Abstract:

It has been widely acknowledged that visual working memory (VWM) only maintains and manipulates a limited amount of visual information (e.g. 3~4 objects). Studies on how VWM stores and processes objects have been effective and have yielded useful results. Previous studies in this field are focused on the processing mechanisms of individual objects. However, the mechanisms for grouping and organization of visual objects have received an increasing amount of attention recently. Among numerous types of organizations, configuration, i.e. the spatial formation of visual objects, plays a vital role in understanding the flexibility of human VWM. Many studies have shown that configuration had significant influence on VWM performance. However, configuration was usually employed as an approach to explore the effect of other factors on the memory system. Only a few studies have directly addressed the mechanism of the configuration VWM. In this study, we attempt to determine the critical factor of the information contained in a given configuration that influences VWM. We then attempt to outline the underlying mechanism of the processing in memory system for a given configuration.

In experiments 1 and 2, we separately controlled the variation of two aspects for a given configuration: the spatial position (i.e. relative position of an object described in left, right, and up/down, will be changed when configuration rotates), and the geometric shape (i.e., shape of the polygon with the objects as its vertexes, irrelevant to rotation). These two factors might vary independently or simultaneously. If the spatial position is the dominant factor in the influence of configuration on VWM, the performance will improve when the spatial position is kept constant. Consequently, broken or rotated geometric shapes will not affect memory performance. If geometric shape is the critical factor, the effects would be in reverse, wherein the performance would improve if the geometric shape is kept constant. The results from our experiments indicate that spatial position is the dominant factor. There was no significant difference on VWM performance between same-position-different- shape conditions (geometric shape varied while spatial position preserved) and the baseline condition (configuration were completely identical). The memory performance declined significantly against the baseline level when spatial position changed, regardless of the geometric shape.

In experiment 3, we systematically considered the rotation of a given configuration. A rotation clue was provided such that mental rotation can be conducted on the previewed configuration. The paradigms were identical to experiments 1 and 2; however, the controlled factor of experiment 3 was the consistency between the rotation clue and the spatial position/geometric shape in test phase. When spatial position is consistent with the rotation clue, the memory performance was better. Once again, the geometric shape showed no significant effect.

These findings suggest the following: First, when a given configuration boost the efficiency of VWM, the primary factor is the spatial position of the object. Second, the overall geometric shape has no significant influence on VWM performance. Third, the role of these two factors maintains the same under mental rotation of a given configuration.

Key words: visual working memory, objects, configuration