Memory and Posttraumatic Stress Disorder

In this issue of the Zeitschrift für Psychologie / Journal of Psychology we have assembled a number of articles and comments that deal with the relationship of trauma and memory, with an emphasis on the role of memory mechanisms in posttraumatic stress disorder (PTSD). There is uniformity in the assumption that memory is severely altered by traumatic events, but memory deficits may also contribute to the development of stress-related disorders. Although there is agreement that in PTSD memory mechanisms are at the core of the disorder, there is disagreement on the nature of the pathology and to what extent additional factors such as appraisal of the deficit are of importance. As Ehlers (2010) notes, there is clear evidence of a preponderance of emotional memories that are often sensory, fast, tied to very aversive moments of the traumatic experience, experienced as immediate rather than past, and related to non-declarative learning processes such as Pavlovian conditioning and priming. Whereas these memories seem to predominate, generalize (see Lissek & Grillon, 2010), and fail to extinguish (Wessa & Flor, 2007), episodic memories of the trauma and life in general seem to be affected, be it by fragmentation (Brewin, 2007), lack of connectedness (Michael, Ehlers, Halligan, & Clark, 2005), or overgeneralization (Schönfeld, Ehlers, Böllinghaus, & Rief, 2007). Early theories about memory mechanisms in PTSD suggested that the documented reduction of hippocampal volume in PTSD might lead to a dissociation of declarative and non-declarative memories (Elzinga & Bremner, 2002). However, these alterations are not as straightforward as it might seem since it could be shown that recall of trauma-related materials is similar in traumatized persons with and without PTSD (Wessa, Jatzko, & Flor, 2006). Retrieval-induced forgetting is also not significantly different in PTSD versus non-PTSD subjects as shown by Koessler and colleagues (2010). Diener, Flor, and Wessa (2010) demonstrate a specific deficit in encoding but not retrieval of episodic memory that is associated with hyperarousal symptoms. This suggests that a very thorough analysis of memory mechanisms is needed and that categories such as declarative/non-declarative or episodic versus semantic need to be further differentiated and specific aspects of these memory processes – such as elaboration, connectedness, or trauma-relevance – need to be considered. For example, the reduced hippocampal volumes could also lead to reduced context conditioning, which is also a form of emotional associative learning, whereas cue conditioning, which is more amygdala-dependent, seems to be enhanced. We (Lang et al., 2009) and others (Alvarez, Biggs, Chen, Pine, & Grillon, 2008; Marschner, Kalisch, Vervliet, Vansteenwegen, & Büchel, 2008) have shown that the hippocampus is involved in context conditioning in human beings and we have preliminary data that indicate impaired context conditioning in persons with small hippocampal volumes (Pohlack, Liebscher, Ridder, Lipinski, & Flor, 2009). A detailed examination of contextual memory in PTSD seems therefore warranted.

Context is also important for extinction learning where the organism has to learn to associate the conditioned stimuli with a certain context, no longer with a danger. Bouton and colleagues (2006) and others have emphasized the important role of context for extinction. Disordered context conditioning might therefore not only contribute to the feeling of “nowness” that characterizes the intrusions of PTSD patients and to the inability to feel safe (since safe and dangerous contexts cannot be properly differentiated) but may also impair the extinction of the cue-conditioned fear response since the extinction context has not been associated to the cues. This reasoning is in line with the proposal by Schauer and Elbert (2010), who argue that it is important to consider the peritraumatic defense processes that are activated and that subgroups of patients can be identified that are at different positions in the defense cascade. They specifically emphasize that memory processes are substantially different if persons are putting on a peritraumatic sympathetic activation versus those who went down the entire defense cascade where a parasympathetic dominance and dissociation prevail. This depends on the type of trauma and possibilities of the organism for optimal response. Extinction learning is especially hampered by dissociative responses (Ebner-Priemer et al., 2009) and must be adapted in these patients. In line with this, the study by Mauchnik, Ebner-Priemer, Bohus, and Schmahl (2010) showed differential conditioning in patients with borderline personality disorder with and without PTSD. This points to the necessity to carefully assess comorbidity with other disorders (depression is, e.g., also important) but also to delineate and consider modulating factors such as dissociation (Ebner-Priemer et al., 2009). The need to enhance extinction learning is also emphasized by Davis (2010), who suggests that pharmacological enhancers such as d-cycloserine, a partial NMDA receptor agonist, might improve extinction also in PTSD patients, perhaps by promoting contextual learning (see also Bouton, Vurbic, & Woods, 2008). The elevated false alarm rate reported in Ugandan PTSD patients in the retrieval-induced forgetting procedure by Koessler et al. (2010) also suggests that the contextual binding needed to differentiate traumatic from nontraumatic items may be disturbed and this would also contribute to the generalization reported by Lissek and colleagues (2008) and Lissek and Grillon (2010). In addition, Frewen and Lanius (2010) observed that emotional imagery triggered more trauma-related episodic memory in PTSD patients and this also elicited stronger responses to the script-driven imagery scenes and was associated with a reduced ventromedial prefrontal response suggesting a reduction in inhibitory power related to trauma-related learning. Interestingly, the vividness of mental imagery did not explain differences between PTSD and non-PTSD patients in a study by Jelinek and colleagues (2010). Mental imagery ability was, however, predictive of recovery from PTSD, with those who recovered showing less vivid mental imagery. How this ability is related to the retrieval of specific trauma memories has not yet been investigated, nor has imagery ability for traumatic versus nontraumatic materials been assessed. Moreover, the interaction of cue and context conditioning and episodic traumatic and nontraumatic memory processes must be further delineated.

The study by Dörfel, Werner, Schaefer, and Karl (2010) found a close association between increased PTSD symptoms and more brain activation in areas relevant for visual mental imagery (e.g., precuneus and occipital cortex) and decreases in areas involved in the orientation toward the environment or keeping the balance between internal and external process monitoring (e.g., supramarginal gyrus). The pattern of brain activations and associations with memory and neuropsychological performance also points to the role of visuo-spatial impairment and insufficient context integration in PTSD that might hamper extinction. An even more dramatic change in memory function is present in the patients described by Staniloiu and Markowitsch (2010) who suffer from dissociative amnesia where a complete shutdown of autobiographic-episodic memory is present as a consequence of stress. The authors report overall reduction in brain metabolism, selective small fiber degeneration as well as specific alterations primarily in right fronto-temporal cortices.

The contributions of this issue show that the area of trauma and memory is in the process of extensive theory building and differentiated empirical research. The use of multiple measures of learning and memory on the subjective, behavioral, and physiological level, the analysis of the time frame of learning and memory, and the consideration of evolutionary theories will greatly enhance our understanding of the role of memory in stress-related disorders such as PTSD.