Available online at www.sciencedirect.com

Psychiatry Research 165 (2009) 68–77 www.elsevier.com/locate/psychres

Verbal learning and impairments in posttraumatic stress disorder: The role of strategies ⁎ Grethe E. Johnsen , Arve E. Asbjørnsen

University of Bergen, Bergen, Norway

Received 13 December 2006; received in revised form 5 July 2007; accepted 1 January 2008

Abstract

The present study examined mechanisms underlying verbal memory impairments in patients with posttraumatic stress disorder (PTSD). Earlier studies have reported that the verbal learning and memory alterations in PTSD are related to impaired encoding, but the use of encoding and organizational strategies in patients with PTSD has not been fully explored. This study examined organizational strategies in 21 refugees/immigrants exposed to war and political violence who fulfilled DSM-IV criteria for chronic PTSD compared with a control sample of 21 refugees/immigrants with similar exposure, but without PTSD. The California Verbal Learning Test was administered to examine differences in organizational strategies and memory. The semantic clustering score was slightly reduced in both groups, but the serial cluster score was significantly impaired in the PTSD group and they also reported more items from the recency region of the list. In addition, intrusive errors were significantly increased in the PTSD group. The data support an assumption of changed memory strategies in patients with PTSD associated with a specific impairment in executive control. However, memory impairment and the use of ineffective learning strategies may not be related to PTSD symptomatology only, but also to self-reported symptoms of depression and general distress. © 2008 Elsevier Ireland Ltd. All rights reserved.

Keywords: CVLT; Depression; Organizational strategies; Refugees

1. Introduction reexperiencing, avoidance behaviors,andhyperarousal (American Psychiatric Association, 1994). There is This study focuses on mechanisms underlying verbal evidence that PTSD is associated with distinct brain memory alterations in refugees/immigrants with posttrau- dysfunction patterns and cognitive impairments (Elzinga matic stress disorder (PTSD) after war exposure. PTSD is and Bremner, 2002), and verbal learning and memory classified as an anxiety disorder and develops in some alteration have consistently been documented (e.g., individuals as a response to traumatic stress and is charac- Bremner et al., 2004; Gilbertson et al., 2001). In addition terized by three main clusters of symptoms: intrusive aspects of and memory dependent on executive control appear to be impaired in individuals with PTSD (Kanagaratnam and Asbjørnsen, 2007; Vasterling et al., 1998). ⁎ Corresponding author. Department of Biological and Medical Psychology, University of Bergen, Jonas Liesvei 91, N-5009 Bergen, Previous studies on the relationship between verbal Norway. Tel.: +47 55582561; fax: +47 55589872. memory impairments and PTSD have reported both E-mail address: [email protected] (G.E. Johnsen). acquisition impairment (Yehuda et al., 2005b), and

0165-1781/$ - see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.psychres.2008.01.001 G.E. Johnsen, A.E. Asbjørnsen / Psychiatry Research 165 (2009) 68–77 69 impairments related to an reduced ability to In a study of Bosnian refugees, comorbidity of PTSD consolidate memory material (e.g., Bremner et al., and depression was related to high rates of psychosocial 1995; Uddo et al., 1993). Lately, studies have been disability (Mollica et al., 1999). Most of the studies designed to examine more explicitly the relation examining cognitive alteration in PTSD have been between stage of processing and memory impairment. conducted on groups with coexisting psychiatric condi- Two recent studies have provided evidence that tions (Vasterling et al., 2002; Yehuda et al., 2005b). impairment in encoding accounted for the verbal Lately, there has been a growing awareness that memory impairments in PTSD patients when control- complex and prolonged trauma may result in many ling for attention and the effect of total learning (Yehuda posttraumatic outcomes (Bremner, 2002; van der Kolk et al., 2005a; Yehuda et al., 2004). In addition, we et al., 2005). Some studies that have applied statistical previously found evidence that learning and memory control for comorbidity or concurrent conditions have impairments in refugees with PTSD were related to the suggested that PTSD makes an independent contribu- ability to elaborate memory material to facilitate a more tion to the group differences (Bremner et al., 2004; efficient encoding (Johnsen et al., 2008). In that study, Gilbertson et al., 2001; Jelinek et al., 2006; Jenkins et wefoundnodifferencesbetweenadiagnosedPTSD al., 1998; Jenkins et al., 2000); other studies have found group and a control group on memory span, but less depressive symptoms in traumatized individuals to be effective learning emerged over trials on a standardized related to the cognitive impairments (Brandes et al., verbal learning test. Controlling for measures of 2002; Johnsen et al., 2008). In a prospective study, early attention ruled out the possibility that the difference in depressive symptoms were found to predict PTSD better verbal learning and memory was secondary to atten- than early PTSD symptoms (Freedman et al., 1999). tional dysfunction. No impairments on tests of recogni- This makes it important to assess concurrent conditions tion were found. This suggested a pattern of memory such as depression, in order to understand if the impairment secondary to difficulties using effective underlying memory impairment and executive failure learning or organizational strategies (Sternberg and are specifically related to PTSD symptoms or to the Tulving, 1977) during encoding. The above-mentioned depressive symptoms connected to it. Depression has studies did not analyze organizational strategies. been linked to impairment on tasks requiring effortful Jenkins et al. (1998) reported that although their processing (Hasher and Zacks, 1979). In addition, PTSD sample did not seem to be impaired on general intellectual functioning has been found to be a acquisition over learning trials, they were impaired at risk factor for PTSD (Macklin et al., 1998)and short and long delay recalls. Follow-up tests revealed controlling for IQ measures is necessary. no differences in semantic and serial organizational To date, neuropsychological studies of PTSD suggest strategies. Two previous studies have analyzed recall that verbal memory impairments in patients with PTSD errors, which may indicate executive failure. Both are related to problems with encoding and executive increased intrusive errors and perseverative responses memory. However, the previous studies have focused were found on of list-learning tests (Uddo mostly on the number of items recalled and to a lesser et al., 1993; Vasterling et al., 1998). extent on qualitative aspects underlying the memory PTSD has been seen as a failure to process and impairments such as intervening cognitive processes, integrate trauma, and is described as an information e.g., organizational strategies and error types. The processing disorder (Brewin et al., 1996). Decreased association between memory impairments and difficul- executive and inhibitory control of trauma , ties in implementing organizational strategies has been emotions, and impulses characterizes the disorder. The documented in other psychiatric disorders and brain verbal memory impairments among trauma samples disease (Gershberg and Shimamura, 1995; Lundervold have been linked both to increased arousal state and et al., 1994; Savage et al., 2000). Thus, the question of intrusive reexperiencing phenomena (Kolb, 1987). If whether the encoding impairments found in patients with intrusive symptoms in PTSD reflect a general failure in PTSD are related to ineffective organizational strategies executive control, with a deficit in the ability to suppress still remains to be answered. Studies documenting involuntary thoughts, it is assumed that this impairment memory impairment in war-related PTSD have mostly also is seen on specific memory tests involving been conducted on US war veteran samples, with only a organizational strategies and executive control. few presenting data from civilian samples and different However, traumatized refugees exhibit a spectrum of cultural groups. War veterans have been specially trained clinical signs, and in addition to PTSD, major for challenging situations, which may affect their coping depressive disorder is common (Mollica et al., 1987). and later symptom development. 70 G.E. Johnsen, A.E. Asbjørnsen / Psychiatry Research 165 (2009) 68–77

