Visual semantics signature

Hedonic value signature

RSA_results

Subcortical results

FLT_results

RSA searchlight validations

Template arrangement

ARTyficial_signatures.scene

FULL TITLE:
Latent neural architecture organizing shared aesthetic evaluations of visual artworks

SPECIES:
Human

ABSTRACT:
Aesthetic experience shapes a wide range of behaviours, from everyday consumer choices to the appreciation of visual art. While theoretical accounts propose that such experiences emerge from interactions among core brain systems during aesthetic evaluations, the principal dimensions organizing this mental process remain unclear. Here we introduce a data-driven framework to characterize the latent neural architecture of shared aesthetic evaluations. Participants viewed traditional paintings during 7T functional MRI and subsequently rated their aesthetic appreciation of each artwork. Drawing on approaches used in online recommender systems, we constructed an aesthetic agreement matrix capturing the similarity structure of participants’ ratings and applied dimensionality-reduction techniques to identify the main axes organizing aesthetic evaluations. Two principal dimensions emerged: visual semantics and hedonic valuation, each predicted by distinct multivariate neural signatures. Category-selective regions along the ventral visual stream predicted variation in visual semantics, whereas medial prefrontal and subcortical valuation circuitry predicted variation in hedonic value. Importantly, individual differences in the neural representation of this two-dimensional latent aesthetic space, particularly within key regions of the default mode network, were systematically associated with expertise in the visual arts. Together, these findings reveal how interacting core brain systems synergistically organize shared aesthetic evaluations of visual artworks while also accounting for idiosyncratic aspects of aesthetic experience.

PUBLICATION:
Nature Communications