August 2012
Volume 12, Issue 9
Free
Vision Sciences Society Annual Meeting Abstract  |   August 2012
Decoding EEG data reveals dynamic spatiotemporal patterns in perceptual processing
Author Affiliations
  • Monica Rosenberg
    Boston Attention and Learning Lab, VA Boston Healthcare System
  • Alexandra List
    Department of Psychology, Northwestern University
  • Aleksandra Sherman
    Department of Psychology, Northwestern University
  • Marcia Grabowecky
    Department of Psychology, Northwestern University\nInterdepartmental Neuroscience Program, Northwestern University
  • Satoru Suzuki
    Department of Psychology, Northwestern University\nInterdepartmental Neuroscience Program, Northwestern University
  • Michael Esterman
    Boston Attention and Learning Lab, VA Boston Healthcare System\nBoston University School of Medicine
Journal of Vision August 2012, Vol.12, 1173. doi:10.1167/12.9.1173
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Monica Rosenberg, Alexandra List, Aleksandra Sherman, Marcia Grabowecky, Satoru Suzuki, Michael Esterman; Decoding EEG data reveals dynamic spatiotemporal patterns in perceptual processing. Journal of Vision 2012;12(9):1173. doi: 10.1167/12.9.1173.

      Download citation file:


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

      ×
  • Supplements
Abstract

We investigated the efficacy of multivariate pattern analysis (MVPA) for revealing EEG correlates of visual perception. EEG was recorded with 64 scalp electrodes from four participants as they passively viewed grayscale stimuli varying along four dimensions: location (left or right visual field), category (face or Gabor), subcategory (male or female faces; high or low spatial frequency Gabors), and orientation (upright or inverted faces; horizontal or vertical Gabors). Group-averaged ERPs showed typical temporal shifts and amplitude differences for both contralateral-versus-ipsilateral stimuli and upright-versus-inverted faces. Using linear support vector machines, we performed MVPA on the same EEG data with approximately 1-ms precision to see if stimulus differences could be classified on a trial-by-trial basis. EEG signals occurring at specific time points (174 ms and 674 ms for left-versus-right classification, 180 ms and 271 ms for face-versus-Gabor classification, and 219 ms for upright-versus-inverted face classification) reliably predicted perceptual differences with accuracy ranging from 67% to 93%. Importantly, the critical time points were virtually identical for the four subjects. Although corresponding group-averaged ERPs differentiated a subset of these conditions, MVPA predicted stimuli on an individual subject level, provided remarkably consistent estimates of the timing of perceptually relevant neural information on a trial-by-trial basis, and revealed additional time windows of discrimination accuracy. MVPA did not reliably classify horizontal vs. vertical Gabors, low vs. high spatial frequency Gabors, or female vs. male faces, suggesting that the highly reliable trial-by-trial predictions described above are not an artifact of our MVPA method. Thus, MVPA of trial-by-trial EEG data is a robust complementary approach to ERPs as it uncovers unique neural correlates of visual 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.

×