The aim of the present study is twofold. The first aim (CAPS; Blake et al., 1998). The CAPS examines all is to examine the use of memory-enhancing strategies the diagnostic criteria for PTSD, including criterion A, such as semantic and serial clustering and to assess the both frequency and intensity of criteria B–D, criteria E implications of such strategies on differences in serial and F. The Life Event Check List in the CAPS, which position effects and recall errors in refugees with PTSD identifies the main trauma of the participants, confirmed compared with exposed refugees without PTSD. The that their PTSD syndromes were related to political second aim is to study how strategy use and recall errors violence and war events. are related to the symptomatology. The current study is Participants were excluded from the study if they had based on the same sample as reported by Johnsen et al. organic brain damage or other neurological diseases. (2008). Based on the previous findings of encoding and Those who had been involved in events that gave reason executive impairment, we hypothesized that learning to expect brain injuries such as blows to the head or loss and memory impairments in patients with PTSD would of consciousness exceeding 30 min were excluded from be mediated by difficulties in using effective semantic the study (Vasterling et al., 1998), as were subjects with clustering strategies during encoding. Second, we a history of alcohol/substance abuse and psychotic hypothesized that the executive impairment would disorders. For the PTSD participants, the diagnosis of impact serial clustering, recall errors, and serial position major depressive disorder was not an exclusion effects. Third, we assumed that ineffective use of criterion, and in the PTSD group all participants were organizational strategies in the PTSD group could be diagnosed with a major depressive episode. Subjects explained by self-reported symptoms of PTSD, depres- with other diagnoses such as bipolar disorder, obses- sion, general distress, and/or estimated IQ. If one of sive–compulsive disorder, generalized anxiety disorder, these conditions were an important contributor to the and panic disorder were excluded. Participants on insufficient organization of the memory material, then tricyclic antidepressant, benzodiazepine, and neurolep- there would be reduced group effects when the variables tic medication were excluded. Participants on selective were introduced as covariates, and there is evidence that serotonin reuptake inhibitors (SSRIs) were included. depressive symptoms, in addition to PTSD symptoms, would be expected to be important contributors. 2.2. Procedure

