The Distractor Positivity Component and the Inhibition of Distracting Stimuli

Our attentional systems are constantly bombarded by salient stimuli that have been designed to attract our attention. From neon construction cones on the roadside to pop-up notifications on our cell phones, our attentional systems must make split-second decisions to determine which stimuli in our environments are relevant to our immediate goals and which are just distractions. For this reason, visual warning signals are often designed to be physically salient, using bright colors or flashing lights to attract attention.

Formal research on attentional capture, however, has painted a more complex picture. There has been a longstanding debate about whether salient stimuli have an ability to involuntarily attract attention (see a review by Luck, Gaspelin, Folk, Remington, & Theeuwes, 2021). As a resolution, it has been suggested that salient stimuli might have the potential power to capture attention, but that capture can be reduced via an inhibitory mechanism (Gaspelin & Luck, 2018c, 2019). One line of support for this inhibitory process has come from ERP studies of the distractor positivity (P_D) component, which has been proposed to measure suppression of salient distracting stimuli. For example, Figure 1 depicts a version of a commonly used task to study visual distraction in which participants searched for a target stimulus and attempted to ignore a salient distractor that was uniquely colored. Behavioral performance indicated that the salient distractor was successfully ignored. As can be seen in the ERPs, there was a positivity occurring in electrode sites of visual cortex that were contralateral to the salient distractor relative to ipsilateral electrodes. This P_D component can be more easily seen in the contralateral-minus-ipsilateral difference waveform. It has been suggested to be a neural marker of an inhibitory process that was used to prevent distraction by the salient distractor.

The P_D component has been studied in a variety of visual search tasks and comes in many different shapes and sizes (see Figure 2). As a result, there are differing viewpoints about what exactly the P_D component indicates and what this means about the underlying architecture of visuospatial attention. This collaborative review provides an overview of the P_D component with a specific emphasis on how this ERP component has been used to study inhibition of distracting stimuli. Authors with a variety of viewpoints outline recent advances that have been made using the P_D component in subtopics related to their specific expertise, allowing the reader to gain a broader perspective of the P_D component than would be possible from a single author. In the following sections, we discuss previous research that has linked the P_D to suppression of salient distractors, as well as insights that the field of attentional capture has gained from this research. Ultimately, we conclude that the P_D component is an ERP measure of attentional orienting that could provide valuable insights into other areas related to cognitive control and inhibition.

Before explaining the P_D component, it will help to first explain the N2 posterior contralateral (N2pc) component. The N2pc has been proposed as a measure of attentional allocation to a visual stimulus in a given hemifield. It appears in posterior electrode sites over visual cortex (e.g., PO7/PO8) about 200–300 msec after the appearance of the search display and is a negative-going deflection that is larger in the hemisphere contralateral to the attended stimulus than in the ipsilateral hemisphere. For example, Luck and Hillyard (1990) had participants search for a triangle target among a set of arrows and report whether the target was present or absent. Eye movements were prohibited in this task, and participants were therefore required to use covert attention to find the target. When the target appeared in the left hemifield, there was a greater negativity in electrode sites over right visual cortex than over left visual cortex. Similarly, when the target appeared in the right hemifield, there was a greater negativity in electrode sites over left visual cortex than over right visual cortex. This N2pc component was interpreted as a measure of a covert attentional process that was used to find the target stimulus in the search displays.

Most now agree that the N2pc reflects some aspect of covert attentional allocation (see review by Luck, 2012). Famously, Luck and Hillyard (1994b) proposed that the N2pc reflects filtering of distractors around the attended location, based upon the observation that the N2pc is often largest when the attended location is surrounded by distractors that must be rejected (Bacigalupo & Luck, 2015; Woodman & Luck, 1999; Luck & Ford, 1998; Luck, Girelli, McDermott, & Ford, 1997; Luck & Hillyard, 1994a). However, this interpretation has been questioned, because of evidence that an N2pc can also occur when distractor filtering is seemingly unnecessary (Mazza, Turatto, & Caramazza, 2009a, 2009b; Eimer, 1996). As a result, other interpretations of the N2pc have been proposed, which mostly assume that the N2pc reflects some other aspect of attentional allocation, such as the shifting of attention (Tan & Wyble, 2015) or the extraction of information at the attended location (Foster, Bsales, & Awh, 2020; Wyble et al., 2020; Zivony, Allon, Luria, & Lamy, 2018).

The P_D component is, in many ways, the opposite of the N2pc component. Rather than measuring attentional allocation to a given stimulus, it has been proposed to measure suppression of a given stimulus. Like the N2pc, it appears at posterior electrode sites over visual cortex (e.g., PO7/PO8). Unlike the N2pc, it is a positive-going deflection (rather than a negative-going deflection) in the contralateral electrode sites compared with ipsilateral electrode sites. For this reason, the P_D has been assumed to be related to attentional suppression, rather than attentional enhancement of a stimulus.

The P_D was first identified by Hickey, Di Lollo, and McDonald (2009) who had participants search for a green diamond and ignore a red line distractor (Figure 3). The stimuli were arranged so that, on some trials, one item appeared on the vertical midline (i.e., directly above or below the central fixation point) and the other was lateralized. Because any item on the vertical midline is equally represented in both hemispheres, any lateralized activity from a midline item will effectively cancel out, allowing one to isolate ERP components to the lateralized stimulus. On target lateral/distractor midline trials, there was a negative-going deflection contralateral to the target (a target negativity [N_T]), which was taken to indicate attentional allocation to the target, similar to the N2pc. On distractor lateral/target midline trials, however, there was a positive-going deflection contralateral to the distractor (a P_D). A control experiment suggested that this lateralized positivity was not because of low-level sensory imbalances caused by presenting a lone item in one visual hemifield. The authors therefore reasoned that this newly discovered P_D component reflected suppression of the distractor stimulus, which was task-irrelevant and needed to be ignored to perform the task (see also Hilimire, Hickey, & Corballis, 2012).¹

The P_D component can occur in the time range between 100 and 500 msec, which is more variable than the timing of the N2pc component (200–300 msec). This might be related to the fact that suppression can occur either before or after the first shift of attention. Specifically, if suppression occurs before the first shift of attention, the P_D component may occur relatively early (e.g., 100–275 msec), either before or during the N2pc time window (e.g., Gaspar & McDonald, 2014; Sawaki & Luck, 2010). This “early” P_D component is often interpreted as a suppressive process that preemptively prevents attentional allocation to a stimulus. If suppression occurs after a stimulus is first attended, however, the P_D may occur at later time windows (e.g., 275–500 msec), after an N2pc is elicited by the stimulus. This “late” P_D component is often interpreted as a suppressive process involved in terminating attentional allocation to a stimulus (Drisdelle, Corriveau, Fortier-Gauthier, & Jolicoeur, 2023; Liesefeld, Liesefeld, Töllner, & Müller, 2017; Sawaki & Luck, 2013; Hilimire, Mounts, Parks, & Corballis, 2011).² In short, the P_D component can have a variable latency, and as is shown in the following sections, this is important to consider when interpreting this component.

