Generating an item pool for translational social cognition research: Methodology and initial validation

摘要

Existing sets of social and emotional stimuli suitable for social cognition research are limited in many ways, including size, unimodal stimulus delivery, and restriction to major universal emotions. Existing measures of social cognition could be improved by taking advantage of item response theory and adaptive testing technology to develop instruments that obtain more efficient measures of multimodal social cognition. However, for this to be possible, large pools of emotional stimuli must be obtained and validated. We present the development of a large, high-quality multimedia stimulus set produced by professional adult and child actors (ages 5 to 74) containing both visual and vocal emotional expressions. We obtained over 74,000 audiovisual recordings of a wide array of emotional and social behaviors, including the main universal emotions (happiness, sadness, anger, fear, disgust, and surprise), as well as more complex social expressions (pride, affection, sarcasm, jealousy, and shame). The actors generated a high quantity of technically superior, ecologically valid stimuli that were digitized, archived, and rated for accuracy and intensity of expressions. A subset of these facial and vocal expressions of emotion and social behavior were submitted for quantitative ratings to generate parameters for validity and discriminability. These stimuli are suitable for affective neuroscience-based psychometric tests, functional neuroimaging, and social cognitive rehabilitation programs. The purposes of this report are to describe themethod of obtaining and validating this database and to make it accessible to the scientific community.