Abstract
Using adaptation, we studied a previously under appreciated form of perceptual bistability associated with viewing point-light (PL) animations portraying biological motion. Without depth cues, conventional PL sequences are ambiguous with respect to depth ordering of the limbs. For some animations, this ambiguity is perceptually evident: one experiences alternations over time between two conspicuously different directions of heading of a PL walker, the alternatives being mirror reversals of one another in depth (e.g., either 30° or 150° with respect to an observer's line of sight). We had observers track reversals in a bistable walker's perceived heading direction following 90s exposure to a walker whose limb positions, and hence heading, were unambiguously defined by stereo-disparity (oriented unambiguously either at 30° or 150°). Following stereo adaptation, observers experienced a strong, consistent aftereffect: the bistable walker appeared to face predominantly in the opposite, unadapted direction. Aftereffects were not found, however, following adaptation to a phase-scrambled walker, in which individual dots maintained the same disparity-defined trajectories as in the intact display, but lacked global temporal coordination. In contrast, aftereffects did occur when adapting and test walkers differed in size. Evidently, the adapted representation is tied closely to the underlying human form and cannot be explained by local, disparity-specific motion adaptation. Aftereffects also occurred when adapting and test walkers appeared on opposite sides of fixation, and fell off with increasing differences between adapt/test heading directions. These results underscore the importance of kinetic depth in the perception of PL animations, and imply strong interactions among signals specifying form, motion and depth in the representation of biological motion. Models of biological motion perception based on an analysis of the kinematics or postures portrayed by PL displays should incorporate mechanisms to disambiguate the inherently ambiguous depth relations that exist in these stimuli.
NIH EY13358 (RB), IRCSET graduate scholarship/Seed Funding UCD (SJ).