台大心理系

Visual working memory plays critical roles in our understanding of visual world, by integrating perceptual present and immediate past. The mechanism by which non-spatial features, such as color and shape, are bound in visual working memory, and the role of those features' location in their binding, remains unknown. Considering the fact that objects moves continuously, feature binding mediated by spatiotemporal locations may be beneficial. In contrast, since movements of our eyes, head, and body often induce unpredictable changes in retinal image including disappearance of objects, feature binding unbound to locations may be preferred. To investigate these issues I modified a redundancy-gain paradigm. A set of features was presented in a two-object memory display, followed by a single object probe. Participants judged whether the probe contained any features of the memory display, regardless of its location. Response time distributions revealed feature coactivation only when both features of a single object in the memory display appeared together in the probe, suggesting that a shared location is necessary in the formation of bound representations but unnecessary in their maintenance. Electroencephalography data showed that amplitude modulations reflecting location-unbound feature coactivation were different from those reflecting the location-sharing benefit, consistent with the behavioral finding. These findings are generalizable to moving objects and letters. Taken together, location unbound color-shape binding representations plays a role in visual working memory.