Purchase this article with an account.
Jiaying Zhao, Nicholas B. Turk-Browne; The timecourse of the attentional bias to regularities. Journal of Vision 2014;14(10):10. doi: https://doi.org/10.1167/14.10.10.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Knowledge about the structure of an environment can help us perceive and act more efficiently. But how do we find such regularities in the first place when they are embedded in noisy and complex visual input? We recently showed that attention is naturally biased to regularities, prioritizing locations and features where they can be found (Zhao, Al-Aidroos, & Turk-Browne, 2013). A normative explanation of this effect is that it speeds the extraction of these regularities, such that attention can be released to support processing of information about which we have greater uncertainty. Accordingly, we hypothesized that the bias for regularities should dissipate over time. Experiment 1 examined the timecourse of spatial attention to regularities. Observers were presented with multiple streams of shapes that were interrupted occasionally by visual search arrays. One of the streams was generated from shape triplet regularities. During the first of three epochs of exposure, we found as before that attention prioritized the structured location, as indexed by facilitated search performance at that location. However, by the final epoch of exposure, this benefit had disappeared. Experiment 2 generalized this effect to feature-based attention. Observers were now presented with a single stream of shapes of different colors that was interrupted by visual search arrays with a color singleton. The shapes in one color again appeared in triplet regularities. We found as before that attention prioritized the structured color, as measured by attentional capture for singletons of that color. This time, the bias was most evident in the second of three epochs, but critically, this benefit also disappeared by the final epoch. In sum, the attentional bias for regularities is transient, but of sufficient duration to allow for rapid statistical learning. Such attentional disengagement may allow for the exploration and acquisition of more complex or new forms of structure.
Meeting abstract presented at VSS 2014
This PDF is available to Subscribers Only