May 2016
Volume 16, Issue 7
Open Access
Article  |   May 2016
Accessory stimuli speed reaction times and reduce distraction in a target-distractor task
Author Affiliations
  • Luke Tudge
    Department of General Psychology, Humboldt Universität zu Berlin, Berlin, Germany
    Berlin School of Mind and Brain, Humboldt Universität zu Berlin, Berlin, Germany
    luke.tudge@hu-berlin.de
    sites.google.com/site/luketudge
  • Torsten Schubert
    Department of General Psychology, Humboldt Universität zu Berlin, Berlin, Germany
    Berlin School of Mind and Brain, Humboldt Universität zu Berlin, Berlin, Germany
    torsten.schubert@hu-berlin.de
Journal of Vision May 2016, Vol.16, 11. doi:10.1167/16.7.11
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to Subscribers Only
      Sign In or Create an Account ×
    • Get Citation

      Luke Tudge, Torsten Schubert; Accessory stimuli speed reaction times and reduce distraction in a target-distractor task. Journal of Vision 2016;16(7):11. doi: 10.1167/16.7.11.

      Download citation file:


      © 2017 Association for Research in Vision and Ophthalmology.

      ×
  • Supplements
Abstract

Eye movements in a visual search task are drawn towards items irrelevant to the search (distractors). Advance information about the position or features of distractors can reduce this effect, by speeding the resolution of conflict between search target and distractor. The present study investigated whether this can also be achieved by a prime that merely warns of an impending task without providing any other information (an accessory stimulus). We found that accessory stimuli speed the initiation of a saccade to the target, but also speed the resolution of target-distractor conflict. This finding suggests that the oculomotor system can be prepared to counteract distraction in advance of task onset, without requiring information about a specific spatial location or feature.

Introduction
When a new item appears in the periphery of our field of view, we can make a rapid orienting eye movement (saccade) in order to examine the item more clearly. However, the trajectories of these saccades usually do not describe a straight line from our current point of fixation to the target location (Erkelens & Sloot, 1995; Viviani, Berthoz, & Tracey, 1977). Moreover, the extent to which a saccade trajectory deviates from this straight line is systematically influenced by nearby, nonselected items, or distractors (McPeek & Keller, 2001; McPeek, Han, & Keller, 2003; Walker & McSorley, 2008). 
Distractor effects on saccade trajectories may be measured in a simple visual search task. A participant views a display and is given the task of looking at a particular target, for example a cross, whenever it appears. Simultaneously with the appearance of the target, a distractor appears, for example a circle, which is not relevant to the task. Saccade deviation may be measured as the extent to which the trajectory of the saccade departs from the ideal straight line between initial fixation and target position, and instead moves closer to the distractor. 
The extent to which a saccade deviates towards a distractor is negatively associated with saccade latency; saccades executed shortly after stimulus onset deviate the most, and later saccades show less deviation. Eventually, under certain conditions, later saccades may show no deviation at all or even deviate away from the distractor (McSorley, Haggard, & Walker, 2004; McSorley, Haggard, & Walker, 2006; Theeuwes & Godijn, 2004). This phenomenon has been attributed to the action of two processes on a spatial map that encodes the motor plan for a saccade. First, a fast, stimulus-driven process activates this map at both target and distractor locations (Aizawa & Wurtz, 1998; McPeek & Keller, 2001; McPeek et al., 2003; Port & Wurtz, 2003; Quaia, Aizawa, Optican, & Wurtz, 1998; Sheliga, Riggio, & Rizzolatti, 1994). Second, a slower, task-sensitive process inhibits activity at the irrelevant location (McSorley et al., 2004; McSorley, Haggard, & Walker, 2005; Tipper, Howard, & Paul, 2001). Because saccade trajectory is based on the spatial average of activity in the map at the time of saccade execution (McPeek & Keller, 2001; McPeek et al., 2003; Port & Wurtz, 2003; Tipper et al., 2001), saccades executed before distractor inhibition is complete will show more deviation towards the distractor. 
Saccade deviation can thus serve as an index of the visual system's success in resolving the conflict between the saccade target and an irrelevant stimulus. If conflict has not been resolved, activity at the location of the distractor will remain high, and the saccade will deviate more towards the distractor. If conflict has been resolved, this distractor activity will have been suppressed, and will have less influence on the trajectory of the saccade. Lower deviation towards the distractor therefore indicates better resolution of target-distractor conflict. 
Different factors are known to improve resolution of target-distractor conflict. When information about the location of the impending target is given, for example in the form of a cue preceding the task, saccade deviation towards the distractor is reduced, and becomes deviation away, which suggests that spatial attention to the target location allows other locations to be inhibited more effectively (Al-Aidroos & Pratt, 2010; Walker, McSorley, & Haggard, 2006). A similar effect has been demonstrated for information about the location of the distractor. If a cue indicates that a distractor will probably appear at a certain location, saccades deviate away from that location, even if no distractor actually appears, which suggests that the process of spatial inhibition can begin before the onset of distraction (Van der Stigchel & Theeuwes, 2006). If the target and the distractor are of different colors, deviation towards the distractor is reduced compared to when it is the same color as the target (Al-Aidroos & Pratt, 2010; Ludwig & Gilchrist, 2003; Mulckhuyse, Van der Stigchel, & Theeuwes, 2009). This finding has been taken as evidence of the deployment of an attentional control set (Folk, Remington, & Johnston, 1992; Folk, Remington, & Wright, 1994) to inhibit items that do not match target features. Finally, in experiments involving both distractor-present and distractor-absent trials, saccades are more likely to land on target if the participant knows in advance that the trial will involve a distractor (Moher, Abrams, Egeth, Yantis, & Stuphorn, 2011). 
Taken together, these findings suggest that top-down advance preparation for the task is an important factor in the resolution of target-distractor conflict (see Van der Stigchel, 2010, for a review). However, in all of these studies, advance preparation involved information about the spatial characteristics or other visual features of the stimuli in the impending task. It may also be the case that simple task readiness plays a role in reducing distraction, even in the absence of any prior information about the spatial or other characteristics of the task. Studies in which a warning signal, also termed an “accessory stimulus,” notifies participants of an impending task have found that this signal leads to faster responses on the task, an effect that has been attributed to enhanced readiness for the task (Niemi & Näätänen, 1981). In addition to speeding reaction times, an accessory stimulus can enhance the benefit of a nonconscious visual prime, suggesting a general enhancement of visual attention (Fischer, Schubert, & Liepelt, 2007; Schubert, Palazova, & Hutt, 2013). Accessory stimuli have also been found to enhance performance in a binary button-press choice task (Bernstein, Rose, & Ashe, 1970). 
A choice between two manual motor programs, such as left and right button presses, is conceptually similar to the selection required in a target-distractor saccade task. In both cases, one of two competing responses must be selected, and the other discarded or inhibited. It is therefore reasonable to expect that the task readiness prompted by accessory stimuli will speed saccadic reactions as it does manual reactions. 
However, the likely effect of accessory stimuli on saccade deviations towards the distractor is less clear. On the one hand, the evidence from the eye movement studies cited above suggests that the resolution of target-distractor conflict may benefit from advance preparation, although, as noted, so far this has only been shown for situations in which the preparation involved spatial or other feature information about the task. On the other hand, several arguments give reason to question whether this would be the case. 
First, it may be that, in the absence of any advance information about the specific location of target or distractor, accessory stimuli will serve to enhance bottom-up visual capture by both target and distractor, resulting in no net change in saccade deviations. Evidence for this hypothesis comes from experiments with accessory stimuli in the Simon task. In this task, participants must respond to a particular stimulus, for example an X, by pressing a left key, and to a different stimulus, for example an O, by pressing a right key. When a stimulus is presented on a side of the visual field opposite to its allocated response (incongruent trial), reaction times are slower than when it is presented on the same side (congruent trial). An accessory stimulus speeds responses in general, but also increases the slowing effect of incongruence (Böckler, Alpay, & Stürmer, 2011; Fischer, Plessow, & Kiesel, 2010; Soutschek, Müller, & Schubert, 2012). This finding has been taken as evidence that an accessory stimulus enhances automatic, spatially-mapped stimulus responses, but does not enhance the resolution of conflicts among responses. This leads to quicker reactions but also to more conflict when an automatic spatial mapping contradicts the desired response (Fischer, Plessow, & Kiesel, 2012). 
The target and distractor in the saccade task described above may also elicit two competing, spatially-mapped responses, akin to the button presses in the Simon task. Therefore, if accessory stimuli act on the oculomotor system in the target-distractor task in the same way as they do on other motor responses in the Simon task, they should speed saccadic responses but not necessarily enhance conflict resolution. As noted above, deviation towards a distractor is negatively associated with saccade latency. Therefore, accessory stimuli may cause saccades to be executed before conflict resolution is complete, leading to a decrease in saccade latency but an increase in deviation towards the distractor. We have termed this hypothesis the response execution hypothesis (left panel in Figure 1, below), as it posits speeded execution of saccades, but no effect on conflict resolution. 
Figure 1
 
