experimental index of suppression, leading to the conclusion that suppression may be linked to the “apparent inactive state of depressive cognitions during remission” (p. 1). More broadly, the relationship between cognitive reactivity and other cognitive features of depression, such as depressive rumination, have not been explored. The response styles theory of depression (Just & Alloy, 1997; Nolen-Hoeksema, 1991) contends that the duration and course of depression are determined by how one responds to depressive symptoms. That is, individuals who respond to low mood and associated symptoms by engaging in rumination about their causes and consequences are more likely to remain depressed for longer. Prospective studies have confirmed the role of rumination in the onset (Nolen-Hoeksema, Morrow, & Fredrickson, 1993) and maintenance of depression (Nolen-Hoeksema, 2000). In addition, experimental manipulations of rumination in dysphoric and clinically depressed samples have demonstrated a number of deleterious outcomes, including poor problem solving (Lyubomirsky & Nolen-Hoeksema, 1995; Watkins & Moulds, 2005), overgeneral autobiographical memory (Watkins & Teasdale, 2004; Watkins, Teasdale, & Williams, 2000), and negative cognition (Lavender & Watkins, 2004). These convergent findings support the role of rumination in the maintenance of depression-related deficits. Rumination and cognitive reactivity both represent ways in which individuals respond to low mood. Moreover, both rumination (Nolen-Hoeksema, 2000) and cognitive reactivity (Segal et al., 1999) are established vulnerability factors for recurrence of depression. Surprisingly, however, their relationship has not been explored. Van der Does (2002a, 2005) has called for research to examine the relationship between the LEIDS and the well-established gold standard measure of rumination, the Ruminative Response Style subscale of the Response Style Questionnaire (NolenHoeksema & Morrow, 1991). On the basis that depressive rumination and cognitive reactivity are both dysfunctional cognitive processes activated in response to low mood, we hypothesized that cognitive reactivity (indexed by the LEIDS-R) and depressive rumination (measured by the RRS) would be correlated in our nonclinical sample after controlling for current depression symptoms. Specifically, it was predicted that depressive rumination would be positively corre-
67
lated with Hopelessness/Suicidality, Aggression, Control/Perfectionism, Risk Aversion and Rumination, but negatively associated with Acceptance/ Coping (Van der Does, 2002a). Second, in order to examine the capacity of the LEIDS to index cognitive reactivity, and so to discriminate between recovered and never-depressed individuals, we administered a structured interview to a subset of participants to index depression status (i.e., currently depressed, previously depressed, neverdepressed). In line with the focus of the differential activation hypothesis, our predictions were made in terms of recovered depressed and never-depressed participants. Specifically, we predicted that recovered depressed participants would report higher levels of cognitive reactivity (i.e., higher scores across all subscales of the LEIDS-R) and rumination than their never-depressed counterparts and that these variables would be positively correlated after current depression symptoms were partialled out.
Method participants Participants were 324 undergraduate students in the School of Psychology at The University of New South Wales, who participated in the study in exchange for course credit. Participants had a mean age of 20.07 years (SD = 4.73), and the total sample was comprised of 215 females and 109 males. A subset of 130 participants was administered the mood module of the Structured Clinical Interview for DSM-IV (SCID; Spitzer, Williams, Gibbon, & First, 1996) in order to establish current mood status. The SCID was administered by a master’slevel clinical psychologist (EK) with extensive experience in the use of this instrument in clinical research settings. measures Leiden Index of Depression Sensitivity–Revised (LEIDS-R; Van der Does & Williams, 2003). The LEIDS-R is a self-report measure that indexes cognitive reactivity in response to low mood. Participants are instructed to think about the last time they felt “somewhat sad,” and to indicate the degree to which a list of statements describe their typical cognitions and behaviors in response to sad mood. The LEIDS has good psychometric properties (Van der Does, 2002a), and responses on this measure correlate with a mood induction procedure employed to index cognitive reactivity (e.g., Segal et al., 1999). Internal consistency in the current sample was strong (.88). Beck Depression Inventory-II (BDI-II; Beck, Steer, & Brown, 1996). The BDI-II is a self-report
68
moulds et al.
measure of depressive symptoms that possesses strong psychometric properties, including internal consistency of .92 in a sample of psychiatric outpatients and .93 in college students (Beck et al., 1996). Ruminative Response Scale of the Response Styles Questionnaire (RRS; Nolen-Hoeksema & Morrow, 1993). The RRS is a self-report measure that indexes the tendency to ruminate in response to depressed mood. The RRS contains symptomfocused and self-focused items. The RRS possesses good internal consistency and validity (NolenHoeksema & Morrow, 1991).
procedure After providing written informed consent, participants completed a package of self-report questionnaires (described above) in a counterbalanced order, so as to avoid potential order effects. Participants were then thoroughly debriefed and informed about the aims of the research. Participants whose responses indicated high levels of anxiety or depression were provided with information about relevant clinical services.
Results Preliminary screening of the data (prompted by the large standard deviations for BDI-II and LEIDS-R scores) indicated that 6 participants’ scores on these measures were outliers (i.e., more than two standard deviations above the mean). These participants were excluded from the analyses. Specifically, 3 never-depressed participants were excluded on the basis of their LEIDS-R score, 1 recovered depressed participant was excluded owing to an extreme BDI-II score, and 1 never-depressed and 1 recovered depressed participant were excluded owing to outlying scores on both measures. Thus, the analyses reported below were carried out with the data of 318 participants. Mean scores on the questionnaire measures across the total sample are presented in Table 1. Partial correlations controlling for BDI-II scores were conducted to explore the association between rumination and cognitive reactivity. As shown in Table 2, significant positive correlations were detected between the RRS and the LEIDS-R Total and subscale scores after controlling for current depressive symptoms, confirming that the observed relationships between these measures are independent of current mood state. The hypothesized inverse correlation between RRS and Acceptance/ Coping was not observed (r = .01, p N .05). In the second part of the study, participants were classified according to their responses on the major
Table 1 Mean scores (and standard deviations) on study measures (N = 318) BDI-II RRS LEIDS-R Total LEIDS-R Hopelessness LEIDS-R Acceptance LEIDS-R Aggression LEIDS-R Control LEIDS-R Risk Aversion LEIDS-R Rumination
8.74 (6.82) 41.98 (11.24) 46.61 (17.20) 4.90 (4.23) 3.53 (3.20) 8.44 (4.59) 7.77 (3.85) 10.60 (4.00) 11.48 (4.68)
Note. BDI-II = Beck Depression Inventory-II; RRS = Ruminative Response Scale; LEIDS-R Total = Leiden Index of Depression Sensitivity Total; LEIDS-R Hopelessness = Hopelessness/Suicidality subscale; LEIDS-R Acceptance = Acceptance/Coping subscale; LEIDS-R Aggression = Aggression subscale; LEIDS-R Control = Control/Perfectionism subscale; LEIDS-R Risk Aversion = Risk Aversion subscale; LEIDS-R Rumination = Rumination subscale.