In summary, the N2pc and P_D are highly related ERP components. They both are visually evoked components that occur at posterior electrode sites over visual cortex and are related to attentional processes that occur in vision. Whereas the N2pc is presumed to reflect attentional allocation to a stimulus, the P_D is presumed to reflect attentional suppression of a stimulus.

In attention research, there has been a longstanding debate about whether salient stimuli can involuntarily distract us even when they are task irrelevant. Here, a salient stimulus is defined as an object that contrasts with neighboring objects (or the background) on a low-level feature dimension, such as color or orientation (Nothdurft, 1993). For example, a lone red object among a set of homogenously colored green objects (called a color singleton) would be considered highly salient. Historically, the attentional capture debate was divided into two opposing viewpoints: stimulus-driven and goal-driven accounts.

Stimulus-driven accounts proposed that certain kinds of salient stimuli will automatically capture attention regardless of the observer's goals or intentions (Franconeri & Simons, 2003; Theeuwes, 1992; Jonides & Yantis, 1988; Yantis & Jonides, 1984). For example, Theeuwes (1992, Exp. 1) devised the additional singleton paradigm, in which participants searched for a circle target among diamonds and reported the orientation of a line inside the target (Figure 4A). On some trials, one of the diamonds was uniquely colored (i.e., a color singleton). RTs were slower when the singleton distractor was present compared with when it was absent. This interference effect was taken to indicate that the task-irrelevant color singleton automatically captured attention, slowing detection of the target when it was present.

Goal-driven accounts, however, proposed that salient stimuli have no automatic power to attract attention and that attentional selection is instead driven by the intentions of the observer (Folk, Remington, & Johnston, 1992). According to this account, participants first establish an attentional template for the target features and only stimuli matching this attentional template capture attention. Initial support for this account largely came from a modified spatial cueing paradigm, in which participants searched displays for a target of a specific color (e.g., red letter) and attempted to ignore a salient cue that appeared before the search display. Importantly, this cue could either match or mismatch the target color. Several studies showed that matching cues produced cue validity effects (i.e., better performance when the target appeared at the cue location vs. elsewhere), indicative of capture, whereas mismatching cues did not (e.g., Folk & Remington, 1998; Folk et al., 1992; see also Becker, Folk, & Remington, 2010). This contingent capture effect has been taken to suggest that salient stimuli do not have inherent power to attract attention and will only capture attention if they match the immediate goals of the observer.

Furthermore, goal-driven accounts suggested that previous studies supporting stimulus-driven accounts may have encouraged a strategy whereby participants established an attentional template for any kind of feature singleton (a strategy called singleton-detection mode; Bacon & Egeth, 1994). For example, in the study shown in Figure 4A, the target was a shape singleton, and this may have encouraged participants to look for any unique “popout” stimulus. As evidence of this, several studies have shown that when participants are instead encouraged to search for a specific shape rather than for a singleton (a strategy called feature-search mode; Figure 4B), interference effects on manual RT from color-singleton distractors can be largely eliminated (e.g., Leber & Egeth, 2006; Lamy & Egeth, 2003; Bacon & Egeth, 1994). Such findings were taken as additional evidence that attentional capture may be limited to situations in which the salient stimulus matches the intentions of the observer, consistent with goal-driven models.

Initial ERP studies of attentional capture largely paralleled the behavioral studies of attentional capture. These studies tested whether salient distractors elicited an N2pc component, as a measure of whether they captured attention. For example, Hickey, McDonald, and Theeuwes (2006) used a task similar to Theeuwes (1992) in which participants searched for a popout target on the shape dimension and attempted to ignore a color-singleton distractor (Figure 4A). Color-singleton distractors elicited an N2pc component, suggesting that they captured attention, and this finding was taken to support stimulus-driven accounts (but see McDonald, Green, Jannati, & Di Lollo, 2013). Other ERP studies, however, produced opposite results. For example, Lien, Ruthruff, Goodin, and Remington (2008) used a spatial-cueing paradigm in which participants searched for a target of a specific color and ignored a salient but irrelevant cue. Salient cues that matched the target color elicited an N2pc component, whereas salient cues that mismatched the target color did not. These findings were taken to suggest that salient stimuli capture attention only when they match the attentional template of the observer, consistent with goal-driven accounts (see also Lien, Ruthruff, & Cornett, 2010). Altogether, the N2pc-component studies resulted in the same kind of stalemate as the behavioral studies.

In summary, the attentional capture debate has been extraordinarily challenging to resolve. Each theoretical camp has been equally well-supported empirically and has typically relied on a unique set of tasks, making it difficult to pinpoint why opposing results were obtained. This state of affairs has laid the groundwork for a debate that has lasted several decades without a coherent resolution.

As an attempt to resolve this debate, Sawaki and Luck (2010) proposed that suppression could provide a framework for understanding when salient distractors will automatically capture attention and when they will be ignored. Specifically, they proposed the signal suppression hypothesis, whereby salient stimuli produce a bottom–up “attend-to-me” signal that competes for attention, consistent with stimulus-driven accounts, yet this salience signal can be overridden by an inhibitory process that prevents attentional capture, consistent with goal-driven accounts. They suggested that suppression of salient distractors could be measured by the P_D component.

To test this hypothesis, Sawaki and Luck (2010) had participants search for a target letter (e.g., a small green A) among green distractor letters (Figure 5). On some trials, the target was present. On other trials, the target was absent, and a distractor appeared at a lateralized location. This distractor could either be a target-similar distractor, which was the target letter in the wrong font size (e.g., large green A), or a singleton distractor, which was a random letter in a unique color (e.g., a red X). Targets elicited an N2pc component suggesting that they were attended. Target-similar distractors also elicited an N2pc component, suggesting that they captured attention consistent with a goal-driven account. Crucially, singleton distractors did not elicit an N2pc component and instead elicited a P_D component. Given the previous association of the P_D component with distractor suppression (Hickey et al., 2009), this pattern of results was taken to indicate that the singleton distractor was suppressed preemptively (i.e., before the first shift of attention) to prevent attentional capture. Follow-up experiments replicated this basic pattern of results and ruled out the possibility that the lateralized positivity (i.e., the P_D) was because of an imbalance in sensory energy.

Several subsequent ERP studies provided additional evidence that the P_D component reflects suppression of salient distractors to prevent attentional capture. For example, Gaspar and McDonald (2014, Exp. 1) had participants search for a yellow target among green items, while ignoring a red singleton distractor. Singleton distractors elicited a P_D component and no corresponding N2pc component, suggesting that the distractor was suppressed to prevent attentional capture (similar to Sawaki & Luck, 2010). In addition, the magnitude of the P_D component was found to be larger on fast-response trials than on slow-response trials, suggesting that successful suppression of the salient distractor allowed the target to be found more quickly (see also Jannati, Gaspar, & McDonald, 2013). Other studies found that the P_D component elicited by salient distractors emerged only when the search array duration was short (200 msec instead of until response), suggesting that the pressure to quickly locate the target incentivized participants to suppress the salient distractor (Kiss, Grubert, Petersen, & Eimer, 2012). All of these findings are consistent with the idea that the P_D component indicates some kind of process involved in the suppression of salient distractors.

