Tests of the dipole model of perceived movement in apertures

Power and Moulden (1992) have proposed a dipole model to account for the apparent movement of gratings in apertures. This includes movement orthogonal to the orientation of the grating, and the barber pole illusion: the illusion that a grating drifting diagonally across a narrow aperture appears to be moving along it. The essence of the model is that movement is signalled by a large number of dipoles, of many orientations and lengths. These dipoles respond if, and only if, one end is stimulated, and then the other. Three experiments intended to test predictions from the model are reported here. In each case a horizontal grating drifted across an aperture and subjects fixated outside the aperture. In experiment 1 subjects fixated just above or below the aperture, and reported the motion aftereffect (MAE) shown by a set of test spots. As predicted by the model, the spots further from the fixation point showed a strong MAE. Experiment 2 combined both viewing conditions of experiment 1, so that test spots above and below the fixation point were viewed simultaneously. The predictions were confirmed, since test spots further from the fixation point exhibited a stronger MAE than test spots closer to the fixation point. In experiment 3 the fixation point in all conditions was below the aperture, and, as predicted, the MAE of a spot near the bottom of the aperture was diagonally upward, although stimulation was horizontal. Again, as predicted, the MAE of a spot in the middle of the aperture appeared to move horizontally. Finally, it was predicted that a test spot at the top of the aperture would appear to move diagonally downwards, but subjects were unable to report unequivocally the direction of motion, since the MAE was occurring too far from the fovea for clear vision. Overall, then, the predictions from the model were confirmed, although there are associated phenomena the model cannot as yet account for.