Acta Psychologica 153 (2014) 66–73

Contents lists available at ScienceDirect

Acta Psychologica journal homepage: www.elsevier.com/ locate/actpsy

Perceptual and motor contributions to the negative compatibility effect Peng Liu, Yonghui Wang ⁎ School of Psychology, Shaanxi Normal University, Xi'an 710062, China Shaanxi Provincial Key Laboratory of Behavior & Cognitive Neuroscience, Xi'an 710062, China

a r t i c l e

i n f o

Article history: Received 24 February 2014 Received in revised form 22 September 2014 Accepted 24 September 2014 Available online xxxx PsycINFO classification: 2323 Visual Perception 2330 Motor Processes Keywords: Negative compatibility effect Stimulus–reaction link Object-updating Self-inhibition

a b s t r a c t The current study focused on contributions to the negative compatibility effect (NCE) from object-updating at the perceptual level and self-inhibition (i.e., automatic motor inhibition) at the response (motor) level. We hypothesized that contributions to the NCE from these two levels were moderated by the strength of stimulus and reaction (S–R) link: object-updating should have greater impact on the NCE with weak S–R links, but both object-updating and self-inhibition should impact on the NCE when the S–R links became strong. To test this hypothesis, in two experiments we used a novel type of stimuli and manipulated mask relevance (relevant versus irrelevant mask) and prime/target compatibility. Participants performed two tasks over three successive days. Results showed that under equivalent masking effectiveness between the two conditions of mask, a significant NCE was observed only in the relevant mask condition when the S–R links were weak, but both close to equal NCE sizes were observed in the two mask conditions when the S–R links were acquired through practice. The results indicated that perceptual and motor contributions to the NCE were moderated by the strength of S–R links, if the strength of the links was too weak to trigger the following inhibition, the NCE primarily originated from object-updating at the perceptual level; if the strength of the links was reinforced by practice, which exceeded the inhibitory threshold, the NCE originated from both object-updating and self-inhibition, but the latter was primarily responsible. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The negative compatibility effect (NCE), first reported by Eimer and Schlaghecken (1998), is an unexpected phenomenon, in which a masked stimulus influences behavior in a way contrary to what one might expect. Eimer and Schlaghecken presented a left or right pointing arrow prime (presented 16 ms), followed by a pattern mask that was eventually replaced by another arrow target (both mask and target were presented 100 ms). The task requires a participant to determine the orientation of the target. The typical NCE finding is that viewers' responses to targets show quantifiable gains (e.g., fast RTs) when the targets are preceded by incompatible primes (i.e., cueing the opposite responses to the targets), but their responses are delayed if the targets are preceded by compatible primes (i.e., cueing the same responses to the targets). To account for their results, Eimer and Schlaghecken proposed that the initial motor activation elicited by a subliminal prime is automatically followed by inhibition (Eimer, 1999; Eimer & Schlaghecken, 1998, 2002; Klapp, 2005; Klapp & Hinkley, 2002; Schlaghecken & Eimer, 2002, 2004, 2006). In this view, inhibition occurs only if a mask terminates perceptual evidence of the prime (Eimer & Schlaghecken, 2002) ⁎ Corresponding author at: School of Psychology, Shaanxi Normal University, Xi'an 710062, China. E-mail address: [email protected] (Y. Wang).

http://dx.doi.org/10.1016/j.actpsy.2014.09.011 0001-6918/© 2014 Elsevier B.V. All rights reserved.

