اثرات نشخوار بر روی خلق و خوی و خاطرات مزاحم پس از مواجهه با مواد آسیب زا: یک مطالعه تجربی

Abstract Correlational studies have shown that trauma-related rumination predicts chronic post-traumatic stress disorder (PTSD). This study aimed to experimentally test the hypothesis that rumination is causally involved in the development and maintenance of PTSD symptoms. A video depicting the aftermath of serious road traffic accidents was used as an analogue stressor. After having watched the video, N = 101 healthy participants were randomly assigned to a guided thinking task designed to induce (a) rumination, (b) memory integration and (c) distraction. In line with the hypotheses, rumination led to less recovery from sad mood triggered by the video than the other two conditions. In addition, self-reported state levels of rumination during the guided thinking task predicted subsequent intrusive memories in the session. However, no significant main effect of the experimental manipulation on intrusive memories of the video was found. Results of exploratory analyses suggested possible sex differences in the way the processing manipulations were effective. Taken together, the results partially support the hypothesis that rumination is involved in the maintenance of negative mood and post-traumatic stress symptoms.

Introduction The majority of individuals who experience traumatic events such as a violent assault or severe road traffic accidents report symptoms of distress in the immediate aftermath (e.g. Kleim et al., 2007 and Murray et al., 2002). While some trauma survivors then go on to experience persistent symptoms of sufficient severity to warrant a diagnosis of post-traumatic stress disorder (PTSD), a significant proportion recover after a few weeks or months. For example, in the Detroit Area Survey of Trauma 25% of individuals meeting PTSD criteria (except duration criterion) in the weeks after the trauma had recovered at 6 months follow-up whereas 70% still met criteria after a year (Breslau et al., 1998). This raises the question of what factors contribute to the maintenance of the disorder. A number of theorists have suggested that rumination about the trauma and/or its consequences is an important factor involved in the development and maintenance of post-traumatic stress symptoms (Ehlers and Clark, 2000, Ehlers and Steil, 1995, Joseph et al., 1995 and Wells, 2000). In line with this view, a cross-sectional study of ambulance service workers found rumination to be significantly related to PTSD symptom severity as well as a measure of general mental health (Clohessy & Ehlers, 1999; for similar results see Ehring, Frank, et al., 2008, Michael et al., 2007 and Steil and Ehlers, 2000). In addition, a series of prospective longitudinal studies with survivors of road traffic accidents or assaults identified rumination in the immediate aftermath of the trauma as one of the strongest predictors of PTSD symptom severity at 6 months and 1 year, even when initial symptom levels were controlled (Ehlers et al., 1998, Ehring, Ehlers, et al., 2008, Kleim et al., 2007, Michael et al., 2007 and Murray et al., 2002). Interestingly, Michael et al. (2007) found that rumination was not unique to assault survivors with PTSD but was also reported by individuals without the disorder. However, the authors were able to identify specific characteristics of rumination that were closely linked to PTSD severity, namely engagement in “why” and “what if” type questions as well as unproductive circular thinking. In summary, there is evidence suggesting that rumination exacerbates post-traumatic stress symptoms. However, all of the above results are based on correlational data. Therefore, experimental studies are needed to test the hypothesized causal relationship. Two earlier experimental studies investigated the effect of worry on intrusive images in the context of general anxiety disorder (Butler et al., 1995 and Wells and Papageorgiou, 1995). In both studies, student participants were presented a distressing film and then asked to either worry about the film and its implications in a verbal form, engage in imagery about the film and its implications, or merely settle down. In both studies, worrying led to significantly more intrusive images about the film than the control conditions. As worry and rumination have been found to share a number of important process characteristics (Ehring & Watkins, 2008), results from these studies can be seen as indirect evidence for a causal effect of rumination on intrusive memories. However, additional analyses in the Butler et al. (1995) study revealed that there were no significant group differences in the time spent worrying during the induction period and that participants in the worry group actually reported more time imaging the content of the film than participants in the imagery or control group. It is thus not clear whether the induction of worry was successful and which processes were responsible for the differences in intrusive memories. In order to clarify the relationship between rumination and PTSD symptoms, studies are needed that induce trauma-related rumination with similar