AbstractLearning to avoid threats often occurs by observing the behavior of others. Most previous research on observational fear learning in humans has used pre-recorded stimuli as social cues. Here, we aimed to enhance the ecological validity of the learning situation: the ‘observer’ watched their friend (‘demonstrator’) performing a differential fear-conditioning task in real time. During the task, one conditioned stimulus (CS+) was repeatedly linked with electric stimulation (US) while another one (CS-) was always safe. Subsequently, the observer was presented with the CS+ and CS- directly but without receiving any shocks. Skin conductance (SCR) and fear-potentiated startle (FPS) responses were measured in observers throughout the whole experiment. While the US applied to the demonstrator elicited strong SCR in the observers, subsequent differential SCR to CSs (CS+ vs. CS-) presented directly were dependent on declarative knowledge of the CS+/US contingency. Contingency-aware observers also showed elevated FPS during both CS+ and CS- compared to intertrial intervals. We conclude that observational fear learning involves two components: an automatic emotional reaction to the response of the demonstrator and learning to predict stimulus contingency (CS+/US pairing). Ecological modifications proposed offer new perspectives on studying social learning of emotions.
Human fear conditioning is moderated by stimulus contingency instructions
