Abstract
Hock, Gilroy & Harnett (2002) have shown that counter-changing luminance provides the informational basis for the perception of luminance defined, single-element apparent motion (AM). That is, irrespective of the presence of Fourier-based, 1st-order motion energy, AM is perceived only when the luminance at one element location changes toward the luminance of the background and the luminance at a second element location changes away from the luminance of its background. The purpose of the experiments reported here was to determine whether the effects of the oppositional luminance changes at each element location are combined multiplicatively, as in various forms of Reichardt's correlational model, or additively, as in Adelson & Bergen's motion energy model. The experiments are based on the generalized version of back-and-forth AM between two locations; elements are simultaneously visible at both locations and their luminance values change in opposite background-relative directions. Whether or not motion is perceived depends on the magnitude of the background-relative luminance change (BRLC) (Hock, Kogan & Espinoza, 1997). In this study, the BRLC value at each element location was varied independently; e.g., it might be larger for one element than the other during some trials, similar in magnitude for both elements during other trials, and so on. The likelihood of motion being perceived was best predicted by the product of the BRLC values, much less so by their sum. This evidence for perception of AM depending on the multiplicative combination of activation changes at nearby element locations is consistent with the Reichardt correlational formulation, but the perception of motion when luminance simultaneously changes at both locations (in opposite background-relative directions) is inconsistent with the “delay-and-multiply” principle that defines standard- and elaborated-Reichardt models.