Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2023 Mar 12.
Published in final edited form as: Behav Res Ther. 2007 May 10;45(12):3069–3076. doi: 10.1016/j.brat.2007.05.001

Revealing negative thinking in recovered major depression: a preliminary investigation

ER Watkins a,*, M Moulds b
PMCID: PMC7614305  EMSID: EMS171049  PMID: 17572380

Abstract

Previous research suggests that formerly dysphoric individuals engage in effortful strategies (e.g., thought suppression) that may mask underlying depressive thinking. The addition of a cognitive load, such as recalling a 6-digit number, which interferes with effortful mental control, reveals depressive thinking in formerly dysphoric individuals. This preliminary study tested whether this effect of cognitive load on revealing negative thinking generalizes to formerly clinically depressed patients. Currently depressed patients, recovered depressed patients and never depressed patients unscrambled sentences that could form either positive or negative statements, after random allocation to either cognitive load or no cognitive load conditions. The number of negative statements unscrambled was used as an index of negative thinking. Without a load, recovered depressed patients did not differ from never-depressed controls: both groups completed fewer negative statements than currently depressed patients. However, the cognitive load increased negative statements in the recovered depressed group, making them resemble the currently depressed group more than the never depressed group. These preliminary findings extend previous demonstrations of cognitive load unmasking negative thinking in dysphoric students to a clinical population, suggesting that formerly depressed patients utilize effortful strategies to minimize the report of negative thinking, which is undermined by the addition of a cognitive load.

Keywords: depression, negative thinking, mental control, thought suppression, effortful strategies

1. Introduction

Recent research into depression has focused on trying to find reliable cognitive vulnerabilities in the form of persistent negative thinking biases that can help explain the onset and recurrence of depression. Such vulnerabilities are predicted by cognitive accounts of depression, which maintain that stable, underlying depressogenic beliefs and schemas underpin risk for depression (Clark & Beck, 1999). Given the very high rates of relapse and recurrence in formerly depressed people (Judd, 1997) one would expect cognitive vulnerabilities to be found in recovered depressed patients.

However, such cognitive vulnerabilities are not typically found in recovered depressed patients on self-report measures in the absence of depressed mood (Beevers, 2005; Ingram et al. 1998; Mathews & MacLeod, 2005; Scher, et al. 2005). In contrast, negative thinking in recovered depressed patients is reported following negative mood inductions, consistent with a view that underlying latent cognitive biases can be activated by shifts in mood state (Ingram et al. 1998; Scher et al. 2005).

One account for the frequent failure to find depressive biases in at-risk individuals, such as the formerly depressed, is that these individuals may use effortful mental strategies that reduce the report of unwanted negative thinking. These effortful control strategies could include deliberate suppression of negative thinking, i.e., trying to prevent negative thoughts coming to mind (see Wegner, 1994 for review). Alternatively, these effortful strategies could involve deliberate attempts to minimize the elaboration and expression of negative thoughts. For example, in response to an initial negative thought, an individual may deliberately seek out more positive interpretations and report these secondary appraisals rather than the initial negative response (Sheppard & Teasdale, 2000). Thus effortful mental control in at-risk individuals could hide a genuine depressive bias. However, the introduction of a cognitive load (e.g. concurrently remembering a 6-digit number) interferes with effortful attempts at mental control. For example, in experimental studies, the addition of a cognitive load interferes with attempts to suppress an unwanted thought, resulting in a greater frequency of the unwanted thought than if suppression had not been attempted (see Wenzlaff & Wegner, 2000 for review). Thus, if the use of effortful mental strategies is concealing an underlying depressive bias in at-risk groups, the addition of a cognitive load should reveal this underlying bias. The hypothesis that the use of effortful strategies explains the failure to observe depressive biases in symptom-free, formerly depressed individuals, thus predicts that (1) Under normal circumstances, when in a euthymic mood, formerly depressed individuals would not differ from never-depressed controls on measures of depressive thinking; (2) When a cognitive load is introduced to disrupt effortful mental control, formerly depressed individuals would display greater depressive bias than never-depressed controls.

