Abstract
It is useful to be able to detect objects that move in an Earth-fixed, allocentric reference frame as well as those that move with respect to oneself. It is known that moving objects are detected more efficiently than stationary ones (Ivry and Cohen, 1992 JEP:HPP 18 1045–1057); however, since these experiments were conducted on immobile observers, it is not known whether it is egocentric or allocentric motion that is detected. We dissociated the two types of reference frame by working with observers that moved the head or the eyes, and studying the detection of objects whose motion was yoked to that of the observer. Thus, we were able to create motion selectively in several different reference frames. Preliminary studies (ECVP, 02) with the head moving and the eyes stationary in the head showed that allocentric motion is detected as efficiently as egocentric (motion relative to head end eyes). New experiments conducted with an observer performing smooth pursuit (but with the head stationary) corroborate these results: subjects were able to detect allocentric motion (with minimized retinal slip) as well as egocentric. However, efficient detection of allocentric motion takes place only if the presentation time is greater than about 130–150 ms.
Jacques Droulez