and if the prime's activation strength is sufficiently great that it triggers an inhibitory mechanism (Schlaghecken & Eimer, 2002). According to this self-inhibition account, the prime automatically causes an activation of the motor mechanisms associated with it, but the appearance of the mask removes the prime-induced perceptual evidence which, in turn, leads to automatic inhibition of the initial motor activation (Bowman, Schlaghecken, & Eimer, 2006). Then, if a compatible target emerges, the required response is still inhibited, resulting in performance costs and the negative compatibility effect (NCE). A different view of the NCE mechanisms was suggested by Lleras and Enns (2004) (also see Verleger, Jaśkowski, Aydemir, van der Lubbe, & Groen, 2004). They hypothesized that the NCE results from objectupdating triggered by a perceptual interaction between prime and relevant mask. The relevant mask is defined as a stimulus that shares features with the prime (e.g., two overlaid opposite pointing double arrows). According to this object-updating account, the rapid serial visual presentation of prime and mask is interpreted by the visual system as a changing object. An initial representation of the object is established when the prime emerges, but once the mask appears, the object is updated with the new attributes of the representation (Lleras & Moore, 2003). If the mask is composed of two overlaid double arrows, the updated elements (prime incongruent elements) call for the alternative response to that induced by the prime. This leads to the NCE. The present study attempts to investigate contributions to the NCE from object-updating and from self-inhibition, with these contributions

P. Liu, Y. Wang / Acta Psychologica 153 (2014) 66–73

measured by the effect size calculated as mean reaction time (RT) in compatible minus incompatible trials. Klapp (2005) differentiated NCE processing based upon 1) perceptual updating, which only occurred with relevant masks at the perceptual level (i.e., NCE occurred at the perceptual level, NCE-P), and 2) response inhibition, which occurs with irrelevant (sharing no feature with the prime) as well as relevant masks at the response level (i.e., NCE occurred at the response level, NCE-NP). Klapp also suggested that both types of NCE might co-exist. Subsequent researches on this topic have been reported by Schlaghecken and Eimer (2006) as well as Sumner (2008). Originally, Schlaghecken and Eimer (2006) used double arrows as primes and targets; they manipulated the mask relevance, with random diagonal lines defining the relevant mask versus random vertical and horizontal lines defining the irrelevant mask. They also orthogonally varied the compatibility of prime and target, with compatible trials defined by same responses between these two stimuli and incompatible trials defined by opposite responses between prime and target. The goal was to assess whether the NCE would occur even in the unequivocally irrelevant mask condition. The results showed that both the relevant (containing both NCE-P and NCE-NP processes) and irrelevant (containing only NCE-NP process) mask conditions yielded significant NCEs. Furthermore, a comparison of the NCE's size as a function of masking type (relevant versus irrelevant) indicated that the effect size was greater (21 ms) with the relevant mask than that with the irrelevant one (10 ms). These results implied that the NCE might devolve into two sub-forms: the NCE-P (the effect size in the relevant mask condition minus that in the irrelevant mask condition: 21 ms − 10 ms = 11 ms) and the NCE-NP (the effect size in the irrelevant mask condition: 10 ms). However, the irrelevant mask was also significantly less effective in reducing prime visibility according to the forced choice1 results, which would result in reduced NCE sizes compared to a more effective mask (i.e., relevant mask) (Eimer & Schlaghecken, 2002; Klapp & Hinkley, 2002). Consequently, it may be invalid to compare the NCE's size if the masks under comparison differ in their masking effectiveness, and one might argue that the smaller size of the NCE-NP in the irrelevant mask condition might be due to the increased prime visibility. Sumner (2008) further investigated contributions to the NCE from object-updating and from self-inhibition with equivalent prime visibility, i.e., masking effectiveness. He used two vertical and two horizontal lines as primes and targets, and manipulated factors similar to those in Schlaghecken and Eimer (2006). The results showed close to equal NCE sizes with both types of mask (relevant versus irrelevant). This pattern of results was interpreted to mean that self-inhibition (NCE-NP) dominated the NCE processing. However, it is still arguable that the relevant masks used in this study were not as relevant as in the previous studies, which were not sufficiently relevant to support strong objectupdating in Sumner (2008).