These predictions have been confirmed in students who have recovered from dysphoria across a range of different measures of depressive thinking (Wenzlaff & Bates, 1998; Wenzlaff & Eisenberg, 2001; Wenzlaff et al. 2002; Wenzlaff et al. 2001). For example, Wenzlaff and Bates (1998) gave participants sets of scrambled sentences that could be arranged to form positive or negative statements. In the no-load condition, there was no difference between the at-risk and never-depressed groups, but in a load condition, the at-risk group produced significantly more negative completions than the never-depressed controls. These studies defined the at-risk group as having had moderate levels of depression (Beck Depression Inventory short-form > 7), which significantly reduced (BDI-SF < 4) over 4-6 weeks.

Because the at-risk individuals also demonstrated elevated levels of chronic thought suppression, Wenzlaff and Bates (1998, p.1559) interpreted these results as evidence that chronic “thought suppression masks a cognitive vulnerability to depression that can be revealed when mental control is disabled”. However, these studies did not include any direct assessment of attempts at thought suppression during the tasks, nor did they examine the consequences on negative thinking of attempted thought suppression with or without a cognitive load. Without these, it is not possible to be certain that the observed effects were due to the cognitive load interfering with effective thought suppression, as hypothesized by Wenzlaff and Bates (1998): changes in secondary appraisals or the ability to control the choice of completion reported during the scrambled sentence task in response to a cognitive load could also account for the observed findings. Nonetheless, these studies demonstrated that the addition of cognitive load altered the tendency of a group at increased risk for dysphoria to report negative thinking, consistent with the effortful strategy hypothesis. However, it is important to note that the groups were based on measures of dysphoria, rather than diagnostic criteria for major depression, leading Wenzlaff and colleagues to note that any clinical generalizations remain tentative until research is conducted within a clinical sample.

Within clinical samples, Hedlund and Rude (1995) also used the scrambled sentence task, finding that currently depressed and recovered depressed individuals produced more negative completions than never-depressed individuals following a self-focus manipulation. Rude, Covich, Jarrold, Hedlund and Zentner (2001) found that recovered depressed individuals show more negative thinking on this task than controls, but failed to determine the role of self-focus, as an attempted manipulation of self versus external focus was not successful. These studies differed from the Wenzlaff studies in a) including the self-focus manipulation; b)not explicitly assessing thought suppression; c) not replicating the cognitive load methodology used by Wenzlaff and Bates (1998). These methodological differences makes it hard to interpret the implications of these studies for the hypothesis that effortful mental strategies conceal underlying depressive thinking in recovered depressed individuals. In particular, the self-focus manipulation may act to exacerbate negative thinking by directing attention towards thoughts with negative meaning. Ruminative self-focus has been shown to exacerbate negative thinking (Nolen-Hoeksema, 1991; Rimes & Watkins, 2005). Furthermore, ruminative self-focus has been shown to impair subsequent controlled processing (Hertel, 1998; Watkins & Brown, 2002), and thus may interfere with effortful mental strategies.

Therefore, the aim of the current study was to directly test whether the effects of the addition of a cognitive load on negative thinking found by Wenzlaff and Bates (1998) in dysphoric students could be replicated in a clinical sample similar to that examined by Hedlund and Rude (1995). Wenzlaff & Wegner (2000) predict that the results found in at-risk students will generalize to recovered depressed patients, such that cognitive biases are hidden until a cognitive load is added. To test these hypotheses, we explicitly used Wenzlaff and Bates (1998) methodology in patients meeting diagnostic criteria for current and past major depression. In this way, we were able for the first time to test whether a cognitive load influenced the report of negative thinking in a clinical sample.

We predicted that in the absence of a cognitive load, recovered depressed individuals would not differ from never-depressed controls on measures of depressive thinking, but that when a cognitive load is introduced to disrupt effortful mental strategies, recovered depressed individuals would display greater depressive bias than never-depressed controls. We further predicted that currently depressed patients would show elevated negative thinking as their depressive affect will directly interfere with effortful mental strategies and increase the accessibility of negative thoughts.

2. Method

2.1. Participants

An experienced clinical psychologist assessed all participants using the Structured Clinical Interview for DSM-IV (SCID; Spitzer, et al. 1996) for Axis I disorders to determine current and past diagnoses. Inter-rater reliability was high: a second independent interviewer, experienced in administering the SCID and blind to the other assessment, had 100% agreement for diagnostic status of mood disorders on 5 clinical participants, selected by chance.

