Abstract
Abstract: :
Purpose: how do humans combine the velocity information from two moving gratings to detect pattern motion direction. Three previously reported plaid motion effects were investigated with a view to examining the possibility that they could be predicted by a single hypothesis. Methods: three experiments were carried out representing the three effects. All three effects were reproduced using a single two-component moving stimulus where only the speed of one of the components was varied to produce all three effects. In each experiment the stimulus representing one effect only was randomly rotated in the plane so that the stimulus moved in 90 different orientations. Subjects were asked to make a forced choice as to whether the stimulus moved in the same direction or not as an oriented line. The orientation of the line was randomly generated from the predicted direction of one of three different hypotheses: 1. the intersection of constraints rule (Adelson and Movshon, 1982) 2. the vector sum (Yo and Wilson, 1992); and 3. pattern motion predicted by components introduced after squaring the pattern (Derrington et al., 1992; Lu and Sperling, 1995). Results: the method produced very clear results, and facilitated a complete mathematical description of all possible stimulus change showing how only hypothesis 3 'squaring' could account for all of the main results. However, in one of the experiments the results were not completely invariant under rotation, and hypothesis 3 would not predict this. Conclusion: squaring the plaid introduces two very salient components that predict three results not normally attributed to squaring. A different model proposed by Bowns (1999, 2001) also unifies the results without squaring, and includes information corresponding to the two directions reported in the experiment where the results are not invariant to rotation.