Category effects in temporal visual search.

Abstract

Abstract When subjects identify a target during rapid serial visual presentation (RSVP), they show a reduced ability to detect a subsequent probe stimulus relative to when they ignore the target. The present study demonstrated an effect of target/probe categorical relation upon this probe detection deficit (attentional blink). Experiment 1 used letters for both target and probe, replicating the general methods and results of Raymond, Shapiro, and Arnell (1992). Experiment 2 varied target/probe categorical relation via instructional set: The target was a letter; for some subjects, the probe stimulus o was referred to as the letter oh, whereas for other subjects it was referred to as the number zero. Treating o as a number attenuated the probe detection deficit. This different-category attenuation was confirmed in Experiment 3 where the target was a letter and the probe was a number. The observed category effect suggests that the probe detection deficit in RSVP may map a time course for spreading intra-category inhibition following temporal target selection. In rapid serial visual presentation (RSVP), pictures, words, letters, or other visual stimuli are presented in the same spatial location at a rate of 6-20 items per second (e.g., Broadbent & Broadbent, 1987; Intraub, 1985; Kanwisher, 1987; Lawrence, 1971; Reeve & Sperling, 1986; Weichselgartner & Sperling, 1987). By asking subjects to monitor the RSVP stream for target items, it is possible to examine the temporal limits governing visual and attentional processing independently of saccadic eye movements. In the context of a dual-task RSVP paradigm, Raymond, Shapiro and Arnell (1992, see also Shapiro, Raymond, & Arnell, 1994) presented subjects with a white target letter in a stream of black letters. This target could be followed at various intervals by a black probe x. In an experimental condition, subjects identified the target and detected the presence/absence of the probe; in a control condition, subjects ignored the target and detected only the presence/absence of the probe. Relative to the control condition, when subjects directed attention to the target, they showed a decrement in the detection of probes that occurred 100-450 ms post-target (see also, Broadbent & Broadbent, 1987; Reeves & Sperling, 1986; Weichselgartner & Sperling, 1987). Because this probe deficit did not occur when subjects ignored the target (control condition), Raymond et al. (1992) argued that the deficit is attentionally-based rather than perceptually-based (hence, their use of the term attentional blink). Two boundary conditions determine whether the probe detection deficit is observed. First, the deficit occurs only when the target is followed immediately by a patterned stimulus: When the first post-target item is replaced by a temporal gap, no deficit occurs; however, the elimination of the deficit is not simply a result of reduced perceptual load because the effect is not attenuated by the imposition of a temporal gap at any other post-target interval (Raymond et al., 1992). Second, the deficit is observed whether the target task requires detection or discrimination and regardless of the difficulty of target selection (Shapiro et al., 1994; for a review of the literature and results see Shapiro & Raymond, 1994), yet it does not occur if the target is defined by the absence of a stimulus (i.e., a temporal gap; Shapiro et al., 1994). On this basis, Shapiro et al. (1994) suggested that a probe detection deficit will occur any time attention is directed to a patterned target that is followed immediately by a patterned post-target stimulus. Ward, Duncan, and Shapiro (1992) marshalled support for this view with the demonstration that a probe detection deficit occurs even in the absence of an RSVP stream: Presentation of only a masked target and masked probe will result in a probe detection deficit when attention is directed to the target. On the basis of these observations, Shapiro et al. …