Computational Models of Post-traumatic Stress Disorder (PTSD)

Post-traumatic stress disorder (PTSD) is a multifaceted, often chronic syndrome that may develop following an exposure to a highly traumatic event (e.g., natural disaster, military combat, sexual assault). Besides requiring that individuals have experienced trauma, diagnostic criteria also require at least some symptoms from several clusters, including re-experiencing, avoidance, negative changes in cognition and mood, and changes in arousal and reactivity. These symptoms must persist for at least a month, interfere with daily life, and are not best explained by a substance abuse disorder or other medical condition (American Psychiatric Association, 2013). Given the complexity, heterogeneity, and incomplete theoretical description of this disorder, computational models may be an important tool to help understand the mechanisms, and course, of symptom presentation and maintenance in PTSD. A comprehensive computational model of PTSD should explain the mechanisms that underlie all symptom clusters along with individual differences in susceptibility and symptom severity, comorbidity with other disorders, and should generate novel predictions (Brewin, Dalgleish, & Joseph, 1996; Jones & Barlow, 1990). For example, not all individuals exposed to trauma develop PTSD, and a number of physiological, cognitive, and social factors could contribute to vulnerability. In the general population, women are twice as likely as men to develop PTSD (Altemus, Sarvaiya, & Epperson, 2014). Individuals with particular symptom profiles may also be more vulnerable. In particular, avoidance symptoms are associated with greater susceptibility (North et al., 2004) and a more chronic course (Maes et al., 1998). This chapter will review several computational models and their contributions to understanding PTSD, and highlight important future directions. None of the models to date has provided a comprehensive account of PTSD—most focus on particular symptoms. The subsequent sections will discuss computational models focused on fear learning and expression, changes in arousal and reactivity, avoidance, changes in cognition and mood, and intrusive recollection.