The role of selective attention in matching observed and executed actions

Substantial evidence suggests that observed actions can engage their corresponding motor representations within the observer. It is currently believed that this process of observation–execution matching occurs relatively automatically, without the need for top-down control. In this study we tested the susceptibility of the observation–execution matching process to selective attention. We used a Go/NoGo paradigm to investigate the phenomenon of ‘automatic imitation’, in which participants are faster to initiate a hand movement that is congruent with a concurrently observed action, relative to one that is incongruent. First, we replicated previous findings of automatic imitation, and excluded the possibility that spatial compatibility effects might explain these results (Experiment 1). We then presented participants with the same goal-directed actions while directing their attention to an imperative stimulus that spatially overlapped, but was distinct from, the observed actions (Experiment 2). Crucially, automatic imitation no longer occurred when participants directed their attention away from the displayed actions and towards the spatially overlapping stimulus. In a final experiment, we examined whether the automatic imitation of grasp persists when participants attend to an irrelevant feature of the observed action, such as whether it is performed by a left or right hand (Experiment 3). Here we found that automatic imitation is contingent on participants attending to the feature of the observed hand that was relevant to their responses. Together these findings demonstrate the importance of selective mechanisms in the filtering of task-irrelevant actions, and indicate a role for top-down control in limiting the motoric simulation of observed actions.