Purchase this article with an account.
Hashim Hanif, Rana Khalil, George Malcolm, Jason Barton; Predicting perceptual expertise from semantic knowledge: An indexed car test for prosopagnosic patients. Journal of Vision 2008;8(6):185. doi: https://doi.org/10.1167/8.6.185.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Studies of whether the prosopagnosic recognition impairment extends to other objects are confounded by variable expertise of people for other objects. Ideally, performance for non-face object recognition by these patients should be adjusted for premorbid expertise. We explored whether an index of semantic knowledge about cars could predict the performance of healthy subjects in a test of visual recognition of cars. 23 subjects perform three items involving all European, Asian and American cars made from 1950 to the present. First, they used Likert scales to rate their own knowledge of cars from each decade. Second, we administered a semantic questionnaire, asking them to provide the make (manufacturer's name) for all models made in this period. Third, we administered a perceptual test that showed the images of 150 cars, for which they were required to provide decade, model and make if possible. Half of the subjects performed the perceptual test before the semantic, and half the reverse. Semantic knowledge correlated well with perceptual recognition of make (r=.85) and model (r=.90), but less so with recognition of decade of make (r = .30). A combined perceptual index of Make + 4*(Model) + 0.2*(Decade) yielded the optimum correlation of perceptual knowledge with semantic knowledge (r = .93). Self-ratings correlated moderately with semantic knowledge (r=.57), and with perceptual recognition of make (r=.59), model (r=.54), and decade of make (r = .33). A combined perceptual index of Make + 2.1*(Model) + 0.5*(Decade) yielded the optimum correlation of perceptual knowledge with self-rating (r = .61). We conclude that semantic car knowledge but not self-rating is a reasonably accurate predictor of perceptual recognition of cars by make and model. A semantic index may be useful for adjusting perceptual recognition scores for premorbid expertise when studying patients with face or object recognition deficits.
This PDF is available to Subscribers Only