September 2019
Volume 19, Issue 10
Open Access
Vision Sciences Society Annual Meeting Abstract  |   September 2019
Rapid processing of illusory faces in inanimate objects by the human brain
Author Affiliations & Notes
  • Susan G Wardle
    Laboratory of Brain & Cognition, National Institute of Mental Health
  • Jessica Taubert
    Laboratory of Brain & Cognition, National Institute of Mental Health
  • Lina Teichmann
    Department of Cognitive Science, Macquarie University
  • Chris I Baker
    Laboratory of Brain & Cognition, National Institute of Mental Health
Journal of Vision September 2019, Vol.19, 92b. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Susan G Wardle, Jessica Taubert, Lina Teichmann, Chris I Baker; Rapid processing of illusory faces in inanimate objects by the human brain. Journal of Vision 2019;19(10):92b.

      Download citation file:

      © ARVO (1962-2015); The Authors (2016-present)

  • Supplements

The human brain is specialized for face processing, with a complex network of brain regions supporting face perception. Despite this expertise, we sometimes spontaneously perceive illusory faces in inanimate objects, a capacity we share with other non-human primates (Taubert et al., 2017). In order to understand the temporal dynamics of processing for these natural ‘errors’ of face detection, we used event-related fMRI (N=21) and MEG (N=22) to measure the brain activation patterns of human participants in response to 96 photographs including 32 illusory faces in inanimate objects (e.g. coffee cup, bell pepper, mop), 32 matched objects, and 32 human faces. Multivariate pattern analysis of the fMRI data showed differences in the representation of illusory faces compared to human faces and matched objects in category-selective high-level visual cortex. The MEG data supported this distinction but further revealed the dynamic nature of the representation of illusory faces. Representational similarity analysis revealed that at ~130ms after stimulus onset, some examples of illusory faces were represented more similarly to human faces than to matched objects. By ~160ms, three clusters emerged in the whole-brain representation: human faces, illusory faces, and non-face objects. This timepoint corresponded to the first peak in category-level decoding from the MEG activation patterns. Only 100ms later, this representation evolved into a clear distinction between human faces and all other objects regardless of whether they contained an illusory face. Comparison of the brain’s representational structure with that of existing saliency models and deep neural networks pre-trained on object and face recognition was consistent with the recruitment of a broadly-tuned low-level visual template for face detection. Together the results show that illusory faces are processed incredibly rapidly by the human brain, however, the representational structure quickly stabilizes into one organized by object content rather than by face perception.


This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.