art.pics Database: An Open Access Database for Art Stimuli for Experimental Research

Area of Science:

• Neuroscience
• Psychology
• Computer Science

Background:

• Neural mechanisms of art perception and valuation remain underexplored.
• Existing art stimuli lack standardized definitions and normative ratings, hindering research.
• Functional magnetic resonance imaging (fMRI) studies suggest art engages reward pathways similar to primary rewards.

Purpose of the Study:

• To introduce a novel, large, and well-characterized database of art images for neuroimaging and psychological research.
• To provide standardized visual stimuli with normative ratings for consistent experimental design.
• To facilitate the study of neuroesthetics by offering a resource comparable to existing visual stimuli databases.

Main Methods:

• A deep learning algorithm generated 2,332 novel art stimuli by applying various painting styles to common object, animal, and plant images.
• Image characteristics (size, color, complexity, intensity, valence) were analyzed using software.
• A human experimental validation study (n=1,296) collected subjective ratings for liking, wanting, recognizability, arousal, and valence.

Main Results:

• The "art.pics" database offers 2,332 standardized art images with comprehensive normative data.
• Image properties are comparable to established visual stimuli databases.
• Validation study confirmed diverse subjective value ratings and distinct variations across art styles and categories.

Conclusions:

• The "art.pics" database provides a crucial, open-access resource for advancing neuroesthetics research.
• Standardized stimuli and normative data enable robust investigation of art perception and valuation.
• The database supports research comparing art as a reward to other primary rewards.