AN APPLICATION OF THE OPTIC SPHERE THEORY IN DISCRIMINATION OF SLANT WITH MINIMAL INFORMATION

Citation
L. Eriksson et E. Borjesson, AN APPLICATION OF THE OPTIC SPHERE THEORY IN DISCRIMINATION OF SLANT WITH MINIMAL INFORMATION, Scandinavian journal of psychology, 38(4), 1997, pp. 275-287
Citations number
17
ISSN journal
00365564
Volume
38
Issue
4
Year of publication
1997
Pages
275 - 287
Database
ISI
SICI code
0036-5564(1997)38:4<275:AAOTOS>2.0.ZU;2-6
Abstract
In Johansson and Borjesson (1989), a new theory of visual space percep tion-the optic sphere theory-was presented in which the hemispheric sh ape of the retina is utilized for determination of slant of plane surf aces in wide-angle perception. The process of the optic sphere mechani sm can be described as the projection of a translating distal line on the optic sphere, and an extrapolation of this projection to a great c ircle. The determination of the 3-D slant of the distal line is made b y identification of points of no change on the great circle during its rotation. The main objective of the present study was to investigate this process as applied to central stimulation of the retina with redu ced and minimal information of slant or horizontal orientation. Each s timulus pattern consisted of either two continuous lines or two pairs of dots in motion presented on a computer screen. The pairwise lines a nd the pairs of dots defined simulated 3-D slants (or horizontal orien tations) of different magnitude within each pair, and the subjects' ta sk was to discriminate between these simulated slants. It was shown th at the simulations evoke percepts of 3-D slants, and of horizontal ori entations, and that it is possible to discriminate between them even f rom minimal information (pairs of dots). Further, the empirical findin gs of Borjesson (1994) indicated that longer extrapolations of the pro jected are to a point of no change yield less accurate discriminations of slant. We failed to replicate this in Experiment 4, in which case stimulus variables that covaried with extrapolation length were elimin ated or minimized. It is suggested that this raises some doubt about d iscrimination accuracy as dependent on extrapolation length per se. Th e overall conclusion, however, is that the optic sphere theory represe nts a possible explanation of, or analogy to, the process utilized by the visual system for determination of the simulated 3-D slants and ho rizontal orientations in the present study.