Abstract
A horizontally moving grating can be made to appear to move obliquely either upwards or downwards by introduction of appropriate depth-ordering cues (Duncan et al., J. Neurosci, 2000). A widely held explanation of such depth-motion interactions is that depth cues allow identification and suppression of motion signals arising from ‘extrinsic’ line-terminators. We show that depth-cues alter the perceived direction of unoriented moving features, a result not accounted for by terminator-based explanations.
Stochastically moving random dots were viewed through a diamond-shaped aperture. The four panels defining the aperture had one of two depth configurations: 1) Upper-left and lower-right panels in front of the moving stimulus (NEAR), lower-left and upper-right behind (FAR), or 2) vice-versa. Motion was oscillatory with some dots moving along the NEAR-to-NEAR axis and the rest moving along the FAR-to-FAR axis. The NEAR vs. FAR percentage ratio was varied from 86:14 to 14:86. Five naïve observers reported the perceived motion axis in a 2AFC direction discrimination task. A no-depth condition provided comparison.
For all observers, psychometric curves for the depth-ordered stimuli were shifted relative to the no-depth condition. For the completely ambiguous condition (50:50 NEAR to FAR ratio), motion was reported along the NEAR axis on 75% of the trials. The effect of depth cues on perceived motion direction is thus not limited to overcoming the motion ambiguity of oriented features (i.e. the aperture problem). Consequences for models of depth-motion interaction will be discussed.