Abstract

In 1965 (Science 150, 348-350) I first showed that stereoscopic depth contours could be generated across large homogeneous central regions of visual space and that these contours were real in that they reversed directly when the sign of the disparity in the frame also reversed, provided that the inducing frame(s) was of sufficient strength. I left the issue unsettled as to what that strength must consist of. Much work has been done by others since, but no one seems to have advanced significantly on this key issue. It thus seemed worthwhile to devise some new targets and to repeat and extend the early measurements. These contours seem limited to a range of ~5° of visual angle, and they are most robust, paradoxically perhaps, when they pass directly across the foveal region, hence, between the two interhemispherical cortical projection areas. Consequently, neural processes must extend over these ranges, and any adequate model of stereopsis must take account of this impressive low frequency global event. The contrast to high spatial frequency stereoacuity processes seems worthy, also, of reemphasis.

© 1989 Optical Society of America