A memory like an elephant? The consistency of memory for emotional
events.
Giezen, A.E. van
Citation
Giezen, A. E. van. (2007, November 8). A memory like an elephant? The consistency of
memory for emotional events. Retrieved from https://hdl.handle.net/1887/12420
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the
Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/12420
Note: To cite this publication please use the final published version (if applicable).
3
The Relationship of Posttraumatic Stress, Dissociation and Thought
Suppression with Completeness of Memory for Violent Crime
The aim of the present study was to examine the relationship of post-traumatic stress, dissociation, and thought suppression with the recall of different aspects of violent assault. Within two months following the assault, 77 non-treatment seeking victims were assessed with a semi-structured interview. Emotional aspects related to the assault were more often recalled compared to situational aspects and central aspects were better recalled than peripheral aspects. Better recall of emotional aspects and impaired recall of situational aspects was associated with peritraumatic dissociation at the time of trauma exposure. Over and above the effect of dissociation, thought suppression had the paradoxical effect of enhancing the recall of situational aspects of the assault. Future studies must be more attentive for the differential effects of psychological mechanisms such as dissociation and thought suppression on the recall of emotional versus situational aspects of trauma exposure.
Van Giezen, A.E., Arensman, E., & Spinhoven, Ph.
Manuscript submitted for
publication.
Introduction
At least one out of three persons will become a victim of serious physical or sexual assault during their lifetime (Kessler, Sonnega, Bromet, Hughes, & Nelson, 1995). Exposure to violent crime is associated with a wide range of emotional problems. For example, of those exposed to violent crime, over 25% will develop a Post-traumatic Stress Disorder (PTSD), the three core symptoms being intrusions, avoidance and hyper-arousal (American Psychiatric Association, 1994). In individuals diagnosed with PTSD, autobiographical memory is often affected. A main characteristic of PTSD is the presence of recurrent, intrusive memories, which are vivid and rich in sensory detail, nightmares, and flashbacks (Brewin & Holmes, 2003). Intriguingly, exposed individuals experiencing such intrusive recollections often find it difficult to access the specific details of non-trauma related aspects of their autobiography (Vasterling, Brailey, Constans, & Sutker, 1998).
Memories for traumatic events are often disorganized or fragmentary. For example, following exposure to a traumatic event, individuals tend to encode the most important central aspects of the
event, instead of the peripheral aspects (Christianson, 1992; McNally, 2003b). The degree of fragmentation and disorganization of trauma narratives appears to be associated with PTSD symptom severity (Ehlers, Hackman, & Michael, 2004). Trauma narratives can become more organized and more complete over time or in the course of treatment (Foa, Molnar & Cashman, 1995; Harvey &
Bryant, 1999; Mechanic, Resick, & Griffin, 1998).
Cognitive mechanisms, such as dissociation and thought suppression, may be associated with difficulty to encode and consolidate the traumatic event in memory and appear to have predictive value for development of PTSD. Peritraumatic dissociation appears to be one of the most important predictors of PTSD (Koopman, Classen, & Spiegel, 1994; Marmar et al., 1994; Marmar, Weiss, &
Metzler, 1997; Ozer, Best, Lipsey, & Weiss, 2003; Shalev, Peri, Canetti, & Schreiber, 1996; Van der Kolk & Van der Hart, 1989). Peritraumatic dissociation can be conceptualised as a mechanism that reduces awareness of the negative consequences of a traumatic event in the absence of effective mechanisms to cope with trauma exposure (Van der Kolk & Van der Hart, 1989). It involves a disruption of cognitive and perceptual functioning during or immediately after an emotionally distressing event (Marmar et al., 1997). Koopman et al. (1994) argued that dissociation may be adaptive during and immediately following a traumatic event. However, the ongoing use of dissociative mechanisms as a primary strategy of dealing with a traumatic event may lead to a failure to emotionally process the trauma and therefore will result in more severe post trauma reactions.
