خاطرات مزاحم رویدادهای منفی در افسردگی: آیا مرکزیت رویداد مهم است؟

Abstract Background and Objectives Following stressful events, there is evidence that the degree to which the event is perceived as central to one’s identity and forms a reference point for the attribution of meaning to other events (i.e., “centrality of events”) is associated with depression symptoms. However, these findings have primarily come from cross-sectional studies of undergraduate samples, and no past study has investigated whether centrality predicts depression over time. We sought to further examine the role of centrality of negative autobiographical events that were the content of intrusive memories in depression

1. Introduction Depressed individuals often report frequent, distressing intrusive memories of negative life events (Kuyken & Brewin, 1994) that are highly vivid, elicit significant levels of negative emotions, and interfere with daily activities (Birrer, Michael, & Munsch, 2007). Longitudinal studies demonstrate that these memories play an important role in the course of depression: their degree of intrusiveness and avoidance at baseline predicts depressive (Brewin, Reynolds, & Tata, 1999) and anxiety (Brewin, Watson, McCarthy, Hyman, & Dayson, 1998) symptoms six months later. Recent empirical investigations have provided insight into the phenomenological experience of the intrusive memories reported by depressed samples (Patel et al., 2007) and have found strikingly similar features to the intrusive recollections experienced in posttraumatic stress disorder (PTSD) (Reynolds & Brewin, 1999). Similarly, the way in which an individual appraises the experience of an intrusive memory (e.g., “This memory is a sign that I’m going crazy”) also appears to be important in both depression (Newby and Moulds, 2010a and Starr and Moulds, 2006) and PTSD (Ehlers, Mayou, & Bryant, 1998). These important extensions to the literature have improved our understanding of the features and the role of maladaptive appraisals of intrusive memories in depression. However, little research has investigated the degree to which depressed individuals conceptualise the event in their memory as central to their identity. Moreover, there have been no investigations of the degree to which the event (recalled in an intrusive memory) shapes beliefs and expectations about the world and the future, and whether these appraisals of the event play a maladaptive role in depression. Recently, Berntsen and Rubin (2006) argued that the “centrality of an event,” that is, the degree to which a highly stressful negative event becomes central to an individuals’ identity, and forms a reference point for attributing meaning to other events and expectations for future events, might be important in PTSD. In support of this model, the centrality of a stressful event is positively associated with levels of posttraumatic stress symptoms in both university student (Berntsen and Rubin, 2006 and Berntsen and Rubin, 2007) and combat veteran samples (Brown, Antonius, Kramer, Root, & Hirst, 2010). Recent studies have also shown that centrality is correlated with depressive symptoms (e.g., Berntsen and Rubin, 2007 and Boals, 2010), which suggests that, similar to PTSD, the centrality of remembered events might have interesting and important applications to our understanding of depression (see Boelen, 2009 for similar findings in a bereaved sample). There are a number of reasons why centrality might play an important role in depression. Leading cognitive behavioural models of depression emphasise the role of global stable negative beliefs about the self (e.g., Beck, Rush, Shaw, & Emery, 1979). When a depressed individual experiences a negative event (e.g., relationship breakup), if they then view the negative event as defining their identity (e.g., “I’m a failure”), and use the event to attribute meaning to other events and expectations for the future (e.g., “I will fail in relationships in the future”), it may create an unhelpful cycle that will in turn exacerbate depressive cognitions and low mood. As cited above, depressed individuals often recall distressing intrusive autobiographical memories of a range of events including negative interpersonal events, death and/or illness of loved ones, and personal assault and abuse (Patel et al., 2007). Interestingly, dominant models of autobiographical memory (e.g., the Self Memory System, Conway & Pleydell-Pearce, 2000) emphasise that autobiographical memory and identity are closely linked. These models argue that autobiographical memories that are highly accessible are often closely related to an individuals’ current goals (i.e., either concordant or highly discrepant from current goals). Applying this to depression, if, for example, a depressed person has a goal of being competent in relationships yet frequently recalls a memory of a breakup that they believe demonstrates that they are a failure in relationships, the recurrent experience of such a memory (i.e., intrusive recollections) might maintain the memory’s accessibility. When a depressed individual experiences a memory of this event (either deliberately retrieved or via intrusive repetitive recollections), the degree to which the event is central to identity is likely to exacerbate depressed mood by reinforcing its accessibility, as well as reinforcing negative self images and evaluations of the self, in turn perpetuating feelings of hopelessness about the future. Given the potentially