Abstract
Background:
In order to understand binocular vision it is important to recognize what features in the stimulus are matched between the 2 eyes. One issue in matching that is of interest is the luminance polarity of the elements- whether a dark line in one eye can match a bright line in the other eye. Helmholtz (1909) showed that line stereograms with opposite luminance polarity can be fused, but Julesz (1971) reported that random dot patterns with opposite dot polarity could not be fused. Masson et al. (1997) reported the vergence system responds to spots with opposite polarity by going the wrong way. Howard (1997) has suggested that the polarity of edges, rather than spots, may be the key feature. In this study we have measured vergence responses to both lines and edges of matched and mismatched polarity in an effort to clarify this issue.
Methods:
A dual-Purkinje eye tracker was used to measure vergence movements to dichoptic targets with same and opposite luminance polarity. The targets were: randomly placed line segments; a single black/white edge; a single Gaussian spot whose luminance was reversed in one half to form a sharp edge; or multiple such Gaussians. Disparity was changed either sinusoidally with 0.50° amplitude at 0.25 Hz, or in a random, jittery fashion. We stimulated horizontal and vertical vergence in separate recordings, and recorded both matched and mismatched target polarities.
Results:
The targets with opposite polarity showed vergence responses that were reliable, but weaker than targets with same luminance polarity. When present, the vergence response was always in the correct direction. This was true even for a single edge with opposite luminance polarity driving vertical vergence.
Conclusions:
Our results show that vergence responses do occur to targets of opposite polarity under conditions in which they cannot be attributed to matching of same-polarity edges or to voluntary control of vergence.
Supported by R01-EY-12986