Hypotheses for the effect of accessory stimulus on saccade latency and saccade deviation towards the distractor. Plots show a schematic of the negative relationship between latency (x axis) and deviation (y axis). According to the response execution hypothesis, an accessory stimulus affects only the speed of the saccades, but not the resolution of target-distractor conflict. Therefore, the solid line representing the latency distribution of saccades in accessory stimulus absent condition shifts to the left in the accessory stimulus present condition (dotted line). However, a decrease in saccade latency should be accompanied by an increase in saccade deviation. Together, this would result in decreased latencies but increased deviation (a shift towards the top left of the latency/deviation space). In the conflict resolution hypothesis, accessory stimuli can speed both saccade execution and conflict resolution. The result is decreased latencies but no change in deviation (a shift towards the left of the latency/deviation space).
Figure 1
 
Hypotheses for the effect of accessory stimulus on saccade latency and saccade deviation towards the distractor. Plots show a schematic of the negative relationship between latency (x axis) and deviation (y axis). According to the response execution hypothesis, an accessory stimulus affects only the speed of the saccades, but not the resolution of target-distractor conflict. Therefore, the solid line representing the latency distribution of saccades in accessory stimulus absent condition shifts to the left in the accessory stimulus present condition (dotted line). However, a decrease in saccade latency should be accompanied by an increase in saccade deviation. Together, this would result in decreased latencies but increased deviation (a shift towards the top left of the latency/deviation space). In the conflict resolution hypothesis, accessory stimuli can speed both saccade execution and conflict resolution. The result is decreased latencies but no change in deviation (a shift towards the left of the latency/deviation space).
Alternatively, it may be that an accessory stimulus not only speeds the execution of a saccadic response, but also improves conflict resolution, as occurs following spatial cues (Al-Aidroos & Pratt, 2010; Van der Stigchel & Theeuwes, 2006; Walker et al., 2006). If this is the case, then the prediction for saccade latencies would be the same as for the response execution hypothesis, namely that latencies decrease following an accessory stimulus. However, if conflict resolution is also improved, then the amount of saccade deviation should remain the same, despite the fact that saccades are occurring earlier; more efficient target-distractor conflict resolution “keeps pace” with speeded processing of the distractor (right panel in Figure 1). We have termed this the conflict resolution hypothesis, as it posits not only speeded execution of saccades, but also improved resolution of the conflict between target and distractor. 
It is important to note here that the conflict resolution hypothesis involves an enhancement of the resolution of target-distractor conflict that nonetheless does not necessarily result in a decrease in saccade deviation. Rather, it results in a decrease in saccade deviation relative to what is expected given the latency of the saccade. A test of the conflict resolution hypothesis therefore requires first controlling for the overall negative relationship of saccade deviation to latency. If accessory stimuli reduce saccade deviation once this control has been applied, then this result constitutes evidence in favor of the conflict resolution hypothesis. 
A further consideration is important when investigating the effects of an accessory stimulus or other cue. Cues are often more effective at reducing reaction times when they are presented in homogeneous blocks rather than in blocks that mix both cued and uncued trials (Los, 1996; Thiel, Zilles, & Fink, 2004). In homogeneous cued blocks, participants are sure that the cue will appear, and can therefore better prepare themselves to make optimum use of it. In mixed blocks, they are unsure whether the next stimulus they must process will be the cue or the task, and are therefore not always optimally prepared to make use of the cue even when it does appear. 
In order to investigate the effect of accessory stimuli on saccade responses, but independently of spatial or feature information, we gave participants a target-distractor saccade task (McSorley et al., 2006), both with and without a preceding auditory accessory stimulus. Additionally, in order to investigate a possible moderating role of cue expectation, we presented accessory stimuli either in homogeneous blocks or mixed together with nonaccessory trials. 
Methods
Participants
Thirty-four participants, aged 18 to 35 (mean age 22), 32 female and 2 male, were recruited from the participant pool at the Institute of Psychology at the Humboldt-Universität in Berlin. Participants were informed about the details of the experiment in advance and gave their written consent to participate. As compensation they received participation credit towards the completion of their undergraduate degree in psychology. 
Apparatus
The task display was programmed using MatLab with the Psychophysics Toolbox, and shown on a Samsung SyncMaster 2233 monitor with a refresh rate of 60 Hz using the default manufacturer settings for brightness and contrast. Eye movements were recorded from the left eye only, using an SMI iView X Hi-Speed system with a sampling rate of 1250 Hz. The experiment was conducted in a blinded room with a diffuse, dim light source. The participant was seated at a desk facing the display monitor at a distance of approximately 70 cm, with chin resting on the eye tracking system's built-in chin rest. The eye tracking system was controlled from a separate PC at the experimenter's desk nearby. 
Stimuli
Participants completed a simple saccade task in which they were instructed to fixate a central point, marked by a plus sign approximately 0.5 degrees of visual angle (dva) in width, then look at a target object, an X approximately 1 dva in width, as soon as it appeared. This target could appear at any one of the four oblique positions, 45, 135, 225 and 315 angular degrees from vertical, and approximately 10 dva from central fixation. An irrelevant distractor, a circle approximately 1 dva in diameter, appeared simultaneously with the target and at the same distance from fixation but 45 angular degrees either clockwise or anticlockwise from the target. These stimuli replicate approximately those used by McSorley et al. (2006). 
On some trials (AS-present trials), the appearance of target and distractor was preceded by the accessory stimulus, a 650-Hz tone of 200-ms duration, played over speakers placed either side of the monitor. In AS-present trials, the fixation cross was displayed for 2400 ms plus a random interval drawn from an exponential distribution with a mean of 800 ms. Then the accessory stimulus was presented, followed by a further 600 ms before the fixation cross was replaced by target and distractor. In other trials (AS-absent trials), the fixation cross was displayed for 3000 ms plus the random interval, and then replaced by target and distractor. In both trial types, target and distractor were displayed for 1000 ms; then the screen was blanked for 1000 ms before the beginning of the next trial. See Figure 2 for a schematic of the trial procedure. 
Figure 2
 
Illustration of trial procedure and event timings. Stimuli are not shown to scale.
Figure 2
 
