Parallel dual-task processing and task-shielding in older and younger adults: Behavioral and diffusion model results

ABSTRACT

Background/Study Context: The study investigated the Backward Crosstalk Effect (BCE) in dual-task situations, that is, the observation that Task 2 characteristics can even influence Task 1 processing. This observation suggests that the tasks are processed in parallel. Besides determining the existence of a BCE in a group of older adults, the size of the BCE was compared to that in a group of younger adults. Importantly, recent studies yielded unclear results.

Methods: Twenty-four younger and older adults (19–27 and 58–71 years of age, respectively) performed a dual-task experiment, where Task 1 required a left/right manual response and Task 2 required a left/right foot response. The BCE manifests in shorter Task 1 RTs if both responses are given on the same side (compatible) compared to when they are given on different sides (incompatible). Data were analyzed by Analyses of Variance and diffusion modeling.

Results: Both age groups clearly exhibited a BCE, and the BCEs were of the same size. Further, for both age groups, the size of the BCE was similarly modulated by the previous trial’s compatibility status. Diffusion model analyses attribute the BCE to an increased drift rate in compatible compared to incompatible trials, and also revealed no age group differences in any of the analyzed parameters.

Conclusion: The results point to an aspect of cognition that seems not to show age-related deteriorations, similar to, for example, n-2 repetition costs in task-switching situations. Certain response selection-related aspects of task processing are processed in parallel to the same degree in younger and older adults, and both age groups are similarly able to shield Task 1 processing from interfering Task 2 processing in a dual-task situation.

Notes

¹ GPower calculated the total required sample size as n = 46 to reach a power of 0.9 to detect a medium-sized effect of the here critical between-within interaction. Thus we collected data from 48 participants, that is, 24 per group to allow for a full counterbalancing of the stimulus-response mappings.

² Grouped responses occur when participants withhold the Task 1 response until they have selected both responses, which they then emit in rapid succession. This strategy can induce a BCE-like RT1 difference which would actually reflect RT differences in selecting the Task 2 response. A minimum inter-response interval of 120 ms is rather conservative in comparison with other studies (e.g., 50 ms in Hommel, 1998; 100 ms in Janczyk, 2016).

³ The basic results patterns for Task 1 (see below) remained the same when this outlier elimination procedure was applied.