Horizontal vergence can be stimulated binocularly with disparity (disparity vergence) or monocularly with accommodation (accommodative vergence). The latter results from a neural cross-coupling that causes both horizontal vergence and accommodation to respond when either one is stimulated [Alpern, M., & Ellen, P. (1956). American Journal of Ophthalmology, 42, 289-303]. The velocity of disparity and accommodative vergence is enhanced when accompanied by saccades [Enright, J. T. (1984). Journal of Physiology (London) 350, 9-31; Enright, J. T. (1986). Journal of Physiology (London) 371, 69-89]. Based upon the coupling between accommodation and vergence, we predicted that accommodation should also be facilitated by saccades. An SRI Dual Purkinje Eyetracker was used to measure left and right eye position, and the accommodation of the left eye, in response to stimulation. Horizontal saccades were stimulated by targets separated by 2-6 degrees and accommodation was stimulated monocularly over a range of +/- 2 diopters (D). When saccades occurred within 0-400 ms following a monocular step stimulus to accommodation, latency of accommodation decreased and the associated accommodative-vergence response was synchronized with the saccade. Saccades also enhanced the velocity of accommodation and accommodative-vergence, and this facilitation increased with saccade amplitude. Transient vergence responses that are normally associated with saccades [Erkelens, C. J., Steinman, R. M., & Collewijn, H. (1989). Proceedings of the Royal Society of London B. Biological Sciences, 236, 441-465; Maxwell, J. S., & King, W. M. (1992). Journal of Neurophysiology, 68 (4), 1248-1260] did not affect accommodation when it was not stimulated by defocus. Because saccades and accommodation utilize separate plants and final common pathways, the synchronization of saccades and accommodation and the enhanced velocity of accommodation and accommodative-vergence must occur at more central sites. Possibilities include the superior colliculus, which represents both accommodation and saccades [Nagasaka, Y., & Ohtsuka, K., (1998). Investigative Ophthalmology AVRO supplement], vestibular nuclei which project to regions near the oculomotor nuclei [Lang, W., Buttner-Ennever, J. A., & Buttner, U. (1979). Brain Research, 177, 3-17], and interactions between omni pause neurons and near response cells of the mesencephalic reticular formation (MRF) [Mays, L. E., & Gamlin, P. D. R. (1995a). Current Opinions in Neurobiology, 5, 763-768; Mays, L. E., & Gamlin, P. D. R. (1995b). Eye movement research: Mechanisms, processes and applications. New York: Elsevier] which represent both accommodation and vergence [Judge, S. J., & Cumming, B. G. (1986). Journal of Neurophysiology, 55, 915-930; Zhang, Y., Mays, L. E., & Gamli, P. D. R. (1992). Journal of Neurophysiology, 67, 944-960].

The tuning of the transient-stereopsis system to luminance contrast and spatial-frequency (SF) was investigated with narrow-band gabor targets with a constant sigma of 1 degree. They were presented for brief (140 ms) durations and subtended a large (6 degrees) disparity. When dichoptic gabor stimuli were matched in SF (0-5 cpd), transient stereo performance was either uniform across SF or greater at frequencies below 1 cpd. When dichoptic stimuli had unmatched SF (0.5 + 0-5 cpd) and matched contrast (100%), stereo performance was impaired below that of the matched SF condition. Stereo performance with matched SF at 0.5 cpd was impaired when contrast of one eye's image was reduced, demonstrating a contrast-paradox effect (i.e. contrast tuning) for transient stereopsis. Performance with three dichoptic unmatched SF conditions (0.5 and 1.0 cpd; 0.5 and 5.0 cpd; 1.5 and 3.5 cpd) was improved when the contrasts of the low SF gabor was reduced while holding the contrast of the high SF gabor constant at 100%. However stereo performance was not improved by reducing the contrast of a high SF gabor (3.5 cpd) while holding the contrast of the lower SF gabor (1.5 cpd) constant at 100%. We interpret these findings as indicating that transient-stereopsis performance is mediated by a single spatial-channel that has low-pass spatial-frequency sensitivity and which compares the ocular based signals prior to binocular combination so that signals that are not balanced in terms of their strength lead to a weaker binocular signal, as per the model proposed by Kontsevich and Tyler (Vis Res 1994; 3417:2317-2329) for sustained stereopsis.

