A Representation of Perceptual Decision Bias in a Distributed Occipito-Parieto-Frontal Cortical Network

Tao Tu1, Paul Sajda2

  • 1Columbia Univ
  • 2Columbia University

Details

11:30 - 13:30 | Fri 26 May | Emerald III, Rose, Narcissus & Jasmine | FrPS1T1.48

Session: Poster I

Abstract

Perceptual decisions can be biased by a propensity to choose one stimulus alternative over another. Such bias is revealed especially in cases when there is inadequate stimulus evidence. Here we report on a three-alternative forced choice (3-AFC, Face vs. Car vs. House) visual categorization task, where subjects, on average, exhibited a “face bias”: faster response times for faces than for cars, even when degraded (low stimulus evidence) images were presented. We used simultaneous EEG and fMRI to characterize the neural correlates underlying this face bias. Specifically, in the first-level EEG-informed fMRI analysis, we incorporated EEG trial-to-trial variability to temporally identify a brain network related to a late decision process. To link the individual differences in bias with the changes in BOLD activity in the decision network, we used the subject-wise bias, obtained by fitting the subject's behavioral data using the drift diffusion model, as a covariate in the group-level fMRI analysis. Our results showed significant correlation between decision bias and activity in a distributed occipito-parieto-frontal network. Our work shows that some decision related variables may be represented in a distributed fashion across regions in the human brain.