August 2014
Volume 14, Issue 10
Free
Vision Sciences Society Annual Meeting Abstract  |   August 2014
Data driven identification of functional organization
Author Affiliations
  • Jason Webster
    Psychology, University of Washington, Seattle, WA
  • Ione Fine
    Psychology, University of Washington, Seattle, WA
Journal of Vision August 2014, Vol.14, 208. doi:https://doi.org/10.1167/14.10.208
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Jason Webster, Ione Fine; Data driven identification of functional organization. Journal of Vision 2014;14(10):208. https://doi.org/10.1167/14.10.208.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose: Current fMRI methods for examining cortical functional organization require assumptions about either stimulus category boundaries or the region of interest (ROI). Conventional fMRI localizers rely on statistical contrasts between predefined conditions, requiring assumptions about how stimuli should be categorized. Representational similarity analysis, which quantifies the similarity of responses within a region of interest to a set of stimuli, requires a predefined ROI that does not contain cortical areas with diverse selectivity profiles. Here, we describe a method that identifies cortical regions with similar selectivity profiles across a stimulus set without assumptions about either stimulus categorical structure or the ROI. We evaluated our approach by using it to identify known category selective areas in ventral temporal cortex. Methods: Two subjects passively viewed short video clips of faces, bodies, scenes, objects, scrambled objects, and uniformly colored screens. There were 12 stimuli in each category and ten repetitions per stimulus. For each vertex on the cortical surface, we calculated a vector of 72 beta-weights that represented the response to each stimulus. A dissimilarity matrix was constructed for these beta-weights, which was then sorted to cluster vertices with similar beta-weight vectors. Results: When projected onto the cortical surface, vertex clusters formed spatially contiguous regions. Both the spatial extent of these regions and the profile of responses within these ROIs were highly replicable using an independent dataset. A subset of these regions almost perfectly overlapped with conventionally identified category selective regions (e.g. PPA, FFA), while others were novel. Future work with a more diverse stimulus dataset will investigate the selectivity of these novel clusters. Conclusions: Our method successfully identifies ROIs with similar responses across a stimulus set without requiring assumptions about stimulus categorical structure or the region of interest.

Meeting abstract presented at VSS 2014

×
×

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.

×