When predictions take control: the effect of task predictions on task switching performance.

In this paper, we aimed to investigate the role of self-generated predictions in the flexible control of behavior. Therefore, we ran a task switching experiment in which participants were asked to try to predict the upcoming task in three conditions varying in switch rate (30, 50, and 70%). Irrespective of their predictions, the color of the target indicated which task participants had to perform. In line with previous studies (Mayr, 2006; Monsell and Mizon, 2006), the switch cost was attenuated as the switch rate increased. Importantly, a clear task repetition bias was found in all conditions, yet the task repetition prediction rate dropped from 78 over 66 to 49% with increasing switch probability in the three conditions. Irrespective of condition, the switch cost was strongly reduced in expectation of a task alternation compared to the cost of an unexpected task alternation following repetition predictions. Hence, our data suggest that the reduction in the switch cost with increasing switch probability is caused by a diminished expectancy for the task to repeat. Taken together, this paper highlights the importance of predictions in the flexible control of behavior, and suggests a crucial role for task repetition expectancy in the context-sensitive adjusting of task switching performance.