Illustration of trial procedure and event timings. Stimuli are not shown to scale.
The background color of the screen was a midgray of approximately 24 cd/m2 luminance. All stimuli were maximum white of approximately 110 cd/m2 luminance. 
Procedure
Each participant completed 384 trials, in eight blocks of 48 trials each. Each block began with a 13-point calibration. The calibration was then validated, and repeated if validation failed. In order to speed up data collection and obtain enough trials for analysis, no drift correction was performed during the block (but drift was corrected posthoc; see Data processing below). After each block, the participant was given time to rest before continuing. 
Blocks were of two types. In homogeneous blocks, all trials were of the same type, either no-AS or AS. In mixed blocks, the two trial types were mixed together in random order. Each block consisted of six repetitions of the random orthogonal permutation of four target positions (upper left, upper right, lower left, lower right) × 2 distractor positions (clockwise, counterclockwise). The order of trial types was fully randomized within each block. Participants completed four homogeneous blocks (two each of AS-absent and AS-present) and four mixed blocks, all block types mixed together in random order. 
The experiment began with 10 trials of practice for the basic saccade task without an accessory stimulus. At the beginning of each block participants were reminded of the task, and depending on the block type they were told whether they would always, never, or sometimes hear a tone warning them of task onset. 
Data processing
All gaze samples falling outside the dimensions of the stimulus monitor were discarded. Gaze samples that did fall within the dimensions of the monitor were smoothed, in order to average out small-scale sampling noise. This was achieved by replacing the x and y coordinates of each sample with the mean of coordinates from all samples within 5 ms of the current sample (i.e., smoothing with a “rectangular sliding window”). 
For each trial, gaze samples were recentered on the fixation spot to correct for drift in the eye-tracking system. This was accomplished by assuming that the participant was fixating the fixation spot as instructed during the 60 ms prior to the onset of the task display. The median gaze position during this time window was then assumed to be the center of the screen, and all samples for the trial were recentered on this point by rigid body translation. 
To extract the first saccade from the processed samples, we used a “velocity peak method” (e.g., Smeets & Hooge, 2003). This method avoids erroneously categorizing small fluctuations in gaze velocity as saccades, as may occur with a fixed saccade velocity criterion (Nyström & Holmqvist, 2010). The first velocity peak was identified as the first set of contiguous samples with a velocity greater than 100°/s. The start and end points of the saccade were identified by searching from this peak backwards and forwards in time respectively until finding a sample with a velocity below 35°/s and an acceleration below 0°/s2
Trials were excluded from further analysis if saccade latency was less than 80 ms or greater than 600 ms, or if saccade landing point was more than 22.5 angular degrees either side of the target. 
Saccade deviation was measured as the mean clockwise angular deviation of sample points from a straight line drawn from fixation to target once the drift correction described above had been performed. Saccade deviation was not measured relative to saccade landing position, as is the case in some approaches (see Van der Stigchel, Meeter, & Theeuwes, 2006, for some examples). If on a given trial the distractor was counterclockwise of the target, the sign of deviation values was reversed, so that positive values indicate deviation towards the distractor and negative values indicate deviation away. See Ludwig and Gilchrist (2002) and Van der Stigchel et al. (2006), for reviews of saccade deviation and other measures of saccade trajectory deviation. 
Results
Saccade latency
To assess the effect of accessory stimuli on saccade latency, we calculated a mixed-effects linear model, with latency as the dependent variable, participant as a random effect, and accessory stimulus (absent or present) and block type (homogeneous or mixed) as fixed effects. 
Latencies were faster in AS-present trials (198 ms) compared to AS-absent trials (223 ms), F(1, 33) = 128.13, p < 0.01, η2G = 0.16. Latencies were also faster in homogeneous blocks (209 ms) than in mixed blocks (212 ms), F(1, 33) = 6.80, p = 0.01, η2G < 0.01. The effect of accessory stimuli on latency also interacted with block type, F(1, 33) = 10.92, p < 0.01, η2G < 0.01. Specifically, when an accessory stimulus was presented, latencies were equally fast in homogeneous (197 ms) and mixed blocks (198 ms), t(33) = 0.28, p = 0.78. However, with no accessory stimulus, latencies were faster in homogeneous (220 ms) than in mixed blocks (226 ms), t(33) = 4.12, p < 0.01 (see Figure 3). 
Figure 3
 
Interaction plot of saccade latency for accessory stimulus (x axis) by block type (separate lines). Error bars show ±1 standard error of the mean (SEM).
Figure 3
 
Interaction plot of saccade latency for accessory stimulus (x axis) by block type (separate lines). Error bars show ±1 standard error of the mean (SEM).
In order to assess whether the increase in latency observed for AS-absent trials in mixed blocks might be due to an expectation of an accessory stimulus carried-over from the previous trial, we carried out an additional analysis of the mixed blocks only, with accessory stimulus from the current and from the present trial as explanatory factors. In addition to the main effect of accessory stimuli in the current trial (as noted above in the main analysis), we found an interaction of accessory stimulus in the current trial and accessory stimulus in the previous trial on latencies, F(1, 33) = 25.93, p < 0.01, η2G = 0.01. Specifically, when an accessory stimulus was present in the current trial, latencies were slightly faster when the previous trial had included an accessory stimulus (196 ms) than when it had not (200 ms), t(33) = 3.25, p < 0.01, whereas when there was no accessory stimulus in the current trial, the pattern was the opposite; latencies were slower when there had been an accessory stimulus in the previous trial (232 ms) than when there had not (221 ms), t(33) = 5.17, p < 0.01. This suggests that the effects of an accessory stimulus (or its absence) carry over into the next trial, speeding processing if the next trial is of the same type 
Saccade deviation
We calculated the same mixed-effects linear model for saccade deviation as for latency. We did not find any significant effect of accessory stimulus F(1, 33) = 0.78, p = 0.38, or of block type F(1, 33) = 2.29, p = 0.14, or an interaction of accessory stimulus with block type F(1, 33) = 0.46, p = 0.50. 
As noted in the Introduction, it is possible that the absence of an effect of accessory stimuli on saccade deviation, coupled with a decrease in latency, is actually an indication of enhanced resistance to distraction (conflict resolution hypothesis, Figure 1). To examine the effect of accessory stimuli on saccade deviation while controlling for the effect of latency, we conducted a further analysis that corrects for the negative correlation of latency and deviation. 
First, to check that saccade deviation was indeed negatively correlated with latency in the present data, we divided each participant's data into four bins by saccade latency, and plotted the mean deviations across participants for each latency bin (e.g., van Zoest, Donk, & Van der Stigchel, 2012), separately for AS-absent and AS-present trials, and separately for homogeneous and mixed blocks (Figure 4). The clear negative slope of this relationship is reflected in a significant effect of latency bin on deviation, F(3, 99) = 43.93, p < 0.01, η2G = 0.15 (Greenhouse-Geisser corrected). 
Figure 4
 
Mean latencies from four latency bins, plotted against mean deviation within the same latency bin, separately for AS-absent and AS-present trials, and for homogeneous and mixed blocks. Error bars show ±1 SEM.
Figure 4
 
