Inhibition of masked primes as revealed by saccade curvature 1 Frouke Hermens a Petroc Sumner b Robin Walker a,∗ a Department

of Psychology Royal Holloway, University of London b School

of Psychology Cardiff University

Abstract In masked priming, responses are often speeded when primes are similar to targets (‘positive compatibility effect’). However, sometimes similarity of prime and target impairs responses (‘negative compatibility effect’). A similar distinction has been found for the curvature of saccade trajectories. Here, we test whether the same inhibition processes are involved in the two phenomena, by directly comparing response times and saccade curvature within the same masked priming paradigm. Interestingly, we found a dissociation between the directions of masked priming and saccade curvature, which could indicate that multiple types of inhibition are involved in the suppression of unwanted responses. Key words: Masked priming, Masking, Eye movements, Saccade curvature, Motor inhibition

Author’s note This is a preprint of the paper. For the final version, please refer to Elsevier’s Vision Research archive (http://www.journals.elsevier.com/vision-research/). ∗ Corresponding author. Email address: [email protected] (Robin Walker). 1 Dr. Robin Walker, Department of Psychology, Royal Holloway, University of London, Egham, Surrey, UK, TW20 0EX. This work was supported by ESRC grant RES-000-22-2932. We thank Masud Husain for useful discussions that motivated this study.

Preprint submitted to Vision Research

17 January 2012

Introduction

In everyday life, we are confronted with a stream of visual inputs that needs to be processed to extract information that is relevant for our on-going behavior while avoiding responses to irrelevant stimuli. We here investigate the processes that underlie such response selection by comparing two different measures: response times and the curvature of saccade trajectories, both measured in the same masked priming paradigm. Our results suggest that multiple mechanisms might be at work in the selection of an appropriate response to incoming information. While parts of the processes underlying response inhibition are under voluntary control, a large section may occur automatically. One paradigm often used to investigate such automatic activation and inhibition of responses to incoming stimuli is masked priming. In masked priming, a prime preceding the target influences response times and error rates to the target (e.g. Leuthold & Kopp, 1998 ; Neumann & Klotz, 1994). Typically, the prime is presented for a very brief duration, and it is masked, either by a separate mask, or by the target. This masking has the consequence that if participants are asked to report the identity of the prime, they are often close to chance level. For this reason, the influences exerted by masked primes are thought to reflect automatic processing, possibly through direct perceptuo-motor links, providing a direct pathway from sensory to motor processing (e.g. Klotz & Neumann, 1999). The most common effect a prime has on the response to the target is that response times are faster and errors occur less often when the prime is identical or similar to the target, in comparison to when the prime is unrelated to the target or has features associated with the opposite response. However, under certain conditions, for example, when the time between the presentation of the prime and the target is long, the pattern reverses, and response times are slower when the prime and the target are similar or identical (for reviews see Eimer & Schlaghecken, 2003 ; Sumner, 2007). The reduction in response times with compatible primes is often referred to as straight priming (Ja´skowski & Verleger, 2007) or the positive compatibility effect (PCE), and the increase in response times with compatible primes is called inverse priming or the negative compatibility effect (NCE). Negative compatibility effects are found only if the time between the presentation of the prime and the presentation of the target is sufficiently long. Keeping the stimuli and task constant, varying the prime-to-target interstimulus interval (ISI), a shift from a positive to a negative compatibility effect is found at intervals around 80ms (Figure 1 of Bowman, Schlaghecken, & Eimer, 2006). Additionally, NCEs are more often found when the prime is just below the threshold for conscious detection (e.g. Eimer & Schlaghecken, 2002), although they can also be found for visible 2

primes (Klapp & Hinkley, 2002 ; Sumner, Tsai, Yu, & Nachev, 2006). The cause of the negative compatibility effect is highly debated. One explanation, known as the self-inhibition hypothesis, assumes that motor program activated by the prime is automatically suppressed (e.g. Bowman et al., 2006 ; Eimer & Schlaghecken, 1998 ; Klapp & Hinkley, 2002 ; Schlaghecken, Bowman, & Eimer, 2006 ; Schlaghecken & Eimer, 2006). If the target is presented at the time the prime response is still active, a PCE occurs. If the target is presented later, once the prime response has been suppressed, then a NCE is observed. Such an explanation is supported by EEG recordings, showing a triphasic pattern of preparatory motor activation following the sequence of prime, mask and target (Eimer, 1999 ; Eimer & Schlaghecken, 1998) (however, see Verleger, Ja´skowski, Aydemir, Van der Lubbe, & Groen, 2004, for an alternative explanation), in which the first phase corresponds to prime activation, the second one with prime suppression and the third one with target activation (Eimer & Schlaghecken, 2003 ; Praamstra & Seiss, 2005). An alternative explanation was suggested by Verleger and colleagues (2004) and Lleras and Enns (2004, 2005, 2006) and is known as the active mask, or object-updating hypothesis. This theory proposed that, rather than self-inhibition, interactions between the prime and the mask occur, that cause the priming to reverse. Such interactions only occur if elements of the prime are also present in the mask, which means that it cannot account for all NCEs (Ja´skowski, 2009 ; Schlaghecken & Eimer, 2006 ; Sumner & Brandwood, 2008). Finally, there is the ‘masktriggered inhibition’ hypothesis (Boy, Clarke, & Sumner, 2008 ; Ja´skowski & Verleger, 2007 ; Ja´skowski, 2008 ; Ja´skowski & Przekoracka-Krawczyk, 2005) which assumes that inhibition of the prime is caused by another stimulus that requires attention, being the mask, or the target. However, in all three theories, a response to the prime in a typical prime-mask-target sequence is initially activated followed by a relative suppression of this prime related activity. Similarly, studies of saccadic eye movements have shown that the trajectory of an eye movement can be modulated under conditions in which one of the responses has to be inhibited (Doyle & Walker, 2001 ; McSorley, Haggard, & Walker, 2004, 2005, 2006, 2009 ; Mulckhuyse, Van der Stigchel, & Theeuwes, in press ; Van der Stigchel, Meeter, & Theeuwes, 2007a, 2007b ; Van der Stigchel & Theeuwes, 2005, 2006, 2008 ; Van der Stigchel, Mulckhuyse, & Theeuwes, 2009 ; Theeuwes & Van der Stigchel, 2009 ; Van Zoest, Van der Stigchel, & Barton, 2008 ; Walker, McSorley, & Haggard, 2006). These studies have shown that saccade trajectories deviate away from an attended location (Nummenmaa & Hietanen, 2006 ; Sheliga, Riggio, & Rizzolatti, 1994 ; Sheliga, Riggio, Craighero, & Rizzolatti, 1995 ; Van der Stigchel et al., 2007b) and away from the location of an eccentric distractor (Doyle & Walker, 2001 ; McSorley et al., 2004, 2005, 2006, 2009 ; Van Zoest et al., 2008 ; Walker et al., 2006). However, in some situations trajectories tend to deviate towards the distractor (McPeek, Han, & Keller, 2003 ; McPeek & Keller, 2001 ; Walker et 3

