October 2020
Volume 20, Issue 11
Open Access
Vision Sciences Society Annual Meeting Abstract  |   October 2020
Behavioral and Neural Evidence that Visual Statistical Learning is Shaped by Task Demands and Categories
Author Affiliations & Notes
  • Leeland Rogers
    University of Delaware
  • Su-Hyoun Park
    University of Delaware
  • Timothy Vickery
    University of Delaware
  • Footnotes
    Acknowledgements  NSF OIA 1632849, NSF BCS 1558535
Journal of Vision October 2020, Vol.20, 1196. doi:https://doi.org/10.1167/jov.20.11.1196
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Leeland Rogers, Su-Hyoun Park, Timothy Vickery; Behavioral and Neural Evidence that Visual Statistical Learning is Shaped by Task Demands and Categories. Journal of Vision 2020;20(11):1196. doi: https://doi.org/10.1167/jov.20.11.1196.

      Download citation file:

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

  • Supplements

Visual statistical learning (VSL) describes humans’ ability to incidentally extract temporal and spatial statistical regularities from visual information. VSL is often characterized as the result of always-on processes, constantly at work on the objects of attention. Yet, in the real world, task demands change constantly as a function of our immediate goals and the environment. Aside from basic manipulations of visual attention, however, little work has focused on how behavioral or neural markers of VSL are modulated by different task conditions. To examine this problem, we first had participants view sequences of fractal images in which image pairs were embedded (pair members always appeared back-to-back), while subjects were engaged in an arbitrary categorization task. Embedded pairs were either both of the same category or two different categories. In a surprise 2AFC test, Experiment 1 revealed that same-category pairs were learned better than different category pairs (p = .01). Experiment 2 used fMRI to examine whether task during familiarization modulates neural signatures identified by prior research. After being exposed to streams of images under different task demands (either 1-back or arbitrary category-mapping), participants were scanned while passively viewing a stream of the same images in a rapid event-related design. BOLD responses were modulated by both item order and training context; while activity in the Frontoparietal Network and Middle Frontal/Medial Prefrontal areas for images learned during the 1-back task was reminiscent of prior research, activity for images learned during categorization was not: cerebellar activity uniquely differentiated the second image of same/different category pairs. While previous research has overlooked the potential for task and context to influence VSL, we argue that it may dramatically influence behavioral and neural responses to statistical associations. As our task and efforts change throughout the day, incidental learning mechanisms such as VSL also adaptively change.


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.