Abstract
A target presented along the trajectory of two-dot apparent motion (AM) is less visible. This phenomenon of motion masking implies that the mental representation of AM can interfere with processing of other sensory inputs. Motion masking was first introduced as suppression of higher processing such as letter identification (Yantis & Nakayama, 1998), but recently, masking has been reported to increase when the difference in orientation between the AM stimuli and the target decreases and thus has been discussed in relation to suppression at an early stage (Hidaka et al., 2011). Our study aims to examine whether motion masking is based only on the similarity in physical attributes between the AM stimuli and the target or whether the AM direction plays some role. We used a random-noise luminance carrier confined within a Gaussian contrast window as each element of the two-dot AM stimuli, and a Gabor patch as the target. In each trial, two vertically or horizontally aligned elements of the AM stimuli were flashed asynchronously, and the target was either flashed between the two elements or absent. Target detectability was measured against Gabor orientation. If masking was based only on the stimulus similarity, orientation dependency should not be observed because random noise has no predominant orientation. Surprisingly, we observed orientation dependency of detectability associated with the perceived direction of AM. When the AM stimuli appeared to move horizontally, a horizontal patch became less detectable than a vertical patch. When the AM stimuli appeared to move vertically, the vertical patch became less detectable. These results were also confirmed when plaid stimuli were used as the elements of the AM stimuli. To reconcile apparent controversy between our findings and previous notions, multiple levels of object processing and differences in task strategy are discussed in relation to the identification of overlapping objects.
Meeting abstract presented at VSS 2013