2.1.1. Currently Depressed Group

Twenty participants (10 males, 10 females) meeting DSM-IV criteria for current major depressive disorder were recruited from an out-patient mood disorders clinic, a community clinical psychology service, and a national depression self-help group. Participants had recurrent depression (current episode, M = 14.2 months, SD = 20.4; number of episodes, M = 8.1, SD = 7.0). Co-morbid diagnoses were generalised anxiety disorder (n = 6), post-traumatic stress disorder (n = 2), social phobia (n = 3), panic disorder with agoraphobia (n = 4), obsessive-compulsive disorder and bulimia eating disorder (n =1 each).

2.1.2. Recovered Depressed Group

Twenty participants (9 males, 11 females) meeting criteria for past but not current major depressive disorder, and with Beck Depression Inventory scores < 9 were recruited as above. Participants had recurrent depression (number of episodes, M = 3.7, SD = 3.5). Co-morbid diagnoses were social phobia (n = 3) and agoraphobia (n = 1).

2.1.3. Never Depressed Controls

Twenty participants (8 males, 12 females) who had never met criteria for major depression (or any psychiatric disorder) and with BDI < 9 were recruited by press advertisements.

2.2. Materials

2.2.1. Beck Depression Inventory (BDI)

The BDI (Beck & Beamesderfer, 1974) is a 21-item self-report measure of symptoms of depression, with good psychometric properties (Beck, et al. 1988).

2.2.2. White Bear Suppression Inventory (WBSI; Wegner & Zanakos, 1994)

The 15-item WBSI asks participants to endorse their tendency to use thought suppression (e.g. “I always try to put problems out of my mind”) using a 5 point scale (strongly agree-strongly disagree). The WBSI has good reliability and validity (Wegner & Zanakos, 1994).

2.2.3. Scrambled Sentence Task (SST)

Two parallel forms of the SST created by Wenzlaff and Bates (1998) were used, counterbalanced across load and no load conditions. Each form consisted of 20 sets of six scrambled words (e.g. “bright the very dismal looks future”) that could be unscrambled using only five words to form either a negative thought (“the future looks very dismal”) or a positive thought (“the future looks very bright”). All participants were instructed to “Unscramble each sentence to form whatever statement comes to mind first”, i.e., the instruction left it open to each participant to choose the valence of each sentence completion (positive or negative). This instruction was that previously used by Wenzlaff and Bates (1998) when they found a difference in negative completions on the SST between load and no-load conditions in dysphoric groups. The SST is scored by counting the number of positive and negative completions consistent with instructions (i.e. grammatical, only using 5 words). With few response options, scoring is reliable: there was 100% agreement between two independent blind raters on 10% of forms, selected randomly.

2.3. Procedure

We followed Wenzlaff and Bates (1998) paradigm where group (currently depressed, recovered depressed, never depressed) and condition (load or no load) during the SST were between-subjects factors. Participants were tested individually after they gave written informed consent. The SCID was completed before the experimental procedure, either during the same session or on a different day (in the case of the majority of depressed participants). Each participant worked through a SST form.

Within each diagnostic group, participants were randomly assigned to the load and no load conditions, such that half of the participants were asked to retain a six digit number in memory (the cognitive load condition) immediately prior to presentation of the scrambled sentences, and the other half of participants were given no additional instructions or tasks (no load condition). Following the load condition, participants attempted to recall the six-digit number. Following Wenzlaff and Bates (1998), participants were only allowed 4 minutes to complete each set of 20 sentences, and were told that it was important to work as quickly as possible.

Following the design used by Wenzlaff and Bates (1998), on completion of the scrambled sentences, participants rated the frequency (“In the past month how often have you tried to suppress unwanted negative thoughts?” 1 = never to 7 = all the time) and success of thought suppression (“In the past month how successful have you been in suppressing unwanted negative thoughts?” 1 = not at all to 7 = very). Participants completed the BDI and WBSI at the end of the same session.

3. Results

An alpha level of .05 was used for all statistical tests. Analyses of variance (ANOVAs) with group and load condition as independent between-subject variables were calculated to test the principal hypotheses. Scheffé tests were used for the post hoc comparisons. Table 1 displays means and standard deviations for age and the questionnaire measures.

Table 1. Means (SD in parentheses) for symptoms, suppression ratings, and age as a function of depression group and load condition.

