Purchase this article with an account.
Jian Ding, George Sperling; A gain-control theory of binocular combination. Journal of Vision 2002;2(7):327. doi: 10.1167/2.7.327.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
When different stimuli are presented to the left and right eyes, only a single, combined “cyclopean” image is perceived. The numerous theories that have been proposed to describe binocular combination rule typically abstract a parameter from each eye's image (e.g., maximum stimulus contrast or the visual direction of a significant point) and use the two eyes' values to predict the parameter value in the cyclopean image. Here we offer a process theory that applies to any image within a spatial frequency band, and predicts, pixel by pixel, the combined cyclopean image. We propose simply that, in every neighborhood, each eye exerts gain control on the other in proportion to the strength of its own input. We tested the theory with superthreshold sinewave stimuli. Sine waves were horizontal to make the cyclopean image independent of vergence. Subjects judged the position of the dark stripe of the cyclopean image relative to an adjacent bar marker. The sine waves in the left and right eyes were of different contrasts and in different phases. The arithmetic addition of two parallel sine waves produces a sine wave of known phase and amplitude. The data consisted of 96 combinations of contrasts of the left- and right-eye stimuli, and of their relative phases. For stimuli to one eye only, and for equal stimuli to both eyes, the position judgments agreed perfectly with the linear summation model. For all other contrasts in the two eyes, the linear summation model failed. From the judged position, we inferred the strength of the contribution of each eye to the cyclopean image. In general, the stimulus with greater contrast had more weight than predicted by simple linear summation. Without estimating any parameters from the data, the gain-control model accounted for 99% of the variance of the data. We conclude that a simple, robust, physiologically plausible model accurately describes binocular combination within a spatial frequency channel.
This PDF is available to Subscribers Only