July 2013
Volume 13, Issue 9
Free
Vision Sciences Society Annual Meeting Abstract  |   July 2013
Decoding invariant representations in visual working memory
Author Affiliations
  • Thomas Christophel
    Bernstein Center for Computational Neuroscience, Charité Universitätsmedizin, Berlin, Germany\nBerlin Center for Advanced Neuroimaging, Charité Universitätsmedizin, Berlin, Germany
  • Christian Endisch
    Bernstein Center for Computational Neuroscience, Charité Universitätsmedizin, Berlin, Germany\nBerlin Center for Advanced Neuroimaging, Charité Universitätsmedizin, Berlin, Germany
  • John-Dylan Haynes
    Bernstein Center for Computational Neuroscience, Charité Universitätsmedizin, Berlin, Germany\nBerlin Center for Advanced Neuroimaging, Charité Universitätsmedizin, Berlin, Germany
Journal of Vision July 2013, Vol.13, 927. doi:https://doi.org/10.1167/13.9.927
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Thomas Christophel, Christian Endisch, John-Dylan Haynes; Decoding invariant representations in visual working memory. Journal of Vision 2013;13(9):927. https://doi.org/10.1167/13.9.927.

      Download citation file:


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

      ×
  • Supplements
Abstract

Visual shape recognition can exhibit considerable invariance across changes in visual appearance and viewing position (Cichy, Chen, & Haynes, 2011; Edelman, 1997; Grill-Spector, Kourtzi, & Kanwisher, 2001). This raises the question whether also the retention of shape information in visual working memory (Baddeley & Hitch, 1974) exhibits such invariance. Here, we specifically investigated whether objects memorized across brief delays are encoded using a rotation-invariant code. While positioned in an MRI scanner, 22 healthy subjects memorized simple shape stimuli. To prevent subjects from using a semantic code, we chose abstract decagons as stimuli. These are randomly generated ten-sided shapes. In each trial, subjects had to memorize one shape indicated by a retro-cue method controlling for perceptional confounds (see Harrison & Tong, 2009). After a delay of 10 seconds, subjects had to identify which of two test decagons had a more similar shape (see Christophel, Hebart, & Haynes, 2012). Importantly, sample and test decagons were shown in random 2D-rotations to encourage the use of rotation-invariant representations. We used fMRI in combination with time-resolved multivariate searchlight decoding to identify areas that maintained the memorized object during the delay period using a rotation-invariant code (Cichy et al., 2011; Haynes & Rees, 2006; Kriegeskorte, Goebel, & Bandettini, 2006). Testing for classifier generalization between different rotational views of the same shape, we identified three regions that showed significant (p(FWE) <0.05) memory-specific information bilaterally: Lateral occipito-temporal cortex, posterior parietal cortex and the human frontal eye fields (see Petit, Clark, Ingeholm, & Haxby, 1997). These results demonstrate that invariant shape-coding in working memory is prevalent in perception-driven areas across the brain (see Postle, 2006). Importantly, our findings demonstrate that invariant visual memory representations do not require higher-order dorso-lateral prefrontal areas.

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.

×