phenomenological properties as those described in PTSD (e.g., Michael et al., 2007). Little is known about the mechanisms by which rumination exacerbates post-traumatic stress symptoms. It has been suggested that, like worry in generalized anxiety disorder (Borkovec, Shadick, & Hopkins, 1990), rumination might be a form of cognitive avoidance, serving to distract from more distressing cognitions, such as visual memories of the worst moments of the traumatic event (Ehlers & Steil, 1995). This is in line with clinical observations showing that ruminative thoughts often take the form of ‘why’ and ‘what if’ questions rather than focusing on the actual traumatic situation itself. In their cognitive model of PTSD, Ehlers and Clark (2000) suggest three different pathways by which rumination may contribute to PTSD symptoms. First, rumination is thought to prevent an elaboration of the trauma memory and to hinder its integration into preceding and subsequent experiences and other autobiographical memories, which has the consequence that cue-driven retrieval of intrusive trauma memories is not sufficiently inhibited. Secondly, rumination is assumed to strengthen negative appraisals of the trauma and/or its sequelae. Finally, Ehlers and Clark suggest that rumination may also directly trigger symptoms such as feelings of nervous tension, dysphoria and hopelessness, and because it provides internal retrieval cues, may also directly trigger intrusive memories. This proposition is in accordance with recent results from depression research showing that rumination about depressive symptoms exacerbates negative affect, prolongs depressive episodes, and interferes with successful problem solving (Nolen-Hoeksema, 1991, Nolen-Hoeksema and Morrow, 1993 and Thomsen, 2006). In addition, Michael et al. (2007) found that negative affect during and after rumination was linked to PTSD severity. The present study therefore aimed to investigate the effect of rumination on negative affect as well as intrusive memories. Since it has been shown that intrusive memories are often triggered by stimuli bearing similar physical characteristics to stimuli signalling the moments of strongest emotional impact (Ehlers et al., 2002), the present study assessed spontaneously occurring memories as well as memories triggered by matching visual and acoustic stimuli. Effective treatments of PTSD such as trauma-focused cognitive behaviour therapy (TF-CBT) and eye movement desensitization and reprocessing (EMDR) aim to reduce re-experiencing by systematically recounting the traumatic experience or elements of it (see Bisson et al., 2007 for a review). TF-CBT interventions such as repeated imaginal exposure to the trauma (Foa & Rothbaum, 1998) or writing a detailed account of the event (Resick & Schnicke, 1993) aim to systematically transform the fragmented recall of the trauma into a coherent narrative. Similarly, Ehlers and Clark (2000) suggested that in order to reduce re-experiencing, the trauma memory needs to be elaborated. In particular, they suggest that the worst moments of the trauma memory need to be linked in treatment with relevant preceding and subsequent autobiographical experiences that put the meaning of these moments into context (e.g., linking the moment where the person thought that they were never going to see their children again with recent memories of playing with their children that signify to them that they survived the event). In an experimental study, Michael and Ehlers (2007) used a structured writing task to promote such autobiographical memory elaboration and integration after exposure to an analogue stressor. As expected, participants in the memory integration condition subsequently reported fewer intrusive memories than participants in a distraction control condition. The aim of the present study was to experimentally investigate the effects of rumination on analogue PTSD symptoms, using a distressing video depicting real life footage of the aftermath of road traffic accidents as the stressor. Earlier studies have shown the usefulness and validity of the trauma film paradigm for studying the development and modulation of analogue PTSD symptoms (for a review see Holmes & Bourne, 2008). This study further aimed to extend earlier findings on worry by Butler et al. (1995) and Wells and Papageorgiou (1995) by explicitly inducing trauma-related rumination. Furthermore, the study included two control conditions, a distraction condition, and a condition designed to promote autobiographical memory integration. It was hypothesized that (1) participants in the rumination condition will experience less recovery from sad and fearful mood than participants in the distraction or memory integration conditions, (2) participants in the rumination condition will experience more intrusive memories from the video than participants in the distraction or memory integration conditions, whereas participants in the memory integration condition will experience fewer intrusive memories than those in the control condition, and (3) levels of state rumination across all conditions will be positively associated with the number of intrusive memories.