Mean latencies from four latency bins, plotted against mean deviation within the same latency bin, separately for AS-absent and AS-present trials, and for homogeneous and mixed blocks. Error bars show ±1 SEM.
Overall, the results appear to conform qualitatively to the conflict resolution hypothesis, as Figure 4 shows a leftward shift in the latency-deviation curve for the AS-present condition, indicating that when participants receive an accessory stimulus, they achieve the same level of deviation towards the distractor, but at faster latencies (compare Figure 1). If this is really the case, then saccade deviation should be significantly reduced in the AS-present condition once the overall effect of latency has been controlled for. 
In order to test this conclusion quantitatively, we controlled for the linear effect of latency on saccade deviation by calculating a linear regression of deviation on latency separately for each participant using all trials, and subtracting the predicted deviation values from the real values to give a residual deviation score. As a result we can compare the residual deviation values, controlled for differences in saccade latencies, between the different conditions of accessory stimulus present and absent. 
We calculated the same mixed-effects model as before, but with residual deviation as the dependent variable instead of deviation. Now, residual deviation was lower in AS-present trials (− 0.54°) than in AS-absent trials (0.58°), F(1, 33) = 21.20, p < 0.01, η2G = 0.21 (see Figure 5). This reflects the fact that although saccade deviation towards the distractor is approximately the same with or without an accessory stimulus (3.84° and 3.60°, respectively), these same deviations occur at markedly faster latencies with accessory stimuli (198 ms) than without (223 ms). Viewed in another way, when cued by an accessory stimulus, participants achieve at shorter reaction times a level of saccade deviation that they only achieve at longer reaction times without the accessory stimulus. 
Figure 5
 
Residual saccade deviation for AS-absent and AS-present trials, in homogeneous and mixed blocks.
Figure 5
 