2. Method The study was approved by the Regional Committee for Medical Research Ethics. Information about the 2.1. Participants study was given to health personnel in outpatient clinics and at health services for refugees/immigrants. Written The sample consisted of 42 refugees/immigrants, 32 information was posted at local public health services. men and 10 women. They were mainly from the Middle- Participants who had been exposed to war or political East, the former Yugoslavia and Chile, and were violence and who had a tentative PTSD diagnosis were voluntary participants in the study. Twenty-one were either referred by counselors or recruited themselves. diagnosed with chronic PTSD according to DSM-IV Exposed participants for the non-PTSD group were criteria following experiences with actions of war and recruited by the same formal or informal contacts. The political violence (the PTSD group). They were participants gave written informed consent prior to compared with a healthy control group of 21 refugees participation and were informed that participation was with the same exposure, but who had never developed voluntary. All patients went through a diagnostic PTSD (the non-PTSD group). Both groups were assessment administered by a licensed clinical psychol- matched as closely as possible with regard to age, ogist experienced in traumatology. Participants received trauma exposure, and ethnicity profile. Gender distribu- no economic compensation for taking part in the study. tion was 14:7 in the PTSD group and 18:3 in the non- All the participants in the study were interviewed about PTSD group. The controls had no previous history of their exposure to war and political violence, and special psychiatric illness. care was taken in the interviews and assessment Clinical assessment included The Mini-International procedures in order to avoid retraumatization. Neuropsychiatric Interview (M.I.N.I.), version 5.0.0 The clinical self-report scales and the neuropsycho- (Sheehan et al., 1998). The M.I.N.I. was used to logical tests were administered in the same order for all investigate the presence of PTSD and other psychiatric the participants. Other tests were included in the overall disorders. The PTSD diagnosis was given based on the neuropsychological battery, but these are not reported in Clinician Administered PTSD Scale for DSM-IV the current study. All the test material and self-report G.E. Johnsen, A.E. Asbjørnsen / Psychiatry Research 165 (2009) 68–77 71 scales were presented in the participants' native gories. The two lists have two common semantic language, and the interviews and testing were done in categories, but the category items are specific in each collaboration between the clinician, a specially trained list. During the learning trials, the items are spoken test technician, and an authorized interpreter. Efforts aloud by the examiner in a fixed order, and no two were made to use the same interpreters within the same consecutive items are derived from the same semantic language groups. category. In this way an evaluation of the degree to which items are grouped into semantic categories in the 2.3. Clinical assessments first five test trials (semantic clustering; “fruits” and “clothes”) or to which degree recall follows the order in The Impact of Event Scale — Revised (IES-R), a 22- which the items were presented (serial clustering) is item self-report measure, was used to examine the severity possible. In order to assess semantic clustering, of PTSD symptoms the week before testing (Weiss and participants receive scores based on the number of Marmar, 1997). Current depressive symptoms were words they recall from the same semantic category in assessed with the Montgomery and Åsberg Depression succession. This score is corrected for the total number Rating Scale (MADRS), a 10-item questionnaire mea- of words recalled. suring depressive symptoms in the last 3 days (Mont- List A is presented five times, and the participants are gomery and Åsberg, 1979). The MADRS was used as a instructed to recall as many words as possible after each structured interview to examine depressive symptoms. presentation of the list. After the five presentations of General distress was measured with the Symptom list A, a new list of words (list B) is presented to the Checklist-90-Revised (SCL-90-R; Derogatis, 1983). participants, and they are instructed to recall as many The subtests Similarities and Picture Completion words as they can from list B. Participants are then from the Wechsler Adult Intelligence Scale-Revised asked to recall list A again in short delay free recall were used to estimate current IQ level (Wechsler, 1981). (SDFR), and after a 20-minute interval they are asked to Scores on the Similarities subtest have a strong recall list A in long delay free recall (LDFR). The correlation with general intelligence. Picture Comple- standard procedure includes the use of semantic cues to tion is seen as a “hold” test, and is described as relatively enhance recall (cued recall). The CVLT measures and stable and not so easily influenced by external factors. indexes used in this study are presented in Table 2. This test is thus expected to be an indictor of premorbid IQ, and both subscales have high loadings on the g- 2.5. Statistical analyses factor (Lezak, 1995). The War Exposure Questionnaire (WEQ) was used to Group differences in demographic characteristics, assess exposure to war and political violence. The WEQ clinical assessments, and neuropsychological test per- is a 27-item self-report measure assessing potentially formance were analyzed using independent t tests. traumatic events such as threat, loss, injury, deprivation, Differences in CVLT clustering scores were analyzed by grotesque impressions and other aspects of war and analysis of variance (ANOVA). Correlation analyses political violence. The WEQ is a modified version of the were performed to examine the relationship between Childhood War Exposure Questionnaire, caretaker CVLT organizational strategy scores and clinical version (Netland and Kanaaneh, 1989). measures. In addition, an analysis of covariance (ANCOVA) was used to assess the influence of PTSD 2.4. Neuropsychological testing symptoms, depressive symptoms, general distress, and estimated IQ on organizational strategies and recall The California Verbal Learning Test (CVLT; Delis errors. To correct for multiple comparisons, Bonferroni et al., 1987) was administered for testing organizational tests were conducted. strategies, serial position effects, and recall errors. The CVLT is a multi-trial serial learning test, and provides an 3. Results assessment of multiple strategies and processes involved in learning and remembering verbal material. It also 3.1. Characteristics of participants provides a measure of organization and learning strategy. The CVLT identifies two measures of organi- Characteristics of the participants in this study are zational strategies: semantic and serial clustering. The presented in Table 1. The two groups were similar in test consists of two lists (list A and list B), each list terms of age, years of education, estimated IQ, duration containing 16 items in four different semantic cate- of exposure to war and political violence, and time 72 G.E. Johnsen, A.E. Asbjørnsen / Psychiatry Research 165 (2009) 68–77