Currently depressed Recovered Depressed Never Depressed
Load
(n = 9)		 No Load
(n = 11)		 Load
(n = 11)		 No Load
(n = 9)		 Load
(n = 10)		 No Load
(n = 10)
Age 43.56 (15.24) a 42.36 (12.40) a 33.18 (13.48) a 38.78 (14.80) a 29.50 (10.78) a 38.70 (16.13) a
BDI 26.56 (11.62) a 29.82 (10.68) a 5.91 (3.02) b 2.44 (1.81)b 2.30 (2.11) b 2.20 (2.04) b
WBSI 55.78 (7.79) a 56.55 (7.22) a 47.55 (14.56) b 39.00 (11.84) c 29.30 (8.63) c 31.20 (11.25) c
Frequency suppression 4.89 (1.76) a 5.09 (1.92) a 3.91 (1.58) a,b 3.22 (1.20) b 3.00 (1.41) b 3.50 (1.58) b
Success Suppression 2.67 (1.22) a 1.91 (1.14) a 3.90 (1.45) a,c 4.39 (2.23) b,c 5.67 (0.87) b 5.22 (1.39) b
Correct recall of number 44% n/a 82% n/a 80% n/a
Open in a new tab

Note. Frequency suppression is the frequency of thought suppression in the last month ranging from 1 (lowest) to 7 (highest). Success suppression is the rating from 1 (not at all) to 7 (very successful) of how well participant can suppress negative thoughts. BDI is the Beck Depression Inventory. WBSI is the White Bear Suppression Inventory. Means in the same row that do not share subscripts differ at p < .05.

3.1. Background Variables

We calculated separate Group by Cognitive load ANOVAs for age, BDI and WBSI score as the dependent variable. There was a significant effect of Group for BDI score, F (2, 54) = 92.58, p < .01. Post-hoc Scheffé tests revealed that the currently depressed group was more depressed than the recovered depressed group (p < .01) and the never depressed group (p < .01), whereas the recovered depressed group was not significantly more depressed than the never-depressed group (p = .61) (see Table 1 for means and S.D.). There was a significant effect of Group for WBSI scores, F (2, 54) = 29.78, p < .01, which reflected the currently depressed group having significantly elevated scores on the WBSI than the recovered depressed group (p < .01), who in turn had elevated scores compared to the never-depressed group (p < .01). The groups did not significantly differ in age (p = .11). There were no significant main effects or interactions of load condition on age (all F’s < 2.26), WBSI scores (all F’s < 1.5) or BDI scores (all F’s < 1).

Table 1 also reports means and standard deviations for ratings of suppression. We analyzed participants’ ratings of suppression frequency and success using separate ANOVAs with group as a between-subjects factor. There was a reliable main effect for group for both frequency, F (2, 57) = 6.56, p < .005, and success of suppression, F (2, 54) = 25.6, p < .001. The currently depressed participants reported engaging in more thought suppression during the last month than both never depressed (p < .005), and recovered depressed participants (p = .051), whilst there was no difference between the recovered depressed and never depressed groups (p = .61). Currently depressed participants reported less successful thought suppression than both never-depressed (p < .001) and recovered depressed participants (p < .005), whilst the recovered depressed group were less successful at thought suppression than the never depressed group (p < .01).

3.2. Unscrambled sentences

We used the percentage of negative statements as the main dependent variable as this was the index of depressive thinking used by Wenzlaff and Bates (1998)1. An analysis of the total number of unscrambled statements using a 3 X 2 ANOVA with Group and cognitive load (load, no load) as between-subjects factors found a significant main effect of Group, F (2, 54) = 7.35, p < .005, but no main effect of Load, F < 1, or interaction of Group by Load, F (2, 54) = 1.54, p = .22. The main effect of Group reflected a lower production of statements overall in the currently depressed group (M = 14.55, SD = 3.71) relative to the recovered depressed (M = 18.25, SD = 2.63) and never-depressed groups (M = 17.40, SD = 3.27). One advantage of using percentage scores, calculated by dividing the number of negative statements by total number of statements, as the index of depressive thinking, is that this approach controlled for the lower production of statements overall in the currently depressed group. Table 2 displays the number of completed sentences and the number and percentage of negative completions for each group and condition.