al., 2006). Studies of the activity of cells in the monkey superior colliculus (SC) and frontal eye fields (FEFs) have demonstrated that multiple sites in these structure are active prior to curved saccades (McPeek & Keller, 2001 ; McPeek et al., 2003 ; Port & Wurtz, 2003). Specifically, distractor related activity above baseline resulted in a curvature towards the distractor, whereas activity below baseline cause deviation away from the distractor location (McPeek et al., 2003). Moreover, inhibiting activity by injecting the GABA agonist muscimol caused saccades to curve away from the inhibited site (Aizawa & Wurtz, 1998). The prevalent view is that saccade curvature arises when multiple saccade goals are activated in the brain areas in the process of saccade target selection, and that curvature away from a site is found when activity is suppressed below baseline (McSorley et al., 2006). The modulation of saccade trajectory deviation is often regarded as a proxy to the distribution of activation and inhibition in the neural structures involved in encoding potential saccade targets prior to saccade onset (McPeek, 2006). In the present study saccade trajectory deviation is used as an alternative way of probing the automatic response selection processes involved in masked priming. How this might work, is illustrated in Figure 1, in which a hypothesized map of saccade motor activity is shown, with the amount of activity shown along the vertical dimension for different positions in space (shown in the horizontal plane). If the interval between the presentation of the prime and the mask is short (top plot), activation in the map corresponding to the prime location has not yet been suppressed and the mean vector of activation (e.g. Lee, Rohrer, & Sparks, 1988) falls in between the peaks of activation representing the prime and the target location. As it is commonly assumed that saccade direction is determined by the mean vector of activation, the initial direction of the saccade will be in between the prime and the target. This initial direction error is corrected such that the eye lands on the target, with the consequence that the overall trajectory of the saccade will curve towards the direction of the prime. However, if the target is presented after a longer interval (bottom plot), the activity in the map corresponding to the prime has been suppressed below baseline, and the resulting mean vector will curve the saccade away from the prime. Note that this initial activation in the direction of the prime followed by an inhibition below baseline corresponds to the first two phases observed in the EEG data (Eimer, 1999 ; Eimer & Schlaghecken, 1998 ; Verleger et al., 2004). In this respect, measuring saccade curvature could provide a measure of oculomotor preparation, just like the phases of the EEG response are thought to reflect manual response preparation. Alternatively, saccade curvature effects and reaction time priming effects might be dissociated. Such dissociation could arise if the inhibition involved in the saccadic NCE and that involved in saccade curvature takes place at different stages of saccade programming. In ‘inhibition of return’, a phenomenon that could involve similar mechanisms as the NCE in masked priming, evidence for 4

Activation by target

Activation by prime Fixation point Initial saccade direction

Long prime-to-target ISI Activation by target Suppressed activation by prime

Initial saccade direction

Fig. 1. Illustration of the hypothesized activity in the saccade motor map (along the vertical dimension) for different positions in space (horizontal plane) for a short prime to target interval (top) and a long interval (bottom). The gray arrow indicates the initial direction of the saccade, which follows the direction of the mean activation in the map.

such dissociation was obtained (Godijn & Theeuwes, 2004 ; Theeuwes & Van der Stigchel, 2009). Inhibition of return occurs when a response is required to a previously attended stimulus. Typically, an irrelevant stimulus is presented in the periphery, or an item in the periphery is cued with a central cue, both of which participants are asked to ignore. After a delay, the target is presented either at the location of the irrelevant stimulus or at a different location. If the delay is short, responses to the same location are faster. However, for longer delays, slower response times are found when the target is presented at the same position as the ignored stimulus. IOR and masked priming may involve similar processes, because they both show a reversal of congruency effects at longer cue or prime to target intervals. Godijn and Theeuwes (2004) investigated the relation between inhibition of return and saccade curvature by comparing the time-course at which they take place. In addition, the effects of the salience of the cue and the type of cue (exogenous versus endogenous) were studied. Whereas cue salience and cue type had similar effects on response times (IOR) and saccade curvature, differences in the time course of the two effects were found. Significant saccade curvatures were only found for short cue-to-target intervals, while inhibition of return was found just for longer cue-to-target intervals. This was interpreted as evidence that saccade curvature originates from inhibition in the superior 5

colliculus saccade map, whereas inhibition of return might be caused by inhibition preceding the saccade motor map. Theeuwes and Van der Stigchel (2009) came to a similar conclusion with a slightly different paradigm, in which the cue position needed to be processed rather than to be ignored. In the current study, automatic response preparation in masked priming is investigated by determining its influence on saccadic response times as well as on the curvature of saccade trajectories. In the majority of the trials (‘regular trials’), participants were asked to make an eye movement following the direction of a centrally presented arrow. In half of these regular trials, the arrow, indicating the saccade target, was preceded by an identical prime arrow, which was rendered less visible by a pattern mask (‘congruent trials’). In the other half of the regular trials, a mirror image of the target arrow was presented, again masked by a pattern mask (‘incongruent trials’). Two different prime-to-target intervals were used (one short and one long) in order to obtain a positive and a negative compatibility effect on the response times. The effect of the prime on saccade curvature was investigated on the remainder of the trials (probe trials) in which a vertical saccade was made in the direction orthogonal to the briefly presented prime. Assuming that a positive compatibility effect is found for the short prime-to-target interstimulus interval (ISI), a curvature towards the direction indicated by the prime is expected on such trials, as activity in the saccade map corresponding to the prime location has not yet been suppressed. For the long prime-to-target ISI, a negative compatibility effect on regular trials is expected. Such a NCE is expected to result in a curvature away from the prime direction on probe trials, indicative of inhibited prime activity.

Methods

Participants

Twenty participants (four male, 16 female; age range 18 to 40 years), including two of the authors (FH and RW), took part in the experiment. For the final data analysis, the data of two participants were not included in the analyses because of the large number of trials (>50%) that had to be removed on the basis of the exclusion criteria (see ‘data analysis’). Before taking part, all participants read an information sheet, indicating that their participation was voluntary and they could withdraw from the experiment at any time, and they signed an informed consent. Participants received £6,- for their participation. The experimental procedure was approved by local ethics committee. 6

Apparatus The stimuli were presented on a 21 inch CRT screen at a refresh rate of 100Hz, controlled by an AMD Athlon 2400+ PC. Eye movements were recorded using the Eyelink II system (SR Research Osgood, ON, Canada), which was mounted on a chin-and-head-rest at a distance of 57cm from the computer screen. The horizontal and vertical eye positions for both eyes were sampled at 250Hz. Stimuli The stimuli are illustrated in Figure 2, which were similar to those used by Sumner et al. (2007). The fixation screen contained a fixation symbol (‘+’) measuring 1 degree horizontally and vertically flanked by two placeholders with a diameter of 0.6 degrees each, at a horizontal distance of 7 degrees from the center of the screen. The prime consisted either of two leftward or two rightward arrows (‘<<’ and ‘>>’ respectively), which measured 0.5 degrees horizontally each, with a separation of 0.7 degrees, or an ‘X’ symbol of 0.6 degrees in size. The mask was constructed from 30 lines each in a random orientation (although constant across trials), forming a patch extending 2 degrees horizontally and 1.8 degrees vertically. The target either was the identical to the prime, a mirror image of it (‘>>’ or ‘<<’), or a diamond shape was presented (1 degree in size both horizontally and vertically) together with a filled circle with a diameter of 0.5 degrees at a vertical distance of 8 degrees from the center of the screen. Stimuli were presented in white on a black background. Design Two possible stimulus sequences were used (Figure 2a). Either the sequence had a short prime-to-target ISI or a long prime-to-target ISI. On 48 (57%) of the 84 trials of each block, participants were required to follow the direction of the target (regular trials; Figure 2b). On these trials, the prime was identical to the target on half of the trials and pointed in the opposite direction on the other half of the trials. On the remaining 36 (43%) of the trials in each block, the target screen consisted of a centrally presented diamond shape (serving as a control for the arrow target in the regular trials) together with a filled circle presented above fixation. These trials required a vertical saccade to the filled circle (‘probe trials’). On probe trials, either a leftward arrow, a rightward arrow, or the letter ‘X’ was used as the prime, all with equal probability. Participants completed four blocks of eye movement trials in which regular and probe trials were randomly interleaved. The timing of the stimuli was 7