Residual saccade deviation for AS-absent and AS-present trials, in homogeneous and mixed blocks.
Residual deviation was also lower in homogeneous (− 0.22°) than in mixed blocks (0.20°), F(1, 33) = 5.36, p = 0.03, η2G = 0.04. There was no interaction of accessory stimulus with block type, F(1, 33) = 2.01, p = 0.17. 
These results of course depend on the appropriateness of a linear function in controlling for the effect of latency on saccade deviation. A possible alternative would be a quadratic function, which would, for example, allow for changes in deviation to proceed more steeply at faster latencies.1 In order to allow for this possibility, we conducted the same analysis of residual saccade deviation as described above, but using the residual values from a quadratic model instead of a linear one. This analysis yielded almost exactly the same results, including the effects of accessory stimulus, F(1, 33) = 25.16, p < 0.01, η2G = 0.24, and block type, F(1, 33) = 5.19, p = 0.03, η2G = 0.04, and the nonsignificant interaction, F(1, 33) = 2.30, p = 0.14. 
Latency-deviation trade-off
Visually, the results also suggest that the relationship between saccade latency and deviation might be different when an accessory stimulus is present. Specifically, the decline in deviation with increasing latency appears to progress more quickly when an accessory stimulus is present (the steeper slope of the AS-present line in Figure 4). To test this possibility, we returned to the linear models of latency and deviation as described above, this time calculating separate models for the AS-present and AS-absent conditions, separately in homogeneous and in mixed blocks. From each model we took the slope term as a quantification of the steepness with which deviation decreases with increasing latency (i.e., with more negative values indicating a steeper decline in deviation). Slopes were steeper for AS-present trials (−0.08°/ms) than for AS-absent trials (−0.05°/ms), F(1, 33) = 4.93, p = 0.03, η2G = 0.04. There was no effect of block type, F(1, 33) = 0.05, p = 0.83, and no interaction of accessory stimulus with block type, F(1, 33) = 0.40, p = 0.53. 
Discussion
We found that the presence of an accessory stimulus that warns of an impending saccade task improves participants' performance on that task in two different ways. First, the accessory stimulus speeds saccade latency, so that participants initiate a saccade to the target more quickly. Second, this decrease in latency does not lead to an increase in deviation towards the distractor, as would be expected given the negative relationship between the two (McSorley et al., 2004, 2006; Theeuwes & Godijn, 2004). Instead, participants achieve the same level of deviation towards the distractor despite responding earlier. This pattern of results matches the predictions of the conflict resolution hypothesis (see Figure 1), by which a possible increase in saccade deviation caused by earlier execution of saccadic responses is eliminated by earlier resolution of target-distractor conflict. 
If deviation is an indication of stimulus-driven activity that has not yet been inhibited (McSorley et al., 2006; Van der Stigchel, 2010), then the findings suggest that both saccade execution and the process of resolving the conflict between target-related and distractor-related activity are initiated earlier when information about an impending task is available. An important novelty of the present study is the demonstration that this information need not be about the spatial locations (Al-Aidroos & Pratt, 2010; Van der Stigchel & Theeuwes, 2006; Walker et al., 2006) or features (Al-Aidroos & Pratt, 2010; Ludwig & Gilchrist, 2003; Mulckhuyse et al., 2009) of task items; the accessory stimuli are purely temporal cues. 
The effect of accessory stimuli on saccade latencies is consonant with findings from nonsaccade reaction times (Bernstein et al., 1970; Niemi & Näätänen, 1981). However, the results for saccade deviation suggest that the effect of accessory stimuli on conflict resolution in the oculomotor system is different from their effect on conflict between competing manual responses. Whereas accessory stimuli have been found to increase conflict between automatic and desired manual responses (Fischer et al., 2010; Soutschek et al., 2012), our results provide evidence that accessory stimuli improve the resolution of conflict between target and distractor in the oculomotor system. This difference between eye movements and manual responses is in line with the hypothesis that inhibition is more potent for eye movements than for hand movements (Van der Stigchel, Meeter, & Theeuwes, 2007). 
We conclude from these findings that advance warning of task onset, in the form of an accessory stimulus, enhances the resolution of target-distractor conflict, by engaging task-relevant top-down processes that either inhibit activity related to the distractor, promote activity related to the target, or both. An alternative explanation of the findings might be that the accessory stimulus has in fact no direct effect on these “executive” processes, and instead merely speeds their initiation by speeding the completion of an earlier processing stage, for example perceptual discrimination of the target and distractor, that is not itself related to the conflict resolution process, but is a prerequisite to it (the perceptual discrimination hypothesis). 
However, the data from the present study present a problem for this alternative explanation. We found that the negative slope of the saccade deviation/latency relationship was steeper for AS-present trials. Saccade deviation on any given trial represents a “snapshot” of the state of target-distractor competition at the time the saccade was executed. Saccades are executed at different times in different trials. The slope of the saccade deviation /latency relationship over all trials therefore represents the composite of these differently timed snapshots. This composite is the overall speed of the resolution of target-distractor competition once initiated (see Campbell, Al-Aidroos, Pratt, & Hasher, 2009, for an example of deviation/latency slopes used to infer the speed of oculomotor inhibition in this way). Applied to the results of the present study, this interpretation of the deviation/latency slope means that resolution of target-distractor conflict proceeded more efficiently in the condition with AS than in the condition without AS. 
