Abstract
A number of studies have shown that visual working memory (VWM) can hold approximately four objects, but it is not clear whether this capacity of memory is purely object-based or modulated by number and type of bound features. Luck and Vogel (1997) claimed that the unit of memory capacity is an “object” and the number of integrated features does not matter. On the other hand, Wheeler & Treisman (2002) argued that when features come from the same dimension, (e.g. bicolored squares) capacity tends to decrease. Despite of these differences, it is accepted that as long as there is only one feature integrated from each feature dimension, there is no decrease in memory capacity relative to single feature objects. That is to say, the storage capacity does not change, for example, between simple colored patches and colored shapes. Although so far, most studies concentrated on integration of features represented in the ventral pathway, such as color, shape, orientation, size, etc.
The purpose of this study is to investigate whether the upper comment hold when features coded in different visual pathways, say, the ventral and dorsal pathways are integrated in a single object. We performed change-detection tasks using objects defined by colored dynamic random dots. There were 7 conditions, namely, color, shape, direction, color-shape, color-direction, shape-direction and color-shape-direction. We used single probe tests instead of whole-display tests to avoid the problem of interference when testing. Firstly, our experimental results confirmed the previous results where there was no drop in capacity by combining color and shape. To our surprise, the feature “direction” acted quite differently. The capacity dropped significantly when direction was combined with color, shape and color-shape. We predict that there is some interference in visual working memory storage between features from different visual pathways.