September 2011
Volume 11, Issue 11
Vision Sciences Society Annual Meeting Abstract  |   September 2011
Dynamics of feedback-driven visual learning
Author Affiliations
  • Christopher D'Lauro
    Center for the Neural Basis of Cognition, Carnegie Mellon University, USA
  • Yang Xu
    Center for the Neural Basis of Cognition, Carnegie Mellon University, USA
    Machine Learning Department, Carnegie Mellon University, USA
  • Rob Kass
    Center for the Neural Basis of Cognition, Carnegie Mellon University, USA
    Department of Statistics, Carnegie Mellon University, USA
  • Michael J. Tarr
    Center for the Neural Basis of Cognition, Carnegie Mellon University, USA
    Department of Psychology, Carnegie Mellon University, USA
Journal of Vision September 2011, Vol.11, 1002. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Christopher D'Lauro, Yang Xu, Rob Kass, Michael J. Tarr; Dynamics of feedback-driven visual learning. Journal of Vision 2011;11(11):1002.

      Download citation file:

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

  • Supplements

How do modality-specific visual learning and more general learning mechanisms interact in forming representations for new object categories? Prior neuroimaging studies suggest greater sensitivity to object category boundaries in prefrontal cortex (PFC) and relatively lower sensitivity to the same category boundaries in ventral cortex (Jiang et al., 2007). However, such studies have often used familiar objects, and, therefore, do not capture the full time course of learning. Here we address the role of these general learning areas of the brain (e.g., PFC) both earlier and later in learning. The relative timing of neural activity throughout the brain is informative with respect to the contributions of different brain regions as observers learn new object categories. At issue is whether the PFC requires “tuning” through ventral visual learning as a precursor to defining category boundaries for new objects. We used MEG to assess the neural changes concomitant with visual learning in a feedback-corrected categorization task that employed novel objects. In this task, participants learned to classify two different “families” of computer-generated “blob” stimuli as “A” or “B” (as in Krigolson et al., 2009). Participants' performance improved from chance to near-ceiling levels in the course of the single MEG session, providing behavioral and neural data that spans a much wider course of learning relative to most prior studies in this area. Beyond exploring whether ventral visual responses or PFC responses better predict both early and late visual learning, we have developed methods to improve the spatio-temporal resolution of MEG. Therefore, we are better able to associate specific brain areas with behavioral effects.

Funding provided by the Perceptual Expertise Network (#15573-S6), a collaborative award from James S. McDonnell Foundation, by the Temporal Dynamics of Learning Center at UCSD (NSF Science of Learning Center SBE-0542013), by the PA Department of Health, Commonwealth Universal Research Enhancement (C.U.R.E.) program, Formula Award Number 4100050890, 2010, The Richard King Mellon Foundation, and by an NIH EUREKA Award (#1R01MH084195-01) to MJT. 

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.