depression component of the SCID. After the removal of excluded participants (as noted above), 6 participants (4.8%; 3 male, 3 female) met criteria for a current major depressive episode (MDE; American Psychiatric Association, 1994). Of the remaining participants, 40 (32.3%) reported a previous episode of major depression, and 78 (62.9%) denied any current or past MDE. The never-depressed group had a mean age of 19.05 years (SD = 3.28) and was comprised of 53 females and 25 males. Recovered depressed participants had a mean age of 23.33 years (SD = 10.09), and included 30 females and 10 males. As a group, they reported an average of 1.79 (SD = 3.14) previous depressive episodes. On average, their last depressive episode ended 36.93 months (SD = 44.69) prior to participation in the study. The mean duration of the last MDE was 6.10 months (SD = 10.67). Independent samples t-tests demonstrated that the recovered depressed group was older than the never-depressed group, t(43.27) = − 2.61, p b .05, unequal variance estimate. The recovered depressed and never-depressed group did not differ in their current level of depression, as indexed by the BDIII, t(61.00) = − .37, p N .05, unequal variance estimate (see Table 3). Thus, any observed differences cannot be accounted for by between-group differences in current mood. The recovered depressed group scored more highly on the LEIDS-R (Total), t (95.64) = −2.13, p b .05, unequal variance estimate. Not surprisingly, the recovered depressed group reported higher levels of trait rumination on the RRS, t(61.90) = −2.02, p b .05, unequal variance estimate, and on the Rumination subscale of the LEIDS-R, t(116) = −2.37, p b .05. The two groups did not differ significantly on the remaining LEIDSR subscales: Hopelessness, t(116) = − .75, p N .05;
69
cognitive reactivity and rumination Table 2 Correlations between rumination and cognitive reactivity controlling for depression (BDI-II) (N = 318)
RRS LEIDS-R Hopeless Accept Aggress Control Risk Rum
RRS
LEIDS-R
Hopeless
Accept
Aggress
Control
Risk
Rum
– .41⁎⁎⁎ .33⁎⁎⁎ .01 .26⁎⁎⁎ .13⁎ .30⁎⁎⁎ .51⁎⁎⁎
– .61⁎⁎⁎ .40⁎⁎⁎ .70⁎⁎⁎ .60⁎⁎⁎ .73⁎⁎⁎ .77⁎⁎⁎
– .06 .41⁎⁎⁎ .10 .38⁎⁎⁎ .40⁎⁎⁎
– .15⁎⁎ .37⁎⁎⁎ .16⁎⁎ .03
– .23⁎⁎⁎ .30⁎⁎⁎ .51⁎⁎⁎
– .37⁎⁎⁎ .31⁎⁎⁎
– .58⁎⁎⁎
–
Note. BDI-II = Beck Depression Inventory-II; RRS = Ruminative Response Scale; LEIDS-R Total = Leiden Index of Depression Sensitivity Total; Hopeless = LEIDS-R Hopelessness/Suicidality subscale; Accept = LEIDS-R Acceptance/Coping subscale; Aggress = LEIDS-R Aggression subscale; Control = LEIDS-R Control/Perfectionism subscale; Risk = LEIDS-R Risk Aversion subscale; Rum = LEIDS-R Rumination subscale. ⁎p b .05, ⁎⁎p b .01, ⁎⁎⁎p b .001.
Aggression, t(96.15) = −.33, p N .05, unequal variance estimate; Control, t(116) = − 1.45, p N .05; Risk Aversion, t(115) = −1.75, p N .05. An unexpected observation was that the recovered depressed participants scored more highly than their never-depressed counterparts on the Acceptance/Coping subscale, although this difference did not reach statistical significance, t(116) = −1.48, p N .05. To investigate whether cognitive reactivity, as indexed by the LEIDS-R Total score, explained any of the variance in depression beyond depressive rumination (a well-established predictor of depression), we conducted a multiple regression with BDIII scores as the dependent variable, RRS scores entered on the first step, and LEIDS-R total scores entered on the second step. Across the recovered Table 3 Mean participant characteristics for recovered depressed and never-depressed groups Never Recovered depressed (n = 78) depressed (n = 40) Age BDI-II RRS LEIDS-R Total LEIDS-R Hopelessness LEIDS-R Acceptance LEIDS-R Aggression LEIDS-R Control LEIDS-R Risk Aversion LEIDS-R Rumination
19.05 (3.28) 8.44 (5.29) 40.32 (9.25) 42.77 (17.99) 4.47 (4.02) 2.88 (2.85) 8.23 (4.85) 6.94 (4.66) 9.78 (4.07) 10.46 (5.02)
23.33 (10.09) 8.93 (7.27) 44.45 (12.47) 49.18 (13.84) 5.05 (3.80) 3.73 (3.07) 8.53 (3.84) 8.20 (4.10) 11.15 (3.82) 12.60 (4.44)
Note. Standard deviations appear in parentheses. Means are unadjusted. BDI-II = Beck Depression Inventory-II; RRS = Ruminative Response Scale; LEIDS-R Total = Leiden Index of Depression Sensitivity Total; LEIDS-R Hopelessness = Hopelessness/Suicidality subscale; LEIDS-R Acceptance = Acceptance/Coping subscale; LEIDS-R Aggression = Aggression subscale; LEIDS-R Control = Control/Perfectionism subscale; LEIDS-R Risk Aversion = Risk Aversion subscale; LEIDS-R Rumination = Rumination subscale.
depressed and never-depressed participants, RRS scores accounted for 33% of the variance in BDI-II scores (adjusted R2 = .33, β = .57, SE = .03, t = 12.18, p b .001). Interestingly, LEIDS-R Total made a small (but significant) contribution, accounting for 6% of the variance in BDI-II (adjusted R2 = .39, β = .31, SE = .02, t = 5.69, p b .001). Thus, cognitive reactivity made a unique contribution to the prediction of depression, over and above trait level of depressive rumination.
