Trial-to-trial modulation of task-order switch costs survive long inter-trial intervals

Dual-tasking procedures often involve the successive presentation of two different stimuli, requiring participants to execute two tasks in a particular order. Performance in both tasks suffers if the order of the tasks is reversed (i.e., switched) compared to the directly preceding trial. This task-order switch cost is reduced, however, if the preceding trial itself involved a task-order switch compared to a task-order repetition (Strobach in Acta Psychol 217:103328, 2021). Theoretical accounts range from assumptions of top-down implementation of a task-order control set, or passive persistence thereof, to priming based on episodic binding of tasks and temporal positions. Here, we tested these accounts by investigating whether the sequential modulation decays as a function of the inter-trial interval. Task-order switch costs were reliably reduced after a task-order switch (compared to after a task-order repetition) and this reduction did not decrease over inter-trial intervals ranging from 350 ms to 1,400 ms. Also replicating previous findings, for reaction times the reduction was driven by selective slowing in task-order repeat trials, suggesting increased response caution. Our results are consistent with preparatory processes of task-order control or with episodic integration of task-order information but argue against accounts assuming short-lived, decaying task-order sets.