ویژگی شرح حال حافظه و نشانه های سوگ و داغدیدگی پیچیده، افسردگی و اختلال استرس پس از سانحه بدنبال از دست دادن عزیزی

Abstract This study examined the specificity and content of autobiographical memories among bereaved individuals. Self-report measures of bereavement-related distress and a standard and trait version of the Autobiographical Memory Test (AMT) were administered to 109 bereaved people. We examined associations of memory specificity with (a) demographic and loss-related variables and with (b) symptom-levels of complicated grief (CG), depression, and posttraumatic stress disorder (PTSD), (c) associations of the content of memories (related vs. unrelated to the loss/lost person) with symptoms, and (d) the degree to which associations of symptom-levels with memory specificity and content differed between the standard and trait version of the AMT. Findings showed that (a) memory specificity varied as a function of age, education, and kinship; (b) reduced memory specificity was significantly associated with symptom-levels of CG, but not depression and PTSD; (c) symptom-levels of CG and PTSD were associated with a preferential retrieval of specific memories that were related to the loss/lost person on the standard AMT, whereas all three symptom-measures were associated with preferential retrieval of loss-related specific memories on the trait AMT; and (d) on the trait AMT, but not the standard AMT, symptom-measures remained significantly associated with a preferential retrieval of loss-related specific memories, when controlling for relevant background variables. Among other things, these results show that reduced memory specificity is associated with self-reported CG-severity but not depression and PTSD following loss. Moreover, the results are consistent with recent research findings showing that memories tied to the source of an individual’s distress (e.g., loss) are immune to avoidant processes involved in the standard reduced specificity effect.

Introduction Different forms of psychopathology are associated with a reduced ability to recall specific autobiographical memories. Reduced memory specificity is usually assessed with the Autobiographical Memory Test (AMT) developed by Williams and Broadbent (1986). In the AMT, people are instructed to retrieve a specific personal memory in response to positive and negative cue words. Reduced memory specificity (or “overgeneral memory”) is most clearly present in people with a trauma history and people diagnosed with depression or posttraumatic stress disorder (PTSD; Williams et al., 2007). Reduced memory specificity may be the result of at least three psychological processes (Williams et al., 2007). It may result from attempts to avoid distressing memories of traumatic experiences. This “affect regulation hypothesis” explains why overgeneral memory is observed among trauma-victims (Hauer et al., 2006, Henderson et al., 2002 and Kuyken and Brewin, 1995). In addition, it may result from diminished executive control in emotionally disturbed individuals (“execute control hypothesis”). Finally, it could be the consequence of negative self-representations and ruminative processes that keep people stuck at a general level of retrieval (“capture and rumination hypothesis”). Golden, Dalgleish, and Mackintosh (2007) have recently begun to examine overgeneral memory in complicated grief (CG). CG, also termed Prolonged Grief Disorder, is a disorder of grief characterized by persistent yearning, preoccupation with the lost person, intrusive images, and other grief symptoms present to the point of functional impairment (Prigerson et al., 2009). It seems plausible that overgeneral memory plays a role in emotional problems following loss because phenomena associated with overgeneral memory have been observed in people with such problems. For instance, intrusive memories are a key feature of CG (Boelen & Huntjens, 2008) and it seems plausible that the distress evoked by such memories motivates attempts at affect regulation through reduced memory specificity. Other phenomena associated with overgeneral memory such as negative self-views (Boelen, van den Bout, & van den Hout, 2006) and rumination (Nolen-Hoeksema, McBride, & Larson, 1997) have also been observed in people with emotional problems following loss. Golden et al. (2007) administered different versions of the AMT to bereaved individuals with and without a diagnosis of CG. One of their main findings was that, following negative but not positive cue words, CG patients were significantly less specific on the standard AMT than were controls. Interestingly, in a Biographical Memory Test that asked participants to retrieve memories from the life of the lost person, no significant differences in levels of specificity between groups were found. In fact, following negative cues, CG patients tended to retrieve more specific memories on this version of the task. Golden et al. (2007) interpreted these findings as indicating that people with psychopathology relating to a significant stressful life-event have a generalized tendency to avoid specific memories as a form of “functional avoidance”, but that memories tied to the source of a person’s distress (e.g., the death of a relative in people with CG) are “immune” to the processes of affect regulation that underlie overgenerality, because these latter memories are retrieved directly and habitually rather than through generative retrieval processes. Golden et al.’s (2007) study illustrates that examining overgeneral memory in CG is important because it