Peritraumatic dissociation may be associated with incoherence at the time of encoding, resulting in memory fragmentation at the time of recall and may overlap with the processing of sensory and emotional impressions of the event rather than its meaning (Engelhard, Van den Hout, Kindt, Arntz, &
Schouten, 2003). However, other authors suggest that dissociative amnesia may result from deliberate efforts to forget rather than from aberrations in automatic processing (Van den Hout, Merckelbach, &
Pool, 1996).
It has also been observed that traumatized individuals try to suppress thoughts about their aversive experiences (Kuyken & Brewin, 1994) and that thought suppression in traumatized persons is also a risk factor for the development of PTSD (McFarlane, 1988). While the use of dissociative strategies involves primarily emotional suppression, thought suppression involves primarily cognitive suppression. Cognitive avoidance of a traumatic experience, specifically thought suppression however may have a paradoxical effect. Often, attempts to suppress unwanted thoughts produce more frequent and more intrusive thoughts (for a review see Rassin, Merckelbach, & Muris, 2000) and are related to slower recovery from PTSD (Dunmore, Clark & Ehlers, 2001; Ehlers, Mayou, & Bryant, 1998).
According to Van den Hout et al. (1996), dissociation and thought suppression are related. They suggest that dissociative experiences may be due to deliberate efforts to forget. Finally, there is limited evidence that individuals who report peritraumatic dissociation are also more likely to suppress thoughts and show avoidance strategies (Griffin, Resick, & Mechanic, 1997).
However, it remains unclear to what extent the diagnosis of PTSD, posttraumatic stress, (peritraumatic) dissociation and thought suppression are interrelated and to what extent each factor has a unique relationship with the recall of a traumatic event over and above the effect of the other factors. Therefore, the current study examined whether the recall of both emotional and situational
aspects and central and peripheral aspects of a personally experienced violent crime differed between non-treatment seeking victims with PTSD versus no PTSD. It was hypothesized that victims with PTSD would show a more complete recall of the violent crime given their recurrent and intrusive memories. We also expected a more complete recall of central aspects of the event, compared with the peripheral aspects. Furthermore, the association of posttraumatic stress symptoms, peritraumatic dissociation, general dissociative tendencies and thought suppression with the recall of emotional and situational information of the assault was investigated. It was hypothesized that (peritraumatic) dissociation would show a unique positive association with emotional aspects of the assault in particular, because it overlaps with the processing of the more sensory and emotional impressions of the assault. On the other hand thought suppression was hypothesized to be positively related to situational information in particular, because thought suppression primarily involves cognitive suppression and will have the paradoxical effect of producing more frequent thoughts related to assault characteristics.
Method Participants
This study is part of a prospective field study in which victims of assault were interviewed shortly after the traumatic event, including a 3- and a 6-month follow-up. Here, results from the first interview will be presented. Subjects were included in the study if they were a victim of violent assault, either physical or sexual, within the last two months, if they were 18 years or older and had a reasonable level of comprehension of the Dutch or English language. In order to avoid a confounding effect, subjects who showed symptoms of acute psychosis were excluded. Participants were also excluded from the study if the assault had occurred in the context of ongoing domestic violence or incest, or if the circumstances or living environment of the subject posed any potential risk to the interviewer.
Subjects were recruited through Victim Support Services in two regions in the Western part of The Netherlands: Den Haag and Haarlem, and the emergency department of a general hospital. The study was also advertised through information on the internet, posters in public places and advertisements in local newspapers and interviews on local and national television. Victims of a crime were contacted by Victim Support Services according to a standard procedure whereby victims are offered support following a criminal incident which they reported to the police. During the study period the Victim Support Services provided information on the study to victims who were contacted by the service.
Variables and Measurement
Recall of the assault incident was measured using a newly developed semi-structured interview, the Memory of Assault Interview (MAI). This interview consists of 43 items covering different aspects of the event, such as location, circumstances in which the event took place and offender characteristics.
The interview consists of two different sections. The first section comprises 34 questions on situational aspects of the events (e.g. Where did the event happen?, What kind of clothes did the perpetrator wear?, What was the weather like that day?), the second section contains 9 questions on emotional
aspects (e.g. How did you feel during the event? , Did you feel you had any control over the situation?). The MAI was developed by experts in the field of trauma and memory and was piloted among general population subjects for negative and positive events and among victims of assault treated in a general hospital in The Netherlands. Following the pilot study, final adjustments were made with regard to the completeness of items representing situational and emotional aspects of trauma exposure.
