Abstract

In visual search, the presence of a highly salient color singleton can slow or facilitate search for a shape target depending on whether the singleton is a distractor or coincides with the target. This is consistent with an attentional shift (attentional capture) to the salient item. This attentional capture can be driven by bottom–up or top–down processes or both. We investigated the role of the parietal cortex in attentional capture by a singleton using repetitive transcranial magnetic stimulation. Following disruption to the right posterior parietal cortex by sustained transcranial magnetic stimulation, the reaction time (RT) cost of the singleton distractor was reduced. At least part of this lessening of singleton distraction was due to the elimination of priming (top–down) effects between target and distractor singletons on consecutive trials. In Experiment 2, we presented the different conditions in separate blocks meaning any effects of the distractor can most likely be attributed to bottom–up processes. Nevertheless, there was still a decrease in RT interference from the distractor so that a reduction in priming cannot provide a full account of the results. The data are consistent with previous work positing that the right parietal cortex directs attention to salient stimuli (e.g., Constantinidis 2005, Mevorach et al. 2006), while also suggesting a role for the right parietal cortex in the integration of bottom–up salience information with memories for salient features on prior trials.

Introduction

In a complex and dynamic visual world, it is useful to have mechanisms that rapidly draw attention to salient events. For example, such a mechanism may facilitate a rapid avoidance response to a threatening stimulus. The cost of this is that if the stimulus turns out not to be relevant for our current needs, an ongoing task may be disrupted. Psychological evidence that attention may be captured by salient events comes from studies of the effects of salient “singletons” on visual search. Theeuwes (1991), for example, showed that the presence of a highly salient color distractor (e.g., the only red item in an array of green stimuli) interfered with search for a shape target (see Fig. 1 for example displays). He argued that attention is deployed to items in terms of their salience, irrespective of whether these items have features relevant to the task at hand. Subsequently, other researchers have proposed that singleton capture is dependent on the adoption of an attentional set (Folk & Remington, 1992) or singleton detection mode (e.g., Bacon & Egeth, 1994, Yantis & Egeth, 1999, although see Theeuwes and Burger, 1998, Theeuwes, 2004 for counter arguments). Whichever is the case, the data indicate that attention may be captured by a salient event under the appropriate conditions, even when the event is irrelevant to the task at hand.

Figure 1.

A search display showing the target as the shape circle singleton among diamond distractors with the addition of a color singleton distractor.

Figure 1.

A search display showing the target as the shape circle singleton among diamond distractors with the addition of a color singleton distractor.

The neural substrates of singleton capture in search have been investigated recently by de Fockert et al. (2004). Neural activity was measured via functional magnetic resonance imaging (fMRI) as participants viewed similar search displays to Theeuwes (1991). Participants searched for a circle among diamonds, and on 25% of the trials the target was a color singleton, whereas on another 25% of the trials one of the distractors was a color singleton. Although there was no measured neural activity specifically related to the color singleton target, the presence of a color singleton distractor led to bilateral activation within the parietal cortex and left frontal cortex relative to when no color singleton was present. Parietal activity was construed to reflect shifts of attention to the distractor item (e.g., Corbetta and Shulman 2002) and frontal activity to reflect the resolution of subsequent competition between the salient distractor and the target. Consistent with shifts of attention to the distractor item, Hickey, McDonald and Theeuwes (2006) showed that the presence of a salient distractor modulates the N2pc ERP waveform, which is held to reflect the spatial orienting of attention (Luck & Hillyard, 1994). These data suggest that attention is apparently spatially shifted to a salient distractor before being shifted toward a less salient target.

Although the data of de Fockert et al. (2004) indicated that there was bilateral parietal activity related to the appearance of a singleton distractor, other authors have pointed to a difference in the roles of left and right posterior parietal cortices in responding to salient events. Using transcranial magnetic stimulation (TMS) to disrupt activity in these cortical regions, Mevorach Shalev and Humphreys (2006) argued that the right posterior parietal cortex (PPC) may act to draw attention to a salient event whereas the left PPC may act to direct attention toward targets that have relatively low salience. Mevorach et al. based their argument on a repetitive TMS (rTMS) experiment, in which responses to high or low salient dimensions of hierarchical stimuli were examined after rTMS to the left or right PPC. rTMS to right PPC impaired responses to high saliency dimensions, whereas left PPC disrupted responses to low saliency dimensions. More generally, the majority of previous TMS studies (utilizing a wide variety of spatial attention tasks) have shown disruption of spatial attention processes following TMS to the right PPC but not the left (cf. Rushworth and Taylor 2006 for a review). In visual search, for example, Ellison et al. (2003) showed that TMS impaired conjunction search following right parietal TMS. Also consistent with this idea that the left and right parietal lobes subserve different critical functions is a substantial body of data based on unilateral neglect (e.g., Heilman and Valenstein 1979). This syndrome, associated with poor attentional orienting to the affected side, is more commonly found following right parietal as opposed to left lesions. This supports a neuroanatomical model in which a right lateralized frontoparietal network (and particularly the ventral regions of this network) is thought to function as a “circuit breaker,” directing a more dorsal bilateral network to the presence of a salient event (Corbetta and Shulman 2002).

To assess the role of the right parietal cortex, in particular, in responding to a salient visual event, we conducted an interventionist study using rTMS. rTMS was applied to the left or right PPC, and we examined whether this stimulation modulated the effects of attentional capture from a salient singleton distractor. In Experiment 1, a salient singleton stimulus could be either a target or a distractor. Under these conditions, participants may strategically attend to the singleton. In addition, there can be a carryover effect from a singleton target on one trial to a singleton distractor on a following trial (Olivers and Humphreys 2003). Does rTMS to right PPC disrupt selection of the singleton under these conditions, and is this (at least in part) due to its effect on the carryover effect across trials? In Experiment 2, we used conditions in which the singleton, when it occurred, was always a distractor. Did rTMS continue to disrupt performance, as would be the case if the right PPC was involved in the automatic capture of attention by salient stimuli.

