October 2020
Volume 20, Issue 11
Open Access
Vision Sciences Society Annual Meeting Abstract  |   October 2020
Task and Stimulus Dependent Contributions to Ensemble Scene Processing
Author Affiliations & Notes
  • Vignash Tharmaratnam
    University of Toronto Scarborough
  • Jason Haberman
    Rhodes College
  • Jonathan S. Cant
    University of Toronto Scarborough
  • Footnotes
    Acknowledgements  NSERC Discovery Grant to J.S.C
Journal of Vision October 2020, Vol.20, 1292. doi:https://doi.org/10.1167/jov.20.11.1292
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Vignash Tharmaratnam, Jason Haberman, Jonathan S. Cant; Task and Stimulus Dependent Contributions to Ensemble Scene Processing. Journal of Vision 2020;20(11):1292. https://doi.org/10.1167/jov.20.11.1292.

      Download citation file:

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

  • Supplements

We recently demonstrated that summary statistics can be extracted for ensembles of multiple scenes (Tharmaratnam et al., VSS 2019), in addition to groups of faces and objects as shown in past literature. Specifically, participants integrated multiple scenes into their ratings of average scene content (i.e., perceived naturalness or manufacturedness) and spatial boundary (i.e., perceived openness or closedness), two global spatial properties (GSPs) important in scene perception. However, task difficulty increased with increasing set sizes, which was surprising given that ensemble perception usually becomes more efficient as set size increases. In the present study we conducted two experiments to explore whether this finding could be explained by either stimulus or task manipulations. In Experiment 1, scenes were divided into 6 parts (i.e., a 2x3 grid), and 1, 2, 4, or 6 parts were presented (in their corresponding grid positions) on each trial. When rating average GSPs, we again found that task difficulty increased as set size increased, but also found that participants were integrating less information when encoding scene-part ensembles compared with the whole-scene ensembles used previously. In Experiment 2 we used the same whole-scene ensembles utilized previously but changed the task. Specifically, we randomly rotated the orientation of each scene in the ensemble and asked participants to rate the average orientation of the whole-scene ensemble. Similar to our previous results, we found that participants were integrating multiple scenes into their ratings of average orientation, but in contrast, we now found that task difficulty did not increase with increasing set sizes. Together, these results demonstrate that the efficiency of high-level ensemble scene processing is dependent on the nature of the stimuli (i.e., wholes vs. parts) and the type of summary statistic extracted (i.e., average GSPs vs. orientation), with potentially distinct ensemble mechanisms mediating average scene GSP and orientation processing.


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.