(1995), narrative changes from the first to the last exposure session were compared for improved and nonimproved PTSD patients on fragmentation, organization, internal, and external events. Improved (n = 8) and nonimproved (n = 12) patients did not differ regarding changes in fragmentation or organized thoughts. However, improved patients showed a greater decrease in disorganized thoughts during treatment. Furthermore, all patients, independent of improvement, showed significant changes in the same direction; a decrease in disorganized thoughts and external events and an increase in internal events. Although previous results were partly replicated, it is concluded that narrative changes may be due to exposure treatment itself rather than to changes in memory representation.KEY WORDS: PTSD; exposure therapy; narratives; fragmentation; memory.Several studies explored the relationship between trauma narratives during exposure therapy and PTSD symptom severity. Most research thus so far has focused on narrative content (e.g., Pennebaker, 1993). In contrast, Foa, Molnar, and Cashman (1995) explored the process of narrative organization during PTSD treatment. Assuming that traumatic memories differ from other types of memories, Foa and colleagues argued that trauma recovery requires a special type of mental processing.Successful trauma therapy aims at increasing traumatic memory organization and should result in more coherent narratives. To test this, Foa et al. (1995) developed a coding system and examined the cohesiveness of patients' trauma narratives. They hypothesized that successful treatment results in a decrease in fragmentation and an increase in narrative organization. Although the decrease in fragmentation did not reach a significant level, it was significantly related to successful treatment outcome. Moreover, compared to the first session, significantly more organized thoughts were found in the last session, supporting the idea that therapy enhances narrative organization. In addition, at the end of the treatment patients expressed significantly more internal events (e.g., feelings), and, though not significant, slightly fewer external events (e.g., details). This study aims at replicating and extending the findings of Foa and coworkers. Although this previous study is informative with respect to narrative organization processing, an important limitation is that all patients (n = 14) highly improved during therapy. Therefore, it remains unclear whether changes in narratives were exclusively due to the reorganization of traumatic memory. To establish whether these changes reflect an underlying recovery process rather than an epiphenomenon of treatment, we compared trauma narratives in the first and last therapy sessions of improved and nonimproved PTSD patients. If changes in trauma narratives reflect adequate processing of a traumatic event, improved patients should show a larger decrease in fragmentation and a larger increase in narrative organization than nonimprovers. In addition, improvers should sho...