Additional support that the P_D component reflects suppression has come from studies that combine ERPs with other methodologies. For example, Cosman et al. (2018) found evidence from single-unit recordings in nonhuman primates that neurons in the frontal eye fields had lower discharge rates for salient distractors than the baseline rate observed for nonsalient distractors (a suppression effect). This same study also found evidence of a monkey homologue to the P_D component that was observed in ERPs recorded from extrastriate cortex. In addition, Weaver, van Zoest, and Hickey (2017) conducted a study that concurrently measured EEG and eye tracking in humans performing an attentional capture task. They reported a P_D component that occurred before eye movements that were successfully directed away from the salient distractor, and the magnitude of this P_D component was positively correlated with the degree of curvature of eye movements away from the salient distractor. Both findings are consistent with the idea that suppression via covert attention was used to guide eye movements away from the salient distractor.

Further support for the signal suppression hypothesis that does not rely on ERPs has come from a variety of behavioral paradigms. For instance, in the capture-probe paradigm (Gaspelin, Leonard, & Luck, 2015), participants search displays for a target shape (e.g., a green diamond) and attempt to ignore a singleton distractor. On a subset of trials, letters are superimposed over each shape of the search display and participants attempt to report as many letters as they saw (Figure 6A). The key finding is that probe recall is impaired at the location of the singleton distractor compared with nonsingleton distractors, consistent with the idea that a suppressive process was applied to the salient distractor to prevent capture (Figure 6B; see also Lien, Ruthruff, & Hauck, 2022; Stilwell & Gaspelin, 2021; Chang & Egeth, 2019, 2021; Gaspelin & Luck, 2018a, 2018b; but see Oxner et al., 2023). Similarly, other studies measured shifts of gaze and demonstrated that eye movements were directed to the salient distractor at below-baseline levels (Adams, Ruthruff, & Gaspelin, 2023; Stilwell, Adams, Egeth, & Gaspelin, 2023; Hamblin-Frohman, Chang, Egeth, & Becker, 2022; Gaspelin, Gaspar, & Luck, 2019; Gaspelin & Luck, 2018a; Gaspelin, Leonard, & Luck, 2017). In addition, several studies showed that observers gradually learn to suppress salient distractors based upon their specific features (Ramgir & Lamy, 2023; Gaspelin et al., 2019; Stilwell, Bahle, & Vecera, 2019; Gaspelin & Luck, 2018a; Vatterott & Vecera, 2012), spatial locations (Ruthruff & Gaspelin, 2018; Sauter, Liesefeld, Zehetleitner, & Müller, 2018; Wang & Theeuwes, 2018a, 2018b; Goschy, Bakos, Müller, & Zehetleitner, 2014), and general status as a distractor (Ma & Abrams, 2022, 2023; Won & Geng, 2020; Won, Kosoyan, & Geng, 2019; Vatterott, Mozer, & Vecera, 2018). All of these behavioral studies provided converging evidence that some kind of suppressive process can be applied to salient distractors using experimental paradigms very similar to those used to study the P_D component.

Later studies provided a more direct linkage between attentional suppression and the P_D component by combining behavioral approaches to study suppression with ERP approaches (Gaspelin & Luck, 2018b; see also Feldmann-Wüstefeld, Busch, & Schubö, 2020). For example, using the capture-probe paradigm, Gaspelin and Luck (2018b) found below-baseline probe reports at the location of the singleton distractor (Figure 6B), as well as a P_D component associated with the singleton distractor (Figure 6C). Crucially, the magnitude of the P_D component and probe-based suppression effects was correlated: Participants who had large probe-based suppression effects also showed large P_D components. This finding demonstrates a direct association between the P_D component and the suppression of covert shifts of visual attention. Indeed, recent computational models of visual attention that have an inhibitory component are able to simulate a P_D component and several other behavioral measures of suppression (Wyble et al., 2020).

In summary, there is considerable evidence that salient distractors can be suppressed to prevent attentional capture (see also reviews by Gaspelin & Luck, 2018c, 2019). Some of this evidence has come from the P_D component, which is thought to index an inhibitory process in attention (but see section on Alternative Accounts), as well as from several converging lines of evidence from other methodologies (e.g., single-unit recordings, psychophysics, eye tracking, and computational modeling).

As previously reviewed, there is evidence that search strategy can strongly influence suppression of salient distractors.³ Tasks with “popout” targets can induce participants to adopt a more general attentional set for salient stimuli (Stilwell & Gaspelin, 2021; Gaspelin et al., 2015, 2017; Kerzel & Barras, 2016; Leber & Egeth, 2006; Bacon & Egeth, 1994). This singleton-detection mode strategy (Figure 4A) can be prevented by using tasks that encourage a feature-search mode strategy (Figure 4B). Typically, search tasks that encourage singleton-detection mode yield attentional capture by salient distractors, whereas tasks that encourage feature-search mode yield no attentional capture by salient distractors. Although there has been some debate about the underlying mechanisms that lead to this difference (e.g., Gaspelin, Egeth, & Luck, 2023; Liesefeld & Müller, 2023; Theeuwes, 2004, 2023; Leber & Egeth, 2006), it is abundantly clear that tasks in which the target is most often a singleton encourage attentional capture via strategic changes in how participants search displays.

The P_D component has provided some evidence consistent with a difference in capture based on the type of strategy afforded by the task. Several studies have shown that salient distractors elicit a P_D component in heterogeneous displays that encourage feature-search mode (Drisdelle & Eimer, 2021; Gaspelin & Luck, 2018b; Sawaki & Luck, 2010; but see Barras & Kerzel, 2016), and this suppression occurs even when distractors are highly salient (Drisdelle & Eimer, 2023; Stilwell et al., 2022). However, studies using homogenous displays have found more mixed results. If the target shape is predictable (e.g., always a circle), the participant can perform the task correctly by using either feature-search mode or singleton-detection mode. This has been called an “option trial” because the participant has the option to employ either strategy (e.g., Leber & Egeth, 2006). In this case, salient distractors mostly elicit a P_D component (Feldmann-Wüstefeld, Weinberger, & Awh, 2021; van Moorselaar, Daneshtalab, & Slagter, 2021; Barras & Kerzel, 2016; Gaspar & McDonald, 2014; Burra & Kerzel, 2013; Jannati et al., 2013) but may elicit an N2pc component in a subset of trials (McDonald et al., 2013). However, if the target is unpredictable (e.g., a circle among diamond distractors or a diamond among circle distractors, randomly intermixed), it will be disadvantageous to locate the target via feature-search mode, and participants will be forced to use singleton-detection mode to complete the search. In such cases, N2pc components were often found to be elicited by salient distractors (Burra & Kerzel, 2013; Hilimire et al., 2011; Hickey et al., 2006; but see van Moorselaar et al., 2021; McDonald et al., 2013). Therefore, the type of strategy employed by participants is heavily influenced by the type of search display and by the predictability of the target. Both can influence the strategy adopted by observers and should be considered when designing attentional capture studies measuring ERP components.