Discussion Our first goal in this study was to examine the interrelationship of cognitive reactivity and rumination. As predicted, these cognitive processes were positively correlated even when current level of depression was partialled out—confirming an association that is independent of current symptomatology. Consistent with clinical observation, this relationship demonstrates that individuals who respond to depressed mood by ruminating about the causes, meanings, and implications of their symptoms (Nolen-Hoeksema, 1991) are also likely to engage in an array of additional maladaptive cognitive and behavioral responses, including feeling hopeless and suicidal, responding with aggression, and attempting to avoid harm/conflicts. A second goal was to compare cognitive reactivity, as indexed by the LEIDS-R, in recovered and never-depressed individuals. As predicted, the recovered depressed participants indicated greater cognitive reactivity generally (i.e., LEIDS-R Total), and scored more highly on the LEIDS-R Rumination subscale and the RRS. However, the two groups did not differ on the remaining LEIDS-R subscales. Thus although the LEIDS-R Total differentiated how formerly depressed and neverdepressed individuals respond when they experience sad mood, this effect was accounted for mainly by the LEIDS-R Rumination subscale. This finding,
70
moulds et al.
combined with the observation that formerly and never-depressed individuals differed on the basis of RRS scores, accords with extensive empirical support for the role of depressive rumination in the persistence of depression (Nolen-Hoeksema, 2000) and associated cognitive and behavioral deficits (e.g., impoverished problem solving, Watkins & Moulds, 2005; overgeneral autobiographical memory retrieval, Watkins, Teasdale, & Williams, 2000). At this stage, the degree to which the Hopelessness, Acceptance, Aggression, Control, and Risk Aversion subscales are able to differentiate remitted and never-depressed individuals remains unclear. Replication of this investigation is needed for clarification. Importantly, across the sample of recovered and never-depressed participants, the Total LEIDS-R score explained unique variance in depression over and above that accounted for by the RRS. The RRS is the gold standard instrument used to measure depressive rumination that has convincingly predicted depression across multiple longitudinal studies. Our finding therefore underscores the utility of the LEIDS-R as a clinical instrument that gauges responses to low mood that are predictive of concurrent depression. Future studies should investigate whether the LEIDS-R has the capacity to predict depression across time within a longitudinal design. The construct of cognitive reactivity incorporates both depression-related cognitive processes and content that are theorized to be activated in response to sad mood. However, beyond studies that index cognitive content—i.e., maladaptive cognitions— following induced negative mood, experimental work in this area to date has not tested the differential activation hypothesis more broadly and indexed a range of potential dysfunctional processes (cognitive and behavioral) that characterize depressed individuals’ responses to low mood. By measuring a range of cognitive and behavioral processes that an individual may engage in at times of low mood, the LEIDS-R has the capacity to test the differential activation hypothesis more broadly. Moreover, as a clinical tool, responses on the LEIDS-R have the capacity to provide invaluable information about a range of maladaptive responses that clients may engage in, including both dysfunctional cognitive content and processes. These can then serve as key targets in therapy. Related to this, future work would benefit from extending investigations in this area to more broadly incorporate additional well-established cognitive processes that operate in depression; for example, memory disturbances such as overgeneral memory, intrusive memories and biased retrieval of negative material. The possibility that such processes of memory are
activated in response to (and exacerbated by) sad mood, and whether their activation predicts depression relapse, is worthy of examination. A number of limitations warrant acknowledgment. First, we did not employ a treatment-seeking sample. Participants were an unselected sample of undergraduate students. Nonetheless, participants in the second phase of the study were administered a structured clinical interview by a qualified clinical psychologist. Thus, although the sample was not drawn from a clinical population per se, it is noteworthy that their responses on an established structured interview (conducted by an experienced clinician) were used to ascertain their depression diagnostic history. However, future research that involves the recruitment of formerly depressed participants (e.g., posttreatment) would support the generalizability of our findings to treatmentseeking individuals in the wider community. Second, we note that the scores obtained on the selfreport measures for the never-depressed group were relatively high for a healthy participant group. As depression was the focus of our study, we administered only the major depression component of the SCID. Thus, we cannot rule out the possibility that participants also satisfied diagnostic criteria for additional psychological conditions (e.g., anxiety disorders, eating disorders), and thus that the scores observed on self-report measures were an artifact of comorbidity that was not indexed in the current design. Similarly, we did not assess current or additional depressive conditions such as dysthymia and minor or subclinical depression; again, this may have inflated scores in the never-depressed group. Future studies that incorporate a more thorough screening to rule out the influence of comorbidity and include extensive background details (e.g., history of psychological treatment) will provide an important extension of our findings. Our sample included more female than male participants, and future studies should employ gender-balanced samples. An additional aspect of our sample that should be kept in mind when interpreting the results is the age of participants; the never-depressed group was significantly younger than the recovered depressed group. The possibility that some of the never-depressed participants may, in fact, be vulnerable to depression, but had not yet experienced a depressive episode at the time of involvement in the research, cannot be ruled out. Despite potentially possessing features that confer vulnerability to depression, it is plausible that individuals in this younger group had not experienced a sufficiently aversive life event to precipitate a depressive episode. Thus, in comparing the neverdepressed and remitted depressed group, we must
cognitive reactivity and rumination limit our interpretations so as not to draw conclusions about depression-vulnerable versus nondepression-vulnerable individuals. Related to this point, an outstanding question in the field is whether cognitive reactivity encompasses a range of maladaptive cognitive processes that are sequelae of experiencing depression, or whether the presence of these processes confers vulnerability to the onset of an initial depressive episode. As raised by Van der Does (2002a), future studies that investigate cognitive processes in high-risk samples using prospective designs (e.g., Bryant & Guthrie, 2005) are needed in order to disentangle whether cognitive reactivity is a vulnerability factor that increases the likelihood of initial depression onset, as well as recurrence. The development of easily administered self-report instruments that index complex constructs such as cognitive reactivity increases the likelihood that clinical researchers will tackle such research questions in the future. References American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the Beck Depression Inventory–II. San Antonio, TX: Psychological Corporation. Bryant, R. A., & Guthrie, R. M. (2005). Maladaptive appraisals as a risk factor for posttraumatic stress: A study of trainee firefighters. Psychological Science, 16, 749–752. Just, N., & Alloy, L. B. (1997). The response styles theory of depression: Tests and an extension of the theory. Journal of Abnormal Psychology, 106, 221–229. Lavender, A., & Watkins, E. (2004). Rumination and future thinking in depression. British Journal of Clinical Psychology, 43, 129–142. Lyubomirsky, S., & Nolen-Hoeksema, S. (1995). Effects of selffocused rumination on negative thinking and interpersonal problem-solving. Journal of Personality and Social Psychology, 69, 176–190. Merens, W., Booij, L., Markus, R., Zitman, F. G., Onkenhout, W., & Van der Does, A. J. W. (2005). The effects of a diet enriched with α-lactalbumin on mood and cortisol response in unmedicated recovered depressed subjects and controls. British Journal of Nutrition, 94, 415–422. Miranda, J., & Persons, J. B. (1988). Dysfunctional attitudes are mood-state dependent. Journal of Abnormal Psychology, 97, 76–79. Nolen-Hoeksema, S. (1991). Responses to depression and their effects on the duration of depressive episodes. Journal of Abnormal Psychology, 100, 569–582. Nolen-Hoeksema, S. (2000). The role of rumination in depressive disorders and mixed anxiety/depressive symptoms. Journal of Abnormal Psychology, 109, 504–511. Nolen-Hoeksema, S., & Morrow, J. (1989). A prospective
