The effect of disparity/vertical-scaling conflict in a stereoacuity task

Purpose. The precision with which the relative depth of two vertical lines is judged, can be as little as 5 arcseconds. However, McKee (Vision Research 23,1983) showed difficulties in detecting relative depth when the vertical lines are perceptually linked in some fashion; for example, when they form part of a real or imaginary square. McKee's stimuli were of a constant physical length, and thus not scaled according to the changes in depth signalled by horizontal disparity. Thus it is possible that the two depth cues conflicted. Methods. Two vertical lines or horizontally connected vertical lines were viewed in a modified Wheatstone stereoscope and stereoacuity was measured using a 2AFC paradigm. Each observation interval was 1 s. Two sizes, either 25 or 3? were used, and three conditions were tested: a) both the horizontal disparity and the vertical scaling associated with real rotation in space were available to the observers, b) vertical scaling consistent with the rotation was removed, c) vertical scaling alone was available. Results. 1) When disparity and vertical scaling provided mutually consistent information about relative depth, stereoacuity for lines was only slightly better than that for the closed configuration. 2) Conflicting vertical scaling information had little effect on stereoacuity for lines, but it overrode disparity information when closed configurations were used. 3) Vertical scaling alone proved an effective depth cue when the lines were connected, but not when they were separate. 4) The impact of vertical-scaling manipulations was bigger for larger stimuli. Conclusions. A disparity\vertical-scaling conflict, introduced when the physical length of the vertical lines is kept constant despite changing horizontal disparity, may provide a possible explanation for the marked deterioration in stereoacuity for closed configurations.