Credit: CC0 Public Domain
A new study challenges the longstanding belief that fear is primarily communicated through facial expressions, showing instead that context plays the dominant role in real-life fear recognition.
By analyzing real-life fear reactions in videos, researchers found that facial expressions alone fail to reliably signal fear, whereas situational context—such as the environment and body posture—allows for clear and accurate fear perception.
These findings have major implications for psychology, neuroscience, and artificial intelligence, suggesting that current emotion recognition models, including AI-based systems, need to incorporate contextual information rather than relying solely on facial cues. Recognizing fear in others is crucial for survival, but how is this achieved?
A new study published in PNAS, led by Professor Hillel Aviezer and Ph.D. student Maya Lecker from the department of Psychology at Hebrew University, challenges the widely accepted notion that fear is primarily communicated through facial expressions. Instead, the research finds that context, rather than facial reactions, plays a critical role in fear recognition.
For decades, emotion research has centered on the idea that facial expressions provide a clear, universal signal of fear. However, these studies often rely on posed expressions rather than genuine fear responses.
Overcoming significant practical challenges, Professor Aviezer and his team conducted an innovative study analyzing real-life reactions captured in videos during intense fear-inducing situations such as height jumps, physical attacks, and exposure to phobia triggers.
The study, which involved 12 preregistered experiments with a total of 4,180 participants, examined how people perceive fear in real-life scenarios. Participants were shown different versions of video clips: faces alone, context without faces, and full videos combining both elements. The researchers used various methods to assess fear perception, including forced-choice labeling, open-ended responses, and valence-arousal ratings.
The results revealed a striking contrast between commonly held assumptions and real-world observations:
Facial expressions alone failed to communicate fear reliably.
Context without faces, as well as full videos with context, led to clear and robust recognition of fear.
Many current models of emotion recognition—including AI-based emotion detection systems—rely heavily on facial cues. However, this study suggests that these systems may need to incorporate broader contextual understanding to accurately interpret emotions.
“Our research shows that despite the widely held belief that fear is expressed through a distinctive facial expression, real-life fear reactions tell a different story,” said Professor Aviezer. “Facial expressions alone carry minimal diagnostic value for fear recognition, while situational context plays a crucial role.”
This study paves the way for a more nuanced understanding of how humans perceive emotions, urging researchers, clinicians, and technology developers to reconsider existing models of emotion recognition.
More information:
Maya Lecker et al, Real-life intense fear is communicated through context, not facial expressions, Proceedings of the National Academy of Sciences (2025). DOI: 10.1073/pnas.2414677122
Provided by
Hebrew University of Jerusalem
Citation:
AI emotion detection may fall short: Study finds real-life fear is communicated through context, not facial cues (2025, March 13)
retrieved 15 March 2025
from https://medicalxpress.com/news/2025-03-ai-emotion-fall-short-real.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no
part may be reproduced without the written permission. The content is provided for information purposes only.