varied across blocks. Participants either performed one block of long ISI trials followed by two blocks of short ISI trials, followed by one more block of long ISI trials, or vice versa. This particular sequence (‘A-B-B-A’) of short and long ISIs was chosen to counteract effects of fatigue and practice. After the four blocks of eye movement trials, participants performed two additional blocks (one for the short and one for the long ISI) in which they were asked to report the identity of the prime by pressing one of three buttons. If they thought the prime was equal to ‘>>’, they pressed the ‘M’ key on the keyboard, for a ‘<<’ they used the ‘Z’ key, and for an ‘X’, they pressed the space bar. These blocks were 56 trials in length with 32 trials with either a ‘>>’ or ‘<<’ prime, and 24 trials with a neutral prime (the letter ‘X’). Procedure Before the experiment, participants were instructed that they would be presented with sequences of stimuli, and that their task would be to make eye movements following the last stimulus in the sequence, ignoring the other, briefly presented items. If the last item in the sequence was an arrow (‘>>’ or ‘<<’), they were asked to make an eye movement to the placeholder to which this arrow was pointing. If the last item was a diamond shape, a saccade towards the target above fixation was required (see Figure 2 for an illustration of the stimuli). Participants completed four blocks of 84 trials each. For each condition, the sequence started with a fixation symbol flanked by the placeholders for 500ms followed by just the placeholders for 300ms, and ended with the target screen for 1000ms in which participants were required to make their response, followed by a blank screen for 1000ms (Figure 2a). For short ISI, the prime was presented for 80ms, followed by the mask for 20ms. For long ISI, the prime was presented for 20ms, followed by a blank screen (just the placeholders) for 10ms followed by the mask for 100ms, followed by another blank screen (placeholders) for 50ms. Before each block, a calibration procedure for the eye-tracker was performed, in which participants were asked to fixate targets presented in a random sequence on a 3 by 3 grid. Calibration was repeated until all fixations were aligned to a similarly shaped grid for both eyes with the first and the last fixation (both aimed at a target at the center of the screen) at close proximity. Following the calibration a drift correction was performed in which participants fixated a central dot and the experimenter pressed the space-bar which corrected for drifts in the eye movement recordings due to small head movements after the calibration procedure. In the eye movement task, participants were asked to respond as quickly as 8

a) Stimulus sequence Fixation 500ms

>>

+

b) Target screens

Blank 300ms

>>

>>

Regular left

Regular right

Prime 80ms / 20ms Probe

Time

Blank 0ms / 10ms Mask 20ms / 100ms Blank 0ms / 50ms Target 1000ms

>>

Blank 1000ms

Fig. 2. a) Stimulus sequence, shown in reverse contrast (stimuli were presented in white on a black background). Two prime-to-target ISIs were used (short and long), of which the timing is indicated by the two numbers next to each screen. For the short ISI, the prime was presented for 80ms immediately followed by the mask for 20ms. This combination is expected to produce a PCE. For the long ISI, a briefly presented (20ms) prime was used, followed by a blank interval for 10ms and a mask presented for 100ms, followed by another blank interval of 50ms. This combination is expected to produce a NCE. b) The possible target screens. On 57% of the trials (‘regular trials’), participants were asked to make a saccade to the placeholder in the direction of the target (regular trials, left or right), in which the prime was identical to the target (‘congruent condition’) or it pointed in the opposite direction (‘incongruent condition’). On remaining 43% of the trials (‘probe trials’) a diamond was presented together with a peripheral target located 8 degrees above fixation. On these probe trials participants were required to make a vertical saccade to the peripheral target.

possible, but at the same time trying to avoid making saccades in the wrong direction. After the four blocks requiring saccadic responses, two additional blocks were performed in which participants were asked to identify the prime by pressing 9

one of three buttons on a computer keyboard. Participants were asked to be as accurate as possible, but no instruction was given about the response speed. Data analysis Although both eyes were tracked, only the results for the right eye will be shown, as analyses for the left eye yielded similar results. Saccadic eye movements were detected on the basis of a 22 deg/sec velocity and 8000 deg/sec2 acceleration criterion. Trials in which the first saccade after target onset was in the wrong direction were removed from the analysis as well as trials which the first saccade was of insufficient amplitude (<3.8 degrees). In addition, trials were removed in which the saccade was initiated less than 100ms after target onset and more than 2.5 times the standard deviation after the mean response time, as well as probe trials with a peak deviation of more than 50% of the saccade amplitude (i.e., turn-around saccades), and trials in which a blink occurred during the saccade. For two participants, these elimination criteria resulted in a large number of trials (more than 50%) to be removed compared to the other participants (in part due to problems in the data recording). Because of this large number of excluded trials, the data of these two participants were removed from the data set. For the remaining participants, the exclusion criteria led to 13.5% of the trials to be removed from the analysis. For probe trials, requiring a vertical saccade, the curvature of the saccade was computed as the peak deviation of the saccade trajectory from the straight line connecting the start and the end of the saccade, divided by the amplitude of the saccade (for an overview of saccade curvature measures, see Ludwig & Gilchrist, 2002 ; Van der Stigchel, Meeter, & Theeuwes, 2006). Saccade curvatures was compared to baseline, providing a measure of relative curvature by subtracting the mean curvature for the neutral prime (‘X’) from the mean curvature for each directional prime (‘>>’ and ‘<<’) for each participant.

Results

Response times Figure 3 compares the reaction times for short and long ISIs for the different trial types (congruent and incongruent prime and target, and probe trials). 10

For both ISIs, shorter response times are found on the probe trials (short ISI: F(1,17)=6.07, p=0.025; long ISI: F(1,17)=134.49, p<0.001; difference contrast with respect to probe trials), possibly because in these trials a new stimulus appeared (the filled circle above fixation), which could serve as a target for an exogenous saccade. In addition, these probe trials showed faster responses for long ISIs than for short ISIs (t(17)=-7.78, p<0.001). In agreement with earlier observations, trials with a short ISI showed faster response times on the congruent trials compared to incongruent trials (t(17)=-6.79, p<0.001) - the positive compatibility effect (PCE). For long ISIs this effect reverses and congruent trials now show longer response times than incongruent trials (t(17)=4.69, p<0.001) - the negative compatibility effect (NCE). Note that the response times on the congruent trials are almost identical for the short and long ISIs (t(17)=-0.85, p=0.41) , meaning that the prime to target interval only affected the incongruent trial latencies. On average, response times were shorter for the longer ISI, which is in agreement with earlier observations (e.g. Eimer & Schlaghecken, 2003) and could possibly relate to the prime acting as a warning signal.

