Abstract
Bottom-up attention is driven by stimulus features, however most studies have not distinguished whether it is occurring at the level of the object or the feature itself. This is because those studies employed spatially separate stimuli. To study object-based selection, we superimposed two surfaces (random dot kinetograms, RDKs) to control for location-based mechanisms. Another advantage of using this paradigm is that RDKs have a set of well defined parameters which allows us to vary different features (e.g. speed) systematically. We performed 3 experiments to investigate the effect of speed on surface selection. In experiment 1, subjects fixated a central dot and an aperture with a single surface of dots moving left or right appeared in the periphery. After a random period of time, the fixation spot disappeared which was the cue for the subjects to saccade to the aperture. The speed of the surface varied trial-by-trial, from 0.6–24 deg/sec. Saccading to the surface resulted in an automatic pursuit of that surface. Pursuit speed was proportional to surface speed. In experiment 2, a second surface was placed in the aperture, moving at a constant speed in the opposite direction. The other surface varied in speed, and pursuit was again measured. In contrast to theories suggesting that higher speeds are more salient, subjects had no preferential selection for a faster moving surface when presented with 2 superimposed surfaces moving in opposite directions. Experiment 3 varied one surface's speed while the other surface was static. At slow speeds, automatic pursuit was not detected. However, at fast speeds, the moving surface was pursued even in the presence of a static surface. Overall, these findings suggest that motion is salient when presented alone or when presented against a static surface but not when presented against opposing motions.
Funded by NSERC Discovery Grant to M.F.