Encoding of modality-specific face engrams promotes distinct recruitment of mnemonic processing mechanisms: A mobile-EEG study comparing encoding and retrieval of 2D and 3D avatars under virtual reality conditions

• Mobile-EEG VR study compares face recognition from 2D images and 3D avatars • Recognition performance is increased for faces encoded in the dynamic 3D condition • 3D face retrieval elicited distinct ERP old/new effects (N400f, P600f) • Induced Theta and alpha oscillations reveal modality-specific mnemonic mechanisms • Retrieval of 3D face engrams engaged a specific interplay of identity conceptualization, memory access, and attentional engagement The processing of faces under realistic virtual conditions engages distinct cognitive mechanisms that are attuned to their inherent conceptual and socially relevant characteristics. To more closely approximate real-life circumstances of personal encounters, the degree of realism is increased by situational context and further varied with respect to the stimuli’s dimensionality and dynamics. To this end, a mobile-EEG study was conducted in which the participants encountered dynamic 3D virtual avatars and static 2D images of the avatars and were then asked to recognize their faces in a 2D old/new task, all within the same virtual environment. Stimulus processing during encoding and mnemonic mechanisms engaged during retrieval were investigated based on relevant event-related potentials (ERPs) as well as induced oscillatory responses. Although cognitive load did not differ between processing 2D and 3D faces, as indexed by comparable induced theta-band responses, recognition performance during retrieval showed that 3D faces were recognized more accurately. Investigation of the underlying cognitive processing revealed distinct recruitment of mnemonic mechanisms for the retrieval of 3D encoded faces, reflected in the face-specific frontal and centro-parietal ERP old/new effects (N400f, P600f) and characteristic responses in the induced theta and alpha bands. Specifically, the reactivation of the 3D face representations was characterized by concurrent identity-related and semantic recognition processing as well as increased attentional resource allocation. The realistic characteristics of the 3D faces, in terms of visual complexity and social quality, led to the formation of modality-specific engrams, which ultimately enabled better recognition of these faces than their 2D counterparts.