Transient stimulation does not aid visual search: Implications for the role of saccades

Abstract

Previous demonstrations that performance on visual search tasks with briefly flashed presentations declines over time after the initial onset imply that saccades might be necessary for efficient acquisition of visual information. We imposed an additional abrupt onset and a displacement on character arrays containing letters and one numeral while subjects searched for the numeral. Presentations were always followed by a visual noise field. Presentation time varied (50-800 msec). We found that performance with presentations containing one onset and remaining continuously visible was better than performance with presentations containing two onsets and containing displacements. Furthermore, information acquired near the onset of a continuous presentation was as effective as information acquired later. Our results demonstrate that abrupt onsets and displacements do not improve information acquisition with displays of alphanumeric characters. This finding is consistent with earlier reports that saccade-like retinal image motion does not contribute either to the maintenance of target visibility or to visual acuity.Saccadic eye movements are obviously necessary for efficient visual information processing because they bring retinal images to the fovea where visual acuity is best. But this may not be the only useful role for saccades. Saccades can also aid visual information processing when visual details are already foveal and clearly distinct.For example, Kowler and Steinman (1977) found that saccades were helpful for counting items that were haphazardly arranged in a 2-deg-diam display.In their experiment, subjects were required to count the items while either using saccades to scan around the display or not using saccades and maintaining the line of sight exclusively with slow eye movements at the center of the display. Counting accuracy was found to be 50010 higher when saccades were used. But the reason that saccades were so helpful for counting was not that directing the line of sight at, or very near, items was necessary in order to see them distinctly. On the contrary, subjects noted that all the items could be easily distinguished when the This research was supported by NIH Postdoctoral Fellowship SF32EY05235 to E. Kowler. We thank B. Fischer for serving as a subject, P. Burt, J. C. Falmagne, G. Iverson, M. Pavel, and R. M. Steinman for their useful suggestions, and 1. Nicholson for typing the manuscript. Requests for reprints should be addressed to E. Kowler,

A theory of visual information acquisition and visual memory with special application to intensity-duration trade-offs.

“…8 Notably, Bundesen (1990), Kowler and Sperling (1980), Loftus and Kallman (1979), Massaro (1970), Rumelhart (1969), and Townsend (1981). 9 In previous formulations of the model, this a(t) function was defined somewhat arbitrarily-that is, on the basis of intuitionrather than generated on the basis of fundamental principles.…”

Providing a sensory basis for models of visual information acquisition

“…Unlike, say, a simple threshold-detection task, digit recall is complex enough to require the fundamental processes enumerated above. At the same time, the task involves a stimulus that should not exceed short-term memory capacity and thus does not require assumptions about subjects' strategies about what to transfer to short-term memory (see, e.g., Rundus, 1971); involves a stimulus that is immediately masked, so that it does not require assumptions about iconic decay (see, e.g., Averbach & Sperling, 1961); involves items of relatively homogeneous difficulty, so that it does not require corrections for averaging artifacts (see, e.g., Kowler & Sperling, 1980); and does not involve partial report, so that it does not require assumptions about probe detection and utilization (see, e.g., Sperling, 1960).…”

“…Correspondence should be addressed to G. R. Loftus, Department of Psychology, University of Washington, Seattle, WA 98195 (e-mail: gloftus@u.washington.edu). Kowler & Sperling, 1980;Loftus & Kallman, 1979;Massaro, 1970;Rumelhart, 1970;Shibuya & Bundesen 1988;Townsend, 1981). By what we will refer to as the basic random-sampling model, a stimulus is composed of a set of relatively homogeneous features.…”

Abrupt onsets do not aid visual search

“…To examine the effect of abrupt onsets on visual information acquisition, we compared performance in the visual search task we had used earlier (Kowler & Sperling, 1980) for displays with gradual and with…”

“…In a previous paper (Kowler & Sperling, 1980), we found that visual search for a single-numeral target, embedded in an array of letters, was not aided by abrupt onsets. This conclusion was supported by three main results: (1) Visual search with displays containing two abrupt onsets (two brief flashes) was not better than visual search with displays containing a single abrupt onset (one long flash); (2) an abrupt onset created by imposing a rapid, lateral displacement on the stimulus was not useful-visual search with stationary displays was as good as visual search with displays containing a rapid displacement; and (3) visual search was consistent with a model in which information was acquired at a constant rate throughout a stimulus presentation.…”