67

1.1.1. Effectiveness of relevant and irrelevant masks A typical NCE paradigm uses arrow stimuli as both primes and targets. Thus the NCE-NP was usually investigated using irrelevant masks that consisted of vertical and horizontal lines in order to avoid triggering a feature updating process. However, it is difficult for such irrelevant masks to cover the after-image of the primes; in this case, these masks would not immediately cease activation of the primes. If the mask fails to occlude the after-image of the prime, this could violate a theoretical precondition of the NCE-NP, namely that perceptual evidence of the prime must be immediately ceased by the mask. 1.1.2. Reduced effectiveness of updating with a less relevant mask A second concern is that less relevant masks might lead to attenuated strength of object-updating. The most effective relevant mask should be one that would produce complementary stimuli of the primes through updating. Therefore, contributions to the NCE from the NCE-P and NCE-NP need to be investigated in a context in which the masks not only cover the after-image of the primes, but also avoid feature updating with masks in the irrelevant mask condition. Meanwhile, a highly relevant mask should be used in the relevant mask condition to avoid a less effective updating with a less relevant mask. 1.2. The present study The current study used numerical stimuli comprising short lines of equal length (shown in Fig. 1). Use of these stimuli has three main advantages. First, the mask covers the after-image of all the primes. Thus, in theory it should cease the activation of primes immediately in both masking conditions. Second, object-updating with the mask produces complementary stimuli of the primes in the relevant mask condition, which matches the theoretical precondition of the NCE-P. This is because the complementary stimuli of the primes can improve performance to the targets' responses in incompatible trials. However, in the irrelevant mask condition, updating with the mask only produces ruleless lines. In theory it mismatches the precondition of the NCE-P, because these updated features (i.e., ruleless lines) could not improve performance to the target responses both in compatible and incompatible trials. Third, comparisons of the NCE's sizes between relevant (containing both NCE-P and NCE-NP processes) and irrelevant masks (containing only NCE-NP process) can shed light upon the respective contribution from the NCE-P and NCE-NP to the NCE. Furthermore, these comparisons will have greater validity if both the mask conditions have the same physical characteristics.

Prime and target

Mask

Object-updating

1.1. Potentially influential factors in the previous studies Relevant

Two potentially influential factors mentioned above, that have not been explored in prior research, motivated the present design. Respectively, these involve 1) the differences in the effectiveness of relevant and irrelevant masks in reducing visibility of the primes (Schlaghecken & Eimer, 2006) and 2) the reduced effectiveness of updating with a less relevant mask (Sumner, 2008). These will be discussed in turn.

mask

Irrelevant mask

1

The visibility of a prime can be influenced by a mask. Prime visibility is low when a mask is effective. Since theoretically this is a factor in self-inhibition hypothesis, it is important to assess prime visibility. In this research this was accomplished using a forced choice task which was conducted using primes and masks as traditional NCE procedure, except no target was presented. Participants were asked to respond to the directions of the primes after the masks' presentation. If the accuracy of the forced choice task does not significantly differ from chance level (50%), this indicates that the masks effectively cease the initial activation of the primes.

Fig. 1. The stimuli used in the current study. Note that mask relevance is varied by changing the prime and target stimuli, the relevant mask condition is defined as “1” and “3” as primes and targets, and the irrelevant mask condition is defined as “2” and “4” as primes and targets.