Besides counting the total number of recalled (situational and emotional) items on the MAI a completeness index was calculated by counting the number of recalled items and dividing it by the number of questions: 43. The same formula was used to compute a completeness index for recalled situational (34 questions) and emotional (9 questions) aspects of the assault. Furthermore, to explore the completeness of the recall of central vs. peripheral aspects, two raters (AvG, PhS) independently rated the situational items of the MAI with respect to centrality. Six questions were identified by both raters as referring to central information (i.e. gender of the perpetrator) and six to peripheral information (i.e. weather at the time of the assault). A completeness index was also computed for the central and peripheral questions.
The PTSD module of the Mini International Neuropsychiatric Interview (MINI; Sheehan et al., 1998) was used to assess the presence of a PTSD diagnosis. The MINI is a semi-structured interview with good reliability and validity (Sheehan et al., 1998).
For the assessment of PTSD symptoms, we used the Impact of Event Scale (IES; Horowitz, Wilner, and Alvarez, 1979). The IES is a 14-item self-report questionnaire which consists of two subscales: Intrusion and Avoidance. Both reliability and validity are good (Sundin & Horowitz, 2003).
Peritraumatic dissociation was assessed by the Peritraumatic Dissociative Experiences Scale (PDEQ; Marmar et al., 1997) answered with respect to experiences during the index event. This 10- item self-report questionnaire has high validity and reliability (Marmar et al., 1994).
General dissociative tendencies were measured with the Dissociative Experiences Scale (DES; Bernstein & Putnam, 1986). The DES is a 28-item self-report measure of dissociative experiences in every day life. The DES has very good validity and reliability, and good overall psychometric properties (Bernstein & Putnam, 1986).
Thought suppression was assessed with the White Bear Suppression Inventory (WBSI;
Wegner & Zakanos, 1994). The WBSI is a 15-item scale to measure stable individual differences in the tendency to suppress thoughts across a variety of situations and thought topics. Both reliability and validity are satisfactory (Muris, Merckelbach, & Horselenberg, 1996).
Procedure
Information on victims of assault was obtained through the Victim Support Services, through internet or by phone. Subjects were then contacted by one of the interviewers to provide information on the study and to schedule an appointment. The study was approved by the Medical Ethical Committee of Medical Centre The Hague. Participants were interviewed by extensively trained and supervised interviewers who had at least a BA in Clinical Psychology. The interview took place either at the home of the subject or at the university, or at another place that was most convenient for the victim. After an
introduction about the study, those subjects who agreed to participate provided written informed consent. The assessment consisted of a semi-structured interview and validated self-report questionnaires and lasted an average of 90 minutes (range 90-240 minutes). The semi-structured interview and questionnaires were presented in a fixed order to avoid any bias in responding to the assault items: MAI, IES, WBSI, PDEQ, DES and MINI.
Results
Subject Characteristics
Ninety-six victims were contacted by a member of the research team. Nineteen subjects were ineligible for the study. Seven subjects turned out to be victims of ongoing violence, and in 4 subjects the crime had occurred too long ago. Moreover, in 3 cases the situation was considered not safe for the interviewer, 3 subjects showed signs of acute psychosis and 2 did not wish to participate. Seventy- seven subjects completed the interview within 2 months of the assault (M=53.6 days post-crime). Most participants (n=52, 67.5%) were referred by the Victim Support Services, 13 victims (16.9%) reacted on posters or the website, 6 victims (7.8%) were referred by a general hospital and another 6 victims (7.8%) chose to participate in response to an item on television.