71
study of depression and posttraumatic stress symptoms after a natural disaster: The 1989 Loma Prieta earthquake. Journal of Personality and Social Psychology, 61, 115–121. Nolen-Hoeksema, S., & Morrow, J. (1993). Effects of rumination and distraction on naturally occurring depressed mood. Cognition and Emotion, 7, 561–570. Nolen-Hoeksema, S., Morrow, J., & Fredrickson, B. L. (1993). Response styles and the duration of episodes of depressed mood. Journal of Abnormal Psychology, 102, 20–28. Segal, Z. V., Gemar, M., & Williams, S. (1999). Differential cognitive response to a mood challenge following successful cognitive therapy or pharmacotherapy for unipolar depression. Journal of Abnormal Psychology, 108, 3–10. Segal, Z. V., Williams, J. M. G., & Teasdale, J. D. (2002). Mindfulness-based cognitive therapy for depression. New York: The Guilford Press. Segal, Z. V., Williams, J. M., Teasdale, J. D., & Gemar, M. (1996). A cognitive science perspective on kindling and episode sensitization in recurrent affective disorder. Psychological Medicine, 26, 371–380. Spitzer, R. L., Williams, J. B. W., Gibbon, M., & First, M. B. (1996). Structured Clinical Interview for DSM-IV (SCID). Washington DC: American Psychiatric Press. Teasdale, J. D. (1988). Cognitive vulnerability to persistent depression. Cognition and Emotion, 2, 247–274. Van der Does, W. (2002). Cognitive reactivity to sad mood: Structure and validity of a new measure. Behaviour Research and Therapy, 40, 105–120. Van der Does, W. (2002). Different types of experimentally induced sad mood? Behavior Therapy, 33, 551–561. Van der Does, W. (2005). Thought suppression and cognitive vulnerability to depression. British Journal of Clinical Psychology, 44, 1–14. Van der Does, A. J. W., & Williams, J. M. G. (2003). Leiden Index of Depression Sensitivity–Revised (LEIDS-R). Retrieved September 4, 2007, from http://www.dousa.nl/publications_ depression.htm#LEIDS. Watkins, E., & Moulds, M. (2005). Distinct modes of ruminative self-focus: Impact of abstract versus concrete rumination on problem solving in depression. Emotion, 5, 319–328. Watkins, E., & Teasdale, J. D. (2004). Adaptive and maladaptive self-focus in depression. Journal of Affective Disorders, 82, 1–8. Watkins, E., Teasdale, J. D., & Williams, R. M. (2000). Decentering and distraction reduce overgeneral autobiographical memory in depression. Psychological Medicine, 30, 911–920. Weissman, A., Beck, A. T. (1978 November). The Dysfunctional Attitudes Scale. Paper presented at the annual meeting of the Association for Advancement of Behavior Therapy, Chicago, IL. Williams, J. M. G., VanderDoes, A. J. W., Barnhofer, T., Crane, C., & Segal, Z. V. (in press). Cognitive reactivity, suicidal ideation and future fluency: Investigating a differential activation theory of suicidality. Cognitive Therapy and Research.
R E C E I V E D : July 2, 2006 A C C E P T E D : May 1, 2007 Available online 26 October 2007
Behavior Therapy 39 (2008) 65 – 71
www.elsevier.com/locate/bt
An Investigation of the Relationship Between Cognitive Reactivity and Rumination Michelle L. Moulds, Eva Kandris, Alishia D. Williams, Tamara Lang, Carol Yap, Karolin Hoffmeister The University of New South Wales, Sydney
Teasdale’s (Teasdale, J.D. (1988). Cognitive vulnerability to persistent depression. Cognition and Emotion, 2, 247–274) differential activation hypothesis refers to the ease with which maladaptive cognitive processes are triggered by mild dysphoria as cognitive reactivity. Supporting this model is evidence of a differential association between sad mood and dysfunctional cognitions in formerly depressed and neverdepressed individuals and the finding that cognitive reactivity predicts depression recurrence in remitted depressives. The Leiden Index of Depression Sensitivity–Revised (LEIDS-R; Van der Does, A.J.W., Williams, J.M.G. (2003). Leiden Index of Depression Sensitivity–Revised (LEIDS-R). Retrieved September 4, 2007, from http://www.dousa.nl/ publications_depression.htm#LEIDS) is a recently developed self-report measure that provides clinicians and researchers with a time-efficient means by which to assess cognitive reactivity. This study investigated the relationship between cognitive reactivity (indexed by the LEIDS-R) and depressive rumination in a nonclinical sample (N = 324). As predicted, partial correlations between the LEIDS-R (subscale and Total scores) and the Ruminative Response Scale (RRS; Nolen-Hoeksema, S., and Morrow, J. (1991). A prospective study of depression and posttraumatic stress symptoms after a natural disaster: The 1989 Loma Prieta earthquake. Journal of Personality and Social Psychology, 61, 115–121) were significant after controlling for current depressive symptoms. A subsample of participants (n= 130)
This research was supported by a grant from the Australian Research Council awarded to the first author. We thank Dr. Willem Van der Does for kindly providing the Leiden Index of Depression Sensitivity and for his helpful correspondence regarding this measure. Address correspondence to Michelle L. Moulds, School of Psychology, The University of New South Wales, NSW 2052 Australia; e-mail: [email protected]. 0005-7894/07/0065–0071$1.00/0 © 2007 Association for Behavioral and Cognitive Therapies. Published by Elsevier Ltd. All rights reserved.
was administered a structured interview to determine current and past depression diagnostic status. Recovered depressed individuals scored more highly on the LEIDS-R Total and LEIDS-R Rumination subscale; however, the groups did not differ on the remaining subscales. Regression analyses indicated that (across all participants) LEIDS-R Total made a unique contribution to the prediction of depression over and above trait level of depressive rumination. Overall, the LEIDS-R is a time-efficient self-report index of cognitive reactivity that demonstrates promise in distinguishing recovered and never-depressed individuals.
BIOLOGICAL , BEHAVIORAL , AND COGNITIVE MODELS have been advanced to account for the processes and features that underpin vulnerability to the onset and recurrence of depression. Teasdale’s (1988) differential activation hypothesis offers a cognitive account of depression vulnerability. The central premise of this model is that the cognitive processes and content that are activated when an individual experiences sad mood are the critical determinants of whether the mood disturbance is transient or, alternatively, whether it persists and develops into an episode of clinical depression. The degree to which negative cognitive content and maladaptive patterns of cognitive processing are triggered by mild dysphoric states is referred to as cognitive reactivity. Teasdale (1988) posits that mild or transient dysphoria will persist and become clinical depression if “a vicious cycle based on the reciprocal relationship between the depressed state and negative thinking” (Teasdale, 1988, p. 254) has been established. Preliminary support for this model comes from the finding that, whereas initial depressive episodes are typically precipitated by negative life events, the relationship between environmental stressors and
66
moulds et al.
