Behind and Beyond the Screen: Neural Differences Between Live vs. Video-Based Action Perception Under Attentional Load

Abstract

Perceiving others actions is essential for survival, interaction, and communication, yet most neuroscience studies rely on 2D videos or images that lack the presence and social affordances of real actions. This limits our understanding of real-world action perception and the development of neurally grounded models. Here, we directly compare behavioral and neural responses to real (live) versus video-based actions. Using a novel experimental setup (Pekcetin et al. 2023), we conducted a two-session EEG study (N = 26) in which participants viewed peripheral actions presented live or via video while performing a central task under low and high attentional load. We examined behavioral performance, mass-univariate ERPs, time-frequency responses, and time-resolved representational similarity (RSA). Behaviorally, real actions imposed a greater cognitive cost than video actions, with the largest "Realness Effect" under high load. ERPs showed reliable Live-Video differences within 150-450 ms after action onset. Time-frequency analyses over occipital and parietal regions revealed weaker alpha (8-12 Hz) and beta (15-25 Hz) suppression for video actions, indicating reduced perceptual engagement. Time-resolved RSA also robustly separated live and video conditions between 250-750 ms. Together, these results show that live actions engage perceptual systems more strongly than their video-based counterparts, underscoring the limitations of screen-mediated paradigms and motivating more ecologically grounded approaches in social and action perception research.