Retinal and extraretinal contributions to visual stationarity

There are two ways in which moving observers could judge whether threedimensional objects are stationary in an earth-fixed—rather than observerrelative—reference frame. An observer may determine the relative motion of an object as well as his or her own motion relative to an earth-fixed reference frame, and then compare the two: the extraretinal criterion. Alternatively, the observer could heuristically assume that the background in the image is stationary, and decide the stationarity of a foreground object on purely retinal criteria. We performed two experiments to determine which of these criteria is actually applied to make stationarity judgments. In the first experiment we show that both retinal and extraretinal criteria for stationarity are in themselves sufficient, but that retinal criteria yield stationarity judgments that are accurate but imprecise, extraretinal criteria are precise but sometimes inaccurate, and that the combination of the two is both accurate and precise. In the second experiment, we show that when both types of stationarity criteria are available, both are utilized; the relative weights, however, have high inter-individual variations.