Steering through task uncertainty: Evidence for activation modulation over dynamic binding

Living in a fast-changing environment requires a dynamic updating of task representations. Previous research has shown that people can adaptively prepare and unprepare for a given task, when the likelihood of a task switch varied in time during preparation. However, the underlying mechanism of this process remains unclear. We proposed two potential mechanisms that could support this flexible task control process: people may either downregulate the current task representation (i.e., the activation modulation account) or try to unbind the task components (i.e., the dynamic (un)binding hypothesis) when expecting a task switch. To distinguish between these mechanisms, we devised a novel paradigm in which, during task preparation, the to-be-performed task could suddenly switch to a new one. Critically, this switch could be either partial (changing only one of the two task components) or complete. We predicted that, if activation modulation were the dominant mechanism, participants would show stronger inhibition when expecting a complete switch compared with a partial one. Conversely, if people relied mostly on dynamic (un)binding, participants would only need to unbind when expecting a partial switch. Our results showed that participants inhibited task representations more with increasing probability of a complete, as opposed to a partial, switch. This finding supports an activation modulation account, suggesting that people primarily regulate the activation strength of task representations rather than the binding strength between their task components.