To make this a little clearer, consider what should normally occur in an AS-absent trial. Target and distractor appear, they each generate activity in the oculomotor system, they are then perceptually discriminated from one another, and finally the distractor's activity is gradually inhibited (or the target's activity enhanced) over the course of a few hundred milliseconds. The slope relating saccade deviation to latency reflects the speed of this final stage, as explained above. Now consider what the perceptual discrimination hypothesis predicts for an AS-absent trial. The perceptual discrimination stage should be shortened, and as a consequence the subsequent stage is brought forward in time. This results in the same “leftward shift” of the deviation/latency curve that the conflict resolution hypothesis predicts (see Figure 1). Crucially, however, the perceptual discrimination hypothesis predicts nothing about the speed of the final stage, only how quickly it is initiated. To also speed target-distractor conflict resolution, and thus produce the steepening of deviation/latency slopes observed in the present study, it must be the case that the accessory stimulus enhances target-distractor conflict resolution, as the conflict resolution hypothesis asserts. This may be achieved, for example, by prompting the inhibitory signal that is applied to the distractor to gather intensity before it is applied, resulting in a faster elimination of the distractor signal once inhibition is engaged. It may of course also be the case that accessory stimuli speed perceptual discrimination, but this alone is not sufficient to explain the present results. 
The credibility of the findings is enhanced by the fact that they hold true for two different types of block in which accessory stimuli are presented, in homogeneous or in mixed blocks. This means that there is no evidence that the effect of the accessory stimulus on saccade deviation is due to a strategic expectation that an accessory stimulus will occur (Fischer et al., 2007). However, the effect of the accessory stimulus on saccade latency was not entirely independent of the presentation context. The pattern of results suggests that the absence of an accessory stimulus in a context in which it is sometimes present carries a reaction time penalty, specifically when an accessory stimulus was experienced recently (i.e., in the preceding trial). A similar phenomenon has been observed for reaching movements (Song & Nakayama, 2007). 
Jaffard, Benraiss, Longcamp, Velay, and Boulinguez's (2007) account of “mixing costs” could explain this finding. It may be that the accessory stimulus itself captures attention, as has been demonstrated for other auditory cues (Dalton & Lavie, 2004). This automatic attentional capture is inhibited using some of the same resources that are subsequently used in the saccade task. In homogeneous non-AS blocks, no resources are required for inhibition of the accessory stimulus, so maximal resources are available for the saccade task. In non-AS trials in mixed blocks, however, participants are unsure whether it will be necessary to inhibit attention to the accessory stimulus, and may therefore hold some resources in reserve until the onset of the saccade task, delaying initiation of the task response while resources are freed. 
This account could also explain why the increased latencies in mixed blocks are not accompanied by a decrease in saccade deviation according to the usual negative association (e.g., McSorley et al., 2006). The mixing costs incurred in mixed blocks may tax resources that are necessary for resolving target-distractor conflict, in which case these resources are brought online later in mixed blocks, resulting in no net change in deviation when saccade execution is also delayed. This is in fact merely a reverse formulation of the conflict resolution hypothesis; whereas accessory stimuli speed both saccade execution and conflict resolution, the mixing costs delay both.2 Consistent with this account, the analysis of residual deviation showed that deviation in mixed blocks was actually greater than expected once the linear association with latency was accounted for. 
An open question is why certain other task-onset cues used in saccade tasks do not produce an enhancement of conflict resolution like that seen in the present study. Most notably, the fixation gap paradigm (Ross & Ross, 1980; Saslow, 1967) involves the disappearance of the central fixation point shortly before or after task onset. When this occurs before task onset, it could conceivably be used by participants as a warning that the task is imminent. Fixation gaps can reliably be used to introduce differences in saccade latencies (e.g., Campbell et al., 2009; van Zoest et al., 2012), but these differences also entail differences in saccade trajectory deviation towards the distractor (greater deviation for faster latencies, less deviation for slower latencies), unlike the accessory stimuli in the present study, which modulated latency without affecting deviation. 
There are three important differences between the accessory stimulus used in the present study and the fixation offset cue. The first is that the fixation gap is a visual, not an auditory, cue and therefore may interfere with the visual processing of the target and distractor in a way that an auditory accessory stimulus does not. Second, fixation gap cues are typically placed very close in time to task onset (<200 ms, lower than the 250 ms that Fischer et al., 2007, found to be optimal for enhancing attention in their task). This may therefore not allow sufficient time for task readiness to accumulate to a useful level. Third, the duration between fixation offset cues and task onset is typically varied at random over several different possible values, whereas in the present study the accessory stimulus was a completely reliable indicator of task onset time. If this complete temporal reliability is necessary for the enhancement of conflict resolution found in the present study, then this would explain why fixation offset cues do not produce a similar effect. Other studies have found that different time intervals between accessory stimulus and task lead to different levels of performance enhancement (Bernstein et al., 1970; Fischer et al., 2007; Niemi & Näätänen, 1981), suggesting that preparedness takes some time to accumulate following a cue (Karlin, 1966). Further studies could therefore investigate whether the enhancement of target-distractor conflict resolution found in the present study also accumulates to an optimal level following the accessory stimulus. 
In summary, forewarning of an impending saccade task appears to increase the speed of resolution of target-distractor conflict, to an extent that keeps pace with a faster initiation of eye movement responses. This enhancement of conflict resolution does not require any information about target or distractor. 
Acknowledgments
Thanks to Antonia Papadakis for her help with data collection. Luke Tudge is supported by the Berlin School of Mind and Brain PhD scholarship. 
Commercial relationships: none. 
Corresponding author: Luke Tudge. 
Email: luke.tudge@hu-berlin.de. 
Address: Department of General Psychology, Humboldt Universität zu Berlin, Berlin, Germany. 
References
Aizawa H., Wurtz R. (1998). Reversible inactivation of monkey superior colliculus. I. Curvature of saccadic trajectory. Journal of Neurophysiology, 79 (4), 2082–2096.
Al-Aidroos N., Pratt J. (2010). Top-down control in time and space: Evidence from saccadic latencies and trajectories. Visual Cognition, 8 (1), 26–49.
Bernstein I., Rose R., Ashe V. (1970). Preparatory state effects in intersensory facilitation. Psychononmic Science, 19 (2), 113–114.
Böckler A., Alpay G., Stürmer B. (2011). Accessory stimuli affect the emergence of conflict, not conflict control. A Simon-task ERP study. Experimental Psychology, 58 (2), 102–109.
