A critical test of the dual-representation account of PTSD

尽管创伤后应激障碍（PTSD）的发病率很高，其后果使人衰弱，但目前的心理治疗并不是对每个人都有效。很明显，有必要揭示引发持续性、侵入性恐惧记忆的心理机制。我们的建议解决了研究中的一个裂痕，其中PTSD的主导理论-双重表征帐户-似乎与正常记忆的既定认知计算模型相冲突，并且受到其经验证据基础差异的挑战。我们建议对这一理论进行两次严格的检验。双重代表说提出了两个基本主张。首先，创伤使得很难将情感刺激与其时空背景联系起来，因此，创伤的情感方面与同一事件的中性元素之间的联系被削弱了。第二，侵入性记忆产生于一种独立的机制，而不是对同一创伤内容的自主记忆--这是第一种说法的结果。我们通过使用创伤的实验室模型来测试这两个核心主张，挑战双重表征帐户，其中参与者呈现描绘令人厌恶的内容（如受伤和死亡）的图片和电影剪辑。到目前为止，第一种说法的实验室测试涉及的研究是向参与者提供情感和中性刺激，然后测试他们对这些刺激之间联系的记忆。这种测试的解释是复杂的，因为结果不仅取决于编码过程中这些刺激的关键结合，而且还取决于维持和提取过程中情绪的后续影响。我们将通过定位编码阶段来消除这些潜在的混淆。我们采用了一个久经考验的学习任务来研究情感对联想绑定的影响，而不会污染后续记忆过程的测量。第二种说法的实验室证据是自愿和非自愿记忆之间的单一行为分离，当参与者在创伤暴露后积极参与视觉空间操作时，只有后者会减少。虽然这种分离已经被多次证明，但以前的实验测量自愿和非自愿记忆的方式之间存在差异，这可以解释这种差异。我们提供了一个更严格控制的实验设计来证实这种解离。然后，我们介绍了一个操纵的潜力，以增加自愿记忆。如果成功的话，视觉空间任务和我们的新操作将相当于自愿和非自愿记忆之间的双重分离，从而为双重表征解释提供更有力的支持。到目前为止，双重代表制账户大多是用文字描述的。这使得人们很难准确地理解所提出的认知计算机制。数学模型增加了理论的清晰度，使人们有可能做出和测试定量预测。如前所述，目前对双重表征的描述似乎与记忆的主流理论不一致。我们提出了一种方法，以适应现有的认知计算模型的中性和情绪记忆，使修订后的模型作出单独的预测自愿和非自愿的记忆。总之，我们的建议建立在既定的记忆模型的角度，这似乎不同意创伤后应激障碍的主导理论。我们使用这个角度来测试的双重表征帐户的经验，并开发一个正式的认知计算模型。如果我们提出的模型能够模拟我们获得的经验数据，包括我们可能观察到的自愿和非自愿记忆之间的双重分离，我们将调和一个重要的理论差距，从而促进对PTSD的理解。

Despite the high prevalence of post-traumatic stress disorder (PTSD) and its debilitating consequences, current psychological treatments are not effective for everyone. There is a clear need to uncover the psychological mechanisms that trigger persistent, intrusive fear memories. Our proposal tackles a rift in research, where a dominant theory of PTSD - the dual representation account - appears to clash with established cognitive-computational models of normal memory, and is challenged by discrepancies in its empirical evidence base. We propose to put the theory to two critical tests. The dual representation account makes two fundamental claims. First, that trauma makes it hard to bind emotional stimuli to their spatiotemporal context, and therefore, the associations between emotional aspects of the trauma and neutral elements of the same episode are weakened. Second, that intrusive memories arise from a separate mechanism than do voluntary memories of the same trauma content - a consequence of the first claim. We challenge the dual representation account by testing these two core claims using laboratory models of trauma, where participants are presented with pictures and film clips that depict aversive content, such as injury and death. Until now, laboratory tests of the first claim involved studies where participants were presented with emotional and neutral stimuli, and then tested for their memory for the association between these stimuli. The interpretation of such tests is complicated because the results depend not only on the critical binding of these stimuli during encoding, but also on subsequent effects of emotion during maintenance and retrieval. We will eliminate these potential confounds by targeting the encoding stage. We adapt a tried-and-true learning task to study the effect of emotion on associative binding without contaminating the measurement by subsequent memory processes. Laboratory evidence for the second claim is a single behavioural dissociation between voluntary and involuntary memory, where only the latter is decreased when participants actively engage in visuospatial manipulations following trauma exposure. Although this dissociation has been demonstrated multiple times, there are differences between the way that previous experiments measured voluntary and involuntary memory that could explain the discrepancy. We offer a more tightly controlled experimental design to confirm this dissociation. We then introduce a manipulation with potential to increase voluntary memory only. If successful, the visuospatial task and our new manipulation will amount to a double dissociation between voluntary and involuntary memory and thus offer stronger support for the dual representation account. To date, the dual-representation account has been mostly been described in words. This makes it difficult to understand the proposed cognitive-computational mechanisms precisely. Mathematical models increase the clarity of theories and make it possible to make, and test, quantitative predictions. As aforementioned, current descriptions of the dual representation account appear to disagree with mainstream theories of memory. We propose a way to adapt existing cognitive-computational models of neutral and emotional memory so that the revised model makes separate predictions for voluntary and involuntary memory. In summary, our proposal builds on the perspective of established models of memory, which appear to disagree with a dominant theory of PTSD. We use this perspective to test the dual-representation account empirically and develop a formal cognitive-computational model of this account. If our proposed model can simulate the empirical data we acquire, including the possible double dissociation we may observe between voluntary and involuntary memory, we will reconcile an important theoretical gap and thereby advance understanding of PTSD.