Emotion Ratings of Motion Captured Human Gait

Read full text
Authors: Liam M Crowley, Hedwig Eisenbarth
Cited by: 0

Abstract

This study investigates how people perceive emotional expression from different regions of the body during gait. Prior research has shown that humans perceive emotions from not only the face, but also the body (and specficially gait). We also know that even when viewing emotional expressions of even isolated body parts such as an arm, participants are able to perceive the expressed emotion. What is not clear, however, is how much the upper versus lower body contributes to perceived emotion during gait, and whether those relative contributions are emotion-specific (e.g., some emotions may be carried more by torso/arms than by legs/stride, or vice versa). We know emotion can be read from gait, and even isolated body parts, but what is yet to be known is where in the moving body that information is most diagnostic, even when different regions convey competing signals. Participants will view short videos of human gait rendered onto a digital avatar. The gait in each video will express one of five emotions (joy, sadness, fear, anger, disgust) or a combination of the two. On a given trial either the entire body will express a single emotion (congruent) or the top half will express one, while the bottom half expresses another (incongruent). Participants will rate the videos on how much they think the avatar expresses each of the five emotions (0-100). We will also derive a “selected emotion” based on which emotion was rated as highest on a given trial. In the case of a tie, participants will be prompted to select which emotion they believe the avatar is expressing the most. The primary goal of this research is to quantify the relative contributions of the top and bottom half of the body for the perception of expressed emotion through gait, as well as whether these contributions differ by emotion.

Fields of Study & Subjects

Health Education and Physical Culture Face Recognition and Perception Emotion and Mood Recognition

Publication details

Journal / source
Open Science Framework
Publication date
1 January 2026
Language
EN