Table 1 Demographic and clinical characteristics and estimated intelligence for the PTSD group and the non-PTSD group PTSD group (n=21) Non-PTSD group (n=21) M (S.D.) M (S.D.) t-value Effect size (d) P Age 36.90 (8.77) 39.24 (10.06) −0.80 −0.25 n.s. Education (years) 12.12 (2.86) 14.05 (4.12) −1.76 −0.56 n.s. Exposure range (months) 83.14 (66.87) 100.81 (71.62) −0.83 −0.26 n.s. Time since exposure (months) 107.43 (96.00) 166.27 (109.66) −1.71 −0.59 n.s. MADRS a 26.20 (6.93) 5.85 (4.21) 11.23 3.64 0.000 CAPS 73.76 (18.55) n.a. IES-R b Avoidance 18.29 (5.83) 8.10 (8.36) 4.54 1.45 0.000 IES-R Intrusion 20.81 (4.34) 8.15 (7.41) 6.71 2.15 0.000 IES-R Hyperarousal 20.29 (3.42) 6.60 (6.54) 8.45 2.71 0.000 IES-R Total score 59.38 (9.89) 22.85 (21.22) 7.12 2.28 0.000 IES-R GSI d 2.70 (0.45) 1.09 (0.94) 7.03 2.25 0.000 SCL-90-R c 190.95 (56.55) 76.75 (52.35) 6.28 2.12 0.000 SCL-90-R GSI d 2.12 (0.63) 0.85 (0.58) 6.27 2.12 0.000 Estimated IQ e 82.19 (13.00) 91.90 (17.90) −2.01 −0.64 n.s. a Montgomery and Åsberg Depression Rating Scale. b The Impact of Event Scale — Revised. c Symptom Checklist — 90-Revised. d GSI = General Symptom Index. e Estimated IQ from the WAIS-R subtests Picture Completion and Similarities.

passed since their exposure. As would be expected, the serial coding (F(1,36) =7.70, P=0.008, see Fig. 1), but no PTSD group had higher scores on the measures of IES- interaction effect. The effect was rejected when using R, MADRS and SCL-90-R. Bonferroni tests. The Least Significant Difference test showed significant differences between serial and semantic 3.2. Organizational strategies, serial position effects, organization strategies within the non-PTSD group and recall errors (P=0.02). When testing the effect of gender, a significant groupbygender,interactionwasfound(F(1,34) =4.30, Pearson product-moment correlations revealed no P=0.05,seeFig. 2), and a nearly a three-way interaction by significant correlation between age and education and strategy, grouping, and gender (F(1,34) =3.99, P=0.053). the cognitive measures. The performance of the two The next set of analyses showed that on the serial groups on the CVLT, organizational strategies, serial position effect of the CVLT list the PTSD group recalled position effects, learning slope, recall consistency and significantly more recency words. The PTSD group was recall errors are presented in Table 2. characterized by an increased number of intrusive errors. An ANOVA with group as the between-groups factor, The effect sizes, using Cohen's d for serial clustering, and coding strategy as a within-subjects factor revealed a recency, and intrusions, were medium to large (Cohen, significant main effect of groups, with the highest score for 1969). These findings are detailed in Table 2.

Table 2 Cognitive measures for the PTSD group and the non-PTSD group PTSD group (n=21) Non-PTSD group (n=21) M (S.D.) M (S.D.) t-value Effect size (d) P CVLT 1–5 41.75 (13.61) 52.83 (7.26) −3.08 −1.03 0.004 Semantic clustering 1.44 (0.66) 1.65 (0.68) −0.99 −0.33 n.s. Serial clustering 1.68 (1.03) 2.49 (1.28) −2.17 −0.72 0.03 Learning slope 1.02 (0.56) 1.31 (0.60) −1.48 −0.49 n.s. Consistency 73.66 (19.22) 81.63 (8.87) −1.61 −0.54 n.s. Perseverations 3.10 (2.59) 3.61 (3.79) −0.49 −0.16 n.s. Intrusions 2.45 (3.50) 0.61 (1.46) 2.07 0.69 0.05 Primacy % 27.25 (6.73) 23.67 (5.03) 1.84 0.61 n.s. Middle % 41.96 (10.52) 50.26 (3.90) −3.15 −1.05 0.003 Recency % 30.79 (7.23) 26.06 (4.40) 2.40 0.80 0.02 G.E. Johnsen, A.E. Asbjørnsen / Psychiatry Research 165 (2009) 68–77 73

90-R (r=0.34, Pb0.05), and estimated IQ (r=0.42, Pb0.05). To analyze the relative importance of self-reported symptoms of PTSD (IES-R), depression symptoms (MADRS), general distress (SCL-90-R) and estimated IQ on semantic and serial clustering, a covariate analysis (ANCOVA) was performed with these variables as covariates. When controlling for IES-R, the group differences for semantic and serial clustering disap- peared (F(1,35) =1.46, P=0.23). This analysis revealed that self-reported PTSD symptoms had the greatest effect on serial clustering in the PTSD group. When MADRS scores were controlled for, the differences between the groups on semantic and serial clustering disappeared (F(1,34) =0.05, P =0.83), and the self- reported depressive symptoms had a stronger effect than PTSD symptoms on the serial coding in the PTSD Fig. 1. Performance on semantic and serial clustering on the CVLT for group. When SCL-90-R scores were controlled for, the the PTSD group and the non-PTSD group. (Small bars depict the 95% group effect was reduced (F(1,32) =2.60, P=0.12), and confidence interval.) the general distress symptoms had the strongest effect on semantic clustering in the non-PTSD group. When 3.3. Relationship between encoding problems and estimated IQ was controlled for, the differences between symptomatology the groups persisted (F(1,35) =4.91, Pb0.03). In order to test the relative importance of the three Pearson product-moment correlations revealed no PTSD symptom clusters on organizational clustering, significant correlations between semantic clustering, we used covariance analyses. When the effects of the clinical measures (IES-R, MADRS, SCL-90-R), and Intrusion and the Arousal scales were removed, the estimated IQ. Serial clustering correlated with IES-R between-groups differences of organizational clustering (r=0.39, Pb0.05), MADRS (r=0.40, Pb0.05), SCL- disappeared (F(1,35) =0.61, P=0.44 and F(1,35) =1.78,

