Abstract
The Eriksen flanker effect refers to the slowing of responses to targets with known location when flanked by items associated with a different response. Typically, this effect is explained through response competition resulting from the attentional integration of perceptual features of incongruent flankers. Since the delay is more pronounced when target and flankers are closer together, the attentional integration is assumed to operate like a zoom lens. Here, we present a novel flanker paradigm where the stimuli consist of random dot kinematograms (RDKs). Using RDKs, the strength of the perceptual evidence can be manipulated by modifying the percentage of coherently moving dots. While we were able to replicate the classic flanker effect with low coherent RDKs, the effect decreased with increasing coherence, suggesting that the flanker effect is not only due to competition at the response stage but also involves a perceptual stage. Surprisingly, and contrary to the zoom lens model, we found no flanker effect in a second experiment where target-flanker spacing was reduced. We interpret these results as evidence for the involvement of perceptual grouping in the attentional integration of flankers. We assume that responses are facilitated when flankers can be grouped and perceptually segregated from the target. Following the Gestalt principles, this segregation is most effective when target and flankers are dissimilar. In our experiments, lower coherence increases target-flanker similarity, leading to less efficient segregation and in turn a stronger flanker effect. Furthermore, the beneficial effect of reduced target-flanker spacing is explained through the Gestalt law of grouping by proximity, where close flankers are more strongly grouped leading to a reduction of the flanker effect. Together, these results suggest attention can be guided by perceptual groups capable of moderating the detrimental influence of incongruent flankers on response execution.