Integration of alternating cues to slant

Several studies showed that cue integration is close to optimal when two or more cues are available simultaneously. However, most of these studies consider only constant cues. Here we investigate how different depth cues interact when they are not presented simultaneously but they are alternating. We ask whether there is fusion of cues in time and how the interaction between cues depends on the frequency of alternation. To study this, we presented two surfaces in alternation at six different frequencies (from 0.8 to 15 Hz). One surface was defined by a random-dot pattern displayed in stereo (disparity-defined surface); the other was defined by a monocularly viewed regular texture (texture-defined surface). The angle between the two surfaces was always +20 or −20 degrees. Participants had to indicate whether the texture-defined surface was slanted to the left or to the right. The orientation of the two surfaces was varied jointly using a double staircase procedure to find the orientation at which the texture-defined surface appeared frontoparallel. Results indicate that there is a significant interaction between the cues depending on frequency. That is, the orientation of the stimulus needed to see the texture-defined surface as frontoparallel depended on the sign of conflict and the frequency of alternation. At high frequencies (above 6 Hz) there was a perceptual bias of the texture-defined surface in the direction of the disparity-defined surface, indicating integration of the signals. At low frequency (0.8 Hz), however, this interaction did not only disappear, it reversed in the opposite direction, indicating a contrast effect. This contrast may be explained as an aftereffect resulting from adaptation to the disparity-cue slant. We conclude that simultaneity between cues is not necessary for integration to occur. There seems to be a temporal window for integration in the order of 150 ms