Fig. 2. Performance on semantic and serial clustering on the CVLT for males and females separate for the PTSD group and the non-PTSD group. 74 G.E. Johnsen, A.E. Asbjørnsen / Psychiatry Research 165 (2009) 68–77

P=0.19, respectively). Removing the effect of the The increased recency score in the PTSD group Avoidance scale had no major impact on the overall compared with the non-PTSD group could indicate a group differences (F(1,35) =4.28, P=0.04). tendency to use a passive recall strategy of echoing back Since depressive symptoms had a strong effect on the the last items presented. In addition intrusive responses use of organizational strategies, we also performed a dominated the PTSD patients, so the results confirmed covariate analysis with MADRS as the control variable on previous studies in war veterans (Vasterling et al., 1998), intrusive errors. The analysis yielded an interaction effect suggesting that intrusion in PTSD is not limited to trauma- (F(1,34) =5.13, P=0.03), with reduced intrusions in the related cognitions, but reflects a more general pattern of PTSD group, and increased intrusions in the non-PTSD impaired inhibition and cognitive control. In addition, the group when the effect of the MADRS score was removed. depression score after war exposure explains an extensive part of the variation in intrusive errors in the PTSD group. 4. Discussion Monitoring describes the ability to control the verbal output, and this function was affected differently in the The main findings in this study were less serial two groups, and was also differently related to comorbid organization, more intrusive errors, and more pro- depression. When controlling for the MADRS score was nounced recency effect in patients with PTSD compared controlled for, intrusive errors were reduced in the PTSD with exposed controls during learning and memorizing group, but increased in the control group. This could verbal material. This contributes to our understanding of suggest that the control group maintained executive impaired verbal learning and memory associated with control in the test situation, whereas those who developed PTSD. Age, education, and estimated intelligence did PTSD as a clinical syndrome did not. Maintained not explain the findings, since the participants were executive control may act as a protective mechanism matched on these variables. These results disputes after trauma exposure. Deficits in verbal memory shortly Jenkins et al's (1998) study showing no differences on after the traumatic event have been related to a greater risk organizational strategies using the CVLT. for developing PTSD (Bustamante et al., 2001). The hypothesis of less effective semantic coding in the Furthermore, the covariance analysis suggested that PTSD group was not supported. Compared with norms IES-R and MADRS scores had a relatively large impact for the CVLT, both groups were below average for on serial clustering in the PTSD group, especially the semantic clustering (Delis et al., 1987). Since both groups MADRS score. In addition, the SCL-90-R score had an are slightly impaired, we cannot say if they have an intact important effect on semantic clustering in the control semantic organization. The reduced scores seen may group. This suggests that encoding deficits are mainly partly be related to the different use of the semantic related to depressive symptoms, and not directly to categories within the different languages used by the symptoms unique to PTSD. This finding is consistent participants in this study. Both a Norwegian validation of with the findings of Brandes et al. (2002), who reported the norms (Egeland et al., 2005)andtheAmericannorms that cognitive impairments in PTSD were related to (Delis et al., 1987) indicate more serial coding in men and depressive symptoms, although other studies have more semantic coding in women. This may also have reported that PTSD makes an independent contribution influenced our results as our samples include more men. to the memory impairments (e.g., Bremner et al., 2004; Both the males and the females in the control group Gilbertson et al., 2001; Jenkins et al., 1998). These had scores on encoding strategies as expected, with seemingly contradictory findings may be related to males using more serial clustering and females more methodological differences, as tasks, samples, and semantic clustering. Interestingly, the females in the diagnostic criteria vary. The PTSD group in the present PTSD group used more serial clustering, which study had relatively high scores on current depressive indicates that women under stress use less sophisticated symptoms (MADRS). In their country of origin they were encoding strategies (serial clustering) than the controls. all exposed to complex and prolonged trauma, involving The men in the PTSD group had difficulties with both threat to life as well as loss. Complicated adaptations to strategies. To our knowledge there are no published severe and prolonged trauma have been described as both studies on gender differences in verbal memory or complex trauma (Herman, 1992) and Disorders of encoding strategies in PTSD samples, but women tend Extreme Stress Not Otherwise Specified (DESNOS; to perform better on verbal learning and memory in Roth et al., 1997) where posttraumatic problems other normal samples (Kramer et al., 1988). Our finding than PTSD are included. Van der Kolk et al. (2005) have related to gender is significant despite low power, and it been critical of the PTSD literature, which discusses should be tested