Congruent Incongruent Probe

Response time (ms)

340

300

260

220

180

Short ISI

Long ISI

Fig. 3. Mean saccadic response times for the two different primeto-target ISIs and for the different trial types (congruent and incongruent prime and target trials, and probe trials) across 18 participants. The error bars show the standard error of the mean.

11

Saccade curvature

Figure 4 plots the curvature of the vertical saccades on probe trials for short and long prime-to-target ISIs, separately for primes pointing to the left and to the right. Negative values indicate a curvature to the left, positive values a curvature to the right. Both ISIs show a curvature away from the prime, meaning that for leftward primes, saccades curve to the right and vice versa. This finding agrees with the prediction for the long ISI in which an NCE was found, and thus curvature away from the primed location was expected. However the results go against the prediction for the short ISI condition, in which a PCE was found, and thus curvature towards the prime direction was expected. Interestingly, the largest curvature effect was observed for the short ISI condition, which is the condition in which the observed effects are in the opposite direction from the predictions. A 2x2 (ISI x prime direction) repeated measures ANOVA confirms the main effect of prime direction (F(1,17)=8.63, p=0.009) and also suggests that curvature effects may be different for short and long ISIs, showing an interaction effect that approaches significance (F(1,17)=3.28, p=0.088). Paired comparisons show a significant curvature away for the short ISI (t(17)=2.91, p=0.01) and a curvature away that approaches significance for the long ISI (t(17)=1.95, p=0.068). As we will discuss later, these smaller curvature effects for the long ISI condition are most likely to be the result of the shorter latencies of the responses.

Curvature as a function of response time

Several studies have shown that curvature varies as a function of saccadic reaction time (McSorley et al., 2006, 2009 ; Mulckhuyse et al., in press ; Van der Stigchel & Theeuwes, 2007 ; Van Zoest et al., 2008). Typically, saccades curve towards the distractor for fast responses (<200 ms), whereas they curve away for slower responses (>200 ms). This time course of saccade curvature is thought to reflect the evolution of activity in the saccade motor map at the site representing the distractor (e.g., Trappenberg, Dorris, Munoz, & Klein, 2001). Initially, the distractor is coded in the map due to exogenous (stimulus driven) input into the map, resulting in activity above baseline. If saccades are initiated at this point, a curvature towards the distractor is expected. However, this peak of activation has to be suppressed to avoid making saccades to the distractor, resulting in endogenous (top-down) input entering the system. This endogenous input causes the activity at the distractor site to drop below baseline, and if saccades are initiated at this point a curvature away from the distractor will most likely be found (e.g. Godijn & Theeuwes, 2004). 12

Leftward prime Rightward prime

Short ISI

Long ISI

−0.015

−0.01

−0.005

0 0.005 Peak Deviation

0.01

0.015

Fig. 4. Mean peak deviation (as a proportion of saccade amplitude) for the two target-to-prime ISIs (short and long) and the two prime directions (left and right). Negative values represent a curvature to the left, positive values a curvature to the right. Error bars show the standard error of the mean across participants.

Because of the link between saccade curvature and response times, it is possible that the apparent dissociation between the compatibility effect and curvature reported above may, in part, be due to differences in response times. For example, the mean PCE on response times may have arisen entirely from saccades with relatively short latencies (on regular trials), while the negative trajectory curvature may have arisen entirely from saccades with long latencies (on probe trials), and thus the two effects may not actually dissociate in saccades of similar latency. Note that such an explanation is unlikely on the basis of the mean response times, in which faster responses were found on probe trials compared to regular trials. However, to fully understand the relationship between saccade curvature and response times, it is important to look at the response time distributions and the corresponding trajectory curvatures. Figure 5 shows the latency distributions for congruent, incongruent and probe trials in the top plots, and the time course of the mean curvature in the bottom plots. For the purpose of illustration, the reaction time distribution for the probe trials is plotted downwards, so that it can more easily be compared with the curvature measures. The dotted lines (going from top and bottom) connect the center of the reaction time quartiles (top) and the corresponding mean curvature (bottom). We here focus on the data of the short ISI condition for which a dissociation between the directions of the mean compatibility effect and mean 13

Proportion of trials

Short ISI

Long ISI

0.3

0.3

0.2

0.2

0.1

0.1

0

0

−0.1

−0.1

−0.2

−0.2

100

200

300

400

500

100

Probe Congr Incongr

200

300

400

500

Response time Mean curvature in bin

Short ISI

Long ISI

0

0

−0.01

−0.01

−0.02

−0.02

−0.03

−0.03

−0.04 100

200

300

400

500

−0.04 100

200

300

400

500

RT bin center (ms)

Fig. 5. Reaction time distributions for congruent (top left) and incongruent (top right) regular trials (plotted upward) and probe trials (plotted downward). Dotted lines connect the centers of the quartiles of the probe trial distribution to the average curvature in that bin (bottom row).

curvature was found. Two things are apparent: first, it is clear that the PCE arose from a shift of the whole latency distribution between congruent and incongruent trials and was not driven by only the early saccades or by only the late saccades. Second, it is clear that the mean curvature was away from the prime for all four quartiles of the probe latency distribution. Thus the dissociation between the mean PCE and mean curvature cannot be explained by differences in the latencies of saccades driving the two effects. It is also worth noting that overall, our curvature data conforms to previous reports showing that curvature tends to get more negative with longer latencies. This pattern is apparent for the short prime-to-target ISI, and between the short and long ISIs. This also means that, presumably, if the latencies of probe trials had been equal to the other trials (in other words, if we had not included an exogenous saccade target on probe trials), then curvature would have been more negative in both prime conditions, enhancing our reported dissociation between the compatibility effect and curvature for the short ISI. 14

Effect of the previous trial

On a relatively large proportion of trials, participants were required to make a vertical saccade. Possibly, curvature effects would have been stronger if a higher proportion of regular trials would have been used, because this could possibly lead to a stronger association between the direction of the arrows (either prime or target) and leftward and rightward saccades. If such repeated stimulus-response pairings are important, recent encounters with regular trials would also be expected to lead to stronger curvature effects on probe trials compared to when no recent arrow-horizontal saccade pairing took place. To test this latter prediction, the data was split into trials on which the trial prior to the probe trial was a regular trial, requiring a response to the arrow and trials on which the previous trial did not show an arrow as the target (probe trial). Figure 6 shows the pattern of saccade trajectory deviations for a previous horizontal saccade (top left) and a previous vertical saccade (top right). This analysis shows that the previous trials did not have a strong effect on the saccade curvature, arguing against immediate effects of previous responses to arrows. In fact, if any effects are present at all, larger deviations away are obtained when the previous trial was a probe trial, suggesting that to obtain a large saccade curvature, it is actually better not to have responded to an arrow recently. It should be noted, however, that the difference between previous probe and previous regular trials was not significant (the 2x2 interaction of previous trial type and prime-to-target ISI did not reach significance: F(1,17)=0.12; p=0.73, neither did the main effect of the previous trial (F(1,17)=1.77, p=0.201). As before, however, the main effect of ISI, did reach significance (F(1,17)=4.67, p=0.045).