68

P. Liu, Y. Wang / Acta Psychologica 153 (2014) 66–73

However, the strength of the response links with the numerical stimuli may be weaker than that with the traditional arrow stimuli which contain natural directional information. These attenuated response links with the present stimuli might result in the prime triggering motor activation that fails to exceed the requirements of inhibitory threshold, and thus the NCE-NP restricted to self-inhibition might not occur. In order to slowly build up the new links between numerical stimuli and their corresponding responses, the present experiments were conducted over three successive days aiming to get extensive practice. Effectiveness of this manipulation was supported by a number of studies. Some of them demonstrate that practice is a useful way to strengthen the stimulus and response (S–R) link and thus results in a robust and persistent effect (Klapp, 2005; Klapp & Haas, 2005; Klapp & Hinkley, 2002; Neill, Terry, & Valdes, 1994; Sumner, 2008; Tipper, Weaver, Cameron, Brehaut, & Bastedo, 1991), others confirm that an NCE can develop gradually through incidental learning of a stimulus response (S–R) link (Boy & Sumner, 2010; Schlaghecken, Blagrove, & Maylor, 2007), once the S–R link becomes strong enough that it exceeds the inhibitory threshold, the NCE would occur (Schlaghecken & Eimer, 2002). Moreover, Klapp and his colleagues claim that non-conscious processing (e.g., NCE) based on a direct S–R link without deep processing is in terms of automaticity (Klapp & Greenberg, 2009; Klapp & Haas, 2005). Consequently, it should be safe to reinforce the S–R link by practice while avoiding algorithmic processing (e.g., strategy) in our experiments. 2. Overview of experiments and the hypothesis According to Klapp (2005), who suggested that both perceptual updating (NCE-P) and response inhibition (NCE-NP) might contribute to the NCE, we hypothesize that when the initial S–R links are weak in the early days of a three-day sequence, the initial activation in the motor cortex triggered by the prime should be too weak to trigger the NCE-NP according to the precondition of self-inhibition. Accordingly, the NCE-P is predicted to contribute significantly to the NCE; however, when the strength of the S–R links is reinforced by practice occurring during the later days, the prime triggered activation should be strong enough to motivate the inhibitory mechanisms, and then this inhibition will automatically affect NCE processing. Accordingly, both NCE-P and NCE-NP are predicted to contribute to the NCE. To test this hypothesis, two experiments were conducted. Experiment 1 was designed to investigate contributions to the NCE from NCE-P and NCE-NP as a function of mask relevance (relevant versus irrelevant mask). Each level of mask relevance was assigned in one of two blocks within a session (order counterbalanced) in order to accelerate acquisition of the S–R links; this was intended to ensure that triggering motor activation via the primes exceeded the inhibitory threshold as early as possible. In this case, the NCE-NP should take effect during the NCE processing. Another variable was the compatibility of prime and target. In both experiments the compatible and incompatible trials occurred randomly with equal probability within each of the two trial blocks. The third variable was test day (three successive test days: Day 1, Day 2, and Day 3). However, the blocked design of mask relevance might lead to expectancy effects. Specifically, if a whole block of trials consistently calls for updating, given relevant mask, this could induce unwanted preparation of updating as trials progress, which might interfere with the size of NCE. Accordingly, Experiment 2 examined the same three variables but randomized mask relevance in two trial blocks to eliminate possible expectancy effects. The relevant and irrelevant mask conditions occurred equally often in both blocks across days in Experiment 2. Ideally if the pattern of results observed in Experiment 1 is replicated in Experiment 2, this will rule out any interference due to expectancy effects. Two predictions about the NCE are based upon the hypothesis outlined above. First, we predicted that early in practice (e.g., Day 1), a significant NCE should emerge in the relevant mask condition. This is

because updating with the relevant mask could produce complementary digits to the primes, and these complementary stimuli should improve performance with the identical targets (on incompatible trials: prime and target differed in features but the complementary digits to the primes were the same as the targets). For the irrelevant mask condition, no significant NCE should appear. This is because the S–R links should be too weak to trigger subsequent inhibition. The second prediction was that significant NCEs should appear in both mask relevance conditions in later days. This is because the motor activation via the prime should exceed the inhibitory threshold over practice. Moreover, the NCE's size in the relevant mask condition should be larger than that in the irrelevant mask condition. The latter outcome is anticipated because both the NCE-P and the NCE-NP should contribute to the NCE in the relevant mask condition, but only the NCE-NP should impact on the NCE in the irrelevant mask condition. 3. Experiment 1 The aim of Experiment 1 was to investigate contributions to the NCE from NCE-P and NCE-NP. Each level of the mask relevance was assigned in one of two blocks within a session, which aimed to accelerate acquisition of the S–R links. 3.1. Method 3.1.1. Participants Sixteen paid college students (6 males) ranging in age from 18 to 22 years (mean age = 19.50 years) took part in the experiment. Two participants were left-handed and the others were right-handed; all the participants had normal or corrected-to-normal vision. 3.1.2. Stimuli and apparatus All stimuli were composed of similar straight lines subtending a visual angle of 1° horizontally and 2° vertically. These are shown in Fig. 1. Within this group, “1” and “3” as well as “2” and “4” were used as primes and targets, “8” was used as mask. All stimuli were presented on the central fixation point with white stimuli and black background. The experiment was conducted by PC running at 60 Hz linked to a 17in CRT Monitor. 3.1.3. Procedure The participants were seated in a dimly lit chamber, a computer screen was placed 50 cm in front of their eyes, and the screen center was in the center of their horizontal sightline at a fixed straight-ahead level. Experiment 1 was divided into two tasks: a primary task (NCE task) and a forced-choice task (testing prime visibility). The trial procedure of the primary task is shown in Fig. 2. Each trial began with a central fixation cross (300 ms), and followed by a blank (200 ms), then a prime (i.e., “1” and “3” in one block, “2” and “4” in

