Standardized database of 400 complex abstract fractals

Area of Science:

• Cognitive psychology
• Neuroscience
• Computer science

Background:

• Stimulus characteristics significantly impact cognitive processes in experimental research.
• A lack of standardized data exists for complex, abstract stimuli, hindering experimental control.
• Existing databases often lack comprehensive feature information for abstract stimuli.

Purpose of the Study:

• To develop and validate a standardized database of 400 abstract and complex fractal stimuli.
• To provide normative data on key stimulus features for use in experimental settings.
• To establish external validation for the collected stimulus feature norms.

Main Methods:

• Collected ratings from 512 participants on features like abstractness, complexity, and memorability for 400 fractal stimuli.
• Obtained naming data from 111 participants to calculate naming agreement and modal names.
• Validated stimulus features using a deep neural network classifier, image compression rates, and a recognition-memory task.

Main Results:

• The fractal stimuli were confirmed as highly abstract and difficult to verbalize.
• External validation demonstrated correlations between feature ratings and objective measures (e.g., complexity and compression rate).
• Memorability ratings positively correlated with performance in a recognition-memory task.

Conclusions:

• The study successfully created a validated database of abstract stimuli with comprehensive normative data.
• This database serves as a valuable resource for researchers needing to control or manipulate stimulus features in cognitive experiments.
• The validated norms enable precise experimental control over factors like verbalizability, novelty, and familiarity in abstract stimulus research.