Of the 77 subjects who completed the first interview, 41 (53.2%) were female. The mean age of participants at the time of the interview was 38.2 (SD=14.4), with a range from 18 to 81. Over one third (37.7%) of the sample were married or cohabiting, 50.6% were single, 11.7% were divorced or widowed. With respect to highest educational achievement, 23.4% were college graduates, 67.5%
were high school graduates, and 9.1% did not graduate from high school. All subjects were victims of violent assault, such as being threatened, intimidated, sexually assaulted, physically attacked or beaten up. Nineteen subjects (24.7%) experienced a loss of consciousness during or after the assault.
No differences were found between those who did and did not lose consciousness with respect to demographic variables, psychopathology variables and completeness of memory.
Completeness of Memory
The two different sections of the MAI, situational and emotional aspects, were not significantly correlated (r=.16, ns.), indicating that they refer to different memory aspects. Overall, victims of recent assault show a relatively high percentage of recall. They responded to 89% of the items of the MAI.
The victims provided significantly more information in relation to emotional aspects compared to situational aspects of the assault. Ninety-five percent of the emotional items were recalled, while 86%
of the situational items were recalled, t(74)= -6.1, p<.01. With respect to recall of central vs. peripheral information, victims responded significantly more [t(76 )= 5.80, p<.001] to the central aspects (95.7 %), compared to the peripheral aspects (81.6 %). Two victims (2.7%) were able to give information on all items of the semi-structured interview. With regard to gender differences in completeness of recall, the results indicated that women tended to recall more emotional aspects of the assault than men, t(73)=- 1.76, p=.08. No gender differences were found for situational aspects and central and peripheral aspects. Age and educational level were not significantly associated with recall of emotional or situational aspects of the assault and recall of central or peripheral aspects.
Descriptives for PTSD Diagnosis
In nineteen cases a PTSD diagnosis could not be assessed, because the interview took place within the first month following the violent assault. Of the remaining 57 cases, 18 (31%) of the participants were diagnosed as having PTSD, according to the diagnostic criteria of the DSM-IV as assessed with the MINI. Victims with a diagnosis of PTSD reported significantly higher scores on thought suppression, peritraumatic dissociation, general dissociative tendencies, intrusions, avoidance and total IES-scores compared to those without a PTSD diagnosis (Table 1). Participants with a diagnosis of PTSD had lower levels of education [χ²(2)=8.71; p<.01], but did not differ in terms of gender or age from those without a PTSD diagnosis.
Descriptives for IES, PDEQ, DES, WBSI
Participants had a mean total score of 30.4 (SD=19.5) on the IES. They reported significantly more intrusions (M=17.1, SD=10.9) than avoidance symptoms (M=13.2, SD=10.3) in reaction to the assault, t(75)=3.96, p<.01. With regard to the total scores on the IES, women had a borderline significant higher total score and higher avoidance score than men (t[74]=-1.82, p=.07; t[74]=1.79, p=.08).
The total mean scores on the PDEQ and DES were 28.7 (SD=10.2) and 12.0 (SD=14.5) respectively. The mean score on thought suppression was 47.2 (SD=13.8). There were no gender differences with regard to dissociation and thought suppression.
Participants with a low educational level had significantly higher scores on intrusions [F(2,73)=
3.34; p<.05], total IES-scores [F(2,73)= 3.44; p<.05] and peritraumatic dissociation [F(2,70)= 3.36;
p<.05]. A trend towards significance was found for avoidance [F(2,73)= 2.46; p=.09] and thought suppression [F(2,70)= 2.64; p=.08].
Assault Characteristics
Table 2 shows the assault characteristics for those with and without a PTSD-diagnosis. Assault characteristics are also presented for those with high and low PDEQ scores, and high and low thought suppression scores. Cut-off scores on the PDEQ and WBSI are based on median split. Results show that none of the assault characteristics differentiated between participants with or without PTSD (all p
> .10). In addition, most of the assault characteristics did not differentiate between participants from the high and low PDEQ and high and low WBSI group. However, participants with high scores on peritraumatic dissociation reported significantly higher levels of life threat and a lower number of assailants involved in the crime. Participants with high scores on thought suppression reported a higher number of assailants.