Table 2. Mean number and mean percentage of negative statements and mean number of total statements (SD in parentheses) unscrambled by condition and group.

Currently depressed Recovered Depressed Never Depressed
Load
(n = 9)		 No Load
(n = 11)		 Load
(n = 11)		 No Load
(n = 9)		 Load
(n = 10)		 No Load
(n = 10)
Number of negative statements 5.00 (4.09) 8.36 (4.32) 3.73 (3.26) 1.56 (2.65) 0.90 (1.10) 0.30 (0.48)
% of negative statements 33.87 (25.38) 57.82 (21.05) 21.80 (17.41) 8.76 (15.46) 4.77 (5.57) 1.75(2.85)
Total statements completed 14.56 (2.96) 14.54 (4.37) 17.64 (3.23) 19.00 (1.50) 18.50 (2.72) 16.30 (3.52)
Open in a new tab

Note: The maximum number of statements that can be unscrambled is 20.

We analyzed the percentage of negative sentences unscrambled using a 3 X 2 ANOVA with Group (currently depressed, recovered depressed, never depressed) and cognitive load (load, no load) as between-subjects factors. There was a main effect of Group, F (2, 54) = 34.74, p < .001, secondary to the predicted significant interaction of Group by Load, F (2, 54) = 6.58, p < .005. There was no main effect of Load, F (2, 54) < 1. (Similar significant patterns of results were found when we repeated the analysis with the number of negative sentences completed as the dependent variable).

To explore the Group by Load interaction, we analyzed the effect of group on percentage of negative completions for load versus no load conditions separately. There was a significant effect of group for both the no load condition, F (2, 27) = 41.79, p <.001, and the load condition, F (2, 27) = 6.51, p < .005. For both conditions, there were a greater percentage of negative completions in the currently depressed group than the never depressed group (p < .005). The difference between conditions centered on the recovered depressed group: In the no load condition, the recovered depressed group had a lower percentage of negative completions than the currently depressed group (p < .001) and did not differ from the never depressed group (p = .609). However, in the load condition, the recovered depressed group did not differ from the currently depressed group (p = .332) and showed a trend towards a greater percentage of negative completions than the never depressed group (p = .054 one-tailed, using a conservative Scheffé test). That is, whereas recovered depressed participants in the no load condition behaved like the never depressed group, recovered depressed participants in the load condition behaved like the currently depressed group2.

4. Discussion

To our knowledge, this study is the first attempt to explicitly replicate Wenzlaff and Bates (1998) cognitive load methodology in patients with major depression and patients recovered from major depression. The study was designed to determine whether effortful mental strategies might play a role in concealing cognitive biases in recovered depression (as predicted by Wenzlaff & Wegner, 2000), by testing whether a cognitive load would unmask negative thinking in recovered depressed patients.

The results are broadly consistent with the predictions of Wenzlaff and Wegner (2000) and extend the findings of Wenzlaff and Bates (1998) from dysphoric students to a clinical group. First, in the no load condition, recovered depressed patients did not significantly differ from never-depressed controls in percentage of negative completions, with both groups completing fewer negative statements than current depressed patients. This finding is consistent with our prediction that the recovered depressed group would not differ from the never-depressed group under normal circumstances. Second, as predicted, recovered depressed patients in the load condition did not significantly differ from the currently depressed group in percentage of negative completions, producing more negative completions than recovered depressed patients in the no load condition. Thus, whereas recovered depressed participants in the no load condition generated a similar number of negative completions to the never depressed group, recovered depressed participants in the load condition generated a similar number of negative completions to the currently depressed group. Third, as expected, the currently depressed group reported more negative sentence completions, consistent with predictions that currently depressed patients would have increased accessibility of negative thinking and impaired use of effortful mental strategies, presumably as a result of increased depressed affect.

When examining the reported extent of thought suppression, we found that the recovered depressed group did report more suppression on the WBSI and less success at thought suppression than the never depressed group. This finding is consistent with Wenzlaff and Wegner’s (2000) proposal that chronic thought suppression may be the effortful strategy concealing depressive biases in recovered depressed patients. However, as noted in the introduction, there is no direct evidence for a causal role of thought suppression in concealing negative thinking, nor that the addition of a cognitive load interferes with thought suppression in the recovered depressed patients. Thus, the results are equally consistent with other forms of effortful strategies preventing the report of negative thinking, such as an individual deliberately focusing on more positive secondary appraisals and/or not reporting the initial negative thoughts.

