Mor Sasi¹, Daniel Toledano¹, Shlomit Yuval-Greenberg¹, Dominique Lamy^1,2; ¹Tel-Aviv University, ²Sagol School of Neuroscience, Tel Aviv University

Most visual search theories assume that attention is automatically allocated to the location with the highest priority at any given moment. The Priority Accumulation Framework (PAF) challenges this assumption by suggesting that priority weights accumulate across time until action-relevant information signals the appropriate moment for deploying attention to the location with the highest accumulated priority. This account suggests a distinction between attentional capture - by which attention is deployed at the wrong moment, and attentional-priority bias - by which attention is deployed at the right time but biased towards the locations of previous irrelevant events. In previous work, we recorded eye movements in a free-viewing spatial-cueing task. In line with PAF’s predictions, we found that only a minority of first saccades occurred early, in response to the onset cue (attentional capture), whereas most first saccades occurred later, after processing of the search display had already started, but were nevertheless biased towards the cue location. This bias was stronger the more difficult the search (attentional-priority bias). Here, we replicate these findings when the cue-target interval is longer, thereby precluding the possibility that the impact of the cue on late saccades resulted from delayed capture of overt attention. In addition, we show a dissociation between early and late saccades, which we take to uncover a qualitative difference between exogenously and endogenously triggered saccades. Specifically, we show that unlike for endogenously triggered saccades, the identity of the object (target or distractor) on which an exogenously triggered saccade lands, does not determine how long this object remains fixated or whether or not an additional saccade is initiated. These findings, which were replicated in three experiments, suggest that processing is shallower at the locus of a stimulus-driven saccade than at the locus of endogenously triggered saccades.

Acknowledgements: Support was provided by the Israel Science Foundation (ISF)