It is worth noting that the adoption of these control settings might be flexible. In tasks that use option trials, some participants might adopt either feature-search or singleton-detection mode for the entire experiment, or participants might fluctuate between search modes across trials. In either case, ambiguities about search strategies can lead to unclear interpretations of an observed P_D component, culminating in false conclusions concerning whether salient stimuli capture attention.

The ability to prevent attentional capture has been shown to critically depend on learning from distractor regularities as well as on short-term influences from recent experience, both thought to be implicit and inflexible (often collectively referred to as selection history; Theeuwes, Bogaerts, & van Moorselaar, 2022; Slagter & van Moorselaar, 2021; van Moorselaar & Slagter, 2020; Chelazzi, Marini, Pascucci, & Turatto, 2019; Noonan, Crittenden, Jensen, & Stokes, 2018). Several studies have shown that observers can learn to ignore salient distractors based upon their prior features (e.g., Stilwell et al., 2019; Gaspelin & Luck, 2018a; Vatterott & Vecera, 2012), prior locations (e.g., Failing, Feldmann-Wüstefeld, Wang, Olivers, & Theeuwes, 2019; Ferrante et al., 2018; Sauter et al., 2018; Wang & Theeuwes, 2018a, 2018b; Noonan et al., 2016), and the probability of their presence (Won & Geng, 2018, 2020; Moher, Abrams, Egeth, Yantis, & Stuphorn, 2011; Geyer, Müller, & Krummenacher, 2008). Furthermore, many studies showing suppression of salient distractors used experimental designs in which the color of the singleton distractor was fixed across trials (e.g., Gaspelin & Luck, 2018b). These findings are consistent with the notion that selection history may play a key role in eliminating attentional capture (Luck et al., 2021).

Statistical learning of distractor features and locations—and the corresponding reduction in capture—has been associated with an “early” P_D component (e.g., 100–200 msec) that occurs before the typical N2pc time window. For example, Wang, van Driel, Ort, and Theeuwes (2019) found that salient color singletons elicited a P_D emerging around 100 msec when these distractors appeared at a high-probability location, a finding that was presumed to reflect suppression of the distractor as a result of implicit learning. To further explore how learning shapes the time course of distractor suppression, van Moorselaar and colleagues (2021) compared distractor processing in a condition without any regularity to conditions where participants could learn to predict the upcoming distractor's location, its features (i.e., shape and color), or both. They found that both feature and location learning shifted the onset of the P_D earlier in time than in the baseline condition. Thus, statistical regularities of distractors across trials, whether spatial or feature based, seem to facilitate suppression of distractors to prevent attentional capture.

Interestingly, some studies have also reported a learning-based reduction in amplitude of a “late” P_D (250–350 msec) that occurred after an initial N2pc. This has mostly occurred in search tasks in which the target and distractors were defined on the same feature dimension (e.g., van Moorselaar, Lampers, Cordesius, & Slagter, 2020; van Moorselaar & Slagter, 2019). These findings suggest that when salient distractors are difficult to distinguish from the target stimulus, statistical learning may only influence reactive suppression after initial capture. When the distinction is easy (e.g., when the salient distractor is defined by color and the target is defined by shape), statistical learning may allow participants to proactively suppress the salient distractor to prevent attentional capture (see also the section on The Dimension Weighting Account).

Learning of distractor regularities necessarily involves repetition. For example, if a salient distractor appears at a given location with a high probability, there will necessarily be many trials in which the salient distractor repeats its location from the previous trial. This is important because several studies have shown that the location and feature of search items from the previous trial can automatically influence attentional allocation (called intertrial priming; Talcott & Gaspelin, 2020; Maljkovic & Nakayama, 1994, 1996; see review by Lamy & Kristjánsson, 2013). Yet, to date, there is little evidence that intertrial priming modulates the P_D component. For example, Feldmann-Wüstefeld and Schubö (2016) examined the effect of distractor-color repetition when the distractor's color varied unpredictably or alternated in triplet sequences. They reported no distractor-elicited P_D component in any of the conditions, suggesting that three distractor-color repetitions do not suffice to elicit a P_D component. Moreover, van Moorselaar and colleagues (2021) tracked the emergence of the early P_D in a statistical learning study and found a P_D only after 20 feature repetitions. It thus appears that the modulation of the P_D reflects learning across longer time scales and may not be because of intertrial priming, per se, but future research is necessary to establish how much learning is necessary to modulate the P_D.

Interestingly, the experimental task on the previous trial does strongly affect distractor suppression as evidenced by the P_D component. For example, Feldmann-Wüstefeld, Uengoer, and Schubö (2015) had participants perform a categorization task that was randomly intermixed with a search task. When the categorization task involved categorizing a color singleton as blue or green, the attentional priority of color transferred to a red singleton in a search task on subsequent trials: The salient distractor captured attention (as evidenced by an N2pc) before it was suppressed (as indicated by a subsequent P_D). When the categorization task involved classifying a shape singleton (hexagon vs. triangle), however, the same color singleton did not capture attention (i.e., it elicited a P_D with no N2pc). Later studies suggested that this effect of selection history could not be overcome by explicit cueing of the upcoming trial type and required almost 200 trials for extinction, indicating that the bias resulted from prior selection episodes rather than the observer's intention (Kadel, Feldmann-Wüstefeld, & Schubö, 2017; Feldmann-Wüstefeld et al., 2015; see also Berggren & Eimer, 2016). However, the attention bias could be reduced either by using a variant of the paradigm that encouraged voluntary choice on the next trial (Henare, Kadel, & Schubö, 2020) or by making the task sequence perfectly predictable (Abbasi, Henare, Kadel, & Schubö, 2023). Both manipulations eliminated attentional capture (no N2pc), whereas a P_D was still observed.

In summary, it is clear that selection history can influence the magnitude, presence, and latency of the P_D component. These findings are generally consistent with the interpretation that the P_D component measures a suppressive process and that the ability to ignore distractors becomes stronger as participants gain experience with them.

