July 2013
Volume 13, Issue 9
Free
Vision Sciences Society Annual Meeting Abstract  |   July 2013
What’s Feedback Got To Do With It? Examining Learning Rate and Generalization in Cross-scene Statistical Learning With and Without Feedback
Author Affiliations
  • Lauren Emberson
    Brain and Cognitive Sciences Department, University of Rochester
  • Patricia Reeder
    Brain and Cognitive Sciences Department, University of Rochester
  • Richard Aslin
    Brain and Cognitive Sciences Department, University of Rochester
  • Daphne Bavelier
    Brain and Cognitive Sciences Department, University of Rochester\nFPSE, University of Geneva
Journal of Vision July 2013, Vol.13, 330. doi:10.1167/13.9.330
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Lauren Emberson, Patricia Reeder, Richard Aslin, Daphne Bavelier; What’s Feedback Got To Do With It? Examining Learning Rate and Generalization in Cross-scene Statistical Learning With and Without Feedback. Journal of Vision 2013;13(9):330. doi: 10.1167/13.9.330.

      Download citation file:


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

      ×
  • Supplements
Abstract

Most models of learning predict that the presence of feedback facilitates the rate of learning, but can inhibit generalization by emphasizing attested exemplars rather than underlying rules. In a task where participants can learn statistical information without feedback (statistical learning), we examined what additional role feedback plays in learning and generalization. Following Shohamy and Wagner (2008), participants were taught arbitrary face-scene associations. On every trial, participants saw simultaneous presentation of a face and two scenes. After selecting one of the two scenes in a fixed response interval, participants were either given feedback (N=16, picture of a thumbs-up or -down) or not (N=19). Unbeknownst to participants, every two faces (Face1, Face2) were reciprocally paired with two scenes (Scene1, Scene2) creating a family of 4 overlapping associations (e.g. Face1—Scene1, Face1—Scene2, Face2—Scene1, Face2—Scene2). During training (4 blocks), participants were exposed to 3 of the 4 possible associations for each of the 16 families (trained associations). At test, feedback was always withheld to test for trained associations and untrained associations (generalization associations). There were robust and nearly identical learning rates for trained associations across feedback conditions. Both groups readily generalized but with marginally more generalization in the feedback condition (p <0.1). We find no significant difference in the frequency of non-learners across groups (defined as participants who fail to exhibit significant learning in any training block, feedback: N=2, no-feedback: N=6). Overall, these results show that learning of scene-pair associations and generalization to unseen, overlapping associations is minimally affected by the presence of feedback. In contrast to previous models, these results suggest that in the presence of salient statistical information, feedback may not modulate the time-course and outcomes of learning. Future work will examine how individual differences in the exposure phase affect learning rates and the factors that predict non-learners.

Meeting abstract presented at VSS 2013

×
×

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.

×