Does visual format matter in brand name recognition? ERP evidence challenging abstractionist models

Visual word recognition is resilient to changes in surface-level features (e.g., font, letter case, color, size). This resilience has led to the common assumption that lexical memory relies on abstract representations. However, recent behavioral studies suggest that this may not hold for brand names-a type of words usually encountered in highly consistent visual formats. The present experiment examined the electrophysiological signature of font and letter case deviations in brand names embedded in logotypes, focusing on how these modifications, relative to their original format, modulate the neural dynamics of lexico-semantic processing. To this end, we recorded event-related potentials (ERPs) while participants performed a semantic categorization task on brand names ("Is the brand related to transportation?") presented in their original format or with a modified font or case. Letter case modifications reduced P200 amplitudes, reflecting early perceptual costs. Critically, both font and case manipulations also modulated the P300 and N400 amplitudes. These findings challenge strong abstractionist models in which visual format is fully normalized before lexical access and instead favor a hybrid account in which format-specific information is retained for familiar brand names and continues to influence lexico-semantic processing.