Affective and computational determinants of threat extinction biases

Pavlovian threat acquisition and extinction are fundamental processes by which individuals learn about threat and safety in their environment. Research has shown that humans learn more rapidly and persistently to associate threatening and—somewhat counterintuitively— positive rewarding stimuli with aversive events, beyond their inherent threat value. However, the computational mechanisms underlying these Pavlovian learning biases remain unclear. Here, I examined the affective and computational determinants of Pavlovian threat acquisition and extinction biases for threat-related and positive emotional stimuli. I combined and reanalyzed data from four experiments (N = 247) using a differential Pavlovian threat conditioning paradigm. Computational modeling techniques were applied to identify signatures of these learning biases. Threat-relevant (angry faces, snakes), positive-relevant (baby faces, happy faces, erotic stimuli), and neutral (neutral faces, colored squares) stimuli were used as conditioned stimuli, with skin conductance response serving as an index of learning. Model comparison indicated that a reinforcement-learning model differentiating between excitatory (e.g., learning from reinforcement) and inhibitory (e.g., learning from the absence of reinforcement) learning was the best explanatory model for the observed data. Whereas no evidence for differences in excitatory learning rates was found between stimulus categories, both threat- and positive-relevant stimuli exhibited a lower inhibitory learning rate compared to neutral stimuli, contributing to the persistence of the conditioned response during extinction. These findings provide insights into the mechanisms underlying Pavlovian threat extinction biases, thereby advancing our understanding of how humans attribute and update affective value to their environment.