Effects of prime visibility

To be able to conclude that the obtained effects occur without conscious awareness, studies typically take much care to ensure that participants do not perceive the prime. For example, in their recent study of briefly presented peripheral distractors on saccade curvature, Van der Stigchel and colleagues (2009) excluded the data of those participants who were above chance at reporting the identity of the distractor, to make sure that the effects obtained were subliminal in nature. Here we take a different approach to investigate the effects of prime visibility. This is done by investigating performance in the saccade task across participants, comparing participants who were able to report the identity of the prime above chance with participants who failed at identifying the prime on most trials. Because we used two types of trials, regular and probe, each with 15

Regular previous trial

Probe previous trial Leftward Rightward

Short ISI

Short ISI

Long ISI

Long ISI

−0.02

−0.01

0 0.01 Peak deviation

0.02

−0.02

−0.01

0 0.01 Peak deviation

0.02

−3

Peak deviation away (+ve)

15

x 10

Regular previous Probe previous 10

5

0 Short Long Prime to target interval

Fig. 6. Top plots: Leftward (negative values) and rightward (positive values) curvature for leftward and rightward primes and short and long prime-to-target ISIs when the previous trial required a horizontal saccade (left; ’regular previous trial’) and when the previous trial required a vertical saccade (right; ’probe previous trial’). Bottom plot: Average curvature away (positive values) from the prime direction for short and long ISIs for the two types of previous trials. Error bars show the standard error of the mean.

a different number of primes (two for the regular trials (a leftward (‘<<’) and a rightward (‘>>’) arrow; three for the probe trials, which used an additional neutral prime (‘X’)), two percentages of correct prime identification were obtained. To compute the correlation between the measure of prime visibility and the performance measures in the saccade task, the two percentages were pooled into one mean. Figure 7 shows scatter plots of the mean performance on the prime discrimination task against the response times compatibility effect (top left) and the curvature away from the prime (top right). Because it has been suggested that the number of incorrect saccades following an endogenous cue could be an indicator of automatic response preparation (e.g. Kuhn & Benson, 2007), we also examine the relation between the measure of prime visibility and the number of incorrect saccades following the prime (bottom left). In each plot the best fitting (least squares) regression line is shown. Each dot represents one participant. No consistent correlations were found, suggesting that the effects on response times, accuracy and curvature were largely independent of prime visibility. For the response times congruency effect, a tendency towards a positive correlation 16

was found for the short ISI (r=0.44, p=0.065), meaning that participants who could report the prime were perhaps more likely to have slower response times on incongruent trials. By contrast, no such correlation was found for the long ISI. This latter result supports the finding by Schlaghecken and colleagues (2006) that an increased visibility of the prime after perceptual learning does not change the size of the negative compatibility effect. However, the absence of a correlation for the long ISI is at odds with findings by Klapp and Hinkley (2002), who found a larger NCE for participants for whom the prime was less visible. The lack of a correlation poses an example of a dissociation between a direct measure of the effect of the prime (performance on the prime discrimination task) and an indirect measure (the size of the NCE; although it should be noted that the dissociation in our study arises in a comparison between rather within participants) (see Schmidt, 2006). For the relation between the performance on the prime discrimination task and the peak deviation away, the short ISI showed a slight (not significant) tendency towards a positive correlation (r=0.268, p=0.28), possibly suggesting that the saccades of participants who could identify the prime were less likely to curve away from the primed location (see, however, Cardoso-Leite, Mamassian, & Gorea, 2009). However, no such correlation was found for the long ISI condition. The number of incorrect saccades in the direction of the prime showed a slight (not significant) tendency towards a correlation with prime identification performance for long ISIs. However, this correlation was negative (r=-0.321, p=0.195), meaning that participants who were worse at identifying the prime, may have been more likely to make a saccade in the direction of the prime, which could be an indication that the primes always resulted in an automatic response, but that participants who could identify the prime were better at suppressing it. For the short ISI condition, no such correlation between prime identification performance and incorrect saccades was found, even though there were several participants for whom the prime seemed to be well visible.

Discussion

When a prime precedes a target stimulus, performance on the target is often affected by the identity of the prime, even though discrimination performance on the prime often suggests that the prime is not visible (for an overview and meta-analysis, see Van den Bussche, Van den Noortgate, & Reynvoet, 2009). If the interval between the presentation of the prime and the target is short, performance is often better when the prime and the target are identical or similar, in comparison to when they are linked to opposite responses (‘positive compatibility effect’, PCE). However, for longer prime-to-target ISIs, the pattern of results reverses and responses are initiated more quickly if the prime and the target are linked to incompatible responses (‘negative compatibility 17

0.01 100

Peak deviation away

RT incongr − congr (ms)

150

50 0 −50

0 −0.01 −0.02 −0.03 −0.04

−100 20

40

60

80

100

20

% incorrect saccades to prime

% correct on prime

40

60

80

100

% correct on prime

40

Short ISI Long ISI

30 20 10 0 20

40

60

80

100

% correct on prime

Fig. 7. Scatter plots showing how performance on the prime correlates with the compatibility effect (top left), the peak deviation (top right), and the number of incorrect saccades towards the prime (bottom left). Each dot represents one participant. Solid lines show the best fitting (least-squares, first-order) regression line.

effect’, NCE; e.g. Eimer, 1999 ; Eimer & Schlaghecken, 1998). It has been proposed that this reversal of the compatibility effect is due to the suppression of neural activity following the prime (for reviews, see Eimer & Schlaghecken, 2003 ; Sumner, 2007). For eye movements, an apparently similar explanation has been proposed to explain why sometimes saccades may curve towards an attended location (or distractor location) and sometimes curve away. Specifically, if saccades are initiated quickly, saccades more often curve towards the distractor, whereas saccades with longer response times tend to curve away (e.g. McSorley et al., 2006). This finding was explained by assuming that for fast saccades, the exogenously induced activity in the motor map corresponding to the distractor is still present, resulting in an initial saccade direction aimed in between the distractor and the target (i.e. towards an intermediate location weighted by relative activity). For slowly initiated saccades, the activity in the map at the distractor site has been suppressed below baseline and the saccade starts in a direction away from the distractor. In the current study we investigated whether the two phenomena, the change from a PCE to a NCE on response times for longer prime-to-target ISIs, and 18

the modulation in saccade trajectory towards or away from a (cued) distractor location, are related. To this end, we presented sequences of a prime, a mask, and a target, and asked participants to make saccades following the target, while we investigated the effect of the preceding priming. In the majority of trials (’regular trials’), the target consisted of a leftward or rightward arrow, with the preceding prime either congruent or incongruent with the target, and participants made a saccade following the direction of the target. On a subset of the trials (’probe trials’), a target was presented above fixation which needed to be fixated. On these probe trials, in which a vertical saccade was made in the direction orthogonal to the primes and the targets on regular trials, the curvature of the trajectory towards the target is measured. If the two phenomena of response inhibition in masked priming and saccade curvature are linked, curvature towards the prime direction is expected for a short primeto-target ISI and curvature away for a long ISI. Contrary to this prediction, we found a curvature away for both ISIs, although response times on the regular trials showed that the expected compatibility effects on response times were obtained. This finding suggests that our hypothesis, that the compatibility effect on response times and direction of curvature could originate from the same mechanism, is not correct. We here discuss four possible reasons why this might be.

