Abstract
When an observer moves in a straight line while making an eye movement, he must account for the eye rotation to judge his direction of motion correctly. Previous experiments have led to disparate conclusions as to whether the human visual system accomplishes this from visual information alone, or requires extra-retinal information. The visual stimulus used in many of these experiments was ambiguous and could be interpreted as either motion in a straight line with an eye movement or as motion on a curved path (Royden, 1994). Here we examine what features influence this interpretation. Observers viewed a simulated scene depicting motion over a ground plane, on which there were 45 rectangular objects. The simulated motion consisted of a translation plus a rotation, as if the observer were moving in a straight line while making an eye movement. Observers fixated a stationary central cross during each trial. We varied four experimental parameters: random dots vs. textured surfaces, display size (39 × 51 sq. deg vs. 94 × 77 sq. deg), duration (1.0 vs. 1.5 sec), and the presence or absence of explicit instructions to the observers. At the end of each trial, observers positioned a cursor to indicate their perceived motion direction. In runs with explicit instructions, observers were told that the simulations were of motion in a straight line while rotating, as if they were on a swivel chair. Five to seven observers were tested for each set of conditions. In all cases without explicit instructions, observers showed a bias in the direction of rotation (averaging 17.5 deg for the 8 deg/sec rotation rate, for the textured surface) consistent with perception of motion on a curved path. Screen size, texture, and duration had little effect on this bias. When given explicit instructions, the bias was significantly reduced to 4.4 deg for 8 deg/sec rotation. This shows that people are able to interpret this ambiguous stimulus differently depending on the instructions given.
Supported by NSF Grant #IBN-0196068.