3. Results 3.1. Baseline group differences As to be expected with random assignment, participants in the three conditions did not differ in any of the demographic variables or other measures prior to the experimental manipulation (see Table 1). 3.2. Manipulation checks 3.2.1. Effects of video on mood On average, participants rated the video as moderately distressing; M = 56, SD = 18, on a scale from 0 to 100. Condition did not affect the video-related distress, F(2, 98) = .86, p = .43, η2 = .02. As shown in Table 2, the video significantly increased PANAS-X ratings of sadness, F (1, 97) = 209.2, p < .001, η2 = .68, and fear, F(1, 97) = 45.5, p < .001, η2 = .32. No significant Time × Condition interaction effects were found, showing that participants in the different conditions experienced a similar increase in negative affect, F's(2, 97) < .85, p > .43, η2 < .02. Table 2. Means and standard deviations of dependent variables by condition. Total (N = 101) M (SD) Rumination (n = 32) M (SD) Memory integration (n = 35) M (SD) Control (n = 34) M (SD) Variables assessed during the session Sadness 1. Pre-video 1.08 (.18) 1.04 (15) 1.09 (17) 1.10 (.22) 2. Post-video 1.80 (.60) 1.84 (.63) 1.74 (.63) 1.83 (.54) 3. Post-manipulation 1.38 (.51) 1.58 (.53) 1.34 (.56) 1.24 (.36) Fear 1. Pre-video 1.23 (.28) 1.20 (.24) 1.25 (.32) 1.23 (.28) 2. Post-video 1.61 (.66) 1.61 (.74) 1.53 (51) 1.71 (.73) 3. Post-manipulation 1.27 (.46) 1.35 (.57) 1.25 (.43) 1.22 (.37) Intrusive memories 1. Spontaneous 6.1 (5.84) 6.4 (5.21) 5.9 (5.48) 6.0 (6.83) 2. Triggered 7.6 (6.34) 7.0 (5.71) 8.6 (7.27) 7.2 (5.94) Variables assessed during the following week Intrusive memories 13.5 (13.90) 14.7 (12.04) 12.9 (14.44) 13.0 (15.26) Sadness 1.43 (.52) 1.39 (.48) 1.44 (.51) 1.46 (.57) Fear 1.31 (.46) 1.37 (.61) 1.23 (.27) 1.31 (.43) Table options 3.2.2. Manipulation checks for experimental manipulation Results for the manipulation checks are shown in Table 3. Participants in the three conditions did not differ regarding their concentration on the induction task. However, as expected, perceived distress during the manipulation task was significantly lower in the control condition than in the rumination and memory integration conditions, which did not differ from each other. In addition, participants in the rumination group reported a stronger drive to continue dwelling on the video and higher state rumination scores on the PTQ-S than the control group, and, as a trend, than participants in the memory integration condition. For self-reported levels of memory integration processing assessed with the MIPQ, participants in the memory integration condition reported significantly higher levels of memory integration than participants in both the rumination and the control condition. Table 3. Manipulation check. Total (N = 101) M (SD) Rumination (n = 32) M (SD) Memory integration (n = 35) M (SD) Control (n = 34) M (SD) Statistics Concentration on task 84.0 (12.4) 83.3 (11.3) 82.6 (12.9) 86.0 (13.0) F(2, 95) = 1.79, p = .17, η2 = .04 Manipulation-related distress 24.9 (23.5) 32.2 (20.8) a 29.7 (25.1) a 13.1 (19.8) b F(2, 95) = 5.48, p < .01, η2 = .10 Drive to continue dwelling 2.2 (1.22) 2.5 (1.51) a 2.2 (1.03) b 1.9 (1.04) b F(2, 95) = 3.96, p < .05, η2 = .08 PTQ-S 27.8 (12.0) 32.1 (10.7) a 26.9 (12.6) b 24.5 (11.5) b F(2, 95) = 2.73, p = .07, η2 = .05 MIPQ 19.9 (7.9) 20.4 (5.2) a 25.4 (7.3) b 13.8 (6.3) c F(2, 95) = 27.51, p < .001, η2 = .37 Abstractness1 2.89 (.96) 2.07 (.81) t (59) = −3.65, p < .01 Unproductiveness1 .69 (.92) .06 (.23) t (63) = −3.90, p < .01 Repetitiveness1 .52 (.62) .20 (.38) t (63) = −2.50, p < .05 Catastrophizing1 .56 (.72) .17 (.41) t (59) = −2.70, p < .05 PTQ-S = Perseverative Thinking Questionnaire – State version; MIPQ = Memory Integration Processing Questionnaire; different superscripts denote differences between conditions at p < .05 or p < .10 if in parentheses. 1 Blind evaluations of thoughts reported in the short breaks during the thinking task. Table options As expected, independent raters' blind evaluations of the content of thoughts reported in the short breaks during the thinking task showed that participants in the rumination condition thought about the video scenes in a more abstract, unproductive, repetitive, and catastrophizing way, than participants in the memory integration condition (see Table 3). 3.3. Main analyses 3.3.1. Effects of the experimental manipulation on mood Means and standard deviations for sadness and fear at each assessment point are reported in Table 2. There were significant main effects of Time on sadness, F(1, 98) = 87.96, p < .001, η2 = .47, and fear, F(1, 98) = 58.09, p < .001, η2 = .37, showing that, on average, negative emotions decreased from post-video to post-manipulation. In line with Hypothesis 1, the main effect of Time on sadness was qualified by a significant Condition × Time interaction, F(2, 98) = 5.42, p < .01, η2 = .10, which was followed up by separate ANOVAs for each assessment point: Prior to the experimental manipulation, there was no significant effect of Condition on sadness, F(2, 98) = .30, p = .74, η2 = .01, while there was a significant Condition effect at post-manipulation, F(2, 98) = 4.02, p < .05, η2 = .08. As expected, simple contrasts showed that participants in the rumination condition reported significantly more sadness than those in the control condition, p = .01. There was no significant Condition × Time interaction for fearful mood, F(2, 98) = 1.94, p = .15, η2 = .04, nor a main effect of Condition, F(2, 98) = .08, p = .93, η2 = .00. 