Evidence for attentional processing in spatial localization

0 E. J. Davelaar School of Psychology , Birkbeck, University of London , Malet Street, WC1E 7HX London, UK 1 J. J. Adam (&) A. van der Gouw P. Willems Department of Movement Sciences, University of Maastricht , P.O. Box 616, 6200 MD Maastricht, The Netherlands Using a dual-task methodology, this study examined the involvement of selective attention in spatial localization. Thirty participants located a single, briefly presented, peripheral target stimulus, appearing in one of 50 positions on either side of a central fixation point, with or without the requirement to identify a simultaneously presented central distractor stimulus. Results revealed a robust interference effect in localization performance at short target durations that depended on the number of the to-be-identified distractor items. This outcome provides convergent support for the role of the attentional system in spatial localization. - When viewing an object in a visual scene, an observer may attend to different properties of this object, such as its identity or location. A number of cognitive theories of visual attention assume that properties, such as identity and location, are processed in different feature maps that are located in different brain areas (see for a review, Shipp, 2004). In addition, the information coming from the different feature maps is subsequently combined into a signal that can be used to trigger an eye-movement or saccade towards a certain location. This is particularly important if the observers task is to locate a target stimulus that is presented in the visual periphery, so that by making a saccade the fovea will be aligned with the stimulus and detailed location information can be extracted. As the initiation of a saccade takes place at least 100 ms after the onset of a target stimulus (e.g., Carpenter, 2004; Van Loon & Adam, 2006), it is noteworthy that localization performance increases during the period when no saccade is being initiated (Adam, Ketelaars, Kingma, & Hoek, 1993). This finding suggests that pre-saccadic processing is sufficient for coarse localization. In this paper, we have observers extract multiple types of information from a visual scene. Specifically, we focus on the impact of identifying a number of digits at fixation on the localization performance of a peripheral target. We will first outline the two-process model of object localization proposed and investigated by Adam, Huys, van Loon, Kingma, & Paas (2000), Adam et al. (1993); Adam, Paas, Ekering, & van Loon (1995) (see also Uddin, Ninose, & Nakamizo, 2004). We then present an experiment, using a dual-task method, that supports the assertion that visual attention is critical in localizing objects within the first 100 ms. Adam et al. (1993) investigated the time course of visual object localization using a task in which participants had to locate a single target stimulus presented in one square of an imaginary 25 19 grid that contained 474 possible stimulus locations. They varied the presentation duration of the (masked) stimulus between 33 and 300 ms. Participants used the cursor to indicate the perceived target location. Results showed an initial steep rise in localization accuracy during the first 50 ms of stimulus duration, followed by a further but more gradual improvement from 100 ms onwards, finishing with near-perfect performance at about 300 ms. Adam et al. (1993) interpreted these findings within a two-process model of visual object localization. In this model, a fast attentional process provides coarse localization information and precedes a slower saccadic system that provides more detailed information by aligning the fovea with the target. In support of the role of the saccadic system, Adam et al. (1993) showed that the further improvement in localization after 100 ms is absent when participants are instructed to abstain from making saccades. In addition, when saccades are allowed, eye movement analyses indicated that participants nearly always made a saccade (i.e., in 98.4% of all trials), but the saccadic onset latency was never less than 100 ms, indicating that the initial steep rise in localization performance during the first 50 ms of stimulus duration can not be attributed to the saccadic system. Together these results suggest strongly that the execution of saccades underlies the gradual improvement in localization performance after 100 ms. In support of the view that the attentional system underlies the improvement in performance for the first 50 ms, Adam et al. (1993) cited the results of spatial cuing studies, showing that the largest gains in precuing typically occur within the first 50 ms; this provides an estimate of the time necessary to shift attention (Eriksen, 1990). Similarly, visual search experiments have demonstrated scanning rates, i.e., shifts of attention, in the order of 50 ms/item (e.g., Bergen & Julesz, 1983; Treisman & Gelade, 1980; but see e.g. Ward & Duncan, 1996, for much longer estimates). In addition, Adam et al. (2000) showed that advance knowledge about the possible location(s) of the target improves localization performance. In particular, they showed that localization performance improved with short duration (i.e., 71 ms) spatial precues, which accords with the notion that the spatial precue quickly directs spatial attention to the target area and thus mediates localization performance. Furthermore, localization performance for stimulus durations of less than 100 ms is greatly improved when the target stimulus is not backward masked (Adam et al., 1995). Assuming that the masking stimulus disrupts localization performance by involuntarily capturing attention (e.g., Yantis & Jonides, 1984), this finding too suggests that attention is involved in localizing stimuli. So far, the role of visual attention in object localization is supported by experimental manipulations of events before (Adam et al., 2000) and after (Adam et al., 1995) the target stimulus. In this study, we sought to provide additional, converging evidence for the role of the attentional system in localization performance by examining the effect of a central to-be-identified distractor stimulus on localization performance of a simultaneously presented peripheral target stimulus. Thus, participants were facing a dual-task situation. Generally, in dual-task situations, interference occurs when both tasks need the same mechanism (e.g., Pashler & Johnston, 1998). Furthermore, it is well established that visual identification requires the operation of selective visual attention (e.g., Heinke & Humphreys, 2003; Kawahara, Di Lollo, & Enns, 2001). Hence, if localization needs attention too, then it should be vulnerable to the requirement to first identify the central distractor. If, on the other hand, localization is attentionindependent, then it should not be sensitive to the requirement to first identify the central distractor. We hypothesized that if localization depends on the alloc (...truncated)