Abstract
Most studies concerned with motion perception during smooth pursuit eye movements havefocused on the perception of collinear object motion. Usually, the dependent variable in thosestudies is the perceived speed of the object. In contrast, in the few studies addressing non-col-linear object motion, it is perceived motion direction that has been measured. This makes thetwo classes of studies hard to compare. To address this issue, we developed a method to measureboth perceived motion direction and speed at the same time. A random-dot kinematogram(RDK; direction ranging between 1808and 3608) was first presented during smooth pursuit andin a second interval during fixation. Observers had to match speed and direction in the secondinterval to the perceived velocity in the first one. The results show that the eye movementsaffected only the horizontal component of the perceived velocity. This component was shiftedagainst the pursuit direction, with the shift being approximately constant across RDK motiondirections. Interpreted in terms of a linear model of perceived velocity (perceived velocity equalsestimated retinal velocity plus estimated eye velocity) these results suggest that the errors inperceived velocity are mainly due to underestimation of the eye velocity, not to overestimation of the retinal image velocity.
