Search for multiple targets: Evidence for memory-based control of attention

Abstract

There are two opposing models with regard to the function of memory in visual search: a memorydriven model and a memory-free model. Recently, Horowitz and Wolfe (2001) investigated a multipletarget search task. Participants were required to decide whether or not there were at least n targets present. They demonstrated that the reaction time 3 n function has a positive and accelerated curve. They argued that the memory-free model predicts this curve, whereas the memory-driven model predicts a linear function. In this study, I varied the total set sizes of a multiple-target search task and fitted the models separately for each n condition. The model fit indicated that the memory-driven model is more appropriate than the memory-free model in each n condition. These results suggest that an amnesic process does not cause the positive accelerated curve of the reaction time 3 n function but that it is the result of the time needed to examine each additional n item.

Oculomotor behaviour in visual search for multiple targets

“…Results indicate that curvilinear RT and gaze frequency functions were not produced by amnesic attentional guidance, but by changes in oculomotor behaviour as a function of criterion number. 2 Interestingly, the present data accord nicely with the findings of Takeda (2004), who incorporated a manipulation of total set size (number of targets plus distractors) within a multiple-target search task like that employed here. Data showed a slowing of behavioural search rates as the number of targets being sought increased.…”

“…These considerations suggest that gaze frequencies might increase as a positively accelerated function of criterion number because of a tendency for observers to shift from processing many items per gaze over a wide perceptual span to processing fewer items per gaze over a narrow span. Such a change would be consistent with Takeda's (2004) conclusion that variations in search rate are responsible for deviations from linearity in RT data for multiple-target search; a contracting perceptual span would provide a mechanism to explain the decline in search rate that occurs as criterion number increases. The suggestion that observers were in general searching with a perceptual span of larger than 1 is confirmed by the finding that gaze frequencies were consistently lower than the number of attentional samples predicted by a model in which a single item is processed per sample (compare the gaze frequency data of Figure 2 to the predictions plotted in Figure 1).…”

No need for inhibitory tagging of locations in visual search

“…The number of moving items used (up to 18) falls below the estimate of 20 items reported by Takeda (2004) as the memory capacity of visual search. (Other estimates are lower: According to Snyder & Kingstone, 2000, the limit is 5-6; according to Horowitz & Wolfe, 2001, 3-5.…”

Does the inspector have a memory?

“…To the extent that IOR might tag inspected items in a difficult search task, it would be unrealistic to think that the tags left behind would enable the inspector to sample entirely without replacement *as if ''memory'' for inspected items were perfect. The main reason is that the number of inhibitory tags available for this purpose seems to be between three and six (e.g., Snyder & Kingstone, 2000; but see Takeda, 2004, for a much higher estimate). In this regard, and in agreement with Horowitz (2006 this issue), it makes more sense to ask ''How much memory does the search inspector have?''…”