Neural Dynamics Underlying Repeated Learning of Visual Image Sequences

Abstract

Humans possess a remarkable ability to recognize visual objects with high fidelity, supported by complex neural mechanisms underlying memory retrieval. Event-related potential (ERP) studies have identified two key neural signatures of recognition memory: the parietal old/new effect and the frontal old/new effect. Despite extensive research on these ERP components, the extent to which these components reflect distinct memory processes remains debated. In the present study, we investigated how repetitive learning modulates these ERP components. Participants repeatedly studied a fixed list of 32 real-world images across up to five study-test repetitions while EEG was recorded. Additionally, a separate set size 1 condition served as a proxy for working memory. Our results showed that with increased repetitions, the parietal old/new effect exhibited enhanced amplitude and earlier peak latency, reflecting more efficient retrieval of well-learned memories. In contrast, the frontal old/new effect remained unchanged in both amplitude and timing. These findings suggest that the parietal old/new effect is a sensitive neural marker of learning-related changes in long-term memory representations, while the frontal effect is less influenced by repetition. Additionally, despite similarly high accuracy between the well-practiced set size 32 condition and the set size 1 working memory condition, both parietal and frontal old/new effects peaked significantly earlier for set size 1, suggesting that access to working memory is substantially faster than even well-practiced long-term memory. Together, our results highlight the unique role of the parietal old/new effect, but not the frontal old/new effect, in repetitive learning, despite both components being important for successful recognition of learnt visual stimuli.