One theoretical approach to explain how stimulus-, goal-, and experience-driven influences combine—and that also acknowledges a specific function for signal suppression indexed by the P_D—is the dimension weighting account (DWA) (Found & Müller, 1996; Müller, Heller, & Ziegler, 1995). According to the DWA, experience and task goals modulate how strongly salience signals from various feature dimensions affect visual search. Specifically, attentional weights can be adjusted at the processing stage where (feature-less, but dimension-specific) salience signals are integrated into a search-guiding priority map. For example, a slightly tilted bar among vertical bars by itself exerts only little stimulus-driven guidance because of its weak salience, but its salience signal might be amplified because the observer looks for tilted bars (goal-driven) or because the previous search target was a tilted bar (experience-driven). From this perspective, distractor handling starts before the onset of a search display, either by increasing the weights for salience signals from the target dimension or by reducing the weights for those from the distractor dimension. The DWA has been supported by a wide variety of studies using psychophysics, mental chronometry, computational modeling, eye tracking, EEG, and fMRI (see reviews by Liesefeld, Liesefeld, Pollmann, & Müller, 2018; Krummenacher & Müller, 2012).

One recent line of evidence for the DWA that is relevant to the current article has come from studies using a version of the additional singleton paradigm that allows precise control over the stimuli's salience. In these studies, participants look for a popout target (e.g., a 12° tilted bar in a large array of homogenously oriented vertical bars) and a salient distractor is present on some trials. Critically, two types of distractors are typically used: same-dimension and different-dimension distractors. A same-dimension distractor is a distractor that is salient in the same dimension as the target (e.g., a bar tilted 45° in the opposite direction than the 12° target). Because any advance downweighting of salience signals from the distractor dimension would also reduce target priority and vice versa (because observers can only down-weight salience signals rather than specific feature values), the DWA predicts that such a target/distractor combination cannot be differentially weighted in advance. Consequently, if the same-dimension distractor is more salient than the target, it initially obtains a higher value on the priority map and will therefore reliably capture attention. Confirming these predictions, this design produced the clearest electrophysiological evidence for attentional capture (see Liesefeld & Müller, 2019): The same-dimension distractor elicited an N2pc that occurred before the target N2pc, and the target-elicited N2pc was delayed by distractor presence, suggesting that the distractor captured attention and delayed the allocation of attention to the target (Liesefeld et al., 2017). Of special importance here is that following the distractor-elicited N2pc, a P_D component emerged (Figure 7A). This pattern of results is indicative of a reactive control (Geng, 2014; Braver, 2012), with the P_D potentially indexing a suppressive process involved in recovery from capture, such as disengagement of attention (Sawaki, Geng, & Luck, 2012; Fukuda & Vogel, 2009, 2011) and/or avoidance of revisiting the distractor (inhibition of return, Klein, 2000; e.g., as implemented in the computational model of Moran, Zehetleitner, Müller, & Usher, 2013, which has been adapted to the additional-singleton paradigm by Liesefeld & Müller, 2020).

Other conditions of these studies used different-dimension distractors in which the salient distractor was defined in a different dimension than the target (e.g., a color-singleton distractor and an orientation target). It is clear from these studies that different-dimension distractors are heavily downweighted. First, different-dimension distractors interfere less with behavioral search performance than same-dimension distractors (Liesefeld, Liesefeld, & Müller, 2019). Second, instead of producing an N2pc followed by a P_D (as evoked by same-dimension distractors and indicative of capture), different-dimension distractors elicit only a P_D in an otherwise identical experimental setup (Figure 7B; Liesefeld, Liesefeld, & Müller, 2022).

It is intriguing that in some previous studies, distractors reliably produced residual interference effects, even when they were salient in a different dimension than the target (e.g., Theeuwes, 1992; see Figure 4A). This might indicate that the salience signals are never fully eliminated (i.e., dimensional weights are never brought to zero), so that a different-dimension distractor will typically cause some residual activation on the priority map (at least when search is salience-based; see also the section on Search Strategy: Feature-Search and Singleton-Detection Mode). Liesefeld and colleagues (2022) speculated that in studies in which distractors are salient in a different dimension than the target, as in the large majority of studies reporting a P_D during popout search (e.g., Burra & Kerzel, 2013; Jannati et al., 2013), the P_D reflects a resolution process that is needed to decide between multiple peaks on the priority map. In particular, only when the residual distractor activation on the priority map is suppressed would attention be free to shift to the stimulus with the highest activation on the map (i.e., the less salient, but dimensionally upweighted target in these studies). Thus, the cognitive process reflected by the P_D would effectively implement a winner-takes-all mechanism (see Livingstone, Christie, Wright, & McDonald, 2017, for empirical evidence compatible with that idea). In summary, from the DWA perspective, the P_D likely signals a suppressive mechanism akin to the one proposed by the signal suppression hypothesis. However, this suppressive mechanism either serves to give the final impetus for a shift of attention toward the target when the target already has highest priority (e.g., when salience signals from more salient different-dimension distractors are sufficiently downweighted) or to reactively recover from capture when the distractor initially has the highest priority (which inevitably is the case for same-dimension distractors more salient than the target).

Visual working memory (VWM) is a processing hub that serves ongoing tasks by fulfilling several crucial cognitive functions via the active maintenance of visual information (Liesefeld & Müller, 2019; Luck & Vogel, 1997, 2013). Important for the present purposes, there is now considerable evidence that visuospatial attention may be used to help control what information has access to VWM (Vogel, McCollough, & Machizawa, 2005). As a result, many models of VWM now assume that a common spatial priority map coding for potential relevance at each location in the visual field guides both attentional selection and access to VWM (Liesefeld, Liesefeld, Sauseng, Jacob, & Müller, 2020; Bundesen, 1990). Accordingly, manipulations that have well-established effects on search guidance also have strong effects on VWM performance—for instance, highlighting more relevant stimuli via cues (Emrich, Lockhart, & Al-Aidroos, 2017; Bays, Gorgoraptis, Wee, Marshall, & Husain, 2011), varying stimulus salience (Constant & Liesefeld, 2021, 2023), or introducing statistical regularities (Conn, Becker, & Ravizza, 2020; Umemoto, Scolari, Vogel, & Awh, 2010).

Consistent with a strong overlap in underlying control processes, the results from several studies indicate that the mechanism indexed by the P_D does play a role not only in the deployment of attention but also in encoding into VWM. In particular, Feldmann-Wüstefeld and Vogel (2018) provided a direct demonstration of a link between the P_D and VWM encoding. On each trial, displays contained two to-be-memorized stimuli (colored disks, targets) accompanied by two, four, or six to-be-ignored stimuli (colored squares, distractors). Adapting a method from the attentional capture literature (e.g., Hickey et al., 2009), they presented either of the two stimulus types on the vertical midline, whereas the other type was presented on the horizontal midline. The crucial finding was that the lateralized distractors elicited a P_D component, suggesting that they were suppressed to avoid their encoding into VWM. Importantly, this P_D component's amplitude increased with four relative to two distractors and did not increase further with six distractors. This finding indicates that the mechanism indexed by the P_D is sensitive to the amount of suppression required and may have a fixed, limited capacity. Finally, P_D amplitude showed a positive correlation with general VWM performance, a result that nicely dovetails with earlier observations that P_D amplitude measured in an additional-singleton paradigm correlates with VWM performance (Gaspar, Christie, Prime, Jolicœur, & McDonald, 2016).

