Uncanny valley effect: A qualitative synthesis of empirical research to assess the suitability of using virtual faces in psychological research

Recently, virtual faces are often used as stimuli to replace traditional photographs in human face perception studies. However, despite being increasingly human-like and realistic, they still present flaws in their aspects that might elicit eerie feelings in the observers, known as the Uncanny Valley (UV) effect. The current systematic review offers a qualitative synthesis of empirical studies investigating observers’ subjective experience with virtual compared to real faces to discuss the possible challenges that the UV effect poses when virtual faces are used as stimuli to study face perception. Results: revealed that virtual faces are judged eerier than real faces. Perception of uncanniness represents a challenge in face perception research as it has been associated with negative emotions and avoidance behaviors that might influence observers’ responses to these stimuli. Also, observers perceive virtual faces as more deviating from familiar patterns than real faces. Lower perceptual familiarity might have several implications in face perception research, as virtual faces might be considered as a category of stimuli distinct from real faces and therefore processed less efficiently. In conclusion, our findings suggest that researchers should be cautious in using these stimuli to study face perception.