August 2023
Volume 23, Issue 9
Open Access
Vision Sciences Society Annual Meeting Abstract  |   August 2023
Assessing the Role of the Pulvinar in Feature versus Spatial Attention Control using Deep Neural Networks
Author Affiliations
  • Yun Liang
    University of Florida
  • Sreenivasan Meyyappan
    University of California, Davis
  • Mingzhou Ding
    University of Florida
Journal of Vision August 2023, Vol.23, 4701. doi:https://doi.org/10.1167/jov.23.9.4701
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Yun Liang, Sreenivasan Meyyappan, Mingzhou Ding; Assessing the Role of the Pulvinar in Feature versus Spatial Attention Control using Deep Neural Networks. Journal of Vision 2023;23(9):4701. https://doi.org/10.1167/jov.23.9.4701.

      Download citation file:


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

      ×
  • Supplements
Abstract

Pulvinar is the largest nucleus of the thalamus and has long been implicated in attention-related functions. Prior work has shown that stimulus-evoked neural activities in the pulvinar are modulated by attention. To what extent the pulvinar contributes to attention control remains to be assessed. FMRI data were recorded from humans (n=20) performing a cued visual spatial/feature attention task in which an auditory cue instructed the subject to attend either left or right visual field (spatial attention) or to attend either red or green color (feature attention). Following a random delay, two rectangular stimuli appeared, one in each visual field, and the subjects reported the orientation of the rectangle in the attended location (spatial trials) or having the attended color (feature trials). A deep neural network (DNN) was trained to take cue-evoked fMRI data from the pulvinar as input features to predict the trial labels (i.e., attended information). Specifically, for feature (spatial) attention control, feature (spatial) trial data from 19 subjects were used to train a DNN model, which was then tested on the remaining subject. This process was repeated 20 times, and the 20 decoding accuracies were averaged. We found that the accuracy for decoding feature attention control (cue red vs. cue green) and spatial attention control (cue left vs. cue right) were 65% and 63%, respectively, both significantly above chance level of 50%. Applying an occlusion method, we derived heatmaps from the DNN models, which revealed that the lateral pulvinar is the most significant contributor to the DNN decoding performance. These results (1) support the idea that the pulvinar is an important node in the neural network that controls the deployment of feature and spatial attention in advance of stimulus processing and (2) the DNN is a useful tool for neural pattern analysis that complements other machine learning techniques.

×
×

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.

×