depression progressively diminishes with recurrent episodes (Kendler, Thornton, and Gardner, 2000). Cognitive theorists have proposed that cognitive reactivity can account for the onset of depression in the absence of a precipitating event. That is, maladaptive cognitive processing becomes increasingly likely to be activated as more depressive episodes are experienced, and thus, over time, serves to trigger depressive relapse in the absence of external stressors (Segal, Williams, & Teasdale, 2002; Segal, Williams, Teasdale, & Gemar, 1996). More direct support for the differential activation hypothesis has been provided by studies that demonstrate that although recovered depressed patients and never-depressed controls do not differ on self-report of dysfunctional cognitions (e.g., “If I fail at my work, then I am a failure as a person”) in neutral mood, following a sad mood induction, recovered depressed patients report elevated levels of such cognitions compared to controls (Miranda & Persons, 1988). This finding supports a differential association between sad mood and dysfunctional cognitions for formerly depressed and neverdepressed groups; however, its correlational nature precludes interpretations about the causal role of this mood-cognition relationship in the maintenance of depression. An important validation of Teasdale’s (1988) model was the observation that the degree of activation of dysfunctional beliefs following a mood induction predicted depressive relapse by formerly depressed patients at 30-month follow-up (Segal, Gemar, & Williams, 1999). The index of cognitive reactivity employed by Segal et al. (1999) involved participants completing the Dysfunctional Attitudes Scale (DAS; Weissman & Beck, 1978) prior to and following a sad mood induction procedure. In this mood induction, participants listened to a piece of sad music while recalling a negative autobiographical memory. This methodology is recognized as the gold standard procedure for indexing cognitive reactivity and has been used in numerous studies. Nonetheless, the procedure is time-consuming, and moreover, it involves either repeated administration of the DAS in one testing session or administration of parallel forms of the DAS. The latter option is problematic, given that the degree to which the two forms are comparable has been questioned (Van der Does, 2002a). In order to address these challenges, Van der Does (2002a) developed the Leiden Index of Depression Sensitivity (LEIDS) as an alternative self-report and time-efficient tool with which clinicians and researchers can measure cognitive reactivity, independent of the standard mood induction procedure described above. Examples of
LEIDS items include: “When in a sad mood, I become more bothered by perfectionism”and“When I feel sad, I feel less able to cope with everyday tasks and interests.” The original LEIDS was a 52-item self-report measure. Initial investigation of its psychometric properties indicated that the LEIDS was best described by four factors comprised of 26 items (Negative Self-Evaluation, Acceptance/Coping, Indifference, and Risk Aversion) and that these factors possessed good psychometric properties (Van der Does, 2002a). Importantly, analyses confirmed that, as intended, the 26-item LEIDS predicted cognitive change (as measured by the DAS) in response to the standard mood induction procedure (see also Van der Does, 2002b). By comparison, baseline depression and cognitive dysfunction did not predict cognitive change. Van der Does therefore concluded that the LEIDS “is a promising measure of cognitive reactivity” (2002a, p.105). A more recent form of this measure, the LEIDS-R (Van der Does & Williams, 2003) contains 34 items on six subscales (Hopelessness/Suicidality, Acceptance/Coping, Aggression, Control/Perfectionism, Risk Aversion, and Rumination). This version of the LEIDS-R has now been used in a number of studies (e.g., Merens et al., 2005; Williams, Van der Does, Barnhofer, Crane, & Segal, in press), and Van der Does recommends use of this version (personal communication, May 20, 2005). Despite the advantages of employing a validated self-report instrument to index cognitive reactivity, to date the LEIDS-R has not been used widely. Moreover, it has not been extensively applied to compare formerly and never-depressed individuals. In part one of the initial validation study, participants were asked to indicate whether they had experienced a period of sad mood and/or anhedonia, and to clarify whether it lasted 2 or more weeks. This question provided a probable index of depression history. Under this classification, the formerly depressed and never-depressed groups differed as expected on all of the LEIDS subscales (Van der Does, 2002a). In the second part of this study a structured clinical interview was administered to establish the presence of a history of depression. Eight of the 48 participants in this part of the study reported a previous depressive episode (Van der Does, 2002a). Accordingly, Van der Does highlighted the need for future research to employ the LEIDS to examine cognitive reactivity in these two groups. To our knowledge, only two additional studies in the published literature (Merens et al., 2005; Van der Does, 2005) have compared recovered and never-depressed participants on the LEIDS. Van der Does (2005) reported higher LEIDS
cognitive reactivity and rumination scores for recovered depressed compared to neverdepressed participants. In addition, cognitive reactivity as indexed by the LEIDS was positively correlated with self-reported thought suppression and an experimental index of suppression, leading to the conclusion that suppression may be linked to the “apparent inactive state of depressive cognitions during remission” (p. 1). More broadly, the relationship between cognitive reactivity and other cognitive features of depression, such as depressive rumination, have not been explored. The response styles theory of depression (Just & Alloy, 1997; Nolen-Hoeksema, 1991) contends that the duration and course of depression are determined by how one responds to depressive symptoms. That is, individuals who respond to low mood and associated symptoms by engaging in rumination about their causes and consequences are more likely to remain depressed for longer. Prospective studies have confirmed the role of rumination in the onset (Nolen-Hoeksema, Morrow, & Fredrickson, 1993) and maintenance of depression (Nolen-Hoeksema, 2000). In addition, experimental manipulations of rumination in dysphoric and clinically depressed samples have demonstrated a number of deleterious outcomes, including poor problem solving (Lyubomirsky & Nolen-Hoeksema, 1995; Watkins & Moulds, 2005), overgeneral autobiographical memory (Watkins & Teasdale, 2004; Watkins, Teasdale, & Williams, 2000), and negative cognition (Lavender & Watkins, 2004). These convergent findings support the role of rumination in the maintenance of depression-related deficits. Rumination and cognitive reactivity both represent ways in which individuals respond to low mood. Moreover, both rumination (Nolen-Hoeksema, 2000) and cognitive reactivity (Segal et al., 1999) are established vulnerability factors for recurrence of depression. Surprisingly, however, their relationship has not been explored. Van der Does (2002a, 2005) has called for research to examine the relationship between the LEIDS and the well-established gold standard measure of rumination, the Ruminative Response Style subscale of the Response Style Questionnaire (NolenHoeksema & Morrow, 1991). On the basis that depressive rumination and cognitive reactivity are both dysfunctional cognitive processes activated in response to low mood, we hypothesized that cognitive reactivity (indexed by the LEIDS-R) and depressive rumination (measured by the RRS) would be correlated in our nonclinical sample after controlling for current depression symptoms. Specifically, it was predicted that depressive rumination would be positively corre-
67
lated with Hopelessness/Suicidality, Aggression, Control/Perfectionism, Risk Aversion and Rumination, but negatively associated with Acceptance/ Coping (Van der Does, 2002a). Second, in order to examine the capacity of the LEIDS to index cognitive reactivity, and so to discriminate between recovered and never-depressed individuals, we administered a structured interview to a subset of participants to index depression status (i.e., currently depressed, previously depressed, neverdepressed). In line with the focus of the differential activation hypothesis, our predictions were made in terms of recovered depressed and never-depressed participants. Specifically, we predicted that recovered depressed participants would report higher levels of cognitive reactivity (i.e., higher scores across all subscales of the LEIDS-R) and rumination than their never-depressed counterparts and that these variables would be positively correlated after current depression symptoms were partialled out.
