Contextual Novelty and Visuospatial Memory: A Data-Driven Multivariate EEG Approach

Contextual novelty can enhance access to stored information, a process relevant both to memory and to the flexible recombination of representations required for divergent thinking. We examined whether novelty influences visuospatial memory and creativity, and whether multivariate EEG dynamics predict memory performance. Twenty-four adults were assigned to a Novelty group, encountering the EEG room for the first time during testing, or a Habituation group with prior room exposure. Participants completed the Rey–Osterrieth Complex Figure (ROCF) and Alternate Uses Task (AUT) with high-density EEG. The Novelty group showed significantly better ROCF performance, whereas AUT performance did not differ between groups. Across resting baseline, ROCF recall, and AUT, we extracted band-specific measures of spectral and aperiodic activity, phase and symbolic connectivity, entropy, complexity, and directional information flow across five brain regions. PCA summarized the standardized features. An Elastic Net model predicted ROCF performance from baseline component scores, task-related changes (Δrecall–baseline and ΔAUT–baseline), group, and Group×component interactions, using leave-one-out validation with internal hyperparameter selection. The model achieved R²=0.624, r=0.812, and RMSE=0.069. These findings suggest that novelty-related memory enhancement is associated with multivariate EEG configurations shaped by baseline brain state, task-induced changes, and experimental context.