Other studies further suggest that the P_D may specifically index a process involved in filtering out irrelevant information by preventing its encoding into VWM. Liesefeld, Liesefeld, and Zimmer (2014) cued participants to attend to either the left or the right hemifield and memorize a variable number of targets among a variable number of distractors for subsequent change detection. They found that the to-be-ignored distractors elicited a positivity at the same electrode sites and in the same time range as previous studies of the P_D, which is consistent with the idea that participants were suppressing irrelevant distractors to better recall target stimuli. The latency of this P_D-like component was negatively correlated both with VWM performance (in line with the above-mentioned correlations reported by Feldmann-Wüstefeld & Vogel, 2018; Gaspar et al., 2016) and with the amplitude of a prefrontal signal, with both indices showing a close association with the effect of distractor presence on maintenance-related parietal delay activity. The authors concluded that detection of distractors in posterior brain regions, indexed by the P_D-like component, triggers a prefrontal bias signal that reduces the amount of distractor information encoded in VWM (see also Emrich & Busseri, 2015).

A recent study by Hakim, Feldmann-Wüstefeld, Awh, and Vogel (2021) provided additional evidence for the idea that the P_D indexes processes that control access of spatially attended information into VWM. The authors presented intervening stimuli during the retention interval of a VWM task (either irrelevant distractors or stimuli labeled “task-relevant distractors” that were associated with an additional go/no-go task). They relied on the P_D as a measure of distractor suppression, on contralateral delay activity (e.g., Luria, Balaban, Awh, & Vogel, 2016; Vogel & Machizawa, 2004) as a measure of encoding into VWM, and on lateralized alpha power as a measure of spatial attention allocation (e.g., Peylo, Hilla, & Sauseng, 2021; Foster & Awh, 2019; but see Balestrieri, Michel, & Busch, 2022, for an alternate interpretation of alpha-band lateralization). The main conclusion of this study was that both relevant and irrelevant distractors captured visuospatial attention, but unlike the former, the latter were suppressed (i.e., triggered a P_D) and were not encoded into VWM.

Taken together, these findings suggest that the efficiency of distractor filtering—as indexed by the P_D component—is of crucial relevance for controlling access to VWM and that interindividual differences in this ability are related to VWM performance (see also Awh & Vogel, 2008; Vogel et al., 2005). Thus, the P_D is highly relevant for basic and applied research, because the mechanisms it indexes have a much more profound and long-lasting impact on human cognition than via fleeting attention allocations alone (see also Constant & Liesefeld, 2023).

The previous sections reviewed evidence that the P_D component indicates a suppressive process that is applied to salient distractors to prevent attentional capture. However, not all lateralized positivities in visually evoked ERPs are necessarily a P_D component. A lateralized positivity could instead arise from low-level imbalances in “sensory energy” caused by presenting a feature singleton in one hemifield but not the other. For example, this could cause a positivity posterior contralateral (Ppc) component indicating a generalized salience signal (Barras & Kerzel, 2016, 2017; Corriveau et al., 2012). It could also cause a larger contralateral P1 waveform owing to feature-specific lateral inhibition (Schein & Desimone, 1990) or lower levels of adaptation for the singleton's feature (Luck & Hillyard, 1994a, 1994b). Such sensory-level ERP components could easily be confused with the P_D component. It is therefore worth discussing what these components are and how they can be disambiguated from a P_D component.

One specific component of concern is the Ppc component, which is a positive-going deflection that occurs in posterior electrodes sites (e.g., PO7/PO8) contralateral to a salient stimulus and tends to occur in the time range of 80–150 msec. Importantly, the Ppc component seems to be unrelated to suppression and seems to instead be related to imbalances in low-level sensory properties of the stimuli across hemifields. For example, Corriveau and colleagues (2012) had participants search displays of gray circles for a color singleton target in one color (e.g., a red circle) and ignore a color singleton in another color (e.g., green circle). There was a positive-going deflection at electrode sites contralateral to the singletons that occurred irrespective of whether the singleton was a target or distractor. Because the target should not be suppressed, the authors concluded that this lateralized positivity was not a P_D component. Instead, they suggested this Ppc component reflected a salience signal generated by the color singleton (see also Barras & Kerzel, 2016, 2017; Fortier-Gauthier, Moffat, Dell'Acqua, McDonald, & Jolicœur, 2012; Leblanc, Prime, Jolicœur, & Jolicoeur, 2008). Consistent with this idea, early positivities are sometimes reduced when the distractor is less salient (Drisdelle et al., 2023; Kerzel & Huynh Cong, 2022). Relatedly, the positivity could have also indicated a kind of P1 adaptation effect, whereby the visual system adapted more to gray neutral distractors (of which there were eight in each display) than to color singletons (of which there were only two in each display). In either case, the Ppc component would be caused by low-level sensory properties that are unrelated to attentional suppression.⁴

Some initial studies of the P_D component used control experiments to rule out low-level sensory effects. For example, recall that Sawaki and Luck (2010) had participants search for a target letter while ignoring a color singleton and that the color singleton elicited a P_D component (black line in Figure 8A). To test whether this P_D was because of low-level sensory effects, they used a central fixation control task, which used the same stimuli as the search task, but participants searched a small display of Landolt Cs centered at fixation for a stimulus with a specific orientation (e.g., upward C). This central fixation task should focus visuospatial attention at the center of the display, reducing the need to suppress the salient distractor in the periphery. If the P_D in the search task was actually because of a low-level sensory effect (e.g., a Ppc), the lateralized positivity should appear even when attention is focused at fixation. If the lateralized positivity disappears, however, then it suggests that the previously observed positivity was not because of a low-level sensory effect and instead reflects something about attentional suppression. As can be seen, the P_D was eliminated in the central fixation task, ruling out a low-level sensory effect (see also Donohue, Bartsch, Heinze, Schoenfeld, & Hopf, 2018, Exp. 3).

Another approach is to use a singleton-target control task (Figure 8B). For example, Gaspelin and Luck (2018b) found that a color singleton elicited a P_D component, which they attributed to suppression. To rule out the possibility of a salience signal, they ran a control experiment, suggested by John McDonald, in which the color singleton was the distractor for one half of blocks and was the target for the other half of blocks (similar to Barras & Kerzel, 2017; Corriveau et al., 2012). If the lateralized positivity was because of a low-level sensory effect (e.g., a Ppc), this positivity should occur irrespective of whether the singleton was a target or distractor. In both instances, the singleton should be salient and should therefore generate the low-level sensory effect. If the lateralized positivity was instead because of suppression (i.e., a P_D), this positivity should be eliminated when the color singleton is the target, as the participant cannot suppress the target and still perform the task. Consistent with the latter interpretation, the P_D component was eliminated when the singleton was a target, suggesting it was specifically involved in some process related to the rejection of a salient distractor.