Table 1 Mean sum scores for the MAI, IES, PDEQ, DES and WBSI by presence or absence of PTSD- diagnosis
Variable
PTSD (n=18)
NO PTSD
(n=39) Analysis
t
Total group
M SD M SD t(54) M SD
Recall total (0-43) 37.1 4.3 38.3 3.9 1.1 38.1 3.8
Recall emotional items (0-9) 8.5 .87 8.5 .68 1.1 8.6 .68
Recall situational items (0-43) 28.6 4.2 29.8 3.6 .20 29.5 3.7
Recall central items (0-6) 5.8 .55 5.7 .61 -.51 5.7 .6
Recall peripheral items (0-6) 4.4 1.3 5.1 .89 2.47* 4.9 1.1
Avoidance (IES-A) 18.3 10.7 9.5 8.9 -3.2** 13.2 10.3
Intrusions (IES-I) 21.6 9.8 13.4 10.1 -2.8** 17.1 10.9
Posttraumatic stress (IES-T) 40.0 18.1 22.9 17.5 -3.3** 30.4 19.5
Dissociation (PDEQ) 33.9 10.1 25.7 9.5 -2.8** 28.7 10.2
Dissociation (DES) 22.7 25.4 8.9 9.3 -2.1* 12.0 14.5
Thought suppression (WBSI) 57.5 11.9 41.8 11.5 -4.5** 47.2 13.8
Note. MAI= Memory of Assault Interview; IES= Impact of Event Scale; PDEQ= Peritraumatic Dissociaton Experiences Scale;
DES= Dissociative Experiences Scale; WBSI=White Bear Suppression Inventory; PTSD= Post-traumatic Stress Disorder.
* p< .05. ** p< .01.
Table 2 Assault characteristics in relation to Post-traumatic Stress Disorder, peritraumatic dissociation and thought suppressio n
Note. PTSD=
Post- traumatic Stress Disorder; PD
=
Peritraumatic Dissociation;
TS= Thought Suppression;
n.a=not applicable * p< .10. ** p<
.05. *** p<
.01 Variable
PTSD n=18
%
No PTSD n=39 χ²
High PD n=36
%
Low PD n=37 χ²
High TS N=36
%
Low TS n=37 χ²
Total group n=77
%
Assault type n.a n.a n.a
Sexual 11.1 7.7 88.9 94.6 88.9 94.6 9.1
Physical 88.9 92.3 11.1 5.4 11.1 5.4 90.9
Assailant 0.40 0.59 0.29
Known 64.7 76.9 72.2 77.8 72.2 77.8 74.3
Unknown 35.3 23.1 27.8 22.2 27.8 22.2 25.7
Number of assailants
6.03 8.38
**
4.93
* Single
assailants
47.1 66.7 52.8 58.3 44.4 66.7 55.4
Two assailants 29.4 10.3 33.3 8.3 30.6 11.1 20.3
Multiple assailants
23.5 23.1 13.9 33.3 25.0 22.2 24.3
Duration 7.54 4.11 3.77
Minutes 78.6 91.9 78.8 93.9 79.4 93.8 85.3
Half hour 7.1 0.0 6.1 3.0 8.8 0.0 4.4
One hour 0.0 5.4 9.1 0.0 5.9 3.1 4.4
Two or more hours
14.3 2.7 6.1 3.0 5.9 3.1 5.9
People nearby 2.06 0.06 1.39
No 58.8 43.6 50.0 52.8 58.3 44.4 51.4
Yes 41.2 56.4 50.0 47.2 41.7 55.6 48.6
Injury severity 5.37 0.00 0.53
Mild 23.5 30.8 27.8 27.8 30.6 25.0 27.0
Moderate 29.4 7.7 16.7 16.7 13.9 19.4 16.0
Severe 47.1 61.5 55.6 55.6 55.6 55.6 56.8
Able to Resist 4.40 0.84 0.12
No 52.9 50.0 51.4 41.7 45.7 47.2 45.2
To a certain extent
23.5 23.7 20.0 27.8 22.9 25.0 23.3
To a great extent
23.5 26.3 28.6 30.6 31.4 27.8 31.5
Perceived control
1.02 2.22 2.22
No 82.4 68.4 80.0 66.7 80.0 66.7 72.6
To a certain extent
5.9 10.5 8.6 8.3 8.6 8.3 9.6
To a great extent
11.8 21.1 11.4 25.0 11.4 25.0 17.8
Perceived threat
4.33 9.97
***
3.66
No 18.8 41.0 17.1 51.4 25.7 43.2 32.5
To a certain extent
6.3 10.3 17.1 5.4 8.6 13.5 11.7
To a great extent
75.0 48.7 65.7 43.2 65.7 43.2 51.9
Completeness of Recall in Participants with and without PTSD
To examine our hypothesis that participants with a PTSD diagnosis would show more complete recall of emotional and situational aspects of the assault than participants without a PTSD diagnosis, t-tests for independent groups were calculated. None of the differences proved to be significant (p >.10). With regard to the central and peripheral items, results showed that participants with a PTSD diagnosis were less complete in their recall of the peripheral aspects (see Table 1).
