Comparing episodic memory binding outcomes after emotion induction in virtual reality

Abstract The episodic memory (EM) system maintains coherent representations of events by binding central information and contextual elements into unified episodes. Although emotional events are typically remembered more vividly than neutral ones, this enhancement preferentially benefits central information, often at the cost of contextual details. While prior research has explored this differential effect in simplified events using words or static images, their ecological validity remains limited. Furthermore, while face-name and face-context bindings rely on distinct neurocognitive mechanisms, the differential impact of emotion on these binding types remains poorly understood. This study examines the influence of an induced emotional state on two EM binding measures, face-name and face-context associations. A naturalistic virtual reality-based incidental learning task was administered to 44 participants, to compare performance between emotionally neutral and high-arousal, negative-valence conditions. A novel emotion induction procedure was implemented by combining audiovisual materials across pre-encoding and encoding phases. Twenty-four study items were presented sequentially within two distinct virtual environments, each consisting of a 3D character with a photorealistic face and a unique name. Results revealed that high-arousal negative-valence emotional states significantly increase accuracy of face-name associations, while they attenuate accuracy of face-context associations, particularly in post-encoding measurements after a 24-hour delay. Emotion selectively influenced EM binding, with task-dependent conceptual relations between elements determining how central and contextual event details were differentially encoded.