interesting implications of the role of centrality of remembered events in depression, there are a number of unanswered questions that await investigation. First, no study has investigated whether there is an association between the centrality of intrusively recalled memories and depressive symptoms (the research cited above examined deliberately retrieved memories). Given that negative intrusive memories are a striking feature of depression and are involved in the maintenance of depression, investigations of variables that potentially exacerbate the impact of these memories are needed. Second, the relationship between centrality and depression has been limited to cross-sectional investigations with undergraduate samples, and their results are yet to be replicated in clinically depressed and community samples. Related to this, third, it is unclear whether depressed individuals view intrusively recalled negative events as more central to their identity than do non-depressed individuals. Therefore, it is unclear whether perceiving the event as central to identity is related to depression status, or is instead a universal appraisal of memory content that is unrelated to depression status (but related to other variables such as event severity and memory frequency). Fourth, no study has examined whether formerly depressed individuals (who have recovered from a past episode of depression) endorse similar ratings of centrality of remembered events as depressed individuals. Given the high rates of recurrence of depressive disorders, investigations into the cognitive processes that characterise recovered individuals have scope to improve our understanding of the factors that are potentially involved in relapse. Finally, no study has examined whether centrality of a remembered event predicts symptoms longitudinally. Another interesting pattern that has emerged from previous centrality of events research is the relationship between centrality, memory features and avoidance. In a study of undergraduate students, Boals (2010) found that higher centrality scores (related to an autobiographical memory of a negative event) were positively associated with higher levels of reliving, intensity of negative emotions, and the degree to which the memory was intrusive and avoided. Another study by Robinaugh and McNally (2010) showed that for participants who reported negative memories from a field (i.e., first person) perspective, levels of emotional intensity of the memories were associated with PTSD symptom levels. Interestingly, this association was mediated by the degree to which the event was central to identity. Notably, some of these features (e.g., reliving, visual perspective) and responses (e.g., avoidance) have been implicated in the persistence of PTSD (Michael, Ehlers, Halligan, & Clark, 2005) and depression (Brewin et al., 1999 and Newby and Moulds, submitted for publication). Therefore, together (consistent with arguments put forth by Berntsen & Rubin, 2006), the evidence suggests that event centrality might interact with important memory features and responses, and in turn potentially exacerbate the impact of recalling negative experiences. These important preliminary results await replication in a community sample. In addition, the interrelationships of centrality of events and other memory features (e.g., vividness), as well as specific cognitive behavioural avoidance responses (e.g., suppression and rumination) that have been shown to play an important role in depression (Starr & Moulds, 2006) have not been examined. Thus, these issues require further investigation. This study had the following aims. First, we aimed to compare clinically depressed, recovered depressed and never-depressed groups on their experiences of intrusive memories of negative events, and asked them to rate the degree of centrality of these events using the short version of the Centrality of Events Scale (CES-7, Berntsen & Rubin, 2006) (hereafter referred to as “centrality”). Second, we explored the associations between centrality and depression symptoms, memory characteristics, and avoidance strategies, and third, we aimed to investigate whether centrality predicted depression at six months, after taking into account baseline depression symptoms. This study extended the centrality literature in a number of ways. First, by including clinically depressed, formerly depressed and never-depressed control groups, we were able to explore whether: a) previous results were generalisable to clinical samples, and b) centrality is rated higher by depressed versus non-depressed groups. Second, by exploring the association between memory characteristics and centrality, we were able to a) confirm whether previous associations were generalisable to intrusively recalled events, and b) extend beyond previous literature and examine the relationship between centrality and a range of other characteristics, and responses to memories. Third, we were able to answer the question of whether ratings of the centrality of remembered events predicted depression over time. We expected that centrality would correlate positively with depressive symptoms, memory features (e.g., here and now quality/reliving, emotional intensity, intrusiveness) and avoidance. We also predicted higher ratings of centrality for depressed compared with non-depressed groups. Last, we expected centrality scores to predict depression symptoms at six months, controlling for initial symptom levels.