Zero-order Associations of Recall with Posttraumatic Stress, Dissociation and Thought Suppression
To examine our hypothesis that recall of the assault would be differentially related with (peritraumatic) dissociation and thought suppression, we correlated recall of situational and emotional aspects and central and peripheral aspects of the assault (MAI), PTSD symptoms (IES), suppression (WBSI) and dissociation (PDEQ, DES). Prior to the correlation a log transformation was applied to the DES scores because of extreme positive skewness (Tabachnick & Fidell, 2001).
In accordance with our hypothesis, results showed that peritraumatic dissociation and intrusions appeared to be significantly and positively associated with recall of emotional aspects of the assault, while both general and peritraumatic dissociation were significantly and negatively associated with recall of situational aspects (Table 3). Participants with high PDEQ scores appeared to have more complete recall of emotional aspects compared to recall of situational aspects. Contrary to expectations, thought suppression did not show a significant univariate association with the recall of situational items on the MAI, nor with the recall of emotional aspects of the assault. Scores for thought suppression however proved to be significantly and positively associated with scores for both general and peritraumatic dissociation.
Table 3 Bivariate correlations among the variables: MAI, IES, WBSI, MINI-PTSD, PDEQ, DES
1 2 3 4 5 6 7 8 9 10
1.Recall situational items 2.Recall emotional items .16 3.Recall central items .65** .18 4.Recall peripheral items .35* .10 -.02 5.IES-avoidance -.06 .14 .08 -.16 6.IES-intrusions -.14 .26* -.04 -.22 .68*
7. IES-total -.11 .22 .02 -.21 .91** .92**
8.WBSI .03 .05 .16 -.23 .68** .56** .67**
9. PDEQ -.29* .30** -.15 -.05 .35** .46** .45* .36**
10. DES -.24* .15 -.06 -.15 .30* .25* .29* .50** .37**
* p< .05. ** p< .01.
Multivariate Associations of Recall with Posttraumatic Stress, Suppression and Dissociation
In order to investigate whether PTSD symptoms, peritraumatic dissociation, general dissociative tendencies and thought suppression have a unique relationship with the recall of situational and emotional aspects of assault over and above the effect of the other factors multiple regression analysis was performed. In a backward regression, peritraumatic dissociation emerged as the only significant predictor of emotional recall (β=.30, t=2.60, p<.05), accounting for 7.5% of its variance, F(1,70)=6.75, p<.05. Thought suppression (β=.32, t=2.37, p<.05), general dissociative tendencies (β=-.28, t=-2.15, p<.05), and peritraumatic dissociation (β=-.34, t=-2.69, p<.01), predicted situational recall, accounting for 14% of its variance, F(3,68)=4.82, p<.01.
Discussion
The primary aim of this study was to examine the relationship between completeness of memory for different aspects of violent assault and PTSD, dissociation and thought suppression in a non-treatment seeking population of assault victims. The study results show that recall of the violent assault was fairly complete, with emotional aspects being somewhat better recalled (95%) than situational aspects (86%) and central aspects being better recalled (96%) than peripheral aspects (82%). Peritraumatic dissociation proved to be the most important predictor of completeness of recall, predicting enhanced recall of emotional and impaired recall of situational aspects of the assault. Moreover, thought suppression was associated with enhanced recall of situational aspects of the assault above and independent of (peritraumatic) dissociation.