Experiment 1. The Effect of Parietal rTMS on Responses to Salient Singletons

The design of Experiment 1 matched that of de Fockert et al. (2004) so that the singleton could sometimes be the target and sometimes a distractor. Participants were required to carry out visual search for a shape singleton, a circle among 5 diamond distractors, and on 50% of the trials, a high saliency colour singleton was present in the display (see Fig. 1). On singleton present trials, the color singleton corresponded with the target on half the trials and the distractor on the other half. Effects of the color singleton were measured relative to trials when this item was absent. Differing functional roles for the PPC were assessed by comparing search reaction times (RTs) prior to and following either right or left parietal TMS. The conditions where the singleton was absent and when it was a target or a distractor were intermixed in random order through the experiment. This followed the design of de Fockert et al. (2004) and one consequence of this design is that attentional capture is potentially governed by both top–down and bottom–up processes. One top–down influence may be that on a given trial any particular item is more likely to be a singleton target than a distractor maybe giving participants an incentive to attend to the singleton. This could lead to stronger effects of the singleton distractor. However, the experimental design also meant that the target and the singleton distractor share features across trials (e.g., the target could be a singleton on trial N following by a singleton distractor on trial N + 1, or vice versa). This in turn meant that we could extend the findings of de Fockert et al. (2004) and examine the effects of parietal TMS on across-trial priming for the singleton (Pinto et al. 2005). There is evidence that the effects of singleton distractor are enhanced when on the immediately prior trial, the singleton value has been carried by a target. This indicates that singleton effects are sensitive to the past history of events. We evaluate whether this top–down sensitivity to past history is modulated by the PPC in addition to singleton distractor interference.

Method

Participants

Four females and four males (age range 21–38 y) with no previous history of neurological problems participated in the experiment. All participants had normal or corrected to normal vision and were right handed. All but one had participated in TMS experiments previously.

Apparatus

The experiment was run on a Windows PC, one with a 1GMHZ Pentium III processor, using a Philips 109S monitor. Presentation software (release 9.70) from Neurobehavioral Systems (www.neurobehaviouralsystems.com) was used to display the stimuli and record RTs.

Stimuli and Design

Our visual search displays consisted of 6 shapes arranged on a circle of radius 6.65° from a centrally placed fixation cross (Fig. 1). The target item was always a circle of diameter 2.5°, and the distractor items were a diamond, 2.2° square meaning that the target and nontarget items were approximately equal in terms of area. In the center of each display item was a line of length 1°. These lines were randomly oriented to the horizontal or vertical in each item with the constraint that overall there would be 3 vertical and 3 horizontal lines. All shapes and line elements were gray (Red Green Blue [RGB], 160, 160, 160) apart from the color singleton. RGB values for the color singleton were 0, 180, 0. On half the trials, there was no color singleton present. Singleton targets and singleton distractors were present on 25% of trials respectively. The effect of color singletons was measured by subtracting either the singleton distractor or singleton target conditions from the conditions where the color singleton was absent.

Procedure

A trial consisted of a fixation cross for 500 ms, followed by the search displays that were present until a response was made. Participants responded as to whether the orientation of a line within the target circle was horizontal or vertical using the “n” and “j” keys, respectively. A practice session was run to familiarize participants with the task. Participants were instructed to restrict eye movements throughout the task. Each session consisted of 192 trials.

rTMS Procedure

A Magstim Rapid stimulator with a 70-mm figure of 8 coil was used to deliver TMS to the right and left PPC. The site of stimulation was the same as that of Mevorach et al. (2006), corresponding to points P3 (left parietal) and P4 (right parietal) on the 10–20 electroencephalography coordinate system. Stimulation intensity was set at 10% below the resting motor threshold of each participant. Participants received a 10-min 1-Hz train of pulses which previous studies have shown gives extended cortical inhibition at the site of stimulation (e.g., see Hilgetag, Theoret & Pascual-Leone, 2001 and Mevorach et al., 2005). Initially, the task was performed with no stimulation (baseline). Then participants performed the task 2 more times immediately following stimulation of either the left or right PPC (on 2 different days). The order of stimulation sites was counterbalanced between participants. Finally, a post TMS session ensued, the no TMS results being an average of the initial (baseline) and this last run of the task.

Results

Effects of TMS on Search

RTs were treated for outliers according to a procedure devised by Van Selst and Jolicouer (1994), removing 2.72% of RTs. Mean RTs were calculated as a function of TMS (none, right parietal, left parietal) and singleton condition (singleton absent, singleton distractor present, singleton target present). Table 1 shows the overall mean RTs for TMS and singleton distractor, target, and absent conditions, and Figure 2 shows the RT differences on trials where the singleton was respectively either a target or a distractor, compared with when the singleton was absent. We assessed the effects of TMS on singleton targets and distractors separately.

Table 1

Experiment 1, Mean correct RTs with ± 1 standard error for singleton distractor present, singleton target present, and singleton absent with no TMS, left TMS (P3), and right TMS (P4)

 No TMS Left TMS Right TMS 
Distractor 781.9 (45.6) 741.7 (43.6) 690.0 (42.7) 
Target 613.8 (21.3) 609.6 (14.3) 581.5 (21.1) 
No singleton 620.4 (25.6) 591.0 (18.0) 593.3 (29.3) 
 No TMS Left TMS Right TMS 
Distractor 781.9 (45.6) 741.7 (43.6) 690.0 (42.7) 
Target 613.8 (21.3) 609.6 (14.3) 581.5 (21.1) 
No singleton 620.4 (25.6) 591.0 (18.0) 593.3 (29.3) 
Figure 2.

Difference RTs in Experiment 1 for trials where there was a singleton target or a singleton distractor relative to trials where it was absent, for no TMS, left TMS (P3), and right TMS (P4) sessions.

Figure 2.

Difference RTs in Experiment 1 for trials where there was a singleton target or a singleton distractor relative to trials where it was absent, for no TMS, left TMS (P3), and right TMS (P4) sessions.

