Traumatic events lead to distressing memories, but such memories are made all the worse when they intrude to mind unbidden and recurrently. Intrusive memories and flashbacks after trauma are prominent in several mental disorders, including post-traumatic stress disorder and can persist for years. Critically, the reduction of intrusive memories provides a treatment target. While cognitive and descriptive models for psychological trauma exist, these lack formal quantitative structure and robust empirical validation. Here, using techniques from stochastic process theory, we develop a mechanistically driven, quantitative framework to extend understanding of the temporal dynamic processes of trauma memory. Our approach is to develop a probabilistic description of memory mechanisms to link to the broader goals of trauma treatment. We show how the marginal gains of treatments for intrusive memories can be enhanced as key properties (intervention strength and reminder strength) of the intervention and memory consolidation (probability memories are labile) vary. Parametrizing the framework with empirical data highlights that while emerging interventions to reduce occurrence of intrusive memories can be effective, counterintuitively, weakening multiple reactivation cues may help reduce intrusive memories more than would stronger cues. More broadly, the approach provides a quantitative framework for associating neural mechanisms of memory with broader cognitive processes.
Keywords: Markov chains; intrusive memories; probability; reconsolidation; stochastic dynamics.