Interpreting these results, it is important to make a distinction between objective memory performance and subjective evaluations of memory (meta-memory). In the present study subjects were not asked to evaluate their trauma memories with respect to memory quality or memory fragmentation (meta-memory), but to respond to specific questions in relation to emotional and situational aspects of a recent assault experience.
In accordance with clinical observations (e.g. Van der Kolk & Fisler, 1995) victims had a fairly good recollection of affective and sensory aspects concerning the assault. These results are in line with cognitive theories about information processing (Roediger, 1990) in which data driven (or bottom- up) and conceptually driven (or top-down) processes are distinguished as opposing processing modes. The processing style used will determine the quality of the memory representation. A data- driven processing style as frequently observed during trauma exposure (Ehlers & Clark, 2000) will result in an enhanced memory for emotional and perceptual details. Our study results extend the results of previous studies in which the affective and sensory qualities of trauma memories were evaluated in a subjective way (meta-cognition) (Koss, Figueredo, Bell, Tharan, & Tromp, 1996;
Mechanic et al., 1998) and not by actually asking participants to respond to specific questions in relation to emotional and sensory aspects of their trauma experiences.
In analyzing predictors of completeness of recall, peritraumatic dissociation was positively associated with recall of emotional aspects, but negatively associated with recall of situational aspects.
Apparently, higher levels of state dissociation while exposed to violence may enhance later recall of emotional information and impair recall of situational aspects. How can we explain this pattern of results? As in previous studies, peritraumatic dissociation was related to general dissociative
tendencies, (Marmar et al. 1994), perceived threat during trauma (Griffin et al., 1997; Marmar, Weiss, Metzler, & Delucchi, 1996), severity of posttraumatic stress reactions, in particular repetitive intrusions (Ehlers et al., 1998; Engelhard et al. 2003; Holmes, Brewin, & Hennesy, 2004) and avoidance strategies in coping with traumatic memories, such as thought suppression (Van den Hout, Merckelbach, & Pool, 1996). In summary, the findings suggest that in a high threat situation state dissociation denotes a primarily data-driven processing style (Ehlers & Clark, 2000) resulting in enhanced emotional memory. Apparently, state dissociation denotes a state of overwhelming emotions precluding adequate encoding of more situational aspects. Given the differential relationships of peritraumatic dissociation with recall of situational versus emotional aspects of the assault, it seems worthwhile to further study a possible bias of primarily processing the emotional and bodily aspects of the event at the expense of more external situational aspects due to higher levels of peritraumatic dissociation.
Interestingly, our study findings with regard to the relationship of state dissociation with memory were not in line with those of Kindt, Van den Hout, and Buck (2005), who reported that state dissociation was not related to objective memory fragmentation, but only to the subjective experience of memory fragmentation suggesting that dissociation is confined to the realm of subjective experiences. However, although our study did not address subjective and objective memory fragmentation but completeness of memory, the findings suggest that dissociation may interfere with information processing resulting in enhanced or reduced memory performance and may be more than a subjective evaluation of memory performance.
Only after accounting for the effects of state and trait dissociation on completeness of situational memory, thought suppression predicted the number of situational aspects recalled. Apparently, this positive association only emerges after accounting for the effects of the suppressor variables state and trait dissociation on situational recall. Apparently, independent of and above the effect of dissociation the tendency to suppress negative and unwanted thoughts may have paradoxical effects and actually enhance the frequency of these thoughts (Wenzlaff & Wegner, 2000). Thus, cognitive suppression of the assault probably will lead to recurrent thoughts and images of the assault and this process of repetition may enhance situational memory aspects of the assault.