Firstly, for singleton distractors, we carried out a 2-way analysis of variance (ANOVA) with TMS (none, left, right) and singleton distractor (present or absent) as conditions. Overall, there was a main effect of singleton distractor presence, F1,7 = 19.486, P < 0.005. RTs were 137 ms slower when the singleton distractor was present relative to when it was absent. There was a marginally significant effect of TMS, F2,14 = 3.695, P = 0.051. Generally, participants were faster with TMS applied (701 ms without, 665 ms left, and 641 ms right). More importantly, there was an interaction between TMS and singleton distractor, F2,14 = 4.241, P < 0.05. Whereas the RT cost of a singleton distractor was approximately the same for no TMS versus left TMS (162 vs 152 ms), for right TMS the RT cost of the singleton distractor was reduced to 96 ms. We confirmed this finding using paired sample t-tests on difference RTs computed by subtracting the distractor absent scores from distractor present; t7 = 0.441, P > 0.67 for no TMS versus left TMS and t7 = 2.968, P < 0.025 for no TMS versus right TMS. Note that inspection of singleton absent trials in Table 1 indicates a small nonsignificant effect of approximately 30 ms for the non-TMS sessions versus the left and right TMS (F2,14 = 1.369, P > 0.25). This may be due to practice or nonspecific effects of TMS (see Walsh and Pascual-Leone 2005). However, the particular effect of right TMS on singleton distractor RTs is unlikely due to practice or nonspecific TMS effects, as the magnitude of the RT reduction in singleton distraction is similar between left and right parietal TMS and no TMS and right parietal TMS. If our effect was solely due to nonspecific effects of TMS or practice, we would also expect a reduction in RTs for the distractor condition following left parietal TMS, which was not the case. Indeed, the use of left parietal as a control site is arguably more important than non-TMS comparisons, due to possible aforementioned nonspecific effects of TMS (see Walsh and Pascual-Leone 2005).

We also performed a 2-way ANOVA with singleton targets (present or absent) versus TMS (none, left, or right). Here we found no significant differences between the conditions, with or without TMS. This is consistent with de Fockert et al. (2004) who also found no effects of the singleton target on either behavioral or imaging data. Similar analyses of errors also showed no effects, errors averaging approximately less than 10% (see Table 2).

Table 2

Experiment 1, percentage errors ± 1 standard error for singleton distractor present, singleton target present, and singleton absent with no TMS, left TMS (P3), and right TMS (P4)

 No TMS Left TMS Right TMS 
Distractor 10.2 (1.4) 11.4 (2.1) 10.4 (1.8) 
Target 9.8 (1.1) 9.6 (2.1) 7.5 (1.7) 
No singleton 9 (1.0) 7.7 (1.0) 10.2 (1.6) 
 No TMS Left TMS Right TMS 
Distractor 10.2 (1.4) 11.4 (2.1) 10.4 (1.8) 
Target 9.8 (1.1) 9.6 (2.1) 7.5 (1.7) 
No singleton 9 (1.0) 7.7 (1.0) 10.2 (1.6) 

Finally in this section, we analyzed RTs to singleton distractors as a function of whether they appeared in the left or right visual field, that is, ipsilateral or contralateral to applied TMS. Figure 3 shows the mean correct RTs as a function of TMS and visual hemifield. An ANOVA with TMS and visual field as factors showed no TMS by visual field interaction, F2,14 = 0.576, P > 0.5, indicating that the reduction in RT to the singleton distractor relative to absent following right TMS was irrespective of visual field of presentation.

Figure 3.

Difference RTs in Experiment 1 for singleton distractor trials versus singleton absent, according to whether the target appeared in the left or right visual field, depicted for no TMS, left TMS (P3), and right TMS (P4) sessions.

Figure 3.

Difference RTs in Experiment 1 for singleton distractor trials versus singleton absent, according to whether the target appeared in the left or right visual field, depicted for no TMS, left TMS (P3), and right TMS (P4) sessions.

Effects of TMS on Crosstrial Contingencies

We also evaluated the effects of switching between the different types of trial, that is, singleton distractor present, singleton target present, and singleton absent. In particular, we were interested whether there were particular RT costs or benefits resulting in switching to singleton distractor trials from singleton target trials, relative to switching from singleton absent trials and singleton distractor trials, and we assessed whether these could be affected by TMS. Figure 4 shows the difference RT data for current singleton distractor trials, relative to singleton absent trials, as a function of whether the prior trials contained a singleton target or a distractor or whether the singleton was absent, separated according to the type of TMS stimulation. In this analysis as well as removing error trials, we removed the trials following error trials as it was uncertain whether participants adopted a checking strategy in this last case (as response criteria can shift after an error—see Rabbit and Vyas 1970). This led to the removal of 1 subject who made more than 30% errors in at least 2 conditions. Note that this participant made no more than 15% errors in the analysis reported above.

Figure 4.

Difference RTs for a current trial (trial n) when the singleton distractor was present relative to when it was absent trials, shown as a function of the event on trial n – 1. On the prior trial, there was a singleton target present, a singleton distractor present, or no singleton (a singleton absent trial). The data are depicted for no TMS, left TMS (P3), and right TMS (P4) sessions.

Figure 4.

Difference RTs for a current trial (trial n) when the singleton distractor was present relative to when it was absent trials, shown as a function of the event on trial n – 1. On the prior trial, there was a singleton target present, a singleton distractor present, or no singleton (a singleton absent trial). The data are depicted for no TMS, left TMS (P3), and right TMS (P4) sessions.

A 2-way ANOVA was performed on the difference RTs (singleton distractor present – absent), with the factors being TMS (none, left, or right) and type of previous trial (singleton as distractor, singleton as target, singleton absent). This showed a main effect of the previous trial, F2,12 = 25.142, P < 0.001. For a current trial with a singleton distractor, the RT cost (relative to trials where there was no singleton) was 58 ms greater when the target was a singleton on trial n – 1 than when there was either no singleton on trial N-1 and than when there was previously a singleton distractor present. There was also a main effect of TMS, F2,12 = 5.880, P < 0.05; the costs were smallest with right TMS compared with left TMS and no TMS. More importantly, there was also an interaction between TMS and type of previous trial, F4,24 = 3.327, P < 0.05. Right TMS had a specific effect on the carryover effect across trials, relative to left parietal cortex stimulation and to no TMS. For both the left parietal stimulation and the no TMS conditions, there was a greater cost on singleton distractor trials if the prior trial contained a singleton target than when there was either no singleton present (the cost increased by about 100 ms), t1,6 = 4.528, P < 0.01 for left TMS and t1,6 = 3.573, P < 0.02 for no TMS. In contrast, after TMS to right PPC, there was no significant increase in the RT cost following trials where the singleton was a target compared with when the preceding trial contained no singleton, t1, 6 = −0.479, P > 0.6. A similar analysis on current singleton target present trials showed no effect of crosstrial contingencies (Fig. 5). Finally, as for the search data, there was no significant effect of crosstrial contingencies on errors.

