December 2022
Volume 22, Issue 14
Open Access
Vision Sciences Society Annual Meeting Abstract  |   December 2022
Using EEG to uncover the dynamics of physical expectation violation and resolution
Author Affiliations
  • Halely Balaban
    Massachusetts Institute of Technology (MIT)
  • Kevin A. Smith
    Massachusetts Institute of Technology (MIT)
  • Tomer D. Ullman
    Harvard University
  • Joshua B. Tenenbaum
    Massachusetts Institute of Technology (MIT)
Journal of Vision December 2022, Vol.22, 3386. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Halely Balaban, Kevin A. Smith, Tomer D. Ullman, Joshua B. Tenenbaum; Using EEG to uncover the dynamics of physical expectation violation and resolution. Journal of Vision 2022;22(14):3386.

      Download citation file:

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

  • Supplements

To make sense of the world, we rely on rich expectations about objects, already present in infancy. Even babies are surprised by physically-impossible scenarios, such as an object magically disappearing. Behavioral signatures of surprise reflect only the final outcome of a mental construal process that remains hidden in these classic studies. We tracked the moment-by-moment dynamics of expectation violation and resolution, by embedding physically-possible and -impossible scenes within a working-memory task, and measuring the CDA, an ERP component whose amplitude rises when more items are held in working-memory. Participants watched animations of objects crossing a stage, and were subsequently asked about their shape and color. Animations included an occluder rising and lowering, briefly hiding objects from view. In the control (physically-possible) conditions, one or two objects simply passed behind the occluder. There were two physically-impossible conditions: “Create”, where one object went behind the occluder but two exited, and “Vanish”, where two objects went behind but only one exited. The CDA rose after an object was created, and fell after an object vanished, indicating that participants’ scene representations transformed to add or remove objects. However, this wasn’t the result of smooth representational update: immediately after physically-impossible events, the CDA transiently dropped, and only then leveled to reflect the final number of objects. This indicates a resetting process, triggered by the “impossible” creation or vanishing, in which the original representation was removed from working-memory before the scene could be resolved and correctly represented. Interestingly, resetting was faster for creation than vanishing, as predicted by a computational model of expectation-violation that suggests it takes longer to infer an object’s absence than to perceive an object’s presence. To conclude, the CDA revealed the dynamics of representational change as we perceive and understand the world, and how ongoing representations are driven by physical expectations.


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.