1. Response times and saccade curvature effects reside in different brain structures

One possible reason why different compatibility effects on response times and saccade curvature were found, is that the effects on the two different measures originate from inhibition in different brain structures. Although this interpretation lies contrary to our original hypothesis, it is consistent with previous literature. Curvature has been strongly associated with the SC (Aizawa & Wurtz, 1998 ; McPeek et al., 2003) and the FEFs (McPeek, 2006), while masked priming and the NCE have been associated the basal ganglia (Aron et al., 2003 ; Seiss & Praamstra, 2004), the supplementary motor area (SMA) and supplementary eye fields (Sumner et al., 2007). A similar explanation has been proposed for a related effect, called inhibition of return (IOR), which was found to follow a different time-course to that of saccade curvature (Godijn & Theeuwes, 2004 ; Theeuwes & Van der Stigchel, 2009). The inhibition involved in IOR was investigated by measuring the amount of curvature on probe trials as well as response times on regular trials, showing a dissociation of the time at which the strongest IOR on response times occurred and the interval with the strongest curvature effects. 19

For short cue to target intervals a curvature away from the cue was found, however, the curvature did not differ from baseline for long cue to target intervals. Earlier, a study by Dorris et al. (2002), found no evidence for active inhibition in the superior colliculus during the cue-to-target interval in IOR, which therefore suggests that response time effects in IOR result from competition elsewhere in the system. In contrast, curvature effects are thought to originate in brain structures involved in saccade target selection, such as the superior colliculus (Aizawa & Wurtz, 1998 ; McPeek et al., 2003), and the frontal eye fields (McPeek, 2006). Thus, a similar dissociation between brain areas involved in response times and curvature effects in IOR might underlie our findings in masked priming.

2. The PCE and NCE do not reflect motor activation and inhibition

If the masked priming effect reflects perceptual, rather than motor processes, there is no obvious reason to expect them to be related to saccade curvature. It is known that masked stimuli can prime perceptual processing of target stimuli, as well as activate motor responses (Marcel, 1983). It has also been suggested that the NCE might be explained by perceptual (or attentional) habituation (e.g. Huber, 2008 ; Sohrabi & West, 2008) rather than a suppression of sensori-motor activation. However, recent studies have ruled out perceptual explanations of the PCE and NCE in the type of masked priming paradigm employed here (Boy & Sumner, in press ; Sumner & Brandwood, 2008). Therefore it is very unlikely that this can explain the dissociation between reaction time and curvature effects in the current study. Note that the object updating hypothesis for the NCE (Lleras & Enns, 2004 ; Verleger et al., 2004) places the source of the PCE and NCE in motor activation maps, even though reversal of motor priming is caused by perceptual interactions between prime and mask, rather than motor inhibition. Thus for our purposes, the object updating theory makes the same predictions as self-inhibition (e.g. Eimer & Schlaghecken, 1998) or mask-triggered inhibition (e.g. Ja´skowski & Verleger, 2007) - if the motor maps involved in priming and saccade curvature are the same, then the PCE should be accompanied by curvature in the direction of the prime and the NCE by curvature in the opposite direction.

3. The direction of saccade curvature does not reflect inhibition

Alternatively, the direction of saccade curvature might be related to different processes other than inhibition of populations of neurons in the saccade motor map. Several factors are known to affect the direction of saccade curvature, each of which we will discuss here. First, curvature towards is more 20

often found in monkeys (e.g., McPeek et al., 2003), whereas curvature away is more dominant in humans (for an overview, see Van der Stigchel et al., 2006). It could therefore be argued that human data will almost always show a curvature away, independent of the experimental manipulations. Several studies have, however, shown that it is possible to obtain curvature towards distractors with human participants (e.g. McSorley et al., 2006 ; Walker et al., 2006). Although we might not have used the right stimulus conditions to obtain a curvature towards, these data show that, in principle, it should be possible to find saccades that curve towards the distractor direction in human participants. Second, the predictability of the target position has been shown to affect the direction of saccade curvature, with a curvature towards distrators for unpredictable target locations, and a curvature away for predictable targets (Walker et al., 2006). The difference between predictable and unpredictable targets has been attributed to inhibitory processes before stimulus onset. If the target position is known in advance, top-down inhibition can be applied to increase activity at the target location with respect to activity at the distractor location. Such differential activity before target onset could result in faster inhibition of distractor related activity after target onset when the target location is known in advance. In our experiment, the target for probe trials was always presented above fixation, however, this did not mean that the target was predictable, because probe trials were randomly intermixed with regular trials, resulting in three possible saccade locations (left, right, top) rather than one. As a consequence, the positions of both the target and the distractor were unpredictable on each trial, making it unlikely that stimulus predictability caused the curvature away for both prime to target ISIs. Third, if saccades are initiated quickly, saccades tend to curve towards distractors, whereas for longer saccade latencies, saccades curve away (McSorley et al., 2006). As our analysis in Figure 5 has shown, saccades curved away from the prime direction for all response latencies, indicating that response time was not the cause of the observed deviation away in the PCE condition.

4. Differences in the visibility of the prime resulted in differences in response inhibition

A fourth factor possibly involved in the dissociation between the influence of primes on response times and on saccade curvature is the visibility of the prime. For the short ISI (leading to a PCE) prime discrimination was on average 76.6% accurate, while for the long ISI (yielding a NCE), prime discrimination was close to chance, at 53.9%. In our experiment, the conditions yielding a PCE and an NCE differed in the duration of the prime and the mask as well as in the ISI. This was done to maximize the chances of finding a clear and robust difference between the PCE and NCE conditions, as it is known that longer primes, as well as shorter ISIs, tend to lead to PCEs rather than NCEs. 21

However, it remains possible that the mechanism underlying PCEs with short ISIs and invisible primes is different from the mechanism underlying PCEs with longer visible primes. In the latter case, the measured PCE could reflect a combination of two mechanisms, one acting on visible primes and the other only on invisible ones. Even though a PCE was measured overall, the former mechanism may actually have reversed the initial response activation into inhibition by the time responses were made. This inhibition may have affected the saccade curvature while the reaction time data might have been dominated by the strong positive priming of the ‘visible’ mechanism only, causing the effects on RTs and saccade curvature to be dissociated. It should be noted, however, that this speculation that the PCE may be produced by two different mechanisms does not seem to be supported by the literature so far. To establish the exact mechanisms underlying the relationship between PCE, NCE and the inhibition in saccade curvature, it would be of interest to determine how the response time effects and saccade curvature develop if the ISI between the prime and the target is varied systematically, while keeping the other factors, such as prime duration and mask duration, constant.

