The Integration Of Chinese Characters In Visual Working Memory

Visual working memory is a capacity limited system that is not only responsible for brief storage and manipulation but also actively interacts with online perceptual process.A series of recent studies revealed its critical role in integrating spatio-temporal scattered visual inputs into a coherent representation according to bottom-up Gestalt principles(especially the law of collinearity,closure,and similarity),which provided evidence for the interactive model of perception and VWM.However,there are also objects defined top-down by knowledge such as artificial rules acquired by learning,and whether the learned knowledge also guide the VWM integration remains unknown.This study examined whether VWM integration also occurs based on top-down knowledge acquired by learning.By sequentially presenting Chinese characters for participants to memorize,we found that discrete characters could be integrated into words inside VWM and a spatial object-based-attention effect was induced:when attention was cued to a location occupied by the integrated VWM word,the target presented at the location of that word was perceived as occurring earlier than that presented at the location of a different object(Experiment la,lb).This effect occurred even when the phonological loop was occupied(Experiment 2a,3a),or when the characters were presented in transposed order(Experiment 2b,3b).These results showed that discrete visual contents could be integrated into higher-level unit according to lexical knowledge.Moreover,these findings together suggest that top-down knowledge can govern VWM integration,therefore enhancing visual processing by introducing the power of learned knowledge.