PET Studies of Components of High-Level Vision.

Abstract

We have finished analyzing the data from our first PET experiments and written a report of the results. As noted earlier, we have strong evidence (p = .0001) that primary visual cortex is activated selectively during visual mental imagery. In our next PET experiment we will study how objects are identified when seen from unusual points of view. Warrington and Taylor (1973) found that patients with right-parietal lesions have a very difficult time recognizing objects seen from unusual points of view. Kosslyn, Flynn, Amsterdam and Wang (1990) explain this result by positing a top-down, hypothesis testing mechanism that is called into play when a stimulus does not immediately match a stored memory very well. This mechanism not only relies on processes in the parietal lobe to shift attention, but also on processes in the frontal lobe formulate hypotheses. To test these ideas, subjects will see a series of pictures, either of objects seen from a canonical point of view or of objects seen from an unusual point of view. One sec after seeing a picture, a word will be presented, and the subject will decide whether the word names the pictured object. Counterbalancing will ensure that the same objects and words appear equally often in the two conditions.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 1991
Accession Number
ADA240202

Entities

People

  • Stephen M. Kosslyn

Organizations

  • Harvard University

Tags

DTIC Thesaurus Topics

  • Blood Flow
  • Brain
  • Cognitive Neuroscience
  • Cognitive Science
  • Data Displays
  • Identification
  • Language
  • Military Research
  • Neurosciences
  • Psychology
  • Visual Cortex

Readers

  • Brain and Cognitive Science; Experimental Psychology; Cognitive Neuroscience
  • Computer Vision.
  • Theoretical Analysis.