further. psychiatric symptoms that do not fall within the G.E. Johnsen, A.E. Asbjørnsen / Psychiatry Research 165 (2009) 68–77 75 framework of PTSD as comorbid conditions, as if they authorized translators, this procedure is not fully occur independently from the PTSD symptoms. It is also equivalent to using a validated translation of the test. possible that overlapping symptoms in our PTSD and However, this also strengthens the relevance of the study depression measures contribute to our findings, but in a multi-ethnic clinical setting. Efforts were made to use viewing the depressive symptoms as independent of the the same interpreters within the same language groups, PTSD symptoms may not be valid, since PTSD captures and the study used a between-group design with control only a limited aspect of posttraumatic psychopathology. for ethnicity, which should reduce the probability of a When analyzing the impact of PTSD on cognitive systematic bias in our conclusion. Both groups are in the functioning, one has to discuss what a relevant compar- lower range on measures of IQ, and one can only ison group might be. Our control group was exposed to speculate on the possible explanation for this. The most the same traumatic events as the PTSD group, but had not commonly proposed factors contributing to variance on developed a PTSD diagnosis. In addition, they were IQ measures within different cultural groups are educa- sharing the same strain as being a refugee and therefore tion, characteristics of the test, familiarity with stimuli and could not be assumed to be asymptomatic (Hondius et al., test situations in general (Berry et al., 2002). PTSD has in 2000). According to Söndergaard et al. (2001),refugees addition been associated with low premorbid intelligence are influenced to a great extent by political events and the (Macklin et al., 1998), but the lower intelligence scores in situation of significant others in the home country. the PTSD group may also be connected to the develop- However, this also strengthens the possibility to analyze ment of PTSD and to mild related information-processing the unique cognitive patterns of PTSD, since confounding deficits (Gil et al., 1990). factors to a large extent are controlled for. Our findings suggest that future research with PTSD The finding of impaired organizational strategies in in this area should focus on specific components of the PTSD is intriguing. However, functional neuroimaging memory system, especially executive control of mem- studies are showing a significant relationship between the ory. The implication of the present findings, however, is use of semantic organization during learning and recall that memory impairments in refugee groups might be and frontal lobe activation in healthy controls (Fletcher et associated with one or more strain factors common to al., 1998). Our results suggest that intrusive and arousal various forms of posttraumatic psychopathology. symptoms may account for difficulties in using serial To conclude, we have found that impaired memory in organizational strategies. Organization of information PTSD can be explained by ineffective use of memory- during encoding challenges the executive processes, and enhancing strategies, and is not limited to memory capacity. these functions are associated with the left prefrontal This finding is important to both theory development cortex (Fletcher et al., 1998). Structural and functional and clinical work with patients with PTSD. An inefficient neuroimaging studies have shown dysfunction in both memory system is likely to be a key factor affecting the prefrontal and temporal cortex in patients with PTSD individuals with PTSD's ability to function educationally (Kitayama et al., 2005; Shin et al., 2001), although and occupationally. Learning and memory alteration in support for hippocampus alterations is not found in all refugees may affect their lives in the host country with studies (Pederson et al., 2004; Schuff et al., 2001). Further regard to learning the new language and their ability to studies should assess whether patients with PTSD also participate in education and employment. Memory show impaired use of organizational strategies on other impairment has not been sufficiently focused on in the cognitive tasks in order to determine the generalizability rehabilitation of patients with PTSD, and deficient orga- of the organizational dysfunction. Given the possible nization and strategy use should be targeted within the source of errors connected to the different use of semantic clinical educational work with these patients. categories within the different language and ethnic groups participating in this study, organizational strategies Acknowledgments underlying memory impairments should be examined further in a more homogeneous group with more balanced This study was financially supported by the Research distribution between the sexes. Council of Norway, L. Meltzer Høyskolefond, and the There are some important limitations in this study that University of Bergen. should be mentioned. First, the sample was small, which may limit the generalizability of our findings. The References findings need to be confirmed in larger samples. Second, American Psychiatric Association, 1994. Diagnostic and Statistical it is problematic to use standardized tests with a translator. Manual of Mental Disorders: DSM-IV, 4th ed. Author, Even when the stimulus words are translated by Washington, DC. 76 G.E. Johnsen, A.E. Asbjørnsen / Psychiatry Research 165 (2009) 68–77