sheds light on processes of affect regulation associated with this condition. Additionally, studying overgeneral memory is important because it likely plays a role in the maintenance of post-loss psychopathology. That is, overgeneral memory is associated with poor problem solving and reduced specificity of future images (Williams et al., 2007). As recovery from loss involves resolving problems and revising plans for the future, it seems conceivable that deficits in specific retrieval interfere with recovery. The present study aimed to further our knowledge about the role of memory specificity in emotional problems following loss. To this end, a heterogeneous, subclinical sample of mourners completed a standard version and trait version (McNally, Lasko, Macklin, & Pitman, 1995) of the AMT, together with self-report measures of CG, depression, and PTSD. Specifically, this study had four goals. The first goal was to examine the extent to which memory specificity varied as a function of demographic (i.e. age, gender, and education) and loss-related variables (i.e. kinship to the deceased, cause of loss, and time from loss). Due to a relatively small sample, Golden et al. (2007) were unable to examine this. Yet, it is possible that such variation exists. For instance, given the linkage between memory specificity and trauma exposure (Williams et al., 2007), overgeneral memory could be more prominent in people confronted with violent loss (e.g., due to an accident or suicide) than in those exposed to non-violent loss. The second goal was to examine the linkage between memory specificity and self-reported loss-related distress. Reduced specificity has been found to be more strongly associated with severe levels of depression and PTSD in clinical samples than with less severe symptoms in other groups. For instance, stronger correlations between specificity levels and PTSD have been observed in clinical samples (Kuyken & Brewin, 1995) compared to nonclinical samples (Hauer et al., 2006 and Henderson et al., 2002). Accordingly, we felt it was relevant to explore if reduced memory specificity was associated with self-reported CG-severity among mourners with subclinical levels of distress, apart from being associated with a clinical diagnosis of CG (Golden et al., 2007). In addition, to enhance knowledge about the linkage of memory specificity with different forms of distress, we also explored its association with symptom-levels of depression and PTSD. The third goal was to examine the degree to which loss-related distress was associated with a preferential retrieval of loss-related memories. To this end, memories obtained from the AMTs were coded as being related or unrelated to the loss or the lost person. Then, an index representing the overall “relatedness of specific memories” was calculated (as described below). At least two competing expectations could be formulated concerning the association of this index with loss-related distress. Based on the “affect regulation hypothesis”, it could be expected that stronger CG-severity would coincide with a preferential retrieval of specific memories that are unrelated to the loss/lost person, because people high in CG have a tendency to avoid distressing memories related to the loss ( Boelen et al., 2006). Alternatively, based on Golden et al.’s (2007) findings suggesting that loss-related memories are immune to such affect regulation (a viewpoint that we will refer to as the “immunity hypothesis”), it could be expected that higher levels of CG would coincide with a preferential retrieval of specific memories that are related to the loss/lost person. The fourth and final goal was to examine whether or not associations of loss-related distress with memory specificity and content differed between different versions of the AMT. As noted, apart from the standard AMT, a trait AMT was administered. The trait AMT instructs people to retrieve specific memories of moments that they exemplified particular personality traits (McNally et al., 1995). It could be argued that, compared to the standard AMT that taps memories from the entire autobiographical knowledge base, the trait AMT more strongly taps memories related to one’s self-identity. It was expectable that no differences between both AMTs would emerge in the associations of memory specificity with symptoms, given that reduced specificity is relatively robust across different versions of the AMT (Williams et al., 2007). Conversely, with respect to the “relatedness of specific memories”, differences could be expected. That is, there is some evidence that in people with higher levels of CG, the loss-event is more central to the self-concept (Boelen, 2009 and Maccallum and Bryant, 2008). Accordingly, it could be expected that loss-related distress would be more strongly associated with an enhanced inclusion of the lost person with memory information that is self-relevant than with inclusion of the lost person with more general autobiographical knowledge—and thus that emotional problems would be more strongly associated with the “relatedness of specific memories” in the trait AMT compared to the standard AMT. In sum, we examined (a) associations of memory specificity with demographic and loss-related variables, (b) associations of memory specificity with symptom-levels of CG, depression, and PTSD, (c) associations of the “relatedness” of memories with symptoms, and (d) the degree to which associations of symptoms with memory specificity and “relatedness” differed between the standard and trait version of the AMT.

