When the Eyes Disappear: Neural Dynamics of Emotion Recognition under Eye Occlusion

Abstract Facial emotion recognition relies on the rapid integration of visual cues, with the eye region playing a central role in both perceptual and evaluative stages of processing. While reduced eye visibility is known to affect emotion recognition, different forms of eye occlusion are often treated as equivalent, despite evidence suggesting that they may not alter available information in the same way. The present pilot EEG study examines how natural (closed eyes) versus artificial (sunglasses) eye occlusion modulates the temporal dynamics of emotional face processing. Participants categorised facial expressions (fearful, happy, neutral) while EEG was recorded, allowing the examination of early perceptual (N170), intermediate affective (EPN), and later evaluative (LPP) components. Results indicate that early face-sensitive processing responds differently to natural eye states and artificial occlusion, with N170 amplitudes differing between open and closed eyes relative to sunglasses. In contrast, the EPN showed sensitivity to emotional content but appeared unaffected by eye occlusion. At later stages, LPP amplitudes also differed between natural and artificial occlusion, despite comparable reductions in visual access to the eyes. Together, these patterns suggest that natural eye occlusion convey information that is qualitatively different from that available with open eyes, rather than reflecting a mere loss of information, in line with previous behavioural evidence (Granja & Burra, in prep). By contrast, artificial occlusion may primarily reduce available visual input without introducing a comparable reappraisal of eye-related information. As a pilot study, these findings are not intended to provide conclusions, but rather to support the formulation of targeted hypotheses and the design of a larger, hypothesis-driven experiment aimed at disentangling perceptual and evaluative mechanisms underlying different forms of eye occlusion in emotion recognition.