Campbell K., Al-Aidroos N., Pratt J., Hasher L. (2009). Repelling the young and attracting the old: Examining age-related differences in saccade trajectory deviations. Psychology and Aging, 24 (1), 163–168.
Dalton P., Lavie N. (2004). Auditory attentional capture: Effects of singleton distractor sounds. Journal of Experimental Psychology. Human Perception and Performance, 30 (1), 180–193.
Erkelens C., Sloot O. (1995). Initial directions and landing positions of binocular saccades. Vision Research, 35 (23–24), 3297–3303.
Fischer R., Plessow F., Kiesel A. (2010). Auditory warning signals affect mechanisms of response selection: Evidence from a Simon task. Experimental Psychology, 57 (2), 89–97.
Fischer R., Plessow F., Kiesel A. (2012). The effects of alerting signals in action control: Activation of S-R associations or inhibition of executive control processes? Psychological Research, 76 (3), 317–328.
Fischer R., Schubert T., Liepelt R. (2007). Accessory stimuli modulate effects of nonconscious priming. Perception & Psychophysics, 69 (1), 9–22.
Folk C., Remington R., Johnston J. (1992). Involuntary covert orienting is contingent on attentional control settings. Journal of Experimental Psychology. Human Perception and Performance, 18 (4), 1030–1044.
Folk C., Remington R., Wright J. (1994). The structure of attentional control: Contingent attentional capture by apparent motion, abrupt onset, and color. Journal of Experimental Psychology. Human Perception and Performance, 20 (2), 317–329.
Jaffard M., Benraiss A., Longcamp M., Velay J.-L., Boulinguez P. (2007). Cueing method biases in visual detection studies. Brain Research, 1179, 106–118.
Karlin L. (1966). Development of readiness to respond during short foreperiods. Journal of Experimental Psychology, 72 (4), 505–509.
Los S. (1996). On the origin of mixing costs: Exploring information processing in pure and mixed blocks of trials. Acta Psychologica, 94 (2), 145–188.
Ludwig C., Gilchrist I. (2002). Measuring saccade curvature: A curve-fitting approach. Behavior Research Methods, Instruments, & Computers, 34 (4), 618–624.
Ludwig C., Gilchrist I. (2003). Target similarity affects saccade curvature away from irrelevant onsets. Experimental Brain Research, 152 (1), 60–69.
McPeek R., Han J., Keller E. (2003). Competition between saccade goals in the superior colliculus produces saccade curvature. Journal of Neurophysiology, 89 (5), 2577–2590.
McPeek R., Keller E. (2001). Short-term priming, concurrent processing, and saccade curvature during a target selection task in the monkey. Vision Research, 41 (6), 785–800.
McSorley E., Haggard P., Walker R. (2004). Distractor modulation of saccade trajectories: Spatial separation and symmetry effects. Experimental Brain Research, 155 (3), 320–333.
McSorley E., Haggard P., Walker R. (2005). Spatial and temporal aspects of oculomotor inhibition as revealed by saccade trajectories. Vision Research, 45 (19), 2492–2499.
McSorley E., Haggard P., Walker R. (2006). Time course of oculomotor inhibition revealed by saccade trajectory modulation. Journal of Neurophysiology, 96 (3), 1420–1424.
Moher J., Abrams J., Egeth H., Yantis S., Stuphorn V. (2011). Trial-by-trial adjustments of top-down set modulate oculomotor capture. Psychonomic Bulletin & Review, 18 (5), 897–903.
Mulckhuyse M., Van der Stigchel S., Theeuwes J. (2009). Early and late modulation of saccade deviations by target distractor similarity. Journal of Neurophysiology, 102 (3), 1451–1458.
Niemi P., Näätänen R. (1981). Foreperiod and simple reaction time. Psychological Bulletin, 89 (1), 133–162.
Nyström M., Holmqvist K. (2010). An adaptive algorithm for fixation, saccade, and glissade detection in eyetracking data. Behavior Research Methods, 42 (1), 188–204.
Port N., Wurtz R. (2003). Sequential activity of simultaneously recorded neurons in the superior colliculus during curved saccades. Journal of Neurophysiology, 90 (3), 1887–1903.
Quaia C., Aizawa H., Optican L., Wurtz R. (1998). Reversible inactivation of monkey superior colliculus. II. Maps of saccadic deficits. Journal of Neurophysiology, 79 (4), 2097–2110.
Ross L., Ross S. (1980). Saccade latency and warning signals: Stimulus onset, offset, and change as warning events. Perception & Psychophysics, 27 (3), 251–257.
Saslow M. (1967). Effects of components of displacement-step stimuli upon latency for saccadic eye movement. Journal of the Optical Society of America, 57 (8), 1024–1029.
Schubert T., Palazova M., Hutt A. (2013). The time course of temporal attention effects on nonconscious prime processing. Attention, Perception & Psychophysics, 75 (8), 1667–1686.
Sheliga B., Riggio L., Rizzolatti G. (1994). Orienting of attention and eye movements. Experimental Brain Research, 98 (3), 507–522.
Smeets J., Hooge I. (2003). Nature of variability in saccades. Journal of Neurophysiology, 90, 12–20.
Song J.-H., Nakayama K. (2007). Automatic adjustment of visuomotor readiness. Journal of Vision, 7 (5): 2, 1–9, doi:10.1167/7.5.2. [PubMed] [Article]
Soutschek A., Müller H., Schubert T. (2012). Conflict-specific effects of accessory stimuli on cognitive control in the Stroop task and the Simon task. Experimental Psychology, 60 (2), 140–147.
Theeuwes J., Godijn R. (2004). Inhibition-of-return and oculomotor interference. Vision Research, 44 (12), 1485–1492.
Thiel C., Zilles K., Fink G. (2004). Cerebral correlates of alerting, orienting and reorienting of visuospatial attention: An event-related fMRI study. Neuroimage, 21 (1), 318–328.
Tipper S., Howard L., Paul M. (2001). Reaching affects saccade trajectories. Experimental Brain Research, 136 (2), 241–249.
Van der Stigchel S. (2010). Recent advances in the study of saccade trajectory deviations. Vision Research, 50 (17), 1619–1627.
Van der Stigchel S., Meeter M., Theeuwes J. (2006). Eye movement trajectories and what they tell us. Neuroscience and Biobehavioral Reviews, 30 (5), 666–679.
Van der Stigchel S., Meeter M., Theeuwes J. (2007). Top-down influences make saccades deviate away: The case of endogenous cues. Acta Psychologica, 125 (3), 279–290.
Van der Stigchel S., Theeuwes J. (2006). Our eyes deviate away from a location where a distractor is expected to appear. Experimental Brain Research, 169 (3), 338–349.
van Zoest W., Donk M., Van der Stigchel S. (2012). Stimulus salience and the time course of saccade trajectory deviations. Journal of Vision, 12 (8): 16, 1–13, doi:10.1167/12.8.16. [PubMed] [Article]
Viviani P., Berthoz A., Tracey D. (1977). The curvature of oblique saccades. Vision Research, 17 (5), 661–664.
Walker R., McSorley E. (2008). The influence of distractors on saccade-target selection: Saccade trajectory effects. Journal of Eye Movement Research, 2 (3), 1–13.
Walker R., McSorley E., Haggard P. (2006). The control of saccade trajectories: Direction of curvature depends upon prior knowledge of target location and saccade latency. Perception & Psychophysics, 68 (1), 129–138.
Footnotes
1  We are grateful to an anonymous reviewer for this suggestion.
Footnotes
2  We are grateful to an anonymous reviewer for this suggestion.
Figure 1
 
