Distinguishing what predicts danger and what does not and responding accordingly are essential to survival. Pavlovian conditioning is the process through which humans, and many other animals, learn to navigate these critical issues (Pavlov, 1927). For instance, one readily acquires and maintains fear of conditional stimuli (CS) that have predicted aversive, unconditional stimuli (US). During an encounter with the CS, one’s defensive behavior system is activated (Fanselow, 1994). Based on the perceived level of danger, cognitive, physiological, and behavioral effectors can be recruited to prepare for defense, the coordinated activation of which is considered “fear” (Fanselow & Pennington, 2018). When the CS no longer reliably predicts the US, conditional fear of the CS typically extinguishes, a process called “extinction.” Failure to extinguish conditional fear and effectively update threat appraisal of the CS despite disconfirming evidence may impair functioning. When behavioral, physiological, and cognitive resources remain mobilized toward defense in the absence of danger, or are recruited disproportionately to the real danger, goal-directed behavior including attentional control, decision-making, and avoidance become disrupted (Eysenck et al., 2023; Hartley & Phelps, 2012; Zorowitz et al., 2020). A wide range of fear-based psychopathology, including anxiety and posttraumatic stress disorders, is conceptualized as due, at least in part, to this failure (Kausche et al., 2025). Extinguished fear can resurge through several pathways after extinction, posing a long-standing challenge to the treatment of fear-based psychopathology. Spontaneous recovery, for example, occurs over time without continued experiences of the CS without the US (Rescorla, 2004). Another pathway is reinstatement, which occurs following unexpected experiences of the US alone (Bouton & Bolles, 1979; Rescorla & Heth, 1975). Individuals can also re-acquire conditional fear rapidly, should the CS and US be paired again (i.e., re-acquisition) (Leung et al., 2007). The resurgence of conditional fear has been attributed to the preservation of fear memory and contextual dependence of the retrieval of extinction memory (Bouton, 2002, 2004). According to Bouton’s theory, rather than unlearning the CS-US association, extinction results in new learning of a secondary, inhibitory CS-noUS association. This learning inhibits (momentarily) the expression of conditional fear but is confined to the spatiotemporal context where extinction is perceived to have occurred. Both passage of time after extinction and US presentations, whether unpaired or paired with the CS, which are both absent during extinction, may thus be perceived as a shift in context from extinction, leading to a return of fear. Notably, such return is estimated to occur in up to 62% of individuals (Craske & Mystkowski, 2006). The high return-of-fear rate may explain why exposure therapy, the clinical proxy of extinction and first-line treatment for anxiety and fear-related disorders, results in a long-term remission rate of only about 55%, despite large effect sizes in shorter-term fear reduction (Loerinc et al., 2015). There has been some evidence in the laboratory that gradual extinction is effective at attenuating the return of fear (Gershman et al., 2013; Kennedy et al., 2024; Quintero et al., 2022; Shiban et al., 2015). An alternative approach to extinction based on contemporary theoretical models of associative learning, gradual extinction introduces prediction error in a progressive manner to facilitate updating of original fear learning, instead of maximizing prediction error to encode new, inhibitory learning as conventionally emphasized in standard practices of fear extinction. The effects of gradual extinction might conceivably be heterogeneous due to individual differences. Much of human fear conditioning research focuses on the average effects of experimental manipulations, likely not because individual differences are not important but because often studies are underpowered. There is emerging evidence that there are meaningful individual differences in fear acquisition, extinction, and performance at tests of return of fear, although findings have been mixed and sparse (Kausche et al., 2025; Lonsdorf & Merz, 2017; Sjouwerman et al., 2020). Evaluation of novel methods of augmenting extinction should consider the heterogeneity of effects, to facilitate clinical translation. After all, the audience who would benefit from the clinical translation of this work is highly heterogeneous in composition. Among many factors to consider, severity of internalizing symptoms and trait positive affectivity may be of particular importance and relevance to research assessing the effects of gradual extinction. On one hand, in the therapeutic context, those who receive exposure-based treatment for fear-based psychopathology presumably have more severe internalizing symptoms than their average counterparts in the general population. Prior work suggests that anxiety- and stress-related vulnerability may be particularly associated with altered performance of extinction retention and consequently return of fear, including reduced vmPFC activation during extinction recall in anxiety disorders (Marin et al., 2017), poorer CS+/CS− differentiation following reinstatement among individuals with higher state anxiety (Kuhn et al., 2016), and heightened threat expectancy or negative affect ratings during extinction and extinction recall in anxiety-, obsessive-compulsive-, and trauma-related disorders (Kausche et al., 2025). Furthermore, there is evidence that stress impairs extinction and its recall (Maren & Holmes, 2016; Raio et al., 2014). On the other hand, trait positive affectivity may be a protective factor, facilitating learning processes of key relevance to extinction. There is evidence that trait positive affect predicts faster symptom reduction and lower follow-up symptoms for social anxiety disorder (Taylor et al., 2023). Furthermore, other work demonstrates that the induction of momentary positive affect predicts less re-acquisition of fear and attenuates reinstatement (Zbozinek et al., 2015; Zbozinek & Craske, 2017). Conversely, research evaluating the effects of trait negative affectivity is limited, although related constructs such as neuroticism and intolerance of uncertainty (IU) have received more attention (Lonsdorf & Merz, 2017; Sjouwerman et al., 2020). We include trait negative affectivity as a theoretically relevant comparator, given its conceptual overlap with internalizing symptoms and broader negative emotionality.