Oculomotor behaviour in visual search for multiple targets

Abstract

Analysing response time (RT) data from a novel, multiple-target visual search task, Horowitz and Wolfe (2001) found evidence to suggest that the control of attention during visual search is not guided by memory for which of the items or locations within a display have already been inspected. Here, analysis of eye movement data from a similar experiment suggests that RT effects in the multiple-target search task are primarily due to changes in eye movements, and that effects which appeared to reveal memory-free search were actually produced by changes in oculomotor sampling behaviour.

Acknowledgments

We thank J. Krummenacher, T.S. Horowitz, and C. Olivers for helpful comments on an earlier draft. This work was supported by a National Institute of Health grant to MSP (R01 MH64505).

Notes

¹Nomenclature here differs slightly from that of Zelinsky and Sheinberg (1997), who discussed their data in terms of fixation durations rather than gaze durations. Analysis is focused on gazes rather than fixations in the current work, as gaze was deemed a more appropriate measure of a single oculomotor sample of given object. It seemed inappropriate, for example, to treat the fixation following a corrective saccade as a distinct attentional sample of the fixated object.

²See Kristjánsson (2000), Müller and von Mühlenen (2000), Shore and Klein (2000), and von Mühlenen, Müller, and Müller (2003) for methodological discussion of other evidence offered for the memory-free search hypothesis.

³It possible that the strength of this effect varies by the form of working memory load—visuospatial, verbal, or executive (Baddeley, 1986)—imposed by the search task (Han & Kim, 2004; Oh & Kim, 2004; Woodman & Luck, 2004).