July 2013
Volume 13, Issue 9
Free
Vision Sciences Society Annual Meeting Abstract  |   July 2013
What you find depends on how you look: Category selectivity in frontal cortex revealed by whole-brain correlation analysis
Author Affiliations
  • Yida Wang
    Department of Computer Science, Princeton University
  • Kai Li
    Department of Computer Science, Princeton University
  • Moses Charikar
    Department of Computer Science, Princeton University
  • Jonathan D. Cohen
    Department of Psychology, Princeton University\nPrinceton Neuroscience Institute, Princeton University
  • Nicholas B. Turk-Browne
    Department of Psychology, Princeton University\nPrinceton Neuroscience Institute, Princeton University
Journal of Vision July 2013, Vol.13, 493. doi:https://doi.org/10.1167/13.9.493
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Yida Wang, Kai Li, Moses Charikar, Jonathan D. Cohen, Nicholas B. Turk-Browne; What you find depends on how you look: Category selectivity in frontal cortex revealed by whole-brain correlation analysis. Journal of Vision 2013;13(9):493. https://doi.org/10.1167/13.9.493.

      Download citation file:


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

      ×
  • Supplements
Abstract

Neuroimaging studies typically seek to associate evoked neural activity with specific cognitive processes. While successful, this approach might fail to reveal cognitively meaningful interactions between brain regions. Consider a region that is always engaged during object perception, but whose interactions with other regions depend upon the category of object being perceived. This region might only be identified as category selective based on its pattern of correlations with other brain regions. However, typical approaches for assessing correlations in neuroimaging data could also fail to identify this region, since they require selecting seed regions based on activity differences. Here we take a fresh look at category selectivity by surmounting these limitations, applying multivariate analyses to patterns of correlations rather than activity, and extracting correlations in an unbiased manner over the whole brain rather than among seeds. Subjects viewed blocks of faces or scenes. For each block, we computed the pattern of correlations over time of every voxel with every other voxel. We used a linear classifier to label patterns of correlations as having been obtained during face vs. scene blocks. To measure accuracy, we iteratively trained the classifier on one group of subjects and tested it on a new subject. Correlation patterns were highly discriminative of object category (91%, p<4.2e-11). To identify which correlations were category-selective, we examined the frequency with which each voxel was chosen during correlation-based feature selection. Beyond expected voxels in ventral temporal cortex, a large voxel cluster was obtained in medial prefrontal cortex. Applying the same kind of classifier to patterns of activity produced high classification accuracy, but did not include frontal cortex. These findings illustrate the power of exhaustive analyses of brain interactions, and suggest a broader research program to examine how patterns of correlations, and not just activity, may advance our understanding of human brain function.

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.

×