Figure 5.

Difference RTs for a current trial (trial n) when the singleton target was present relative to when it was absent trials, shown as a function of the event on trial n – 1. On the prior trial, there was a singleton target present, a singleton distractor present, or no singleton (a singleton absent trial). The data are depicted for no TMS, left TMS (P3), and right TMS (P4) sessions.

Figure 5.

Difference RTs for a current trial (trial n) when the singleton target was present relative to when it was absent trials, shown as a function of the event on trial n – 1. On the prior trial, there was a singleton target present, a singleton distractor present, or no singleton (a singleton absent trial). The data are depicted for no TMS, left TMS (P3), and right TMS (P4) sessions.

Discussion

Our data implicate the right PPC in attentional capture by a salient color singleton. Right parietal stimulation reduced the effect of singleton interference relative to when no TMS was applied, and it also reduced carryover effects across trials according to whether there was either no singleton or a singleton target or distractor on the prior trial. Consistent with previous research, when no TMS was applied, a color singleton distractor interfered with search RTs (Theeuwes 1991). On trials where a singleton distractor was present, search RTs were 150 ms slower than when the singleton distractor was absent. TMS on the right PPC reduced this interference effect to approximately 100 ms. This reduction was irrespective of whether the singleton distractor was ipsilateral or contralateral to the TMS site. In contrast, TMS on the left PPC had little effect on search performance (see Beck et al. 2006 for a similar result within the change blindness paradigm).

Of further importance, a selective reduction in across-trial priming effects following TMS to the right PPC accounted for a substantial part of the reduction in RTs to the singleton distractor. With no TMS, the effect of the singleton distractor was much greater if the target was a singleton on the previous trial relative to when there was no singleton or a singleton distractor present. This result is in agreement with previous work in crosstrial priming between target and distractor singletons by Pinto et al. (2005); attentional capture is enhanced if the distractor shares features with the prior target. Interestingly, following right PPC stimulation, the priming effect from a prior singleton target was completely eliminated—RTs were no slower on trials where the singleton distractor was present and it was preceded by a singleton target on the previous trial, compared with when trial N − 1 had a singleton distractor or when the singleton was absent. Note that TMS applied to the left PPC had also no effect on crosstrial priming RTs.

The dissociation between left and right TMS shows that the effects found were not simply due to the application of TMS per se (e.g., due to increased arousal following the click emitted when a TMS pulse was generated or practice effects as one no-TMS block was always applied first). Although there was a small advantage for the TMS conditions over the non-TMS (cf. Table 1), the difference between left and right TMS shows that our effects are specific to right PPC stimulation. We also discount the possibility that our results are due to eye movement artifacts. Participants were expressly instructed to maintain fixation throughout the task, and furthermore, if eye movements were critical then again we should have expected no difference in performance under left versus right TMS. There was also no effect associated with the color singleton being the target, either with or without TMS. This may be due to ceiling effects or due to participants attempting to reduce any attentional weight assigned to singleton values, as this may have entailed performance costs on singleton distractor trials. The pattern of the data, with a stronger disruption from a singleton distractor than any benefit from a singleton target, follows the pattern reported by de Fockert et al. (2004).

The effects of right PPC stimulation are consistent with there being a right lateralized response to salient stimuli, and moreover, this same brain region is implicated in carryover effects across trials, leading to particularly strong “pulls” on attention when the singleton value had previously been carried by a target. In Experiment 2, we further extend our findings by asking whether the effects of right PPC stimulation were due to the target being a singleton that may lead to an incentive to attend to the singleton and consequent carryover effects or whether there were still effects when the singleton was never a target. To assess this, we examined performance when the singleton conditions were blocked, so a trial with a singleton distractor was never preceded by one in which the singleton had been a target. Any attentional capture in this scenario will largely be stimulus-driven.

Experiment 2. Blocked Presentation of Singleton Present and Absent Trials

In the second experiment, we removed the possibility of intertrial priming effects by blocking the presentation of singleton distractor, singleton target, and singleton absent trials. If the disruptive effect of right parietal TMS was specific to a crosstrial priming mechanism, there should no longer be a relative decrease in the cost of singleton distractor presence after right PPC stimulation compared with trials with no TMS or with left PPC TMS. Note that in a block of singleton distractor trials there is never any incentive for participants ever to attend to the singleton value. Evidence here for singleton interference (relative to trials where the singleton is absent) and for the effect being modulated by TMS applied to right PPC would provide strong evidence both for the interference effect being automatic and for the right PPC being linked to an automatic response to highly salient stimuli.

Method

Participants

There were 8 participants, 5 male and 3 female of range 24–38 years. All participants had participated in TMS experiments previously and had no history of any neurological disorders.

Design and Stimuli

The stimuli, design, and TMS procedure were identical to Experiment 1. The only difference was that instead of the singleton absent, singleton target, and singleton distractor trials being presented all within the same block, they were presented within separate blocks.

Results

One participant was removed from the analysis as RTs in at least one condition were more than 2.5 standard deviations from the mean RT for all participants. The RT outlier procedure removed 2.71% of correct RTs, and the remaining RTs were summarized by TMS site (none, P3, and P4) and singleton (distractor, target, or absent) as in Experiment 1. Table 3 shows the overall mean RTs and Figure 6 shows the mean difference RTs for trials where the singleton was either a distractor or a target, relative to when it was absent.

Table 3

Experiment 2, mean correct RTs ± 1 standard error for singleton distractor present, singleton target present, and singleton absent with no TMS, left TMS (P3), and right TMS (P4)

 No TMS Left TMS Right TMS 
Distractor 732.8 (65.1) 700.3 (43.1) 658.3 (28.7) 
Target 644.1 (45.6) 625.8 (37.2) 620.3 (32.7) 
No singleton 688.9 (61.8) 664.8 (48.5) 657.9 (34.7) 
 No TMS Left TMS Right TMS 
Distractor 732.8 (65.1) 700.3 (43.1) 658.3 (28.7) 
Target 644.1 (45.6) 625.8 (37.2) 620.3 (32.7) 
No singleton 688.9 (61.8) 664.8 (48.5) 657.9 (34.7) 
Figure 6.

