Latent neural architecture organising shared aesthetic evaluations of visual artworks.

Abstract

Aesthetic experience shapes behaviours ranging from everyday consumer choices to art appreciation. Although theoretical accounts propose that such experiences emerge from interactions among core brain systems, the dimensions organising this mental process remain unclear. Here, we introduce a data-driven framework to characterise the latent neural architecture of shared aesthetic evaluations. Participants viewed traditional paintings during 7T functional magnetic resonance imaging (fMRI), and we applied dimensionality reduction to the similarity structure of their aesthetic ratings to identify the dominant axes of variation. Two principal dimensions emerged, visual semantics and hedonic valuation, each predicted by dissociable multivariate neural signatures. Category-selective regions along the ventral visual stream tracked variation in visual semantics, whereas medial prefrontal and subcortical circuitry tracked hedonic valuation. Moreover, individual differences in this latent aesthetic space, particularly within default mode network regions, scaled with visual art expertise. Together, these findings reveal how core brain systems synergistically organise shared aesthetic evaluations of visual artworks.