Method participants Participants were 324 undergraduate students in the School of Psychology at The University of New South Wales, who participated in the study in exchange for course credit. Participants had a mean age of 20.07 years (SD = 4.73), and the total sample was comprised of 215 females and 109 males. A subset of 130 participants was administered the mood module of the Structured Clinical Interview for DSM-IV (SCID; Spitzer, Williams, Gibbon, & First, 1996) in order to establish current mood status. The SCID was administered by a master’slevel clinical psychologist (EK) with extensive experience in the use of this instrument in clinical research settings. measures Leiden Index of Depression Sensitivity–Revised (LEIDS-R; Van der Does & Williams, 2003). The LEIDS-R is a self-report measure that indexes cognitive reactivity in response to low mood. Participants are instructed to think about the last time they felt “somewhat sad,” and to indicate the degree to which a list of statements describe their typical cognitions and behaviors in response to sad mood. The LEIDS has good psychometric properties (Van der Does, 2002a), and responses on this measure correlate with a mood induction procedure employed to index cognitive reactivity (e.g., Segal et al., 1999). Internal consistency in the current sample was strong (.88). Beck Depression Inventory-II (BDI-II; Beck, Steer, & Brown, 1996). The BDI-II is a self-report
68
moulds et al.
measure of depressive symptoms that possesses strong psychometric properties, including internal consistency of .92 in a sample of psychiatric outpatients and .93 in college students (Beck et al., 1996). Ruminative Response Scale of the Response Styles Questionnaire (RRS; Nolen-Hoeksema & Morrow, 1993). The RRS is a self-report measure that indexes the tendency to ruminate in response to depressed mood. The RRS contains symptomfocused and self-focused items. The RRS possesses good internal consistency and validity (NolenHoeksema & Morrow, 1991).
procedure After providing written informed consent, participants completed a package of self-report questionnaires (described above) in a counterbalanced order, so as to avoid potential order effects. Participants were then thoroughly debriefed and informed about the aims of the research. Participants whose responses indicated high levels of anxiety or depression were provided with information about relevant clinical services.
Results Preliminary screening of the data (prompted by the large standard deviations for BDI-II and LEIDS-R scores) indicated that 6 participants’ scores on these measures were outliers (i.e., more than two standard deviations above the mean). These participants were excluded from the analyses. Specifically, 3 never-depressed participants were excluded on the basis of their LEIDS-R score, 1 recovered depressed participant was excluded owing to an extreme BDI-II score, and 1 never-depressed and 1 recovered depressed participant were excluded owing to outlying scores on both measures. Thus, the analyses reported below were carried out with the data of 318 participants. Mean scores on the questionnaire measures across the total sample are presented in Table 1. Partial correlations controlling for BDI-II scores were conducted to explore the association between rumination and cognitive reactivity. As shown in Table 2, significant positive correlations were detected between the RRS and the LEIDS-R Total and subscale scores after controlling for current depressive symptoms, confirming that the observed relationships between these measures are independent of current mood state. The hypothesized inverse correlation between RRS and Acceptance/ Coping was not observed (r = .01, p N .05). In the second part of the study, participants were classified according to their responses on the major
Table 1 Mean scores (and standard deviations) on study measures (N = 318) BDI-II RRS LEIDS-R Total LEIDS-R Hopelessness LEIDS-R Acceptance LEIDS-R Aggression LEIDS-R Control LEIDS-R Risk Aversion LEIDS-R Rumination
8.74 (6.82) 41.98 (11.24) 46.61 (17.20) 4.90 (4.23) 3.53 (3.20) 8.44 (4.59) 7.77 (3.85) 10.60 (4.00) 11.48 (4.68)
Note. BDI-II = Beck Depression Inventory-II; RRS = Ruminative Response Scale; LEIDS-R Total = Leiden Index of Depression Sensitivity Total; LEIDS-R Hopelessness = Hopelessness/Suicidality subscale; LEIDS-R Acceptance = Acceptance/Coping subscale; LEIDS-R Aggression = Aggression subscale; LEIDS-R Control = Control/Perfectionism subscale; LEIDS-R Risk Aversion = Risk Aversion subscale; LEIDS-R Rumination = Rumination subscale.
depression component of the SCID. After the removal of excluded participants (as noted above), 6 participants (4.8%; 3 male, 3 female) met criteria for a current major depressive episode (MDE; American Psychiatric Association, 1994). Of the remaining participants, 40 (32.3%) reported a previous episode of major depression, and 78 (62.9%) denied any current or past MDE. The never-depressed group had a mean age of 19.05 years (SD = 3.28) and was comprised of 53 females and 25 males. Recovered depressed participants had a mean age of 23.33 years (SD = 10.09), and included 30 females and 10 males. As a group, they reported an average of 1.79 (SD = 3.14) previous depressive episodes. On average, their last depressive episode ended 36.93 months (SD = 44.69) prior to participation in the study. The mean duration of the last MDE was 6.10 months (SD = 10.67). Independent samples t-tests demonstrated that the recovered depressed group was older than the never-depressed group, t(43.27) = − 2.61, p b .05, unequal variance estimate. The recovered depressed and never-depressed group did not differ in their current level of depression, as indexed by the BDIII, t(61.00) = − .37, p N .05, unequal variance estimate (see Table 3). Thus, any observed differences cannot be accounted for by between-group differences in current mood. The recovered depressed group scored more highly on the LEIDS-R (Total), t (95.64) = −2.13, p b .05, unequal variance estimate. Not surprisingly, the recovered depressed group reported higher levels of trait rumination on the RRS, t(61.90) = −2.02, p b .05, unequal variance estimate, and on the Rumination subscale of the LEIDS-R, t(116) = −2.37, p b .05. The two groups did not differ significantly on the remaining LEIDSR subscales: Hopelessness, t(116) = − .75, p N .05;
69
cognitive reactivity and rumination Table 2 Correlations between rumination and cognitive reactivity controlling for depression (BDI-II) (N = 318)
RRS LEIDS-R Hopeless Accept Aggress Control Risk Rum
RRS
LEIDS-R
Hopeless
Accept
Aggress
Control
Risk
Rum
– .41⁎⁎⁎ .33⁎⁎⁎ .01 .26⁎⁎⁎ .13⁎ .30⁎⁎⁎ .51⁎⁎⁎
– .61⁎⁎⁎ .40⁎⁎⁎ .70⁎⁎⁎ .60⁎⁎⁎ .73⁎⁎⁎ .77⁎⁎⁎
– .06 .41⁎⁎⁎ .10 .38⁎⁎⁎ .40⁎⁎⁎
– .15⁎⁎ .37⁎⁎⁎ .16⁎⁎ .03
– .23⁎⁎⁎ .30⁎⁎⁎ .51⁎⁎⁎
– .37⁎⁎⁎ .31⁎⁎⁎
– .58⁎⁎⁎
–
Note. BDI-II = Beck Depression Inventory-II; RRS = Ruminative Response Scale; LEIDS-R Total = Leiden Index of Depression Sensitivity Total; Hopeless = LEIDS-R Hopelessness/Suicidality subscale; Accept = LEIDS-R Acceptance/Coping subscale; Aggress = LEIDS-R Aggression subscale; Control = LEIDS-R Control/Perfectionism subscale; Risk = LEIDS-R Risk Aversion subscale; Rum = LEIDS-R Rumination subscale. ⁎p b .05, ⁎⁎p b .01, ⁎⁎⁎p b .001.
