Abstract Accurately inferring others’ emotions from whole-body motion is essential for effective social interaction; however, the specific movement patterns that signal distinct emotions, as well as their causal status, remain elusive. This study combined behavioural assessment with kinematic decomposition to isolate and manipulate gait components influencing emotion recognition. In Experiment 1, we created point-light videos using motion-capture data of walking movements. Participants then reported the perceived emotion in the videos. The motion-capture data were subjected to principal component analysis decomposing coordinated movement patterns. We then investigated relationships between these components and perceived emotions. The second principal component (PC2), reflecting the coordinated arm and leg swings, varied systematically with perceived emotion. In Experiment 2, we undertook a similar task using stimuli in which PC2 scores were scaled to resemble the characteristics of anger, sadness, or fear (based on gaits judged to be neutral). Participants’ judgements shifted significantly in the predicted direction, with manipulated gaits more frequently judged as angry, sad or fearful. These findings suggest that specific movement patterns can independently and causally influence emotion recognition. Our approach offers an effective framework for isolating and manipulating dynamic features within complex movements, thereby advancing understanding of emotional, aesthetic and technical evaluations of movement.
Wakabayashi et al. (2026) studied this question.