Logo: to the web site of Uppsala University

Although Pavlovian threat conditioning has proven to be a useful translational model for the development of anxiety disorders, it remains unknown if this procedure can generate intrusive memories - a symptom of many anxiety-related disorders, and whether intrusions persist over time. Social support has been related to better adjustment after trauma however, experimental evidence regarding its effect on the development of anxietyrelated symptoms is sparse. We had two aims: to test whether threat conditioning generates intrusive memories, and whether different social support interactions impacted expression of emotional memories. Non-clinical participants (n = 81) underwent threat conditioning to neutral stimuli. Participants were then assigned to a supportive, unsupportive, or no social interaction group, and asked to report intrusive memories for seven days. As predicted, threat conditioning can generate intrusions, with greater number of intrusions of CS+ (M = 2.35, SD = 3.09) than CS- (M = 1.39, SD = 2.17). Contrary to predictions, compared to no social interaction, supportive social interaction did not reduce, and unsupportive interaction did not increase skin conductance of learned threat or number of intrusions. Unsupportive interaction resulted in a relative difference in number of intrusions to CS + vs CS-, suggesting that unsupportive interaction might have increased image-based threat memories. Intrusions were still measurable one year after conditioning (one-year follow-up; n = 54), when individuals with higher trait anxiety and greater number of previous trauma experiences reported more intrusions. Our findings show that threat conditioning can create long-lasting intrusions, offering a novel experimental psychopathology model of intrusive memories with implications for both research on learning and clinical applications.