Evidence for modality-independent order coding in working memory

Abstract

The aim of the present study was to investigate the representation of serial order in working memory, more specifically whether serial order is coded by means of a modality-dependent or a modality-independent order code. This was investigated by means of a series of four experiments based on a dual-task methodology in which one short-term memory task was embedded between the presentation and recall of another short-term memory task. Two aspects were varied in these memory tasks—namely, the modality of the stimulus materials (verbal or visuo-spatial) and the presence of an order component in the task (an order or an item memory task). The results of this study showed impaired primary-task recognition performance when both the primary and the embedded task included an order component, irrespective of the modality of the stimulus materials. If one or both of the tasks did not contain an order component, less interference was found. The results of this study support the existence of a modality-independent order code.

Keywords:

• Working memory
• Serial recall
• Visuo-spatial memory

The research reported in this article was supported by Grant 10251101 of the Special Research Fund of Ghent University to the second author. The authors thank Arnaud Szmalec and three anonymous reviewers for their critical comments on an earlier version of this article.

Notes

1 In the first order task, filler items were included (two in the single-task condition and four in the dual-task condition). In these filler trials, the first or last square was switched in the test phase. These trials were not included in the data analysis. The reason for including these is to vary the positions probed in the test. However, the first and last items in a list are less confusable because they have only one neighbour (end anchor effect, see, e.g., Smyth & Scholey, 1996). These items are remembered better, and when they are included in the test phase these trials are easier than the trials comprising the items in-between. To minimize the chance of ceiling effects, these trials were not used in the analysis.

2 Because the planned contrasts used were orthogonal also r ² is added, because this indicates the percentage of variance of the overall effect that is explained by the contrast.

3 Fillers similar to those used in Experiment 1 were included; these fillers were not analysed.