Academic journal article Psychonomic Bulletin & Review

A New Estimation of the Duration of Attentional Dwell Time

Academic journal article Psychonomic Bulletin & Review

A New Estimation of the Duration of Attentional Dwell Time

Article excerpt

How rapidly can attention move from one object to the next? Previous studies in which the dwell time paradigm was used have estimated attentional switch times of 200-500 msec, results incompatible with the search rate estimates of 25-50 msec shown in numerous visual search studies. It has been argued that dwell times are so long in the dwell time paradigm because the attentional shifts measured are unlike those used in visual search. In the present experiment, a variation of a visual search task was used, in which serial endogenous (volitional) deployments of attention were measured directly by means of a probe reaction time task. The experiment revealed a dwell time of about 250 msec, consistent with the faster estimates from other dwell time studies. This result suggests that endogenous shifts of attention may be relatively slow and that the faster attentional shifts estimated from visual search tasks may be due to the involvement of bottom-up processes.

In visual search, observers search through a variable set of nontarget items for a specific target item that may or may not be present. Reaction time (RT) is measured as a function of the number of items in the display (set size), and the rate at which attention is shifted from item to item is inferred from the slopes of RT × set size functions. Typically, search slopes are 5-10 msec/item for easy (parallel) search and 40-50 msec/item for difficult (serial) search (Wolfe, 1994). Thus, when attention is moved serially from one item to the next, it is generally estimated that attention switches every 50 msec (e.g., Treisman & Gelade, 1980). Moreover, there is good evidence that even the slowest of visual searches are quite fast. An exhaustive literature review in which 2,500 sessions involving 1,000,000 trials over a 10-year period were examined indicated that all slopes are less than 150 msec/item (Wolfe, 1998).

Although indirectly estimating the rate at which attention shifts from one item to the next on the basis of search slopes is common in the visual search literature, an alternative paradigm developed by Duncan, Ward, and Shapiro (1994; Moore, Egeth, Berglan, & Luck, 1996; Ward, Duncan, & Shapiro, 1997) provided more direct estimates of the attentional switch times of 200 msec (Moore et al., 1996) to 500 msec (Ward et al., 1997) per item. In this so-called dwell time paradigm, observers had to identify two objects presented in close temporal separation. The critical measurement in the dwell time paradigm is how long the first object continues to interfere with the second object. This is thought to represent the time course of the first object's attentional demand, because in order to identify the first object, attention must be engaged at the location of the first object. Performance (percentage correct) on the second target is assessed as a function of the stimulus onset asynchrony (SOA) between the first and the second targets, and this gives an estimate of the attentional dwell time. The results of Duncan et al. (1994) have shown that the interference of the first target with the second target lasts up to 450-500 msec, suggesting that attention moves from one object to the next as much as 10 times more slowly than typical estimates from visual search studies. Moore et al. (1996), using the same paradigm, have shown that this dwell time may be reduced to about 200 msec/item when the targets are not masked (as was the case in Duncan et al., 1994).

Clearly, there is a large discrepancy in the attentional dwell time estimates obtained from the two paradigms. On the basis of search slopes from traditional visual search experiments, it can be estimated that attention switches from one object to the next in about 50 msec/item (with an absolute maximum of 150 msec/item; see Wolfe, 1998), whereas in attentional dwell time studies, identification accuracy of the second target indicates an absolute minimum of 200 msec/item (see Moore et al. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.