Dual-task effects in working memory: Interference between two processing tasks, between two memory demands, and between storage and processing

Abstract

Two experiments with a working-memory updating task investigated dual-task interference between short-term storage of numerical and spatial material, between execution of a numerical and a spatial updating operation, and between storage and processing. Participants memorised a set of digits and a set of spatial positions, updated elements of both sets by a sequence of operations, and then recalled the final values. In Experiment 1, a single element in each memory set had to be updated several times. There was little interference between storage of the two sets, and between storage and processing, but parallel execution of the two operations was not possible. In Experiment 2, all elements in both memory sets were updated in random order. There was substantial interference between memory for the numerical and the spatial sets, and between storage and processing. Parallel execution of two operations was again not possible. Moreover, trying to do two operations simultaneously resulted in impaired memory of final results. The results support the distinction between the activated part of long-term memory that can hold elements currently not needed for processing, and a more central, capacity-limited part of working memory that provides access to its contents for processing.

This research was supported by Deutsche Forschungsgemeinschaft (DFG), grants KL 955/6-1 and OB 121/3.

This research was supported by Deutsche Forschungsgemeinschaft (DFG), grants KL 955/6-1 and OB 121/3.

Acknowledgments

We thank Reinhold Kliegl for many inspiring discussions in all phases of this project.

Notes

This research was supported by Deutsche Forschungsgemeinschaft (DFG), grants KL 955/6-1 and OB 121/3.

¹By a functional bottleneck we mean a restriction due to executive parameters in the cognitive system that are set to serve a function (e.g., to avoid crosstalk between two processes). Under suitable conditions, these parameters can be changed and the bottleneck disappears (cf. Oberauer & Kliegl, 2004).

²The same explanation could be applied to Experiment 1: Here, too, the simultaneous condition forces holding two elements in the direct-access region throughout a trial, whereas only one of them is required there in the sequential condition. In Experiment 1, however, only two elements compete for capacity in the simultaneous condition, whereas in Experiment 2 it can be up to five when both set sizes are large. The continuous competition among one numerical and one spatial element might not be sufficiently severe to generate a noticeable decrease in accuracy in the simultaneous relative to the sequential condition.