Endogenous subliminal distractors

Although several studies have shown that central, endogenous cues to a distractor location cause saccades to curve away from the attended location (Godijn & Theeuwes, 2004 ; Theeuwes & Van der Stigchel, 2009 ; Sheliga et al., 1994, 1995 ; Van der Stigchel et al., 2007b), most of these studies required attention to be directed to the cued location, for example, because the instruction as to where to move the eyes was likely to be presented at that position (e.g. Sheliga et al., 1994). To our knowledge, only one study (Nummenmaa & Hietanen, 2006) showed that saccade curvature can be obtained by central cues that had to be ignored. However, in this study, the central cue was a gaze cue, which might have a special status by directing attention in a reflexive manner (Driver et al., 1999 ; Friesen, Ristic, & Kingstone, 2004). Our priming results show that arrows, presented at fixation, and which had to be ignored, can also make saccades curve away from the primed location. Possibly, such effects were obtained because the arrows that needed to be ignored (the primes) were identical to the possible targets, which, on most of the trials, required a response. If such task set effects exist, they must operate across the entire session instead of being the consequence of recent mappings of a prime-like stimulus onto a response in a particular direction. This is because our analysis of inter-trial effects (Figure 6) showed that curvature effects were slightly larger when the previous trial did not require an eye movement in the direction of the central arrow. Moreover, we could show that similar effects or sometimes larger effects of to be ignored arrow primes were found for people who could not identify the prime compared to those who 22

could (Figure 7). This suggests that not only to-be-ignored arrows can affect the planning of saccades, but also that arrows that do not reach awareness affect saccade planning. Similar effects on saccade curvature have recently been reported for peripheral, unmasked stimuli (Cardoso-Leite et al., 2009 ; Van der Stigchel et al., 2009).

Conclusion In common with numerous studies, we show that briefly presented primes can speed up or slow down responses to a subsequently presented target, depending on the duration of the prime-to-target interval. In contrast, saccade curvature is always found to be away from the direction of the prime, independent of time between prime and target. This dissociation in the effects of the prime on saccadic response times and saccade curvature suggests that inhibition of the response to the prime involves multiple levels of the response preparation sequence.

23

R´ ef´ erences

Aizawa, H., & Wurtz, R. H. (1998). Reversible inactivation of monkey superior colliculus. I. Curvature of saccadic trajectory. Journal of Neurophysiology, 79 , 2082–2096. Aron, A. R., Schlaghecken, F., Fletcher, P. C., Bullmore, E. T., Eimer, M., Barker, R., et al. (2003). Inhibition of subliminally primed responses is mediated by the caudate and thalamus: evidence from functional MRI and Huntington’s disease. Brain, 126 , 713–723. Bowman, H., Schlaghecken, F., & Eimer, M. (2006). A neural network model of inhibitory processing in subliminal priming. Visual Cognition, 13 (4), 401–480. Boy, F., Clarke, K., & Sumner, P. (2008). Mask stimulus triggers inhibition in subliminal visuomotor priming. Experimental Brain Research, 190 , 111–116. Boy, F., & Sumner, P. (in press). Tight coupling between positive and reversed priming in the masked prime paradigm. Journal of Experimental Psychology: Human Perception & Performance. Cardoso-Leite, P., Mamassian, P., & Gorea, A. (2009). Comparison of perceptual and motor latencies via anticipatory and reactive response times. Attention, Perception & Psychophysics, 71 . Dorris, M. C., Klein, R. M., Everling, S., & Munoz, D. P. (2002). Contribution of the primate superior colliculus to inhibition of return. Journal of Cognitive Neuroscience, 14 , 1256–1263. Doyle, M., & Walker, R. (2001). Curved saccade trajectories: voluntary and reflexive saccades curve away from irrelevant distractors. Experimental Brain Research, 139 , 333–344. Driver, J., Davis, G., Ricciardelli, P., Kidd, P., Maxwell, E., & Baron- Cohen, S. (1999). Gaze perception triggers visuospatial orienting by adults in a reflexive manner. Visual Cognition, 6 , 509–540. Eimer, M. (1999). Facilitatory and inhibitory effects of masked prime stimuli on motor activation and behavioural performance. Acta Psychologica, 101 , 293–313. Eimer, M., & Schlaghecken, F. (1998). Effects of masked stimuli on motor activation: behavioral and electrophysiological evidence. Journal of Experimental Psychology: Human Perception & Performance, 24 , 1737– 1747. Eimer, M., & Schlaghecken, F. (2002). Links between conscious awareness and response inhibition: evidence from masked priming. Psychonomic Bulletin and Review , 9 , 514–520. Eimer, M., & Schlaghecken, F. (2003). Response facilitation and inhibition in subliminal priming. Biological Psychology, 64 , 7–26. Friesen, C. K., Ristic, J., & Kingstone, A. (2004). Attentional effects of counterpredictive gaze and arrow cues. Journal of Experimental Psychology: 24

Humand Perception & Performance, 30 , 319–329. Godijn, R., & Theeuwes, J. (2004). The relationship between inhibition of return and saccade trajectory deviations. Journal of Experimental Psychology: Human Perception & Performance, 30 , 538–554. Huber, D. E. (2008). Immediate priming and cognitive aftereffects. Journal of Experimental Psychology: General , 137 , 324–347. Ja´skowski, P. (2008). The negative compatibility effect with nonmasking flankers: a case for mask-triggered inhibition hypothesis. Consciousness and Cognition, 17 , 765–777. Ja´skowski, P. (2009). Negative compatibility effect: the object-updating hypothesis revisited. Experimental Brain Research, 193 , 157–160. Ja´skowski, P., & Przekoracka-Krawczyk, A. (2005). On the role of mask structure in subliminal priming. Acta Neurobiologiae Experimentalis, 65 , 409–417. Ja´skowski, P., & Verleger, R. (2007). What determines the direction of subliminal priming. Advances in Cognitive Psychology, 3 , 181–192. Klapp, S. T., & Hinkley, L. B. (2002). The negative compatibility effect: unconscious inhibition influences reaction time and response selection. Journal of Experimental Psychology: General , 131 , 255–269. Klotz, W., & Neumann, O. (1999). Motor activation without conscious discrimination in metacontrast masking. Journal of Experimental Psychology: Human Perception and Performance, 25 (4), 976–992. Kuhn, G., & Benson, V. (2007). The influence of eye-gaze and arrow pointing distractor cues on voluntary eye movements. Perception & Psychophysics, 69 , 966–971. Lee, C., Rohrer, W. H., & Sparks, D. L. (1988). Population coding of saccadic eye movements by neurons in the superior colliculus. Nature, 332 , 357– 360. Leuthold, H., & Kopp, B. (1998). Mechanisms of priming by masked stimuli: Inferences from event-related brain potentials. Psychological Science, 9 (4), 263–269. Lleras, A., & Enns, J. T. (2004). Negative compatibility or object updating? A cautionary tale of mask-dependent priming. Journal of Experimental Psychology: General , 133 , 475–493. Lleras, A., & Enns, J. T. (2005). Updating a cautionary tale of masked priming: Reply to Klapp (2005). Journal of Experimental Psychology: General , 134 (3), 436–440. Lleras, A., & Enns, J. T. (2006). How much like a target can a mask be? Geometric, spatial, and temporal similarity in priming: a reply to Schlaghecken and Eimer (2006). Journal of Experimental Psychology: General , 135 , 495–500. Ludwig, C. J., & Gilchrist, I. D. (2002). Measuring saccade curvature: a curve-fitting approach. Behavior Research Methods, 34 , 618–624. Marcel, A. J. (1983). Conscious and unconscious perception: an approach to 25

