Abstract:  During past years, visual working memory (VWM) consolidation has been studied extensively. Consolidation of visual information into VWM is widely considered to occur with capacity limit. Previous studies have demonstrated that two colors could be consolidated in parallel and two oriented gratings could be consolidated only in serial. Some researchers provided a bandwidth hypothesis for these results: because of the difference between the informational demands for color and for oriented grating, two colors could be consolidated in parallel without exceeding the bandwidth limit, whereas two oriented gratings could only be consolidated in serial because of exceeding the bandwidth limit. But other researchers recently realized that different positions of stimulus could affect the VWM consolidation as well. In the present study, we used change detection paradigm and sequential-simultaneous manipulation to examine whether the spatial distance between memory items could affect the VWM consolidation. In Experiment 1, we briefly presented two masked color patches (targets) within one of three levels of spatial distance (4.6°, 8.5° and 11.1°). A probe color then appeared, and participants needed to judge whether the probe color matched any one of the targets. The results showed that no difference was found between sequential and simultaneous presentations in 4.6° and 8.5° spatial distance conditions, while the performance for sequential presentation was better than simultaneous presentation in 11.1° spatial distance condition. These results suggested that two colors were consolidated in parallel in 4.6° and 8.5° spatial distance conditions, but in serial in 11.1° spatial distance condition. In Experiment 2, we presented a color target firstly, and then two masked color patches appeared and the participants needed to judge whether the color target matched any one of the two masked color patches. As a result, no difference was found in the three given spatial distance conditions, eliminating the possibility that perceptual encoding led to the results in Experiment 1. Considering that the consolidation for oriented grating was different from the process for color and color might also be a unique visual feature, we attempted to use oriented grating to replicate the results of Experiment 1. In Experiment 3, we briefly presented two masked oriented gratings (targets) within three levels of spatial distance (4.6°, 8.5° and 11.1°). A probe cue then appeared, and participants needed to judge whether the cued sinusoidal grating had been oriented in a cardinal (horizontal or vertical) or an oblique (±45º) orientation. Despite superior performance for sequential over simultaneous presentations in all three spatial distance conditions, we found that the performance for parallel presentation was better than the accuracy’ in 4.6° spatial distance condition. The results suggested that parallel consolidation of oriented gratings into WVM was possible when the spatial distance between memory items was short enough. According to the results, we speculated that consolidation into VWM might be related to the distribution of the visual spatial attention, and the consolidation bandwidth increased with increasing resources of attention which memory items could get.

Key words:  视觉工作记忆, 巩固, 带宽