Difference RTs in Experiment 2 for trials where a singleton target or a singleton distractor was present minus trials where it was absent, depicted for no TMS, left TMS (P3), and right TMS (P4) sessions.

Figure 6.

Difference RTs in Experiment 2 for trials where a singleton target or a singleton distractor was present minus trials where it was absent, depicted for no TMS, left TMS (P3), and right TMS (P4) sessions.

We first compared RTs for distractor present and absent conditions over the 3 TMS conditions. There was no main effect of TMS (P > 0.2), but there was a marginal effect of singleton distractor F1,6 = 5.536, P < 0.05. Overall, participants were 27 ms slower when the singleton distractor was present. More importantly, there was an interaction between TMS site and distractor presence F2,12 = 4.337, P < 0.05. In the no TMS and left TMS conditions, there were costs when the singleton distractor was present relative to when it was absent (44 and 36 ms respectively). However, in accordance with our findings in Experiment 1, this cost was reduced after right PPC stimulation (the cost was then just 1 ms). Similar to Experiment 1, this reduction was irrespective of the site of stimulation, F2,12 = 0.449, P > 0.6 (see Fig. 7). An analysis for trials where the singleton was a target showed a facilitation effect relative to when the singleton was absent, F1,6 = 5.655, P = 0.055. However, as in Experiment 1, there was no effect of TMS on singleton target trials (P > 0.8). Finally as in the error data shown in Table 4, there were no statistically significant effects.

Table 4

Experiment 2, percentage errors ± 1 standard error for singleton distractor present, singleton target present, and singleton absent with no TMS, left TMS (P3), and right TMS (P4)

 No TMS Left TMS Right TMS 
Distractor 11.2 (2.0) 10.6 (1.9) 11.6 (1.7) 
Target 6.8 (1.7) 9.2 (1.6) 8.3 (1.1) 
No singleton 11.6 (1.1) 11.9 (2.2) 10.7 (2.5) 
 No TMS Left TMS Right TMS 
Distractor 11.2 (2.0) 10.6 (1.9) 11.6 (1.7) 
Target 6.8 (1.7) 9.2 (1.6) 8.3 (1.1) 
No singleton 11.6 (1.1) 11.9 (2.2) 10.7 (2.5) 
Figure 7.

Difference RTs in Experiment 2 for singleton distractor trials relative to singleton absent, according to whether the target appeared in the left or right visual field; depicted for no TMS, left TMS (P3), and right TMS (P4) sessions.

Figure 7.

Difference RTs in Experiment 2 for singleton distractor trials relative to singleton absent, according to whether the target appeared in the left or right visual field; depicted for no TMS, left TMS (P3), and right TMS (P4) sessions.

Discussion

Once again, we found an effect of TMS on attentional capture by a salient singleton distractor. Consistent with Experiment 1, this was specific to stimulation of the right parietal cortex and again equivalent for both contralateral and ipsilateral visual fields. We also note that as with Experiment 1, RTs in the singleton absent condition were some 30 ms quicker following TMS. However, once again the effect of stimulation on singleton distractor present RTs was confined to the right PPC, the left PPC functioning as a control site. One consequence of presenting our experimental conditions over blocks of trials was that overall the effect of the singleton distractor was much smaller relative to Experiment 1 (30 ms here as opposed to 150 ms in Experiment 1). This result has been observed previously in the literature (e.g., Theeuwes 1994). Nevertheless, there was still a significant reduction in the cost of a singleton distractor following right PPC TMS (indeed the cost was eradicated in this condition). The data indicate both that at least part of the singleton interference effect occurs automatically (even when the singleton feature is never associated with a target) and that this automatic component is affected by TMS applied to the right PPC. We return to this point in our discussion below.

In contrast to Experiment 1, there was a marginally significant benefit in performance for the singleton target condition relative to the no singleton condition. Rather than the singleton target data in Experiment 1 reflecting a floor effect, this would indicate that participants attempt to restrict attentional weight to the singleton, not only reducing RT costs of distraction but also any benefits of it being the target. However, as in Experiment 1, there was no effect of TMS on target singleton trials. Finally, also consistent with Experiment 1, there was no effect of left TMS on search behavior. Our pattern of data cannot be explained by TMS irrespective of stimulation site.

General Discussion

The Role of the Right PPC in Attentional Capture

In 2 experiments, we have found that stimulation of the right PPC reduces the interference effect of a salient singleton in a visual search display, relative to when the distractor is absent. This was the case when singleton distraction could be driven by a combination of top–down and bottom–up factors, that is, when singleton distractor, singleton target, and singleton absent trials were intermixed within blocks of trials as in Experiment 1 or when these differing conditions were presented in separate blocks of trials (Experiment 2). In addition, the data in Experiment 1 showed that there were additional RT costs when a singleton distractor was present if the preceding trial contained a singleton target (relative to when the preceding trial contained a singleton distractor and to when the singleton was previously absent), but these additional costs were eliminated following right TMS. It seems both the attentional capture effect from a singleton distractor and the priming from a prior singleton target (cf. Pinto et al. 2005) are dependent on the operation of the right PPC.

The finding that stimulation of the right PPC disrupted attentional capture by the singleton distractor extends the research of de Fockert et al. (2004). These authors found that bilateral parietal and left frontal activity was associated with the presence of a singleton distractor. Our data suggest that the right PPC in particular is important for responding to saliences and is necessary for biasing attention to salient items in the display—in this case, even when they are detrimental to the task. Further, this finding still held in Experiment 2, in which we presented singleton target and distractor trials in separate blocks, meaning there was no clear top–down incentive to attend to the salient item. This is consistent with previous research by Mevorach et al. (2006), where right parietal TMS disrupted participants’ ability to select the more salient dimension of a hierarchical (global/local) stimulus. A deficit in biasing attention toward salient items would also lead to reduced attentional capture. This argument is lent credence by Fuggetta et al. (2006), who showed that TMS over the right PPC disrupts the early phase of the N2pc ERP waveform, corresponding to attentional orienting to a target location in conjunction search. Interestingly, this is the same component that was modulated by the presence of a salient singleton distractor in the study of Hickey et al. (2006). These 2 findings are consistent with the notion that singleton distractors can generate the early capture of attention, and this is modulated by the right PPC. This argument, for impaired attentional capture after TMS to right PPC can also be applied to other results in the literature. Beck et al. (2006), for example, report a deficit in change blindness following TMS to this region. Further, similar to our data, they also showed no effect of left PPC TMS even though left parietal activation had shown up in the fMRI study on which their stimulation sites were based. Beck et al. (2006) argued that this result was consistent with right hemisphere dominance theory (e.g. Kinsbourne 1987) and this would follow here if the right PPC modulates attention to salient events (e.g., a dynamic change in a display). We return to this point below.