3.3.2. Effects of the experimental manipulation on intrusive memories during the session Means and standard deviations for the reported number of spontaneous and triggered intrusive memories are shown in Table 2. In contrast to Hypothesis 2, there were no main effects of Condition on the number of spontaneous or triggered intrusive memories, both F's < .47, both p's > .63, both η2's < .01. In exploratory analyses, the ANOVAs were re-run with Sex as an additional between-subjects factor. Results showed significant Condition × Sex interactions for both spontaneous intrusive memories, F(2, 95) = 4.69, p < .05, η2 = .09, and triggered intrusive memories, F(2, 95) = 7.08, p < .01, η2 = .13. Follow-up tests were conducted for male and female participants separately. Results showed that men in the rumination condition experienced significantly more intrusive memories (spontaneous memories: M = 7.3, SD = 5.13; triggered memories: M = 9.9, SD = 8.06), than those in the memory integration condition (spontaneous memories: M = 2.1, SD = 2.97; p < .01; triggered memories: M = 4.1, SD = 3.48, p < .05). Furthermore, men in the memory integration condition tended to experience fewer intrusive memories than those in the control condition (spontaneous memories: M = 5.5, SD = 4.55, p < .05; triggered memories: M = 7.4, SD = 4.43, p = .07). Men in the rumination and control conditions did not report different numbers of intrusive memories, p's > .37. In contrast, women in the memory integration condition reported more triggered intrusive memories (M = 10.9, SD = 7.67) than those in the rumination (M = 6, SD = 4.49) or control conditions (M = 7.1, SD = 6.56), p's < .05, which did not differ from each other, p = .67. There were no group differences for spontaneous intrusive memories in women, F(2, 68) = 1.54, p = .22, η2 = .04. 3.3.3. Effects of the experimental manipulation on symptoms in the subsequent week There were no significant Condition or Time × Condition effects on any dependent variable in the week following the session, all F's < 1, p's > .65, all η2's < .03. 3.3.4. Relationship between processing styles and intrusive memories Table 4shows the Spearman rank correlations between processing styles during the experimental manipulation and intrusive memories. In line with Hypothesis 3, levels of state rumination across all conditions correlated with the number of spontaneous and triggered intrusive memories in the experimental session. However, there were no significant correlations between self-reports of memory integration processing and intrusive memories in the experimental session. Table 4. Spearman rank correlations between state processing scores and the number of intrusive memories. Spontaneous IM Triggered IM State rumination (PTQ-S) Total .41** .25* Female .46** .13 Male .24 .48** Memory Integration Processing (MIPQ) Total .18 .16 Female .26* .26* Male −.07 −.07 Memory Integration Processing, corrected for rumination z(MIPQ) - z(PTQ-S) Total −.22* −.10 Female −.25* .04 Male −.17 −.35+ * p < .05; ** p < .01; +p < .10. PTQ-S = Perseverative Thinking Questionnaire – State version; MIPQ = Memory Integration Processing Questionnaire; IM = Intrusive Memories. Table options Exploratory analyses again investigated possible sex differences. Results showed that state rumination was significantly associated with spontaneous intrusive memories in females, r = .46, p < .001, and triggered intrusive memories in males, r = .48, p < .001. Unexpectedly, in female participants, but not males, levels of memory integration processing were also positively correlated with spontaneous and triggered intrusive memories, both r's = .26, both p's < .01. Furthermore, memory integration processing and state rumination were unexpectedly positively correlated in women, r = .28, p < .05, but not in men, r = −.18, p = .33. Thus, the memory integration processing questionnaire appeared to not specifically assess adaptive processing. In order to generate a measure of memory integration processing that does not involve elements of abstract, repetitive and uncontrollable thinking, z-transformed PTQ-S mean scores were subtracted from z-transformed MIPQ mean scores for the entire sample. As shown in Table 4, memory integration processing adjusted for rumination was negatively correlated with spontaneous intrusive memories in the experimental session. The same pattern was found in the female subsample. In the male subsample, adjusted memory integration processing showed a marginally significant negative correlation with triggered intrusive memories.