. Results 3.1. Symptom-levels The mean total scores on the ICG-r, SCL Depression scale, and PSS-SR are shown in Table 1. In comparison with reference groups, scores were all in the subclinical range (cf. Arrindell and Ettema, 2003, Boelen et al., 2003 and Engelhard et al., 2007). Normality data indicated that SCL Depression and PSS-SR scores were positively skewed. Therefore, these scores were log-transformed in all analyses described below. Transformations reduced non-normality of the variables. ICG-r scores and indices obtained from the AMTs were all normally distributed. 3.2. Associations of background variables with symptom-measures It was our intention to control for background variables that were associated with symptom-levels in subsequent analyses. Therefore, we examined associations of symptom-levels with demographic (i.e. age, gender, educational level) and loss-related variables (i.e. time from loss, kinship to deceased, and cause of loss). Outcomes are summarized in Table 2. Table 2. Associations of demographic and loss-related background variables with symptom scores and memory indices. Background variables Gender Age Education Kinship Cause Time since loss Dependent variables F = r = F = F = F = r = Symptom-levels Complicated grief (ICG-r) 0.29 .17 1.87 6.26** 1.68 −.08 Depression (SCL depression) 2.41 .07 0.27 0.70 1.95 −.20* Posttraumatic stress disorder (PSS-SR) 0.75 .05 0.29 0.73 2.34 −.24* Standard AMT Number specific memories 1.28 −.37*** 11.82*** 4.70* 0.56 .04 Specificity-relatedness index 0.57 .21* 2.70 5.78** 1.23 −.18 Trait AMT Number specific memories 3.82 −.41*** 14.83*** 8.31*** 1.00 −.04 Specificity-relatedness index 0.17 .06 0.61 2.14 1.79 −.17 Note. Significant associations are depicted in bold. AMT = Autobiographical Memory Test. ICG-r = Inventory of Complicated Grief-revised version. PSS-SR = PTSD Symptom Scale Self-Report version. SCL = Symptom Checklist. *p < .05. **p < .01. ***p < .001. Table options ICG-r scores differed as a function of kinship. Post hoc tests showed that those who lost a partner had higher symptom-levels of CG than those who lost someone from the category “other relative” (p < .01). In addition, time from loss was negatively associated with depression and PTSD severity. No other significant associations were found. 3.3. Associations of background variables with memory specificity To address the first aim of this study, we examined to what extent the number of specific memories on the AMTs differed as a function of demographic and loss-related variables. Outcomes are summarized in Table 2. On both AMTs, the number of specific memories was inversely related with the participant’s age, varied as a function of education (more specific memories for those who had been to college/university compared to those who had had primary or secondary school [ps < .01]), and varied as a function of kinship (less specific memories for bereaved partners compared to those who lost someone from the category “other relative” [p < .05]). As noted, it could be expected that cause of loss influenced memory specificity. However, this was not the case ( Table 2). 3.4. Associations of symptom-levels with memory specificity To address our second aim, we first calculated correlations between specificity levels on the standard AMT and symptom-levels of CG, depression, and PTSD. To control for alpha-inflation, a significance level of .016 (.05/3) was required. Outcomes showed that the number of specific memories on the standard AMT was significantly associated with symptom-levels of CG (r = −.27, p < .01) but not depression (r = −.11, p = .28) and PTSD (r = −.12, p = .22). More severe CG coincided with less specificity. Regression analysis tested whether or not CG-severity continued to be associated with specificity when controlling for the order in which AMTs were administered and relevant background variables. Relevant background variables were demographic and loss-related variables that were associated with symptom-levels and/or memory specificity (see Table 2). All predictors together accounted for 29.8% of the variance in the number of specific memories (F(7, 92) = 5.16, p < .001). After controlling for relevant background variables (i.e. age, educational level, kinship to deceased), CG-levels continued to be associated with memory specificity (ΔR2 = 3.6%, F change (1, 85) = 4.35, p < .05, β = −.20). Next, correlations between specificity levels on the trait AMT and symptom scores were calculated, again setting the p-value for significant findings at p < .016. Outcomes showed that the number of specific memories on the trait AMT was again not significantly associated with depression (r = −.07, p = .48) and PTSD (r = −.08, p = .41). The correlation with CG-severity was r = .21 and failed to reach significance when adjusting the alpha-level (p = .03). 