We further note the recovered depressed group reported similar frequency at thought suppression for the last month on the single item measure compared to the never depressed group, leaving open the question of whether the recovered depressed group uses more suppression than the never depressed group. We believe that the WBSI measure may be a more relevant measure in this study because (i) it is a better validated measure than the single item question; (ii) it assesses dispositional tendency towards suppression whilst the single item measure measures state-levels of suppression in the last month, which is appropriate for the Wenzlaff and Bates (1998) study where the “at risk” group reported elevated dysphoria within the last 4-6 weeks but less appropriate for our sample where on diagnostic criteria recovered depressed patients have been recovered for at least one month, and thus may have had little need for thought suppression.

These findings extend to recovered depressed patients the earlier findings in dysphoric individuals suggesting that effortful mental strategies mask depressive thoughts until cognitive demands undermine these strategies (Wenzlaff & Bates, 1998). These results suggest that the previous findings that recovered depressed individuals display more negative thinking on the scrambled sentence task than never depressed controls (Hedlund & Rude, 1995; Rude et al. 2001) may have resulted from the presence of the self-focused manipulation. Self-focus may act to increase the awareness of negative thoughts and/or interfere with effortful strategies.

The limitations of the study include first a relatively small sample size, such that any null findings may result from limited power, although our cell sizes are similar to Wenzlaff and Bates (1998). Second, we did not directly nor experimentally measure success at thought suppression, although this was a result of matching Wenzlaff and Bates (1998) methodology in order to see if their findings could be replicated in a clinical population. Third, although we used a between-subject design in order to replicate the methodology of Wenzlaff and Bates (1998), we note that a within-subject design comparing load versus no load in each participant would be useful.

In conclusion, the current study found that recovered depressed individuals did not differ in completion of negative sentences from never depressed controls in a no load condition. However, the recovered depressed completed more negative sentences when exposed to a cognitive load, with the percentage of negative completions not significantly differing from that found in currently depressed patients, consistent with Wenzlaff and Bates (1995) finding that cognitive load unmasked previously hidden negative thinking in previously dysphoric students. This pattern of findings is consistent with the hypothesis that the use of effortful mental strategies may limit the report of negative thinking in recovered depressed patients, such that the disruption of these strategies by a cognitive load reveals more negative thinking.

Acknowledgements

This research was supported by a Project Grant from the Wellcome Trust, UK to Dr Edward Watkins.

Footnotes

1

To check on adherence to the load condition, a chi-square test was calculated to test whether recall of the number differed between the groups; there was no significant difference across the 3 groups, χ2= 4.01, p =.135.

2

Interestingly, there was a significant main effect of load within the currently depressed group, F(1, 18) = 5.20, p = .035., reflecting fewer negative completions within the load condition than in the no load condition. We speculate that this effect is due to the cognitive load acting as a distracter to the spontaneously occurring negative thinking found in current depression, in much the same way that distraction reduces depressive rumination (Nolen-Hoeksema, 1991).

