Fig. 5: Predictions for motion and depth perception in the R and R+T viewing geometries.

A Center, stimulus display showing a fixation target (yellow square) and an object (soccer ball shape) moving up and to the right (illustration of the Fixation condition). Meanwhile, background dots (red/green triangles) simulate the R (top) or R+T (bottom) viewing geometries. Top, in the R geometry, a rightward eye velocity (yellow arrow) is added to the image motion of the object, resulting in a rightward bias in motion perception (orange arrow). The object’s depth remains ambiguous due to the absence of reliable depth cues. Bottom, in the R+T geometry, the horizontal component of the object’s image motion can be explained away as motion parallax, such that the object is perceived as having a near depth. The residual vertical motion is then perceived as the object’s motion in the world (blue arrow), leading to a vertical bias in perception. B Summary of the relationships between retinal motion, eye velocity, and perceived object motion in the R (top) and R+T (bottom) geometries. C Same format as A except that the object moves up and to the left on the display (center panel). In the R geometry (top), depth remains ambiguous, and the same rightward eye velocity is added to the object’s motion, again producing a rightward bias. In the R+T geometry (bottom), the horizontal component of the object’s image motion reverses direction, thus causing a far-depth percept. The perceived motion remains vertical. D Same format as B, except for the scenarios depicted in (C).