Abstract
In the double-drift illusion, the carrier and envelope of a Gabor patch have different directions. When this stimulus is viewed in the periphery, the perceived direction and position dramatically diverge from the physical path. This illusory direction may result from the effect of the Gabor’s aperture on the visible features of the carrier. For example, if the Gabor has only a single visible cycle of the carrier, the brightest and darkest points follow the illusory direction as each cycle slides into and out of the orthogonally moving Gabor aperture. Is this motion of the peak the cause of the illusion, or is it instead the combination of internal and external motion vectors, as others have suggested (Heller et al., 2021)? By replacing the Gabor carrier with type-2 plaids, we tested these hypotheses in one stimulus. In a type-2 plaid, the two component directions (for example 45° and 0°) can be radically different from the direction of the pattern they produce and its brightest points. With the external motion parallel to the faster component, the illusion might be generated by combining the external motion with either the component motions or with the pattern motion (the light and dark blobs). Our data rules out the pattern motion and therefore the tracking of the brightest point as the source of the illusion, thus favoring the vector combination model.