Aggression, t(96.15) = −.33, p N .05, unequal variance estimate; Control, t(116) = − 1.45, p N .05; Risk Aversion, t(115) = −1.75, p N .05. An unexpected observation was that the recovered depressed participants scored more highly than their never-depressed counterparts on the Acceptance/Coping subscale, although this difference did not reach statistical significance, t(116) = −1.48, p N .05. To investigate whether cognitive reactivity, as indexed by the LEIDS-R Total score, explained any of the variance in depression beyond depressive rumination (a well-established predictor of depression), we conducted a multiple regression with BDIII scores as the dependent variable, RRS scores entered on the first step, and LEIDS-R total scores entered on the second step. Across the recovered Table 3 Mean participant characteristics for recovered depressed and never-depressed groups Never Recovered depressed (n = 78) depressed (n = 40) Age BDI-II RRS LEIDS-R Total LEIDS-R Hopelessness LEIDS-R Acceptance LEIDS-R Aggression LEIDS-R Control LEIDS-R Risk Aversion LEIDS-R Rumination
19.05 (3.28) 8.44 (5.29) 40.32 (9.25) 42.77 (17.99) 4.47 (4.02) 2.88 (2.85) 8.23 (4.85) 6.94 (4.66) 9.78 (4.07) 10.46 (5.02)
23.33 (10.09) 8.93 (7.27) 44.45 (12.47) 49.18 (13.84) 5.05 (3.80) 3.73 (3.07) 8.53 (3.84) 8.20 (4.10) 11.15 (3.82) 12.60 (4.44)
Note. Standard deviations appear in parentheses. Means are unadjusted. BDI-II = Beck Depression Inventory-II; RRS = Ruminative Response Scale; LEIDS-R Total = Leiden Index of Depression Sensitivity Total; LEIDS-R Hopelessness = Hopelessness/Suicidality subscale; LEIDS-R Acceptance = Acceptance/Coping subscale; LEIDS-R Aggression = Aggression subscale; LEIDS-R Control = Control/Perfectionism subscale; LEIDS-R Risk Aversion = Risk Aversion subscale; LEIDS-R Rumination = Rumination subscale.
depressed and never-depressed participants, RRS scores accounted for 33% of the variance in BDI-II scores (adjusted R2 = .33, β = .57, SE = .03, t = 12.18, p b .001). Interestingly, LEIDS-R Total made a small (but significant) contribution, accounting for 6% of the variance in BDI-II (adjusted R2 = .39, β = .31, SE = .02, t = 5.69, p b .001). Thus, cognitive reactivity made a unique contribution to the prediction of depression, over and above trait level of depressive rumination.
Discussion Our first goal in this study was to examine the interrelationship of cognitive reactivity and rumination. As predicted, these cognitive processes were positively correlated even when current level of depression was partialled out—confirming an association that is independent of current symptomatology. Consistent with clinical observation, this relationship demonstrates that individuals who respond to depressed mood by ruminating about the causes, meanings, and implications of their symptoms (Nolen-Hoeksema, 1991) are also likely to engage in an array of additional maladaptive cognitive and behavioral responses, including feeling hopeless and suicidal, responding with aggression, and attempting to avoid harm/conflicts. A second goal was to compare cognitive reactivity, as indexed by the LEIDS-R, in recovered and never-depressed individuals. As predicted, the recovered depressed participants indicated greater cognitive reactivity generally (i.e., LEIDS-R Total), and scored more highly on the LEIDS-R Rumination subscale and the RRS. However, the two groups did not differ on the remaining LEIDS-R subscales. Thus although the LEIDS-R Total differentiated how formerly depressed and neverdepressed individuals respond when they experience sad mood, this effect was accounted for mainly by the LEIDS-R Rumination subscale. This finding,
70
moulds et al.
combined with the observation that formerly and never-depressed individuals differed on the basis of RRS scores, accords with extensive empirical support for the role of depressive rumination in the persistence of depression (Nolen-Hoeksema, 2000) and associated cognitive and behavioral deficits (e.g., impoverished problem solving, Watkins & Moulds, 2005; overgeneral autobiographical memory retrieval, Watkins, Teasdale, & Williams, 2000). At this stage, the degree to which the Hopelessness, Acceptance, Aggression, Control, and Risk Aversion subscales are able to differentiate remitted and never-depressed individuals remains unclear. Replication of this investigation is needed for clarification. Importantly, across the sample of recovered and never-depressed participants, the Total LEIDS-R score explained unique variance in depression over and above that accounted for by the RRS. The RRS is the gold standard instrument used to measure depressive rumination that has convincingly predicted depression across multiple longitudinal studies. Our finding therefore underscores the utility of the LEIDS-R as a clinical instrument that gauges responses to low mood that are predictive of concurrent depression. Future studies should investigate whether the LEIDS-R has the capacity to predict depression across time within a longitudinal design. The construct of cognitive reactivity incorporates both depression-related cognitive processes and content that are theorized to be activated in response to sad mood. However, beyond studies that index cognitive content—i.e., maladaptive cognitions— following induced negative mood, experimental work in this area to date has not tested the differential activation hypothesis more broadly and indexed a range of potential dysfunctional processes (cognitive and behavioral) that characterize depressed individuals’ responses to low mood. By measuring a range of cognitive and behavioral processes that an individual may engage in at times of low mood, the LEIDS-R has the capacity to test the differential activation hypothesis more broadly. Moreover, as a clinical tool, responses on the LEIDS-R have the capacity to provide invaluable information about a range of maladaptive responses that clients may engage in, including both dysfunctional cognitive content and processes. These can then serve as key targets in therapy. Related to this, future work would benefit from extending investigations in this area to more broadly incorporate additional well-established cognitive processes that operate in depression; for example, memory disturbances such as overgeneral memory, intrusive memories and biased retrieval of negative material. The possibility that such processes of memory are