Results 3.1. Sample characteristics 3.1.1. Demographic characteristics 3.1.1.1. Time 1 The total sample was comprised of 56 female (59.6%) and 38 male (40.4%) participants with a mean age of 24.24 years (range = 18–54, SD = 5.91). There were no significant differences between the groups (i.e., depression status groups) in terms of age (F (2,82) = 2.21, p > 0.05) or gender (χ2(2, N = 85) = 1.85, p > 0.05). Table 1 presents demographic and sample characteristics for each group. The majority of participants reported a specific intrusive memory of a negative life event. Twenty two, 21 and 20 memories for the depressed, recovered and never-depressed groups, respectively, were included in the analyses. 3.1.1.2. Time 2 Independent samples t-tests compared the individuals who participated at Time 2 with those who were lost to attrition. The groups did not differ in age, gender, BAI or IES scores (ps > 0.05). However, the sub-sample of participants who were lost to attrition had lower BDI-II scores (t(1,83) = 2.01, p < 0.05). Twelve participants who were followed up at Time 2 had reported at Time 1 that they had not recently experienced an intrusive memory, leaving 52 who had reported an intrusive memory at baseline (n = 51 had completed the CES-7, and were included in regression analyses). Notably, none of the participants who participated at Time 2 met criteria for PTSD or ASD, or reported even sub-threshold symptoms of either of these disorders. 3.2. Depression status and centrality of events A one-way analysis of variance (ANOVA) demonstrated that the currently depressed, recovered and never-depressed groups did not differ on CES-7 (centrality) scores (F (2, 59) = 1.78, p > .05) (see Table 1 for means). 3.3. Gender and centrality scores Boals (2010) found evidence of a gender difference on centrality scores, such that females reported higher centrality of remembered events (on the CES-7). Before conducting further analyses, we investigated whether there were any gender differences on centrality ratings. No differences emerged (t (1,60) = 1.63, p > 0.05). 3.4. Relationships between centrality and mood, intrusiveness, and avoidance Pearsons r correlations were conducted to explore the relationship between centrality scores, mood measures (BAI and BDI-II), IES scores, intrusive memory features, and avoidance scores. In an attempt to control for inflation of the Type I error rate, we adopted a conservative alpha value of 0.01 (as opposed to 0.05) to test the significance of correlations. Unexpectedly, there was no association between centrality and BDI-II scores (r = 0.15, p > 0.05), nor BAI scores (r = 0.21, p = 0.088). There was, however, a significant association between centrality and IES Intrusion scores (r = 0.38, p < 0.001). With regard to avoidance, there was a trend towards a significant relationship with the RIQ rumination (r = 0.29, p = 0.014) but no association between CES and the RIQ suppression (r = 0.13, p > 0.05) subscale, nor IES Avoidance (r = 0.21, p = 0.083). 3.5. Relationship between centrality and intrusive memory features 3.5.1. Intrusive memory interview (IMI) variables For general memory characteristics on the IMI, there was no association between centrality and (independently coded) severity of the event (r = 0.22, p = 0.067). However, we found significant positive associations between centrality and memory vividness (on the IMI) (r = 0.33, p < 0.01), ratings of memories’ here and now quality/reliving (r = 0.31, p < 0.01), observer perspective (r = 0.33, p < 0.01), uncontrollability (r = 0.45, p < 0.001) and interference (r = 0.43, p < 0.001). There were no significant associations between centrality and intrusion-related distress (r = 0.16, p > 0.05), IMI vantage perspective (r = 0.20, p > 0.05), lack of context (r = 0.10, p > 0.05) and autonoetic awareness (r = −0.24, p = 0.045). 3.5.2. Memory experiences questionnaire (MEQ) variables For MEQ variables, there was a trend towards a significant association between centrality and scales of MEQ Emotional Intensity (r = 0.26, p = 0.03), and a significant association between centrality and MEQ Accessibility (r = 0.34, p = 0.004). There were no associations between centrality and MEQ Sensory Detail (r = 0.24, p = 0.045), MEQ vividness, (r = 0.23, p = 0.076), MEQ Visual Perspective (r = 0.20, p = 0.073), MEQ Coherence (r = 0.02, p > 0.05), and MEQ Time Perspective (r = 0.07, p > 0.05). 3.6. Partial correlations We sought to rule out the possibility that the associations between centrality and other memory variables could be accounted for by the severity of the event. Therefore, we conducted partial correlations between centrality and the variables mentioned above, controlling for event severity. Interestingly, the previously observed associations between centrality MEQ Emotional Intensity was not significant when severity was controlled. All other relationships remained significant. Given the significant association between centrality and IES Intrusion, we also wanted to rule out the possibility that the relationship between centrality and other variables were not simply driven by the frequency with which the memory intruded into mind. We conducted partial correlations between centrality and other variables controlling for the IES Intrusion subscale scores. After controlling for IES Intrusion, there was no longer an association between centrality and RIQ rumination subscale scores, here and now quality/reliving, nor MEQ Emotional Intensity. Next, we conducted additional partial correlations between centrality and memory variables, controlling for both IES Intrusion and severity scores. The associations between centrality and observer perspective, uncontrollability, and interference remained significant, as well as the MEQ Accessibility. All other relationships no longer remained significant. 3.7. The relationship between centrality at baseline and depression at six months Hierarchical multiple regression analyses (MRAs) were conducted to explore whether centrality scores at baseline predicted follow-up depression and anxiety. In the first MRA, follow-up BDI-II was the DV, baseline BDI-II scores were entered on the first step, followed by centrality scores on the second step. For the total sample, baseline BDI-II accounted for 41.6% of the variance in follow-up BDI-II (Adjusted R2 = 0.42; BDI-II: β = 0.47, SE = 0.08, t = 5.90, p < 0.001). After controlling for baseline depressive symptoms, centrality scores did not predict BDI-II at follow-up (p > 0.05). In a second MRA, follow-up BAI was entered as the DV, baseline BAI was entered on the first step, followed by centrality scores entering on the second step. Baseline BAI scores accounted for 35.1% of the variance in follow-up BAI (Adjusted R2 = 0.35; BAI: β = 0.50, SE = 0.10, t = 5.14, p < 0.001). After controlling for baseline BAI, centrality scores did not predict BAI at follow-up (p > 0.05). 2