the relations between phenomenal experience and perceptual processes. Cognitive Psychology, 15 , 238–300. McPeek, R. M. (2006). Incomplete suppression of distractor-related activity in the frontal eye field results in curved saccades. Journal of Neurophysiology, 96 , 2699–2711. McPeek, R. M., Han, J. H., & Keller, E. L. (2003). Competition between saccade goals in the superior colliculus produces saccade curvature. Journal of Neurophysiology, 89 , 2577–2590. McPeek, R. M., & Keller, E. L. (2001). Short-term priming, concurrent processing, and saccade curvature during a target selection task in the monkey. Vision Research, 41 , 785–800. McSorley, E., Haggard, P., & Walker, R. (2004). Distractor modulation of saccade trajectories: spatial separation and symmetry effects. Experimental Brain Research, 155 , 320–333. McSorley, E., Haggard, P., & Walker, R. (2005). Spatial and temporal aspects of oculomotor inhibition as revealed by saccade trajectories. Vision Research, 45 , 2492–2499. McSorley, E., Haggard, P., & Walker, R. (2006). Time course of oculomotor inhibition revealed by saccade trajectory modulation. Journal of Neurophysiology, 96 , 1420–1424. McSorley, E., Haggard, P., & Walker, R. (2009). The spatial and temporal shape of oculomotor inhibition. Vision Research, 608–614. Mulckhuyse, M., Van der Stigchel, S., & Theeuwes, J. (in press). Early and late modulation of saccade deviations by target distractor similarity. Journal of Neurophysiology. Neumann, O., & Klotz, W. (1994). Motor-responses to nonreportable, masked stimuli - Where is the limit of direct parameter specification. In C. Umilt`a & M. Moscovitch (Eds.), Attention and performance XV: Conscious and nonconscious information processing (pp. 123–150). Cambridge, MA : MIT Press. Nummenmaa, L., & Hietanen, J. K. (2006). Gaze distractors influence saccadic curvature: evidence for the role of the oculomotor system in gaze-cued orienting. Vision Research, 46 , 3674–3680. Port, N. L., & Wurtz, R. H. (2003). Sequential activity of simultaneously recorded neurons in the superior colliculus during curved saccades. Journal of Neurophysiology, 90 , 1887–1903. Praamstra, P., & Seiss, E. (2005). The neurophysiology of response competition: motor cortex activation and inhibition following subliminal response priming. Journal of Cognitive Neuroscience, 17 , 483–493. Schlaghecken, F., Bowman, H., & Eimer, M. (2006). Dissociating local and global levels of perceptuo-motor control in masked priming. Journal of Experimental Psychology: Human Perception and Performance, 32 , 618–632. Schlaghecken, F., & Eimer, M. (2006). Active masks and active inhibition: a comment on Lleras and Enns (2004) and on Verleger, Jaskowski, Ay26

demir, van der Lubbe, and Groen (2004). Journal of Experimental Psychology: General , 135 , 484–494. Schmidt, T. (2006). Learning under anesthesia: checking the light in the fridge? Commentary on Deeprose and Andrade (2006). Consciousness & Cognition, 15 , 24–27. Seiss, E., & Praamstra, P. (2004). The basal ganglia and inhibitory mechanisms in response selection: evidence from subliminal priming of motor responses in Parkinson’s disease. Brain, 127 , 330–339. Sheliga, B. M., Riggio, L., Craighero, L., & Rizzolatti, G. (1995). Spatial attention-determined modifications in saccade trajectories. Neuroreport, 6 , 585–588. Sheliga, B. M., Riggio, L., & Rizzolatti, G. (1994). Orienting of attention and eye movements. Experimental Brain Research, 98 , 507–522. Sohrabi, A., & West, R. L. (2008). A biologicallly-plausible cognitive model (bpcm) of positive and negative congruency effects in masked priming. Poster session presented at the annual meeting of the Cognitive Science Society. Washinton DC, USA. Sumner, P. (2007). Negative and positive masked-priming - implications for motor inhibition. Advances in Cognitive Psychology, 3 , 317–326. Sumner, P., & Brandwood, T. (2008). Oscillations in motor priming: positive rebound follows the inhibitory phase in the masked prime paradigm. Journal of Motor Behavior , 40 , 484–489. Sumner, P., Nachev, P., Morris, P., Peters, A. M., Jackson, S. R., Kennard, C., et al. (2007). Human medial frontal cortex mediates unconscious inhibition of voluntary action. Neuron, 54 , 697–711. Sumner, P., Tsai, P.-C., Yu, K., & Nachev, P. (2006). Attentional modulation of sensorimotor processes in the absence of perceptual awareness. Proceedings of the National Academy of Sciences, USA, 103 , 10520–10525. Theeuwes, J., & Van der Stigchel, S. (2009). Saccade trajectory deviations and inhibition-of-return: Measuring the amount of attentional processing. Vision Research, 49 , 1307–1315. Trappenberg, T. P., Dorris, M. C., Munoz, D. P., & Klein, R. M. (2001). A model of saccade initiation based on the competitive integration of exogenous and endogenous signals in the superior colliculus. Journal of Cognitive Neuroscience, 13 , 256–271. Van den Bussche, E., Van den Noortgate, W., & Reynvoet, B. (2009). Mechanisms of masked priming: a meta-analysis. Psychological Bulletin, 135 , 452–477. Van der Stigchel, S., Meeter, M., & Theeuwes, J. (2006). Eye movement trajectories and what they tell us. Neuroscience & Biobehavioral Reviews, 30 , 666–679. Van der Stigchel, S., Meeter, M., & Theeuwes, J. (2007a). The spatial coding of the inhibition evoked by distractors. Vision Research, 47 , 210–218. Van der Stigchel, S., Meeter, M., & Theeuwes, J. (2007b). Top-down influ27

ences make saccades deviate away: the case of endogenous cues. Acta Psychologica, 125 , 279–290. Van der Stigchel, S., Mulckhuyse, M., & Theeuwes, J. (2009). Eye cannot see it: The interference of subliminal distractors on saccade metrics. Vision Research. Van der Stigchel, S., & Theeuwes, J. (2005). Relation between saccade trajectories and spatial distractor locations. Cognitive Brain Research, 25 , 579–582. 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 , 338–349. Van der Stigchel, S., & Theeuwes, J. (2007). The relationship between covert and overt attention in endogenous cuing. Perception & Psychophysics, 69 , 719–731. Van der Stigchel, S., & Theeuwes, J. (2008). Differences in distractor-induced deviation between horizontal and vertical saccade trajectories. Neuroreport, 19 , 251–254. Van Zoest, W., Van der Stigchel, S., & Barton, J. J. (2008). Distractor effects on saccade trajectories: a comparison of prosaccades, antisaccades, and memory-guided saccades. Experimental Brain Research, 186 , 431–442. Verleger, R., Ja´skowski, P., Aydemir, A., Van der Lubbe, R., & Groen, M. (2004). Qualitative differences between conscious and non-conscious processing? On negative and positive priming effects induced by masked arrows. Journal of Experimental Psychology: General , 133 , 494–515. Walker, R., McSorley, E., & Haggard, P. (2006). The control of saccade trajectories: direction of curvature depends on prior knowledge of target location and saccade latency. Perception & Psychophysics, 68 , 129–138.

28