Time-Based Transition Expectancy in Task Switching: Do We Need to Know the Task to Switch to?

Abstract

Recent research has shown that humans are able to implicitly adapt to time-transition contingencies in a task-switching paradigm, indicated by better performance in trials where the task transition (switch vs. repetition) is validly predicted by the pre-target interval compared to trials with invalidly predicted transitions. As participants switched between only two different tasks, not only the transition, but also the specific task was predictable; at least indirectly when taking into account the temporally predicted transition in the current trial together with the task in the previous trial. In order to investigate if the time-based expectancy effect for transition in previous studies was due to a specific task preparation or due to an unspecific transition preparation, three different tasks were used in the present study. One of two possible pre-target intervals (500 and 1500 ms) predicted a task switch in the upcoming trial with 90 % probability, whereas the other interval predicted a task repetition with 90 % probability. Results revealed that participants were able to prepare both upcoming repetition as well as switch requirements based on predictive pre-target intervals. This means that humans seem to be able to prepare a task switch in a rather unspecific manner, most likely by inhibiting the task just performed in the previous trial. By suggesting a two-stage preparation model in which switches as well as repetitions benefit both from time-based transition expectancy, although apparently with different cognitive processes being involved, the present study provides important impulses for future research on the cognitive processes underlying human task-switching behavior.

Keywords: Task switching; preparation; time-based expectancy; timing.

Figure 1

Schematic illustration of probabilities of task A, B and C in trial n when task A has been presented in trial n-1. In the depicted exemplary experimental condition, the long pre-target interval predicts a task repetition, whereas the short pre-target interval predicts a task switch.

Figure 2

Schematic illustration of the experimental between-subject condition “interval of expected repetition”: Whereas one half of participants (group 1) expected a task repetition to occur frequently after 500 ms in trial n, the other half of participants (group 2) expected a task repetition to occur frequently after 1500 ms trial n.

Figure 3

Mean reaction times (RTs in ms; lines) and percentages of errors (PEs in %; bars), depending on predictability of interval – transition combination, are displayed separately for task transition and the between-subject condition “interval of expected task repetition” (500 ms vs. 1500 ms). Error bars represent 1 standard error of the mean.