Final report of DFG-funded projects: A face to be loved or feared? Emotional and social mediation of verbal threat learning (first funding period). A face to be loved or feared? A follow-up on interindividual differences and clinical implications (se

This project investigated how verbal learning about potential threats modulates face and person perception. Various social-emotional factors were examined (e.g., emotional facial expressions, face identity, and pictures of loved ones), either as threat or safety cues or in threatening or safe contextual settings. Perceptual processing, memory encoding and recognition, as well as psychophysiological responding and overt behavior were examined. In addition to replicating previous findings on the processing of threats and faces, several important findings can be highlighted from this project. 1) Verbal threat instructions lead to selective perceptual processing, encoding and activation of psychophysiological defense systems, when confronted with presumed threat cues or contexts. Here, it does not matter that the threat cue is actually never followed by the anticipated aversive event. 2) Verbal threat instructions flexibly change previous face-, person-, and context-related associations. The instructions easily override the implicit affective meaning of facial stimuli, e.g. a smiling person or even pictures of familiar loved ones can easily become new threat cues. 3) Instructed contextual threat or safety changes how we perceive and respond to other people. Context helps to understand ambiguous information (e.g., to better recognize a subtle fearful expression in a threatening environment), and to draw attention to deviant and unexpected information (e.g., a smiling face in an aversive situation). Moreover, simply looking at a picture of a smiling person or a loved one does not impede functional defensive responding to aversive experimental conditions. 4) Verbal threat instructions lead to surprisingly persistent defensive reactions that are less accessible to passive extinction learning, i.e. the proof of safety is particularly difficult if threat has been acquired socially (i.e., without own CS-US experience). 5) These threat-biases likely reflect the workings of healthy functional and adaptive mechanisms that are partially subject to interindividual differences. For example, individuals with adverse childhood experiences showed intact selective threat processing for newly acquired threat associations, and a transdiagnostic sample (including borderline and PTSD patients) showed adaptive threat-potentiated startle to instructed threat cues. 6) Clinical relevance arises in particular with regard to threat-safety reversal learning, which involves both the acquisition of new and concurrent inhibition of previous threat associations. Here, in individuals with subclinical and clinical relevant anxious psychopathology, for example, less neural differentiation between new and old threat cues was observed as well as enhanced distrust to actually safe persons. Overall, the role of social learning in person perception and psychophysiological responding is still under-researched. Here, the project results come together to form an integrative neuroscientific model of face and person perception as a function of social learning, with implications for clinical research.