Purchase this article with an account.
Jun Saiki, Hirofumi Miyatsuji; The role of attention in maintenance of feature binding in visual working memory. Journal of Vision 2004;4(8):148. https://doi.org/10.1167/4.8.148.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
In perceiving a dynamic visual event with occlusions or other disruptions, feature integration between perception and memory is necessary, in addition to integration within perception. Contrary to the widely held view of maintenance of 4 to 5 object representations, we recently found that our capacity to hold feature bindings is more limited (Saiki, 2003a, 2003b). A paradigm called multiple-object permanence tracking (MOPT) investigated memory for feature binding by requiring observers to identify a violation in combination of features of multiple moving objects occurring during an occlusion period. A sequence in MOPT contains constant sets of features and spatiotemporal locations, and changes occur only in their combinations. Using objects defined by shape and color, we investigated the role of object type (shape-color conjunction) and selective attention in the maintenance of object representations. A switch event is either followed by another switch (switch-back), allowing observers a chance to focus attention to a changing object, or with no switch-back. With 4 objects, observers' performance impaired as the motion speed increases, replicating the previous results and suggesting that object type does not facilitate performance. When objects are stationary, identification of a binding violation was accurate only in the switch-back condition. Analyses of response types with mathematical models and additional experiments with combinations of mixed switch types confirmed that the improvement in the switch-back stationary condition was not simply due to the number of switch events, but to the availability of selective attention. These results suggest that object movement is not the sole determinant of the maintenance of feature binding, and that attention is necessary to keep the feature binding in memory. We perceive dynamic events with limited amount of memory representations by interactive integration of memory and perception mediated by visual attention.
This PDF is available to Subscribers Only