In summary, some lateralized positivities may be because of low-level sensory imbalances caused by presenting salient distractors in one hemifield, and these could easily be confused with a P_D component. These sensory components occur mostly early (between 80 and 150 msec poststimulus) and have been referred to as Ppc. A hallmark of the Ppc is that it does not differ between target and distractor stimuli, whereas the P_D does. Thus, to make sure that a positivity does indeed reflect suppression, it is important to conduct control experiments that rule out the possibility of a low-level sensory effect.

Many of the previous sections presented evidence supporting the idea that the P_D component reflects suppression of salient distractors. However, there are alternative accounts of this component. It should be noted that the N2pc and the P_D are calculated by subtracting ipsilateral from contralateral voltages. These difference values might cause some ambiguity because the N2pc and P_D occur in similar time ranges at the same electrodes. As explained above, a P_D to a distractor corresponds to a more positive voltage contralateral to the hemifield containing the distractor. However, it is possible to reinterpret this difference relative to the other hemifield. That is, a more positive voltage contralateral to the distractor can also be viewed as a more negative voltage contralateral to the hemifield without distractor. Thus, a P_D to a distractor in one hemifield is equivalent to an N2pc to a stimulus in the hemifield without distractor.

The recent controversy around distractor-elicited potentials in feature-search mode with small search displays illustrates this ambiguity (see Figure 9). Gaspelin and Luck (2018b) observed a P_D to the distractor, suggesting that this distractor was attentionally suppressed. Consistently, search times were shorter when the distractor was present, as if suppression had reduced the effective set size from four to three stimuli. Kerzel and Burra (2020) replicated the study by Gaspelin and Luck (2018b) and observed that the positivity to the distractor was followed by a negativity (a so-called P-N flip). Although observed in several studies (Stilwell et al., 2022, Exp. 1; Drisdelle & Eimer, 2021; Gaspelin & Luck, 2018b, Exp. 3), the P-N flip has been interpreted in different ways. Kerzel and Burra (2020) suggested that the P-N flip was a sequence of two N2pcs and reflected an idiosyncratic scanning strategy that occurs with small set sizes. Because the salient distractor's color was predictable, participants may have created a template for rejection (Arita, Carlisle, & Woodman, 2012) and directed attention away from the salient distractor. Thus, the initial P_D to the distractor was reinterpreted as an N2pc elicited by the inconspicuous nontarget on the opposite side, whereas the subsequent negativity was seen as an N2pc to the distractor, which was selected despite being irrelevant. The idea of sequential selection of lateral stimuli was supported by shorter RTs for horizontal than for vertical targets. According to this account, then, the apparent P_D reflects upweighting of a nontarget rather than downweighting of the salient distractor (see also Kerzel & Hyunh Cong, 2023). Consistent with this possibility, recent evidence suggests that the P_D is attenuated when target features are no longer fixed but instead vary randomly (van Moorselaar, Huang, & Theeuwes, 2023): In that case, nontargets contralateral to the distractor no longer benefit from the upweighting of the target color. However, in that study the P_D was not eliminated, suggesting that both upweighting of the majority color and downweighting of the salient distractor's color can occur concomitantly (see Chang & Egeth, 2019, for additional evidence).

There has been some evidence against the idea that the P_D reflects attentional allocation to the nonsingleton distractor on the horizontal midline. For example, Drisdelle and Eimer (2021) discouraged any horizontal attentional scanning strategy by ensuring that the target never appeared at lateral positions, yet the P-N flip persisted. The authors concluded that the P-N flip might reflect an initial P_D to the singleton distractor followed by a P_D to the nonsingleton distractor. In addition, Stilwell and colleagues (2022) also showed that the P_D occurred even at high set sizes in which an idiosyncratic search strategy should be discouraged (see also Drisdelle & Eimer, 2023; Gaspar & McDonald, 2014; Sawaki & Luck, 2010). Finally, Tam, Callahan-Flintoft, and Wyble (2022) applied a computational model of attention to simulate the P-N flip and found the P-N flip occurred even in a model where the salient distractor was suppressed to baseline (whereas the nonsingleton distractors were first activated and then suppressed below baseline). However, the conclusions from this study also critically depend on the interpretation of the components as a P_D elicited by a distractor on one side or as an N2pc to a distractor on the other side. To conclude, there is no definitive answer as to what processes the P-N flip might index and further research is therefore needed to answer this question.

In an attempt to resolve the ambiguity between the N2pc and P_D components in ERP studies of distractor suppression, several studies have relied on other measures of stimulus enhancement and inhibition, namely, steady-state visual evoked potential (SSVEPs) and alpha-band oscillations. The SSVEP is an oscillatory electrophysiological response of the visual cortex to a flickering stimulus. The frequency of the SSVEP matches the frequency of the flickering stimulus, and previous research has established that its amplitude increases when a stimulus is attended (Norcia, Appelbaum, Ales, Cottereau, & Rossion, 2015; Müller, Reimann, & Krummenacher, 2003; Morgan, Hansen, & Hillyard, 1996). Alpha oscillations are not locked to the frequency of the stimulus, but occur in a fixed frequency range of 8–12 Hz and their amplitude was found to decrease in the hemisphere contralateral to the attended stimulus (Thut, Nietzel, Brandt, & Pascual-Leone, 2006). Thus, SSVEPs and alpha oscillations measure the deployment of attention in different ways.

To measure SSVEPs, it is necessary to use flickering search displays rather than the typical abrupt-onset search displays. Accordingly, several SSVEP studies presented each search display as a brief change in color and shape within an ongoing stream of flickering gray disks, with the critical stimuli placed on the horizontal and vertical meridian, as in previous studies on the P_D. As shown in Figure 10A, the typical P_D component was observed when the target was vertical and the distractor was lateral (distractor lateral-target vertical), both with four-item displays (Forschack, Gundlach, Hillyard, & Müller, 2022a) and with two-item displays (Forschack, Gundlach, Hillyard, & Müller, 2022b). If the observed P_D indicates attentional suppression of the distractor, the amplitude of the SSVEP corresponding to the distractor frequency should decrease after search display onset and the amplitude of alpha oscillations contralateral to the distractor should increase. However, the opposite pattern of results was observed as is clear from Figure 10B and 10C. Thus, frequency-based measures of attentional deployment suggest that participants were attending the distractor, rather than suppressing it.

Given these results, frequency-based measures of attentional deployment are inconsistent with the idea that the P_D is an index of proactive attentional suppression. A similar inconsistency was reported with regard to findings by Hilimire and colleagues (2012) and Kiss and colleagues (2012). These authors found that the P_D to the distractor was larger when the distractor was shown together with a target than when it was presented alone and concluded that the P_D reflects distractor suppression in the service of target disambiguation. However, Forschack and colleagues (2022a) and Forschack and colleagues (2022b) failed to replicate this effect: They found a larger P_D when the distractor was shown alone. In addition, SSVEP amplitudes at the frequency of the lone distractor were increased relative to the precue baseline, suggesting that it was attended.

