Latent cause inference during extinction learning in trauma-exposed individuals with and without PTSD

Abstract

Problems in learning that sights, sounds, or situations that were once associated with danger have become safe (extinction learning) may explain why some individuals suffer prolonged psychological distress following traumatic experiences. Although simple associative learning models have been unable to provide a convincing account of how and why this learning fails, it has recently been proposed that this may be explained by individual differences in beliefs about the causal structure of the environment. Here, we tested two competing hypotheses as to how differences in causal inference might be related to trauma-related psychopathology, using extinction learning data collected from clinically well-characterized individuals with varying degrees of post-traumatic stress (N=56). Latent cause modelling revealed that individuals with more severe PTSD were more likely to assign observations from conditioning and extinction stages to a single underlying cause. Specifically, multivariate analysis incorporating multiple PTSD and depression symptom dimensions revealed a negative relationship between tendency to infer multiple causes were active in the environment and re-experiencing symptom severity. We interpret these results as providing evidence of a primary deficit in discriminative learning in participants with more severe PTSD re-experiencing symptoms. Specifically, in these individuals, a greater tendency to attribute all stimulus configurations to the same underlying cause resulted in greater uncertainty about stimulus-outcome associations, and impeded learning both that certain stimuli were safe, and that certain stimuli were no longer dangerous. Better understanding of the role of causal inference in trauma-related psychopathology may have relevance for the refinement of cognitive therapies for these disorders.

The Computations of a Traumatized Mind: A Latent Cause Model of Posttraumatic Stress Disorder

“…This idea fits with evidence that PTSD patients are predisposed toward dissociation—that is, toward creating compartmentalized memory traces, as in the case of trauma (Nijenhuis & den Boer, 2009; Nijenhuis et al, 2010) 25,26 . A recent study conducted by Norbury and colleagues, 27 however, has failed to find an enhanced spreading parameter in PTSD patients. Nevertheless, a possible reason for this finding might be the following confound.…”

“…As in other anxiety disorders, 29,30 PTSD patients exhibit greater fear generalization (e.g., they express more fear when presented with CS−) 31–33 . At first glance, this might be taken to suggest a lower, rather than higher, spreading parameter in PTSD: 27 fear generalization might derive from PTSD patients attributing both CS+ and CS− to the same LC, rather than envisaging two separate LCs, as controls would do. An alternative explanation, however, is that enhanced fear generalization arises because PTSD patients have enhanced sensitivity to punishment (i.e., because they attach a more negative value to the US).…”

“…Nevertheless, a possible reason for this finding might be the following confound. A common approach to assess the spreading parameter (also adopted in Norbury and colleagues 27 ) relies on classical conditioning paradigms, where one trial can be broken down into an initial phase and an outcome phase 10,12 . During the initial phase, CS+ or CS− stimuli (the former predicting the US, the latter predicting absence of the US) are presented, whereas during the outcome phase a US may or may not be delivered.…”

“…The assumption is that, at each phase, the brain infers the LC in play during the trial and that the intensity (or probability) of the fear response (manifested during the initial phase) reflects this inference. Specifically, the intensity of the fear response is typically assumed to reflect beliefs about the probability of potential LCs 10,12,27 . For example, if one believes that a US-related LC is more likely than a safe LC, then fear will be more intense.…”

“…When modeling fear responses to assess inference about LCs, including the role of value is fundamental. This role was not considered by Norbury and colleagues, 27 an aspect that might have affected substantially the analyses, especially given evidence that PTSD patients have enhanced sensitivity to punishment 28 . Future research needs to investigate the link between PTSD and the spreading parameter while controlling for the value of any US.…”