Journal of Vision Cover Image for Volume 19, Issue 10
September 2019
Volume 19, Issue 10
Open Access
Vision Sciences Society Annual Meeting Abstract  |   September 2019
Independent mechanisms for implicit ensemble learning and explicit ensemble perception?
Author Affiliations
  • Sabrina Hansmann-Roth
    Icelandic Vision Lab, University of Iceland
  • Árni Kristjánsson
    Icelandic Vision Lab, University of Iceland
    Faculty of Psychology, National Research University, Higher School of Economics, Moscow, Russian Federation
  • David Whitney
    Department of Psychology, The University of California, Berkeley
  • Andrey Chetverikov
    Visual Computation Lab, Center for Cognitive Neuroimaging, Donders Institute for Brain, Cognition and Behavior
Journal of Vision September 2019, Vol.19, 239c. doi:https://doi.org/10.1167/19.10.239c
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Sabrina Hansmann-Roth, Árni Kristjánsson, David Whitney, Andrey Chetverikov; Independent mechanisms for implicit ensemble learning and explicit ensemble perception?. Journal of Vision 2019;19(10):239c. https://doi.org/10.1167/19.10.239c.

      Download citation file:


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

      ×
  • Supplements
Abstract

Features of objects in the environment can be represented as probability distributions or with summary statistics. Previous visual search studies have shown how previously learned properties of distractor distributions influence search times. The underlying distribution shape was assessed through role-reversal effects upon search time. This implicit ensemble statistical learning method has therefore revealed learning of feature distribution shape, while ensemble perception studies have not been able to capture the learning of higher order statistics (Atchley & Anderson, 1994, Dakin & Watt 1996). In this study we directly compared ensemble perception with this new method of implicit ensemble statistical learning for judgments of mean, variance and distribution shape. Observers learned statistical information in a block of 3–4 learning trials and were then presented with two distractor sets of varying mean, variance or distribution shape. They were encouraged to select the set that appeared more similar to the previously presented ones. These results were compared with the results from the implicit method. The explicit comparison resulted in much noisier estimates of representations of mean and variance than implicit distribution learning. Moreover, we were not able to find representations of the distribution shape with the explicit method while the representation of the distribution shape could be assessed through the implicit learning method. Interestingly, both methods showed that variance was largely overestimated for all observers and was not a result of a response bias towards the more variant set. These results highlight the efficiency of the implicit feature learning method and hint at independent mechanisms for the implicit learning of ensemble information and explicit perception of summary statistical information in scenes. We speculate that these differences reflect the functional distinctness of implicit and explicit information: implicit information about objects is crucial for acting in the environment while explicit ensemble perception determines the appearance of objects.

Acknowledgement: Rannis, Icelandic Research Fund 
×
×

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.

×