One outstanding question is why our TMS effects are restricted to the singleton distractor condition and have no impact on the singleton target or singleton absent conditions. In the latter conditions, the target could also be considered a salient singleton, and thus, we might expect right PPC TMS to slow RTs which here they manifestly did not. However, this result has a clear precedent in work comparing right parietal TMS in feature and conjunction search. For example, Ashbridge, Cowey, and Walsh (1997) showed no effect on feature search following right parietal TMS (whereas conjunction search was significantly impaired). As our singleton target and singleton absent can be considered equivalent to a feature search condition, here we would expect no effect of TMS either. In contrast to this, our data are more consistent with findings from single-cell recordings (Constantinidis and Steinmetz 2005, Bisley and Goldberg 2003). For example, Constantinidis and Steinmetz (2005) recorded the responses of neurons in the monkey PPC under conditions in which a salient color singleton either did or did not fall in their receptive field. As in Experiment 2, the singleton was always behaviourally irrelevant (the monkey was trained on a task at fixation). Constantinidis and Steinmetz (2005) found that the neurons increased their firing rate when the salient distractor item fell within their receptive field. Importantly, the neurons were not sensitive to the specific feature values of the display items but responded as to whether items were salient or not. Bisley and Goldberg (2003) showed a similar result using a saccade cuing paradigm rather than an array of visual items. Of particular interest in the study of Bisley and Goldberg (2003), neurons in the PPC showed greater activation when a distractor appeared in their receptive field compared with a target; in our data shown no effects of parietal TMS on the salient target condition. Both these papers support the idea that the PPC encodes the locations of salient visual stimuli, deriving a salience map of the visual field. This salience map may serve to bias competition (Desimone and Duncan 1995) or guide search (Wolfe 1994) to salient distractor stimuli within the visual field. Interestingly, the monkey and human brain differ in that lesions of the right PPC in monkeys are not associated with symptoms of hemineglect as shown in humans (e.g., Husain and Nachev 2007), and indeed, single-cell recordings were carried out bilaterally in the studies described above. Our data then are consistent with any PPC salience map being lateralized to the right hemisphere in humans. TMS to the right PPC (but not the left) may corrupt this salience map and lessen any attentional bias toward salient singletons. The consequence of this is a reduced RT cost of salient singletons as shown here.

An alternative view is that the right PPC is not necessarily involved in attentional orienting but rather in sustaining attention. It has been argued that deficits in sustained attention, following right PPC damage, may contribute to aspects of the neglect syndrome (Manly et al. 1999), whereas impairments in sustained attention could be linked to impairments in change blindness following right PPC simulation (cf. Beck et al. 2006). However, in terms of reduced sustained attention we might have expected an overall slowing of RTs but not a selective reduction in the costs associated with singleton distractors.

Furthermore, problems in biasing attention to salient signals may also explain the pattern of across-trial priming data. Crosstrial priming between singleton targets and distractors has been shown to increase the effect of attentional interference (Pinto et al. 2005). This may be because having a target carry that feature preactivates the relevant feature map, so that a distractor subsequently carrying the same feature has increased salience—because it will then more strongly activate representations for its location in the saliency map. It follows that if attentional biasing toward salient features is disrupted, any enhanced RT cost may also be reduced or eliminated, as our data show.

The increased effects of the singleton distractor, following a singleton target trial, may be either bottom–up or top–down in causation. For example, there may be a bottom–up increase in saliency if a target in general receives more processing than a distractor and if this holds for singleton as well as for nonsingleton trials. More protracted processing of the target could generate the preactivation of the feature map that may underlie the effect. Alternatively, the increased effects could be due to feedback to the feature map from a template for the target, which temporarily holds the color as well as the shape features defining the stimulus on the trial. If a top–down account is maintained, then it might be argued that TMS blocks this feedback process and so reduces singleton interference when the distractor subsequently carries the critical feature. However, this top–down account of TMS fails to explain the data from Experiment 2. In Experiment 2, singleton distractor trials occurred in a single block, and so there should never be feedback from a target template to the critical feature map. Nevertheless, TMS to the right PPC eliminated the singleton interference effect from distractors. Rather than TMS blocking top–down feedback, the results are more consistent with the idea that it is the response to salience itself within the PPC that is disrupted.

If right TMS disrupted the parietal response to salience, we might ask why the singleton interference effect from distractors was eliminated in Experiment 2, whereas in Experiment 1, there remained a robust 100 ms interference effect even when right PPC stimulation took place. One account of the residual cost of the distractor for Experiment 1 is that multiple salience representations are present in the visual system, particularly under conditions where salience is enhanced by priming critical stimulus features. Due to these other components, some effects of salience remain even when the right PPC response is reduced. Indeed previous research has suggested the representation of salience in areas throughout the visual stream, from V1 (Li 2002) to the frontal eye fields (FEFs) (Cornelissen et al. 2002), and these additional areas may still contribute to bias attention towards a salient distractor even when signals from the parietal orienting system are reduced. There is a paradox here, though. This is that the residual costs were the same irrespective of the stimulus preceding a singleton distractor trial (see Fig. 3), yet they were absent in Experiment 2, when possible priming from singleton targets to singleton distractors was prevented. The paradox can be explained if any priming from singleton target trials lasts longer than the immediate next event, so there would remain some reduced priming effects in Experiment 1 even when the singleton distractor trial did not immediately follow a singleton target trial. There are data on carryover effects in visual search lasting across more than one trial (Maljkovic & Nakayama 1994). These residual priming effects may have been sufficient to boost responses in brain regions other than the PPC to bias attention toward singleton distractors. One interesting implication of this argument is that immediate priming effects may be modulated primarily by the right PPC response, whereas longer lasting priming effects may be modulated more by other brain regions (e.g., the FEFs). This proposal needs to be tested.