activated in response to (and exacerbated by) sad mood, and whether their activation predicts depression relapse, is worthy of examination. A number of limitations warrant acknowledgment. First, we did not employ a treatment-seeking sample. Participants were an unselected sample of undergraduate students. Nonetheless, participants in the second phase of the study were administered a structured clinical interview by a qualified clinical psychologist. Thus, although the sample was not drawn from a clinical population per se, it is noteworthy that their responses on an established structured interview (conducted by an experienced clinician) were used to ascertain their depression diagnostic history. However, future research that involves the recruitment of formerly depressed participants (e.g., posttreatment) would support the generalizability of our findings to treatmentseeking individuals in the wider community. Second, we note that the scores obtained on the selfreport measures for the never-depressed group were relatively high for a healthy participant group. As depression was the focus of our study, we administered only the major depression component of the SCID. Thus, we cannot rule out the possibility that participants also satisfied diagnostic criteria for additional psychological conditions (e.g., anxiety disorders, eating disorders), and thus that the scores observed on self-report measures were an artifact of comorbidity that was not indexed in the current design. Similarly, we did not assess current or additional depressive conditions such as dysthymia and minor or subclinical depression; again, this may have inflated scores in the never-depressed group. Future studies that incorporate a more thorough screening to rule out the influence of comorbidity and include extensive background details (e.g., history of psychological treatment) will provide an important extension of our findings. Our sample included more female than male participants, and future studies should employ gender-balanced samples. An additional aspect of our sample that should be kept in mind when interpreting the results is the age of participants; the never-depressed group was significantly younger than the recovered depressed group. The possibility that some of the never-depressed participants may, in fact, be vulnerable to depression, but had not yet experienced a depressive episode at the time of involvement in the research, cannot be ruled out. Despite potentially possessing features that confer vulnerability to depression, it is plausible that individuals in this younger group had not experienced a sufficiently aversive life event to precipitate a depressive episode. Thus, in comparing the neverdepressed and remitted depressed group, we must
cognitive reactivity and rumination limit our interpretations so as not to draw conclusions about depression-vulnerable versus nondepression-vulnerable individuals. Related to this point, an outstanding question in the field is whether cognitive reactivity encompasses a range of maladaptive cognitive processes that are sequelae of experiencing depression, or whether the presence of these processes confers vulnerability to the onset of an initial depressive episode. As raised by Van der Does (2002a), future studies that investigate cognitive processes in high-risk samples using prospective designs (e.g., Bryant & Guthrie, 2005) are needed in order to disentangle whether cognitive reactivity is a vulnerability factor that increases the likelihood of initial depression onset, as well as recurrence. The development of easily administered self-report instruments that index complex constructs such as cognitive reactivity increases the likelihood that clinical researchers will tackle such research questions in the future. References American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the Beck Depression Inventory–II. San Antonio, TX: Psychological Corporation. Bryant, R. A., & Guthrie, R. M. (2005). Maladaptive appraisals as a risk factor for posttraumatic stress: A study of trainee firefighters. Psychological Science, 16, 749–752. Just, N., & Alloy, L. B. (1997). The response styles theory of depression: Tests and an extension of the theory. Journal of Abnormal Psychology, 106, 221–229. Lavender, A., & Watkins, E. (2004). Rumination and future thinking in depression. British Journal of Clinical Psychology, 43, 129–142. Lyubomirsky, S., & Nolen-Hoeksema, S. (1995). Effects of selffocused rumination on negative thinking and interpersonal problem-solving. Journal of Personality and Social Psychology, 69, 176–190. Merens, W., Booij, L., Markus, R., Zitman, F. G., Onkenhout, W., & Van der Does, A. J. W. (2005). The effects of a diet enriched with α-lactalbumin on mood and cortisol response in unmedicated recovered depressed subjects and controls. British Journal of Nutrition, 94, 415–422. Miranda, J., & Persons, J. B. (1988). Dysfunctional attitudes are mood-state dependent. Journal of Abnormal Psychology, 97, 76–79. Nolen-Hoeksema, S. (1991). Responses to depression and their effects on the duration of depressive episodes. Journal of Abnormal Psychology, 100, 569–582. Nolen-Hoeksema, S. (2000). The role of rumination in depressive disorders and mixed anxiety/depressive symptoms. Journal of Abnormal Psychology, 109, 504–511. Nolen-Hoeksema, S., & Morrow, J. (1989). A prospective
71
study of depression and posttraumatic stress symptoms after a natural disaster: The 1989 Loma Prieta earthquake. Journal of Personality and Social Psychology, 61, 115–121. Nolen-Hoeksema, S., & Morrow, J. (1993). Effects of rumination and distraction on naturally occurring depressed mood. Cognition and Emotion, 7, 561–570. Nolen-Hoeksema, S., Morrow, J., & Fredrickson, B. L. (1993). Response styles and the duration of episodes of depressed mood. Journal of Abnormal Psychology, 102, 20–28. Segal, Z. V., Gemar, M., & Williams, S. (1999). Differential cognitive response to a mood challenge following successful cognitive therapy or pharmacotherapy for unipolar depression. Journal of Abnormal Psychology, 108, 3–10. Segal, Z. V., Williams, J. M. G., & Teasdale, J. D. (2002). Mindfulness-based cognitive therapy for depression. New York: The Guilford Press. Segal, Z. V., Williams, J. M., Teasdale, J. D., & Gemar, M. (1996). A cognitive science perspective on kindling and episode sensitization in recurrent affective disorder. Psychological Medicine, 26, 371–380. Spitzer, R. L., Williams, J. B. W., Gibbon, M., & First, M. B. (1996). Structured Clinical Interview for DSM-IV (SCID). Washington DC: American Psychiatric Press. Teasdale, J. D. (1988). Cognitive vulnerability to persistent depression. Cognition and Emotion, 2, 247–274. Van der Does, W. (2002). Cognitive reactivity to sad mood: Structure and validity of a new measure. Behaviour Research and Therapy, 40, 105–120. Van der Does, W. (2002). Different types of experimentally induced sad mood? Behavior Therapy, 33, 551–561. Van der Does, W. (2005). Thought suppression and cognitive vulnerability to depression. British Journal of Clinical Psychology, 44, 1–14. Van der Does, A. J. W., & Williams, J. M. G. (2003). Leiden Index of Depression Sensitivity–Revised (LEIDS-R). Retrieved September 4, 2007, from http://www.dousa.nl/publications_ depression.htm#LEIDS. Watkins, E., & Moulds, M. (2005). Distinct modes of ruminative self-focus: Impact of abstract versus concrete rumination on problem solving in depression. Emotion, 5, 319–328. Watkins, E., & Teasdale, J. D. (2004). Adaptive and maladaptive self-focus in depression. Journal of Affective Disorders, 82, 1–8. Watkins, E., Teasdale, J. D., & Williams, R. M. (2000). Decentering and distraction reduce overgeneral autobiographical memory in depression. Psychological Medicine, 30, 911–920. Weissman, A., Beck, A. T. (1978 November). The Dysfunctional Attitudes Scale. Paper presented at the annual meeting of the Association for Advancement of Behavior Therapy, Chicago, IL. Williams, J. M. G., VanderDoes, A. J. W., Barnhofer, T., Crane, C., & Segal, Z. V. (in press). Cognitive reactivity, suicidal ideation and future fluency: Investigating a differential activation theory of suicidality. Cognitive Therapy and Research.
R E C E I V E D : July 2, 2006 A C C E P T E D : May 1, 2007 Available online 26 October 2007