Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Comparing Activation Typicality and Sparsity in a Deep CNN to Predict Facial Beauty.

Sonia Tieo1, Melvin Bardin2, Roland Bertin-Johannet3

  • 1CEFE, Univ Montpellier, CNRS, EPHE, IRD, Montpellier, France.

Computational Brain & Behavior
|July 16, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Sexual selection and speciation: a meta-analysis of comparative studies.

Evolution letters·2025
Same author

Nest weave pattern in weaverbirds: a sexual signal selected through sensory drive?

Evolution; international journal of organic evolution·2025
Same author

On the biological basis of beauty.

Biological reviews of the Cambridge Philosophical Society·2025
Same author

Checklist, key and latitudinal distribution of the eastern Atlantic marine gobies along the European coast (Actinopterygii: Gobiidae and Oxudercidae).

Zootaxa·2025
Same author

Testing the equivalency of human "predators" and deep neural networks in the detection of cryptic moths.

Journal of evolutionary biology·2024
Same author

Comparative analysis reveals assortative mate preferences in darters independent of sympatry and sex.

Ecology and evolution·2024

Processing fluency, the ease of information processing, is key to aesthetic appreciation. Neuronal activation sparsity is a better predictor of facial attractiveness than statistical typicality, suggesting distinct neural mechanisms for beauty perception.

Area of Science:

  • Cognitive Neuroscience
  • Computational Vision

Background:

  • Processing fluency, the ease of sensory and neural information processing, is a leading theory for aesthetic appreciation and beauty.
  • Two metrics, neuronal activation sparsity and statistical typicality, have been proposed to model processing fluency.

Purpose of the Study:

  • To compare the predictive power of neuronal activation sparsity and statistical typicality for facial attractiveness.
  • To investigate if refining reference representations enhances statistical typicality's explanatory power.

Main Methods:

  • Utilized convolutional neural networks (CNNs) as a computational model of the human visual system.
  • Assessed the ability of sparsity and statistical typicality metrics to explain variance in facial attractiveness ratings.
Keywords:
AttractivenessCNNsEfficient codingFluencySparsityStatistical typicality

Related Experiment Videos

Main Results:

  • Neuronal activation sparsity proved to be a more robust predictor of facial attractiveness than statistical typicality.
  • Refining statistical typicality's reference representation by ethnicity or gender did not improve its explanatory power.
  • Sparsity and statistical typicality predicted facial beauty using different CNN layers, indicating distinct underlying neural mechanisms.

Conclusions:

  • Neuronal activation sparsity is a stronger predictor of facial attractiveness compared to statistical typicality.
  • The findings suggest that different neural mechanisms underlie processing fluency and contribute to the perception of facial beauty.