Interestingly, in a study of priming effects across trials in patients with unilateral neglect, Kristjansson et al. (2005) found that there was priming from items with the same color, even when the critical item on the preceding trial was neglected by the patient. There was no priming when the position of the neglected target was repeated. The critical lesion overlap in the patients was in the white matter under the inferior parietal lobe and the superior temporal gyrus of the right hemisphere, suggesting that feature-based priming was independent of the (ventral) right hemisphere attentional system supporting conscious orienting to stimuli. This would be consistent with a form of feature-based priming occurring outside of the parietal regions damaged in these patients, while location based priming is mediated through the PPC.

We now return to the lack of effects on the singleton distractor condition following left PPC TMS. This was the case even thought there was activation in the left PPC in the de Fockert et al. (2004) study. One possibility builds on the long-standing proposal that the left PPC biases attention to the contra-lateral side of space, while the right PPC is involved in attentional detection on both sides of space (Kinsbourne 1987). Following left TMS, the right PPC then can still orient attention to both sides of space. However, following right TMS the left PPC will show a deficit in orienting to contralateral space. In terms of distractor interference, this would mean that the associated RT cost would differ across the visual field following right TMS. However, as we showed above, an analysis of RTs according to whether the singleton distractor fell in the right or left visual field in both Experiments 1 and 2 showed no differential performance after left TMS. Further, a deficit in covert orienting would not explain our across-trial priming effects, which were preserved after left TMS but disrupted after right. Our data also do not support the idea that a deficit in sustaining attention is responsible for the effects of right TMS as suggested by Beck et al (2006). As in our data, Beck et al. (2006) found no specific contra lateral deficit in change detection following TMS to right PPC; rather detection was impaired in both visual fields. It has been argued that the right PPC is responsible for sustaining attention across a task (Manly et al., 1999). Following right TMS, then, we would expect an overall slowing of RTs, but not a selective speeding in the singleton distractor condition. One explanation of the lack of visual field differences following right PPC TMS is that this is the neural substrate instigating orienting to salient events (Corbetta and Shulman, 2002) even though these events are represented bilaterally. On another account, if right TMS disrupts a salience map, this may decrease any non-spatial filtering cost associated with a salient singleton distractor. Folk and Remington (1998) argue that a salient singleton distractor may not necessitate a shift of spatial attention but may slow search anyhow, as it will compete for selection strongly with the target (but see Theeuwes, 1995). The non-spatial nature of this filtering cost (and any reduction thereof following right PPC TMS) would ensue that RTs would not differ according to the visual field of the singleton distractor. These accounts await further research.

In conclusion, we have shown that TMS to the right PPC reduces interference in search from a salient singleton distractor. The reduction in interference was due (at least in part) to the elimination of ‘top-down’ priming between consecutive occurrences of target and distractor singletons. However, there was also a reduction even when across trial priming was prevented by blocking the singleton conditions, implicating effects of right PPC TMS in reducing stimulus-driven capture effects. We conclude that the right PPC has an important role to play in attentional responses to stimulus salience.

Funding

Biotechnology and Biological Sciences Research Council (UK); Medical Research Council (UK).

Conflict of Interest: None declared.

References

Bacon
WF
Egeth
HE
Overriding stimulus-driven attentional capture
Percept Psychophys.
 , 
1994
, vol. 
55
 (pg. 
485
-
496
)
Beck
DM
Muggleton
N
Walsh
V
Lavie
N
Right parietal cortex plays a critical role in change blindness
Cereb Cortex
 , 
2006
, vol. 
16
 (pg. 
712
-
717
)
Bisley
JW
Goldberg
ME
Neuronal activity in the lateral intra-parietal area and spatial attention
Science
 , 
2003
, vol. 
299
 
5603
(pg. 
81
-
86
)
Constantinidis
C
Steinmetz
MA
Posterior parietal cortex automatically encodes the location of salient stimuli
J Neurosci
 , 
2005
, vol. 
25
 
1
(pg. 
233
-
238
)
Corbetta
M
Shulman
GL
Control of goal-directed and stimulus-driven attention in the brain
Nat Rev Neurosci.
 , 
2002
, vol. 
3
 (pg. 
201
-
215
)
Cornelissen
FW
Kimmig
H
Schira
M
Rutschmann
RM
Maguire
RP
Broerse
A
Den Boer
JA
Greenlee
MW
Event-related fMRI responses in the human frontal eye fields in a randomized pro- and anti-saccade task
Exp Brain Res.
 , 
2002
, vol. 
145
 (pg. 
270
-
274
)
Desimone
R
Duncan
J
Neural mechanisms of selective visual-attention
Annual Review of Neuroscience
 , 
1995
, vol. 
18
 (pg. 
193
-
222
)
de Fockert
J
Rees
G
Frith
C
Lavie
N
Neural correlates of attentional capture in visual search
J Cogn Neurosci.
 , 
2004
, vol. 
16
 (pg. 
751
-
759
)
Ellison
A
Rushworth
M
Walsh
V
The parietal cortex in visual search: a visuomotor hypothesis
Clin Neurophysiol
 , 
2003
, vol. 
56
 
Suppl
(pg. 
321
-
330
)
Folk
CL
Remington
RW
Johnston
JC
Involuntary covert orienting is contingent on attentional control settings
J Exp Psychol Hum Percept Perform.
 , 
1992
, vol. 
18
 (pg. 
1030
-
1044
)
Folk
CL
Remington
RW
Selectivity in distraction by irrelevant featural singletons: Evidence for two forms of attentional capture
J Exp Psychol Hum Percept Perform.
 , 
1998
, vol. 
24
 (pg. 
847
-
858
)
Fuggetta
G
Pavone
EF
Walsh
V
Kiss
M
Eimer
M
Cortico-cortical interactions in spatial attention: A combined ERP/TMS study
J Neurophysiol
 , 
2006
, vol. 
95
 
5
(pg. 
3277
-
3280
)
Heilman
KM
Valenstein
E
Mechanisms underlying hemispatial neglect
Annals of Neurology
 , 
1979
, vol. 
5
 