Berry, J.W., Poortinga, Y.H., Segall, M.H., Dasen, P.R., 2002. Cross- Herman, J., 1992. Complex PTSD: a syndrome in survivors of prolonged Cultural Psychology. Research and Applications, Second ed. and repeated trauma. Journal of Traumatic Stress 5, 377–391. University Press, Cambridge. Hondius, A.J.K., van Willigen, L.H.M., Kleijn, W.C., van der Ploeg, Blake, D.D., Weathers, F.W., Nagy, L.M., Kaloupek, D.G., Charney, D.S., H.M., 2000. Health problems among Latin-American and Middle- Keane, T.M., 1998. Clinician-Administered PTSD Scale for DSM-IV, Eastern refugees in the Netherlands: relations with violence Revised (CAPS): National Center for Posttraumatic Stress Disorder. exposure and ongoing sociopsychological strain. Journal of Behavioral Science Division — Boston VA Medical Center & Traumatic Stress 13, 619–634. Neuroscience Division — West Haven VA Medical Center. Jelinek, L., Jacobsen, D., Kellner, M., Larbig, F., Biesold, K.H., et al., Brandes, D., Ben Schachar, G., Gilboa, A., Bonne, O., Freedman, S., 2006. Verbal and nonverbal memory functioning in posttraumatic Shalev, A.Y., 2002. PTSD symptoms and cognitive performance in stress disorder (PTSD). Journal of Clinical and Experimental recent trauma survivors. Psychiatry Research 110, 231–238. Neuropsychology 28, 940–948. Bremner, J.D., 2002. Does Stress Damage the Brain? Understanding Jenkins, M.A., Langlais, P.J., Delis, D., Cohen, R., 1998. Learning and Trauma-Related Disorders from a Mind–Body Perspective. W. W. memory in rape victims with posttraumatic stress disorder. The Norton & Company, Ltd, New York. American Journal of Psychiatry 155, 278–279. Bremner, J.D., Randall, P., Scott, T.M., Capelli, S., Delaney, R., et al., Jenkins, M.A., Langlais, P.J., Delis, D.A., Cohen, R.A., 2000. 1995. Deficits in short-term memory in adult survivors of Attentional dysfunction associated with posttraumatic stress disorder childhood abuse. Psychiatry Research 59, 97–107. among rape survivors. The Clinical Neuropsychologist 14, 7–12. Bremner, J.D., Vermetten, E., Afzal, N., Vythilingam, M., 2004. Johnsen, G.E., Kanagaratnam, P., Asbjørnsen, A., 2008. Memory Deficits in verbal declarative memory function in women with impairments in posttraumatic stress disorder are related to depression. childhood sexual abuse-related posttraumatic stress disorder. Journal of Anxiety Disorders 22, 464–474. Journal of Nervous and Mental Disease 192, 643–649. Kanagaratnam, P., Asbjørnsen, A., 2007. Executive deficits in chronic Brewin, C.R., Dalgleish, T., Joseph, S., 1996. A dual representation PTSD related to political violence. Journal of Anxiety Disorders theory of posttraumatic stress disorder. Psychological Review 103, 21, 510–525. 670–686. Kitayama, N., Vaccarino, V., Kutner, M., Weiss, P., Bremner, D., 2005. Bustamante, V., Mellman, T.A., David, D., Fins, A.I., 2001. Cognitive Magnetic resonance imaging (MRI) measurement of hippocampal functioning and the early development of PTSD. Journal of volume in posttraumatic stress disorder: A meta-analysis. Journal Traumatic Stress 14, 791–797. of Affective Disorders 88, 79–86. Cohen, J., 1969. Statistical Power Analysis for the Behavioral Kolb, L.C., 1987. A neuropsychological hypothesis explaining Sciences. Academic Press, New York. posttraumatic stress disorders. American Journal of Psychiatry. Delis, D.C., Kramer, J.H., Kaplan, E., Ober, B.A., 1987. The California 144, 989–995. Verbal Learning Test: Research Edition. Psychological Corporation, Kramer, J.H., Delis, D.C., Daniel, M., 1988. Sex differences in verbal New York. learning. Journal of Clinical Psychology 44, 907–915. Derogatis, L.R., 1983. SCL-90-R: Administration, Scoring, and Lezak, M.D., 1995. Neuropsychological Assessment, 3 ed. Oxford Procedures Manual-II. Second edition. Clinical Psychometric University Press, New York. Research, Towson, MD. Lundervold, A.J., Reinvang, I., Lundervold, A., 1994. Characteristic Egeland, J., Sundet, K., Landro, N.I., Rund, B.R., Asbjornsen, A., et al., patterns of verbal memory function in patients with Huntington's 2005. Validering av normer for oversatte tester av oppmerksomhet og disease. Scandinavian Journal of Psychology 35, 38–47. hukommelse i et norsk normalutvalg. {Validation of norms for a Macklin, M.L., Metzger, L.J., Litz, B.T., McNally, R.J., Lasko, N.B., Norwegian version of attention and memory tests}. Tidsskrift for et al., 1998. Lower precombat intelligence is a risk factor for Norsk Psykologforening 42, 99–105. posttraumatic stress disorder. Journal of Consulting and Clinical Elzinga, B.M., Bremner, J.D., 2002. Are the neural substrates of Psychology 66, 323–326. memory the final common pathway in posttraumatic stress Mollica, R.F., Wyshak, G., Lavelle, J., 1987. The psychosocial impact disorder (PTSD)? Journal of Affective Disorders 70, 1–17. of war trauma and torture on Southeast Asian refugees. American Fletcher, P.C., Shallice, T., Dolan, R.J., 1998. The functional roles of Journal of Psychiatry 144, 1567–1572. prefrontal cortex in . I. Encoding. Brain a journal Mollica, R.F., McInnes, K., Sarajlic, N., Lavelle, J., Sarajlic, I., of neurology 121, 1239–1248. Massagli, M.P., 1999. Disability associated with psychiatric Freedman, S.A., Brandes, D., Peri, T., Shalev, A., 1999. Predictors of comorbidity and health status in Bosnian refugees living in chronic post-traumatic stress disorder: a prospective study. British Croatia. Journal of the American Medical Association 282, Journal of Psychiatry. 174, 353–359. 433–439. Gershberg, F.B., Shimamura, A.P., 1995. Impaired use of organiza- Montgomery, S.A., Åsberg, M., 1979. A new depression scale tional strategies in free recall following frontal lobe damage. designed to be sensitive to change. British Journal of Psychiatry Neuropsychologia 33, 1305–1333. 134, 382–389. Gil, T., Calev, A., Greenberg, D., Kugelmass, S., Lerer, B., 1990. Netland, M., & Kanaaneh, M. (1989). Tests and questionnaires for the Cognitive functioning in post-traumatic stress disorder. Journal of research project Children and political violence. Unpublished Traumatic Stress 3, 29–45. manuscript, Bergen, Norway: University of Bergen. Gilbertson, M.W., Gurvits, T.V., Lasko, N.B., Orr, S.P., Pitman, R.K., Pederson, C.L., Maurer, S.H., Kaminski, P.L., Zander, K.A., Peters, 2001. Multivariate assessment of function in C.M., et al., 2004. Hippocampal volume and memory perfor- combat veterans with posttraumatic stress disorder. Journal of mance in a community-based sample of women with posttraumatic Traumatic Stress 14, 413–432. stress disorder secondary to child abuse. Journal of Traumatic Stress Hasher, L., Zacks, R.T., 1979. Automatic and effortful processes in 17, 37–40. memory. Journal of Experimental Psychology: General 108, Roth, S., Newman, E., Pelcovitz, D., van der Kolk, B.A., Mandel, F.S., 356–388. 1997. Complex PTSD in victims exposed to sexual and physical G.E. Johnsen, A.E. Asbjørnsen / Psychiatry Research 165 (2009) 68–77 77

