An explanation for unidirectional motion aftereffects following adaptation to bivectorial transparent motion

Two spatially superimposed fields of dots, moving coherently, but in, say, opposite or orthogonal directions, will appear to observers as two sheets of dots sliding over one another. However, prolonged exposure to such transparent motion does not produce transparent motion aftereffects; instead aftereffects are unitary and bear a direction opposite to the vector sum of the adapting dot fields. Why does adaptation to bivectorial transparent motion give rise to a unidirectional motion aftereffect is an intriguing and important question. We offer an explanation for this phenomenon, based on well established neurophysiological and psychophysical results. It is well known that locally balanced motion signals in opposite directions, i.e. counterphase modulated gratings or locally paired dots moving in opposite directions, do not give rise to transparency, but instead appear as directionless flicker. Additionally, in the case of locally paired orthogonally moving dots, motion again is not transparent, but perceived in the vector-average direction of the two components. After adaptation to bivectorial transparent motion, we argue that the motion system faces the same challenge as when viewing such locally balanced motion signals: different directional signals arise from exactly the same locations in the visual field. In short, bivectorial motion aftereffects do not appear transparent for the same reason locally balanced motion signals do not appear transparent. In both cases, competitive interactions between local direction detectors establish a single local velocity.