Of special note is the finding that patients with or without PTSD did not differ in completeness of memory recall of the assault, except for the less complete recall of peripheral aspects in subjects with PTSD. The vast majority of the study sample had experienced a severe event, of which one third fulfilled the diagnostic criteria for PTSD. This prevalence is in accordance with the literature on development of PTSD following assault (American Psychiatric Association, 1994; Brewin, Andrews, &
Rose, 2003). Moreover, participants with a PTSD diagnosis had significantly higher and clinically elevated scores for posttraumatic stress complaints, state and trait dissociation and thought suppression. Apparently, differences in the completeness of recalling a traumatic event are not associated with PTSD. Possibly, other memory aspects are more relevant for developing PTSD.
Traditionally, the coherence of trauma memories for successful trauma processing has been stressed (Amir, Stafford, Freshman, & Foa, 1998; Foa et al., 1995). Recently, several authors have pointed out that an autobiographical event has to be embedded or contextualized in autobiographical memory
(Ehlers & Clark, 2000; Brewin, 2001). The formation and strengthening of associations between the trauma representation and other associated representations in order to inhibit the activation of traumatic re-experiences has gained more interest (Brewin, 2001). Future studies should address the clinical relevance of completeness of memory, fragmentation of memory and elaboration of memory simultaneously in a more comprehensive way in order to study their predictive value for development of PTSD.
In light of previous studies, the strength of the current study lies within the fact that the study was conducted among a non-clinical, non-treatment seeking group of assault victims, shortly after the event, and comprising an equal gender distribution. It is one of the first studies to systematically assess completeness of memory of a personally experienced traumatic event using a semi-standard interview. Furthermore, a distinction is made between different memory aspects: not only memory for situational aspects was investigated, but also recall of the emotional aspects surrounding the assault.
Several limitations are important to note in interpreting the findings of this study. A first limitation relates to the fact that no independent corroboration could be obtained. Therefore, we were unable to investigate whether completeness of memory also implies accuracy. Since no objective reports were available, we cannot confirm whether the reported information refers to actual memories or confabulations. Consequently, in measuring the completeness of memory of the assault in the absence of objective reports of the assault, the scores on the MAI also contain aspects of meta- memory (Kindt & Van den Hout, 2003). It is recommended that the results of field and laboratory studies are compared to see whether there exists a convergence of findings across laboratory studies with their higher level of internal validity and field studies with their higher level of external validity.
Secondly, the study sample was restricted to those assault victims who volunteered to retrospectively describe their experiences. A third limitation concerns the validity of the MAI. Although the MAI was standardized and piloted prior to this study, the validity has not yet been verified in the absence of data from other samples. It is important to note that the MAI includes questions related to emotions experienced during the assault which may be easier to answer (e.g. did you feel you had any control over the situation?) than questions related to situational aspects. Moreover, some authors even argue that memory for emotions and memory for facts operate through different processes (see Christianson & Safer, 1995). These differences preclude a direct comparison of the recall of emotional and the recall of situational items. Moreover, with regard to the classification of MAI items in central and peripheral items, we cannot be sure if this classification is accurate, since no objective reports of the assaults were available, nor subjective evaluations of aspects of the assault in terms of centrality.
To study recall of central vs. peripheral aspects of a personal experience, future studies should include subjective ratings of centrality by the participants.
Also, no additional measures for memory or meta-memory were included. Therefore, no distinctions could be made between meta-memory and actual memory performance. A fourth limitation concerns the use of the WBSI to asses thought suppression. The WBSI assesses a general tendency to suppress thought. It is not designed to assess suppression of trauma specific information. Fifth, differences in completeness of memory between participants with or without PTSD could only be assessed in those assault victims who were more than four weeks post incident. Finally, no control
condition was incorporated in the study that assessed recall of a neutral or positive event, thus limiting the generalizability of the results.
In conclusion, results from the present study illustrate that retrospective self-reports of violent assault are fairly complete, with emotional aspects of the traumatic event being even more completely recalled than situational aspects. The occurrence of peritraumatic dissociation while exposed to violent assault is significantly associated with more complete recall of the emotional aspects and less complete recall of situational aspects. Moreover, thought suppression may have the paradoxical effect of enhancing the recall of situational aspects of the assault. Future studies must be more attentive for the differential effects of psychological mechanisms such as dissociation and thought suppression on the recall of emotional versus situational aspects of trauma exposure.