Finally, attentional suppression is expected to increase with the saliency of the distractor (Stilwell et al., 2023; but see Ramgir & Lamy, 2023; Wang & Theeuwes, 2020). The reason is that target saliency facilitates attentional selection. Therefore, more suppression is expected to prevent the potentially stronger attentional capture by more salient distractors. However, Forschack, Gundlach, Hillyard, and Müller (2023) found that the distractor-elicited P_D was unaffected by distractor saliency. Similarly, the contralateral alpha amplitude was unaffected. In contrast, target selection was indeed facilitated by increased salience, as evidenced by an earlier N2pc to the target and shorter RTs (Töllner, Zehetleitner, Gramann, & Müller, 2011). Thus, the predicted increase of attentional suppression with the increased distractor saliency was not observed. It should be noted, however, that other studies have found the P_D to increase in magnitude with distractor saliency (Drisdelle & Eimer, 2023; Gaspar & McDonald, 2014). Thus, future research is needed to resolve this apparent discrepancy.

In summary, the findings from typical ERP studies of the P_D and from studies using frequency-dependent measures do not converge. Whereas the former link the P_D to suppression of the salient distractor, the latter report a P_D together with evidence showing that the distractor is attended rather than suppressed, namely, increased SSVEP amplitudes and decreased contralateral alpha-band amplitudes. Possibly, the enhanced distractor processing is related to one of the following processes, notably (1) the reactive disengagement of attention from the distractor location (Klink, Teeuwen, Lorteije, & Roelfsema, 2023; Belopolsky, Schreij, & Theeuwes, 2010; Theeuwes, 2010), (2) the “zooming in” on the target while ignoring the distractor (Forschack et al., 2022a, 2022b; Liesefeld, Liesefeld, & Müller, 2021), (3) the learning of task-irrelevant features (van Moorselaar & Slagter, 2019), (4) the shielding of working memory from highly distracting input (Feldmann-Wustefeld & Vogel, 2018), or (5) the redirecting of attention away from salient distractors (Kerzel & Burra, 2020). In other words, more research is needed to bridge the gap between these studies. In particular, whether the results from frequency-based measures generalize to different sets of stimuli and tasks would be an important next step in this endeavor.

The P_D component could be used as a translational tool to study mental health disorders that involve dysregulations of inhibition and cognitive control. One line of evidence for this has come from studies of individuals with anxiety disorders. For example, Gaspar and McDonald (2018) recently found that individuals with high anxiety had difficulty suppressing salient distractors. High-anxiety individuals showed an initial N2pc followed by a P_D component, whereas low-anxiety individuals showed only a P_D component (i.e., similar to the results in Figure 7, but using different stimulus displays). This finding was taken to indicate that the high-anxiety group had difficulty preemptively suppressing salient distractors to prevent distraction. In addition, Kappenman, Geddert, Farrens, McDonald, and Hajcak (2021) had high- and low-anxiety individuals attempt to ignore a stimulus that was previously associated with shock. In both the high- and the low-anxiety groups, the stimulus elicited an N2pc component indicating that it captured attention. Interestingly, individuals with high anxiety had a larger “late” P_D after the initial N2pc than individuals with low anxiety. This finding suggests that suppression may be enhanced in individuals with anxiety disorders, a mechanism that would presumably help these individuals avoid threatening stimuli after initial distraction.

Distractor suppression may also provide useful tools to understand schizophrenia, a condition that is known to involve attentional impairments (Luck & Gold, 2008). For example, one recent hypothesis is that individuals with schizophrenia may “hyperfocus” their attentional resources and that, as a result, they have difficulty distributing their attention across space and show supranormal control over attention (Luck, Hahn, Leonard, & Gold, 2019). Interestingly, the extent of such hyperfocus was shown to correlate with impairments in a variety of general cognitive functions. Studies relying on the P_D component played a pivotal role in elaborating this hypothesis. For instance, it was shown that relative to control participants, individuals with schizophrenia were (ironically) better able to ignore task-irrelevant distractors and showed an enhanced P_D component in a typical laboratory task to study distractor suppression (Sawaki et al., 2016). Similar results have been found with eye-tracking measures of suppression (Bansal et al., 2021).

In summary, inhibition and cognitive control processes are impaired to some extent in a variety of mental health disorders. It therefore seems likely that the investigation of these disorders could directly benefit from ERP measures of distractor suppression such as the P_D component.

As reviewed, there has been an abundance of research using the P_D component to study inhibition of distracting stimuli. Some of this research has demonstrated that salient distractors can be suppressed to prevent attentional capture. Other research has shown that, in situations where distraction cannot be prevented, suppression can be used to mitigate the effects of distraction (e.g., via suppression that occurs after the initial shift of attention). Furthermore, much of the evidence from the P_D component has suggested that distractor suppression is learned as participants gain experience with the expected features and locations of salient distractors. Future research is needed to delineate the boundary conditions under which salient distractors can and cannot be ignored.

Although the findings from many studies are consistent with the notion that the P_D component measures a suppressive attentional process,⁵ a few studies have questioned this interpretation, mainly based on findings using SSVEPs and alpha-band oscillations as measures of attentional allocation, and further research is therefore needed to clarify the discrepancies between these measures and the P_D.

In conclusion, the P_D has been a powerful tool toward understanding how humans handle distraction and promises to be an important ERP component for future research on this topic. In addition to understanding the basic cognitive neuroscience of attention, the P_D component could also help provide insights in translational work in clinical science.

Reprint requests should be sent to Nicholas Gaspelin, Department of Psychological Sciences, University of Missouri, McAlester Hall, 320 Sixth Street, Columbia, MO, or via e-mail: [email protected].

This was a collaborative review. Each subsection was authored by a separate group of authors. Authorship is denoted under each subsection. Nicholas Gaspelin and Dominique Lamy provided feedback and edited all sections of the article. All authors approved of this article.

This project was made possible by National Science Foundation (https://dx.doi.org/10.13039/100000169), grant number: BCS-2045624 to N. G. and the research group on “Handling Visual Distraction” funded by the Center for Advanced Studies at Ludwig-Maximilians-Universität München.

Retrospective analysis of the citations in every article published in this journal from 2010 to 2021 reveals a persistent pattern of gender imbalance: Although the proportions of authorship teams (categorized by estimated gender identification of first author/last author) publishing in the Journal of Cognitive Neuroscience (JoCN) during this period were M(an)/M = .407, W(oman)/M = .32, M/W = .115, and W/W = .159, the comparable proportions for the articles that these authorship teams cited were M/M = .549, W/M = .257, M/W = .109, and W/W = .085 (Postle and Fulvio, JoCN, 34:1, pp. 1–3). Consequently, JoCN encourages all authors to consider gender balance explicitly when selecting which articles to cite and gives them the opportunity to report their article's gender citation balance. The authors of this article report its proportions of citations by gender category to be: M/M = .7; W/M = .167; M/W = .1; W/W = .033.