August 2012
Volume 12, Issue 9
Vision Sciences Society Annual Meeting Abstract  |   August 2012
Perceptual Crowding in a Neural Model of Feedforward-Feedback Interactions
Author Affiliations
  • Tobias Brosch
    Institute of Neural Information Processing, University of Ulm, Germany
  • Heiko Neumann
    Institute of Neural Information Processing, University of Ulm, Germany
Journal of Vision August 2012, Vol.12, 329. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Tobias Brosch, Heiko Neumann; Perceptual Crowding in a Neural Model of Feedforward-Feedback Interactions. Journal of Vision 2012;12(9):329. doi:

      Download citation file:

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

  • Supplements

Problem. The orientation of peripherally presented items is more difficult to recognize when surrounded by other items with similar features. This phenomenon is called ''crowding'' (Levi, Vis.Res. 48, 2008, Whitney & Levi, TiCS 15, 2011). Despite the large body of literature on crowding, only few quantitative models exist that can reproduce psychometric curves referring to varying target contrast or target-flanker spacing.

Model. We propose a model of feedforward and feedback processing in visual cortical areas V1 and V2 for robust contour grouping and feature detection (Weidenbacher & Neumann, PLoS ONE 4, 2009). The model incorporates an additional stage of center-surround normalization for pooling oriented contrasts and grouping responses.

Results. We employ input configurations with oriented bars and perturbed them by additive Gaussian noise similar as in van den Berg et. al. (PLoS Comput. Biol. 6, 2010). When varying the target contrast and target-flanker spacing we observed a quantitatively good match with psychometric curves derived from human experiments. But in contrast to previous modeling investigations our approach also predicts that crowding is partially released, when the surrounding flankers form a contour. This effect is explained by the enhancement of elongated contours in the recurrent loop of boundary integration. The enhanced activity and the greater extent of an elongated contour results in an increased inhibition at the final pooling stage which in turn reduces the crowding effect of surrounding flankers (configuration effect).

Conclusions. The proposed model provides further evidence for the role of feature integration along the feedforward sweep of processing that is supported by stages of lateral boundary integration and modulating feedback in target detection and object recognition processes. The integration of context relevant features leads to context-sensitive increase in activity that counter-acts the impact of crowding and, thus, acts like an attentional gating at different stages of cortical processing.

Meeting abstract presented at VSS 2012


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.