Human emotional experiences naturally occur while interacting in a spontaneous, dynamic and response contingent fashion with other humans. This resonates with both theoretical considerations as well as neuroimaging findings that illustrate the nexus between the “social” and “emotional” brain suggesting a domain-general organization of the brain. Nevertheless, most knowledge in affective neuroscience stems from studying the brain in isolation from its natural social environment. Whether social interactions are constitutive or not to the understanding of other people's intentions, incorporating such interactions is clearly required for ecological validity. Moreover, since interpersonal interactions may influence emotional experiences and expressions, interactive paradigms may advance the theoretical understanding of what emotions are and what about them is social, and will correspondingly characterize their underlying neural substrates. We highlight the recent conceptual and experimental advances of bringing realistic social interactions into the neuroimaging lab; review emotion-induction paradigms and consider their congruency with features of social interactions; and emphasize the importance of embedding such spontaneous and dynamic interactive paradigms in the field of affective neuroscience.