3.5. Cue word valence and memory specificity As an additional aspect of our second study aim, we examined whether associations of memory specificity with CG-severity differed when responses to positive and negative cue words were examined separately. For the standard AMT, the association of CG-severity with the number of specific memories in response to positive words was r = −.24 (p = .01) and in response to negative words was r = −.23 (p = .02); differences between these correlations were not significant (t = 0.10, p = .46). The associations of symptom-levels of depression and PTSD with specificity levels remained non-significant when responses to positive and negative words were examined separately (all rs < −.13, all ps ≥ .18). For the trait AMT, associations of levels of CG, depression, and PTSD with the number of specific memories all remained non-significant when responses to positive and negative words were examined separately (all rs < −.18, all ps ≥ .07). Altogether, findings did not indicate that symptom-levels were associated with different response patterns to positive and negative cues. 3.6. Associations of symptom-levels with relatedness of specific memories To address our third aim, we examined associations between symptom-levels and the degree to which participants preferentially retrieved specific memories that were related vs. unrelated to the loss or the lost person. To this end, a “relatedness of specific memories” index (henceforth termed “Specificity-Relatedness Index”) was calculated by subtracting the number of specific memories that were unrelated to the loss from the number of specific memories that were related to the loss, divided by the total number of specific memories. A positive sign of this index indicates that specific memories were more often related than unrelated to the loss. A negative sign indicates that specific memories were more often unrelated to the loss. For instance, for someone reporting three specific memories related to the loss/lost person and one specific unrelated memory, the “Specificity-Relatedness Index” would be valued ([3 − 1]/4=) 0.5. For someone reporting two specific related memories and four specific unrelated memories this value would be −0.33. 2 We first calculated correlations between symptom-levels and the Specificity-Relatedness Index for the standard AMT (with p < .016 as a threshold for statistical significance). Outcomes showed that this Index was positively associated with symptom-levels of CG (r = .31, p < .001) and PTSD (r = .26, p < .006), but not depression (r = .16, p = .10). Thus, higher levels of CG and PTSD coincided with a preferential retrieval of specific memories that were related to the loss/lost person. Regression analyses tested if levels of CG and PTSD continued to be associated with the Specificity-Relatedness Index when controlling for background variables that were either related with symptoms or this index (i.e. age, kinship; see Table 2). This was not so: CG-severity and PTSD severity did not explain variance in the Specificity-Relatedness Index beyond relevant background variables (ΔR2s < 3.2%, F changes < 3.51, ps > .07). Next, correlations between symptom-levels and the Specificity-Relatedness Index for the trait AMT were calculated (again with p < .016 as a threshold for statistical significance). Results showed that this Index was positively and significantly associated with symptom-levels of CG (r = .30, p < .002), depression (r = .34, p < .001), and PTSD (r = .25, p < .009). After controlling for relevant background variables ( Table 2), levels of CG (ΔR2 = 4.7%, F change (1, 88) = 4.54, p < .05, β = .23), depression (ΔR2 = 9.4%, F change (1, 103) = 11.19, p < .001, β = .31), and PTSD (ΔR2 = 4.2%, F change (1, 103) = 4.72, p < .05, β = .21) continued to contribute to the explained variance in the Specificity-Relatedness Index. 3.7. Relatedness of non-specific memories Our analyses thus far showed that higher symptom-levels coincided with a relative preferential retrieval of specific memories that were related to the loss/lost person. It was considered important to explore whether this also occurred with non-specific memories. If higher symptom-levels would be associated with both specific and non-specific memories being relatively more often loss-related, this would suggest that the relatively increased accessibility of loss-related specific memories may be due to a generalized preoccupation with thoughts and memories about the loss/lost person implicated in severe grief-related distress. Alternatively, if more severe distress coincided with higher relatedness of specific, but not non-specific memories, this would suggest that the relatively increased accessibility of loss-related specific memories is due to other processes. Thus, we examined associations of symptom-levels with the degree to which non-specific memories were related vs. unrelated to the loss/lost person. Accordingly, a so termed “Non-specificity-Relatedness Index” was calculated for both AMTs by subtracting the number of non-specific memories that were unrelated to the loss from the number of non-specific memories that were related to the loss, divided by the total number of non-specific memories. A positive sign of this index reflects that non-specific memories were more often related than unrelated to the loss. A negative sign indicates that non-specific memories were more often unrelated to the loss. For both AMTs, this Non-specificity-Relatedness Index was unrelated to symptom-levels of CG, depression, and PTSD (rs < .17, ps > .14). This suggests that the preferential retrieval of loss-related specific memories represents a process that is relatively independent from the retrieval of loss-related non-specific memories.