October 2020
Volume 20, Issue 11
Open Access
Vision Sciences Society Annual Meeting Abstract  |   October 2020
Attention to different statistical structures changes over the course of learning
Author Affiliations
  • Tess Allegra Forest
    University of Toronto
  • Noam Siegelman
    Haskins Laboratories
  • Amy Finn
    University of Toronto
Journal of Vision October 2020, Vol.20, 293. doi:https://doi.org/10.1167/jov.20.11.293
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Tess Allegra Forest, Noam Siegelman, Amy Finn; Attention to different statistical structures changes over the course of learning. Journal of Vision 2020;20(11):293. doi: https://doi.org/10.1167/jov.20.11.293.

      Download citation file:


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

      ×
  • Supplements
Abstract

Previous studies have shown that attention allocation can be determined by the statistical structure of our visual environments: infants attend to moderately predictable input over highly predictable or highly unpredictable input (Kidd et. al., 2012), and adults attend to regular over irregular stimuli (Zhao et. al., 2013). While these results show learners are sensitive to how predictable different input is, no study to date has directly examined how attention to differently structured input shifts as a function of experience. Moreover, learners may be able to extract more or less information from a particular part of their world at any given moment, but it remains unknown whether we flexibly shift attention based on how much information could be gained from a particular stimulus. Here, we had adults (n=75) complete a visual statistical learning experiment in which streams of information were presented simultaneously in four locations, in four levels of predictability: (1) completely random, (2) low predictability, (3) medium predictability, and (4) completely predictable. Intermittent search trials measured where participants attended over the course of the experiment by indexing reaction times in each location. We modeled trial-by-trial entropy in each location to measure how much information could be gained from that location at any given point during learning. Our results show that as the experiment progressed, participants shifted from attending to medium predictability location to attending to lower levels of regularity locations (low predictability and random stream). Additionally, trial-by-trial entropy and time interacted strongly to predict where participants attended, such that over the course of learning higher entropy locations were attended more. This provides the first demonstration that as adults learn the environmental regularities, they gradually shift their attention to less predictable sources of information, and that learners are sensitive to how much information they can gain from a particular source.

×
×

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.

×