Abstract
Arguably, the most important determinant of correspondence matches in long-range apparent motion is the relative proximity of the potential matches. That is, all else equal, correspondence matches are preferred between ‘nearest neighbors’ in retinal coordinates. However, a vast body of research has indicated that similarities between motion tokens on dimensions other than retinal position affect correspondence matches as well, though to a lesser degree. It is investigated here whether the effects of relative color, depth, or spatial-frequency content of motion tokens display characteristics typical of ‘controlled’, attentional processes. In experiment 1, observers were presented with a static display containing all motion tokens for a variable period of time before the motion sequence started. It was found that varying relative color, depth and spatial frequency content of the motion tokens indeed creates a bias to match tokens that are alike. However, it is shown that this bias develops over time. In the second experiment, observers were asked to remember a particular configuration of the colors or spatial frequency contents of motion tokens while judging motion in another configuration of motion tokens. It was found again that relative color or spatial frequency content of motion tokens in the actual motion frames biases correspondence matching. However, the to-be-remembered configuration of tokens affected correspondence matching also, but in an idiosyncratic manner. These findings suggest that relative color, depth, and frequency content of motion tokens affect correspondence matching through a relatively high-level, controlled process.