2
(pg. 
166
-
170
)
Hickey
C
McDonald
JJ
Theeuwes
J
Electrophysiological evidence of attentional capture
J Cogn Neurosci.
 , 
2006
, vol. 
18
 (pg. 
604
-
613
)
Hilgetag
CC
Theoret
H
Pascual-Leone
A
Enhanced visual spatial attention ipsilateral to rTMS-induced ‘virtual lesions’ of human parietal cortex
Nat Neurosci
 , 
2001
, vol. 
4
 (pg. 
953
-
957
)
Husain
M
Nachev
P
Space and the parietal cortex
Trends Cogn Sci.
 , 
2007
, vol. 
11
 
1
(pg. 
30
-
36
)
Kinsbourne
M
Jeannerod
M
Mechanisms of unilateral neglect
Neurophysiological and neuropsychological aspects of spatial neglect
 , 
1987
Amsterdam (the Netherlands)
North Holland
(pg. 
69
-
86
)
Kristjansson
A
Vuilleumier
P
Malhorta
P
Husain
M
Driver
J
Priming of color and position during visual search in unilateral spatial neglect
J Cogn Neurosci
 , 
2005
, vol. 
17
 (pg. 
859
-
873
)
Kumada
T
Humphreys
GW
Cross-dimensional interference and cross-trial inhibition
Percept Psychophys
 , 
2002
, vol. 
64
 (pg. 
493
-
503
)
Li
Z
A saliency map in primary visual cortex
Trends Cogn Sci.
 , 
2002
, vol. 
6
 (pg. 
9
-
16
)
Manly
T
Robertson
IH
Galloway
M
Hawkins
K
The absent mind: further investigations of sustained attention to response
Neuropsychologia
 , 
1999
, vol. 
37
 (pg. 
661
-
670
)
Mevorach
C
Humphreys
GW
Shalev
L
Attending to local form while ignoring global aspects depends on handedness: evidence from TMS
Nat Neurosci
 , 
2005
, vol. 
8
 (pg. 
276
-
277
)
Mevorach
C
Humphreys
GW
Shalev
L
Opposite biases in salience-based selection for the left and right posterior parietal cortex
Nat Neurosci.
 , 
2006
, vol. 
9
 (pg. 
740
-
742
)
Muller
HJ
Heller
D
Ziegler
J
Visual search for singleton feature targets within and across feature dimensions
Percept Psychophys
 , 
1995
, vol. 
57
 (pg. 
1
-
17
)
Nobre
AC
Coull
JT
Walsh
V
Frith
CD
Brain activations during visual search: contributions of search efficiency versus feature binding
Neuroimage
 , 
2003
, vol. 
18
 
1
(pg. 
91
-
103
)
Olivers
CNL
Humphreys
GW
Attentional guidance by salient feature singletons depends on intertrial contingencies
Journal of Experimental Psychology-Human Perception and Performance
 , 
2003
, vol. 
29
 
3
(pg. 
650
-
657
)
Pashler
H
Johnston
JC
Ruthruff
E
Attention and performance
Annu Rev Psychol.
 , 
2001
, vol. 
52
 (pg. 
629
-
651
)
Pinto
Y
Olivers
CN
Theeuwes
J
Target uncertainty does not lead to more distraction by singletons: inter-trial priming does
Percept Psychophys.
 , 
2005
, vol. 
67
 (pg. 
1354
-
1361
)
Rabbit
PMA
Vyas
SM
Sanders
AF
An Elementary Preliminary Taxonomy for Some Errors in Laboratory Choice RT Tasks
Attention and Performance 3
 , 
1970
Amsterdam (the Netherlands)
North Holland
(pg. 
56
-
76
)
Rushworth
MF
Ellison
A
Walsh
V
Complementary localization and lateralization of orienting and motor attention
Nat Neurosci
 , 
2001
, vol. 
4
 (pg. 
656
-
661
)
Rushworth
MF
Taylor
PC
TMS in the parietal cortex: updating representations for attention and action
Neuropsychologia.
 , 
2006
, vol. 
44
 
13
(pg. 
2700
-
2716
)
Theeuwes
J
Cross-dimensional perceptual selectivity
Percept Psychophys.
 , 
1991
, vol. 
50
 (pg. 
184
-
193
)
Theeuwes
J
Stimulus-driven capture and attentional set: selective search for color and visual abrupt onsets
J Exp Psychol Hum Percept Perform
 , 
1994
, vol. 
20
 
4
(pg. 
799
-
806
)
Theeuwes
J
Kramer
AF
Coles
MGH
Logan
GD
Perceptual selectivity for color and form: On the nature of the interference effect
Converging Operations in the Study of Visual Attention
 , 
1995
Washington DC
American Psychological Association
(pg. 
297
-
314
)
Theeuwes
J
Top-down search strategies cannot override attentional capture
Psychon Bull Rev.
 , 
2004
, vol. 
11
 (pg. 
65
-
70
)
Theeuwes
J
Burger
R
Attentional control during visual search: the effect of irrelevant singletons
J Exp Psychol Hum Percept Perform.
 , 
1998
, vol. 
24
 (pg. 
1342
-
1353
)
van Zoest
W
Donk
M
Theeuwes
J
The role of stimulus-driven and goal-driven control in saccadic visual selection
J Exp Psychol Hum Percept Perform
 , 
2004
, vol. 
30
 (pg. 
746
-
759
)
Yantis
S
Monsell
S
Driver
J
Control of visual attention
Control of cognitive processes: Attention and performance XVIII
 , 
2000
Cambridge (MA)
Massachusetts Institute of Technology Press
(pg. 
71
-
208
)
Yantis
S
Egeth
HE
On the distinction between visual salience and stimulus-driven attentional capture
J Exp Psychol Hum Percept Perform
 , 
1999
, vol. 
25
 (pg. 
661
-
676
)
Walsh
V
Pascual-Leone
A
Transcranial Magnetic Stimulation: a Neurochronometrics of Mind
 , 
2005
Cambridge (MA)
Massachusetts Institute of Technology Press
Wolfe
JM
Guided search 2.0 - a revised model of visual-search
Psychonomic Bulletin & Review
 , 
1994
, vol. 
1
 
2
(pg. 
202
-
238
)