Abstract
Observations that apparent motion is robust across features changes, such as color and shape, suggest that surface features cannot serve as correspondence cues to disambiguate ambiguous apparent motion. We revisited this question using a modified Ternus display. In typical Ternus displays, three horizontally adjacent discs are presented, followed by the same discs shifted one position. These two displays are cycling continuously. For long inter-stimulus intervals (ISI), the discs tend to be perceived as moving together as a group. At short ISIs, one element is perceived as jumping from one end to the other, while the other discs remain stationary. We added a small dot to each disc, changing its relative position within the discs over time (Boi, Otto & Herzog, 2008). Consistent with the implied object correspondences, group motion caused the central dot to rotate within its disc, whereas element motion caused the central dot to appear as though it were moving up-and-down or side-to-side within its disc. This modified display allowed us to investigate the influence of surface features without relying on subjective judgments of motion quality. We used surface features to bias either group or element motion. If feature cues are used to solve the correspondence problem in apparent motion, then direction-of-motion reports should reflect the feature bias. If they are not, then direction-of-motion reports should reflect only ISI. All surface-feature correspondence cues yielded large differences between the element and the group bias conditions, indicating that feature cues were used to solve the correspondence problem. Feature identity seems to be an especially important cue, although overall configuration might also play a role. More generally our findings strengthen evidence of the importance of surface feature correspondence cues and their role in the establishment and maintenance of persisting object representations.
Supported by NSF grant BCS-0818536 to CMM.