References

  1. Beck A, Beamesderfer A. In: Psychological measurements in psychopharmacology Modern problems in pharmacopsychiatry. Pichot P, editor. Vol. 7. Darger; Basel, Switzerland: 1974. Assessment of depression: the depression inventory; pp. 151–169. [DOI] [PubMed] [Google Scholar]
  2. Beck AT, Steer RA, Garbin MG. P sychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review. 1988;8:77–100. [Google Scholar]
  3. Beevers CG. Cognitive vulnerability to depression: A dual process model. Clinical Psychology Review. 2005;25:975–1002. doi: 10.1016/j.cpr.2005.03.003. [DOI] [PubMed] [Google Scholar]
  4. Clark DA, Beck AT. Scientific foundations of cognitive theory and therapy of depression. John Wiley & Sons; New York: 1999. [Google Scholar]
  5. Hedlund S, Rude SS. Evidence of latent depressive schemas in formerly depressed individuals. Journal of Abnormal Psychology. 1995;104:517–525. doi: 10.1037//0021-843x.104.3.517. [DOI] [PubMed] [Google Scholar]
  6. Hertel PT. Relation between rumination and impaired memory in dysphoric moods. Journal of Abnormal Psychology. 1998;107:166–172. doi: 10.1037//0021-843x.107.1.166. [DOI] [PubMed] [Google Scholar]
  7. Ingram RE, Miranda J, Segal ZV. Cognitive vulnerability to depression. Guilford Press; New York: 1998. [Google Scholar]
  8. Judd LL. The clinical course of unipolar major depressive disorders. Archives of General Psychiatry. 1997;54:989–991. doi: 10.1001/archpsyc.1997.01830230015002. [DOI] [PubMed] [Google Scholar]
  9. Mathews A, MacLeod C. Cognitive vulnerability to emotional disorders. Annual Review of Clinical Psychology. 2005;1:167–195. doi: 10.1146/annurev.clinpsy.1.102803.143916. [DOI] [PubMed] [Google Scholar]
  10. Nolen-Hoeksema S. Responses to Depression and their effects on the duration of depressive episodes. Journal of Abnormal Psychology. 1991;100:569–582. doi: 10.1037//0021-843x.100.4.569. [DOI] [PubMed] [Google Scholar]
  11. Rimes KA, Watkins ER. The effects of self-focused rumination on global negative self-judgements in depression. Behaviour Research and Therapy. 2005;43:1673–1681. doi: 10.1016/j.brat.2004.12.002. [DOI] [PubMed] [Google Scholar]
  12. Rude SS, Covich J, Jarrold W, Hedlund S, Zentner M. Detecting depressive schemata in vulnerable individuals: Questionnaires vs laboratory tasks. Cognitive Therapy and Research. 2001;25:103–116. [Google Scholar]
  13. Scher CD, Ingram RE, Segal ZV. Cognitive reactivity and vulnerability: Empirical evaluation of construct activation and cognitive diatheses in unipolar depression. Clinical Psychology Review. 2005;25:487–510. doi: 10.1016/j.cpr.2005.01.005. [DOI] [PubMed] [Google Scholar]
  14. Sheppard LC, Teasdale JD. Dysfunctional thinking in major depressive disorder: A deficit in metacognitive monitoring? Journal of Abnormal Psychology. 2000;109:768–776. doi: 10.1037//0021-843x.109.4.768. [DOI] [PubMed] [Google Scholar]
  15. Spitzer RL, Williams JBW, Gibbon M, First MB. SCID user’s guide for the Structured Clinical Interview for DSM-IV. American Psychiatric Press, Inc; Washington DC: 1996. [Google Scholar]
  16. Watkins ER, Brown RG. Rumination and executive function in depression: an experimental study. Journal of Neurology, Neurosurgery and Psychiatry. 2002;72:0-2. doi: 10.1136/jnnp.72.3.400. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Wegner DM. Ironic processes of mental control. Psychological Review. 1994;101:34–52. doi: 10.1037/0033-295x.101.1.34. [DOI] [PubMed] [Google Scholar]
  18. Wegner DM, Zanakos S. Chronic thought suppression. Journal of Personality. 1984;62:615–640. doi: 10.1111/j.1467-6494.1994.tb00311.x. [DOI] [PubMed] [Google Scholar]
  19. Wenzlaff RM, Bates DE. Unmasking a cognitive vulnerability to depression: how lapses in mental control reveal depressive thinking. Journal of Personality and Social Psychology. 1998;75:1559–1571. doi: 10.1037//0022-3514.75.6.1559. [DOI] [PubMed] [Google Scholar]
  20. Wenzlaff RM, Eisenberg AR. Mental control after dysphoria: Evidence of a suppressed, depressive bias. Behavior Therapy. 2001;32:27–45. [Google Scholar]
  21. Wenzlaff RM, Meier J, Salas DM. Thought suppression and memory biases during and after depressive moods. Cognition and Emotion. 2002;16:403–422. [Google Scholar]
  22. Wenzlaff RM, Rude SS, Taylor CJ, Stultz CH, Sweatt RA. Beneath the veil of thought suppression: Attentional bias and depression risk. Cognition and Emotion. 2001;15:435–452. [Google Scholar]
  23. Wenzlaff RM, Wegner DM. Thought suppression. Annual Review of Psychology. 2000;51:59–91. doi: 10.1146/annurev.psych.51.1.59. [DOI] [PubMed] [Google Scholar]

RESOURCES