The Scarred Circuitry of Fear: A Computational–Clinical Synthesis of PTSD Neurobiology

Post-traumatic stress disorder (PTSD) can be understood as a disorder of threat inference: after trauma, the brain assigns excessive probability and excessive cost to danger, while failing to consolidate safety when objective contingencies change. This article synthesizes convergent findings across fear-circuit neurobiology—amygdala reactivity, ventromedial prefrontal and anterior cingulate regulatory control, and hippocampal context processing—together with stress-system and neuromodulatory mechanisms (noradrenergic gain control and heterogeneous hypothalamic–pituitary–adrenal axis alterations). The aim is not to restate the circuit model, but to formalize how these components jointly generate hallmark clinical phenomena: cue-triggered intrusions, hyperarousal, avoidance, and fear generalization with context-dependent relapse. To that end, the paper proposes a minimal multi-scale computational framework that links (i) a gain-modulated threat–control dynamical system (capturing defensive attractor dynamics and state-dependent control collapse) with (ii) latent-context learning models of extinction and renewal. Key parameters are mapped to measurable proxies and intervention targets, yielding falsifiable predictions about individual differences in extinction retention, renewal, stress recovery time, and treatment response. Throughout, claims are framed at the level supported by current evidence, emphasizing heterogeneity, moderators, and the limits of biomarker determinism.