The ability of observers to extract depth from opposite luminance-contrast-polarity stimuli was investigated. The stimuli consisted of two dichoptic-pairs of Gaussians, with one of the Gaussians in each pair having a positive contrast-polarity and the other a negative contrast-polarity. Stimulus durations ranging from 0.2 to 4 s were used. This range of durations was employed to reveal stereo mechanisms that were preferentially sensitive to transient or sustained stimuli. Stimuli were presented in a raised-cosine temporal envelope. Performance with stimuli of the same contrast-polarity was also tested. Observers could easily perceive depth with the same-polarity stimuli, at both long and short durations. Depth could be perceived with low-contrast opposite-polarity stimuli only at short durations. However, depth could be perceived with long-duration stimuli presented within a raised cosine temporal-envelope if a high contrast was used. Depth could also be perceived with low-contrast long-duration stimuli if they were presented within a rectangular temporal-envelope. These findings suggest there are separate sustained and transient mechanisms for stereopsis and that the transient-stereoscopic system can extract depth from opposite-contrast stereograms while the sustained system cannot. Further, it is likely that depth perception with opposite-contrast stereograms found in many previous studies was mediated by the transient-stereopsis system.

Large-field stimuli were used to investigate the interaction of first- and second-order pathways in transient-stereo processing. Stimuli consisted of sinewave modulations in either the mean luminance (first-order stimulus) or the contrast (second-order stimulus) of a dynamic-random-dot field. The main results of the present study are that: (1) Depth could be extracted with both the first-order and second-order stimuli; (2) Depth could be extracted from dichoptically mixed first- and second-order stimuli, however, the same stimuli, when presented as a motion sequence, did not result in a motion percept. Based upon these findings we conclude that the transient-stereo system processes both first- and second-order signals, and that these two signals are pooled prior to the extraction of transient depth. This finding of interaction between first- and second-order stereoscopic processing is different from the independence that has been found with the motion system.

Stereo-perception appears to be mediated by at least two systems: a transient system that processes stimuli presented briefly and a sustained one that processes stimuli presented for longer durations. In this paper we investigated the tuning of the transient-stereopsis system to stimulus orientation. Narrowband-gabor targets with a constant envelope size (Gaussian standard deviation of 1 degree) were presented for brief (140 ms) durations at large (from 4 to 8 degrees) disparities. The results were as follows: (1) while observers could extract depth from orthogonally-oriented gabors at above chance levels, their performance was worse than that with gabors of matched orientation; (2) varying the relative contrasts of the two orthogonally oriented gabors of the same spatial frequency resulted in a reduction in performance; (3) varying the relative spatial frequencies of the orthogonally-oriented gabors impaired performance, relative to that for matched frequencies; and (4) varying the relative contrasts of orthogonal gabors that were at different spatial frequencies could improve performance. These results indicate that transient stereo-performance in the orthogonal condition was not mediated by the channels that extracted depth in either the horizontal- or vertically-matched gabor conditions. This apparent lack of orientation tuning is indicative of a second-order pathway. That this performance was mediated by a binocular, as opposed to a monocular channel, is supported by the finding that performance decreased as the contrast of one of the gabors was reduced. The finding that performance with orthogonal gabors of unmatched spatial frequency (0.5 and 4 cpd) could be improved by varying their relative contrasts suggests that the binocular spatial-frequency tuning exhibited by this channel is broadband in nature. Finally, the observation that lowering the contrast of either the high or low spatial-frequency gabor improved performance suggests the presence of at least two broadband channels: one with its peak sensitivity at a low and the other at a high spatial-frequency.