abuse: results from the DSM-IV field trial for posttraumatic stress Vasterling, J.J., Brailey, K., Constans, J.I., Sutker, P.B., 1998. disorder. Journal of Traumatic Stress 10, 539–555. Attentionandmemorydysfunctioninposttraumaticstress Savage, C.R., Deckersbach, T., Wilhelm, S., Rauch, S.L., Baer, L., et al., disorder. Neuropsychology 12, 125–133. 2000. Strategic processing and episodic memory impairment in Vasterling, J.J., Duke, L.M., Brailey, K., Constans, J.I., Allain, A.N., obsessive compulsive disorder. Neuropsychology 14, 141–151. Sutker, P.B., 2002. Attention, learning, and memory performances Schuff, N., Neylan, T.C., Lenoci, M.A., Du, A.T., Weiss, D.S., et al., and intellectual resources in Vietnam veterans: PTSD and no 2001. Decreased hippocampal N-acetylaspartate in the absence of disorder comparisons. Neuropsychology 16, 5–14. atrophy in posttraumatic stress disorder. Biological Psychiatry 50, Wechsler, D., 1981. Wechsler Adult Intelligence Scale — Revised. 952–959. The Psychological Corporation, New York. Sheehan, D.V., Lecrubier, Y., Sheehan, K., Amorim, P., Janavs, J., et al., Weiss, D.S., Marmar, C.R., 1997. The Impact of Event Scale — 1998. The Mini-International Neuropsychiatric Interview (M.I.N.I): Revised. In: Wilson, J.P., Keane, T.M. (Eds.), Assessing the development and validation of a structured diagnostic psychiatric Psychological Trauma and PTSD: a Practitioner's Handbook. interview for DSM-IVand ICD-10. Journal of Clinical Psychiatry 59, Guildford Press, New York, pp. 399–411. 22–33. Yehuda, R., Golier, J.A., Halligan, S.L., Harvey, P.D., 2004. Learning Shin, L.M., Whalen, P.J., Pitman, R.K., Bush, G., Macklin, M.L., et al., and memory in Holocaust survivors with posttraumatic stress 2001. An fMRI study of anterior cingulate function in posttraumatic disorder. Biological Psychiatry 55, 291–295. stress disorder. Biological Psychiatry 50, 932–942. Yehuda, R., Golier, J., Tischler, L., Stavitsky, K., Harvey, P.D., 2005a. Sternberg, R.J., Tulving, E., 1977. The measurement of subjective Learning and memory in aging combat veterans with PTSD. organization in free recall. Psychological Bulletin 84, 539–556. Journal of Clinical and Experimental Neuropsychology 27, Söndergaard, H.P., Ekblad, S., Theorell, T., 2001. Self-reported life 504–515. event patterns and their relation to health among recently resettled Yehuda, R., Stavitsky, K., Tischler, L., Golier, J.A., Harvey, P.D., Iraqi and Kurdish refugees in Sweden. The Journal of Nervous and 2005b. Learning and memory in aging trauma survivors with Mental Disease 189, 838–845. PTSD. In: Vasterling, J.J., Brewin, C.R. (Eds.), Neuropsychology Uddo, M., Vasterling, J.J., Brailey, K., Sutker, P.B., 1993. Memory and of PTSD. Biological, Cognitive and Clinical Perspectives. attention in combat-related post-traumatic stress disorder (PTSD). Guilford Press, New York, NY, US, pp. 208–229. Journal of Psychopathology and Behavioral Assessment 15, 43–52. van der Kolk, B.A., Roth, S., Pelcovitz, D., Sunday, S., Spinazzola, J., 2005. Disorders of extreme stress: the empirical foundation of a complex adaptation to trauma. Journal of Traumatic Stress 18, 389–399.