Hypotheses for the effect of accessory stimulus on saccade latency and saccade deviation towards the distractor. Plots show a schematic of the negative relationship between latency (x axis) and deviation (y axis). According to the response execution hypothesis, an accessory stimulus affects only the speed of the saccades, but not the resolution of target-distractor conflict. Therefore, the solid line representing the latency distribution of saccades in accessory stimulus absent condition shifts to the left in the accessory stimulus present condition (dotted line). However, a decrease in saccade latency should be accompanied by an increase in saccade deviation. Together, this would result in decreased latencies but increased deviation (a shift towards the top left of the latency/deviation space). In the conflict resolution hypothesis, accessory stimuli can speed both saccade execution and conflict resolution. The result is decreased latencies but no change in deviation (a shift towards the left of the latency/deviation space).
Figure 1
 
Hypotheses for the effect of accessory stimulus on saccade latency and saccade deviation towards the distractor. Plots show a schematic of the negative relationship between latency (x axis) and deviation (y axis). According to the response execution hypothesis, an accessory stimulus affects only the speed of the saccades, but not the resolution of target-distractor conflict. Therefore, the solid line representing the latency distribution of saccades in accessory stimulus absent condition shifts to the left in the accessory stimulus present condition (dotted line). However, a decrease in saccade latency should be accompanied by an increase in saccade deviation. Together, this would result in decreased latencies but increased deviation (a shift towards the top left of the latency/deviation space). In the conflict resolution hypothesis, accessory stimuli can speed both saccade execution and conflict resolution. The result is decreased latencies but no change in deviation (a shift towards the left of the latency/deviation space).
Figure 2
 
Illustration of trial procedure and event timings. Stimuli are not shown to scale.
Figure 2
 
Illustration of trial procedure and event timings. Stimuli are not shown to scale.
Figure 3
 
Interaction plot of saccade latency for accessory stimulus (x axis) by block type (separate lines). Error bars show ±1 standard error of the mean (SEM).
Figure 3
 
Interaction plot of saccade latency for accessory stimulus (x axis) by block type (separate lines). Error bars show ±1 standard error of the mean (SEM).
Figure 4
 
Mean latencies from four latency bins, plotted against mean deviation within the same latency bin, separately for AS-absent and AS-present trials, and for homogeneous and mixed blocks. Error bars show ±1 SEM.
Figure 4
 
Mean latencies from four latency bins, plotted against mean deviation within the same latency bin, separately for AS-absent and AS-present trials, and for homogeneous and mixed blocks. Error bars show ±1 SEM.
Figure 5
 
Residual saccade deviation for AS-absent and AS-present trials, in homogeneous and mixed blocks.
Figure 5
 
Residual saccade deviation for AS-absent and AS-present trials, in homogeneous and mixed blocks.
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×