Fig. 2. Schematic representation of trial procedure in Experiment 1.

P. Liu, Y. Wang / Acta Psychologica 153 (2014) 66–73

69

the other block) was presented for 33 ms and was followed by a mask (“8” presented for 100 ms). The mask appeared at the same spatial location as the prime, in order to guarantee the prime was occluded. Finally, a target stimulus was presented for 100 ms. Participants were instructed to maintain central eye fixation and to respond as quickly and accurately as possible to the final stimuli (i.e., the targets) presented in each trial. Using an English keyboard, half of the participants were instructed to use the index finger of their left-hand to respond with a Z key-press to “1” and “2”, and use the index finger of their right hand to respond with an M key-press to “3” and “4”. This was counterbalanced with the other half of participants, who were told to use the index finger of their left-hand to respond with a Z key-press to “3” and “4”, and use the index finger of their right hand to respond with an M key-press to “1” and “2”. In the primary task, the mask relevance was varied by changing the prime and target stimuli, the relevant mask condition is defined as “1” and “3” as primes and targets as well as the irrelevant mask condition is defined as “2” and “4” as primes and targets. These two conditions of mask divided the experiment into 2 blocks (100 trials each) and compatibility was randomized and equiprobable within each block. The sequence of blocks was counterbalanced in order. The inter-trial-interval (ITI) was randomized within 1300–1600 ms. A 16 trial practice phase was presented before the formal experiment, and this formal experiment would not begin unless the correct rate in practice exceeded 90%. Participants were given a break of at least 2 min after each block and were encouraged to take longer breaks whenever necessary. The tasks were completed in three successive days for each participant; this was done in order to strengthen the S–R links over practice. The secondary task was a forced choice (FC) task designed to assess prime visibility, i.e., masking effectiveness. It was scheduled to occur following the primary task on the third day of experiment. Each of the two blocks (50 trials each) was assigned a mask relevance condition; block order was counterbalanced. The trial arrangement was similar to the primary task, except that no target was presented. The ITI in the FC task was identical to that in the primary task. Participants were instructed to react to the primes (i.e., the first digit in each FC trial) as they did to the targets in the primary task; they were told that all responses should be executed after the mask presentation, and encouraged to guess if they could not clearly identify the prime.

the two-way interaction was significant in test day × compatibility, F (1.33, 20)g = 5.46, p = .02, MSE = 240, η2p = .27, and was marginally significant in mask relevance × compatibility, F (1, 15) = 3.47, p = .08, MSE = 397, η2p = .19. However, the two-way interaction between test day and mask relevance was not significant, F (1.89, 28.39)g = 1.75, p = .19, MSE = 544. We then looked separately at the interaction between mask relevance and compatibility by day. Fig. 3 details this relation across the three experiment days. In Day 1, the two-way interaction between mask relevance and compatibility was significant, F (1, 15) = 11.35, p = .004, MSE = 207, η2p = .43. Subsequent paired t-tests confirmed a highly significant NCE in the relevant mask condition, t (15) = 3.49, p = .003, SE = 5.59, but no compatibility effect in the irrelevant mask condition, t (15) = −0.69, p = .50, SE = 6.77. In Day 2, the two-way interaction between mask relevance and compatibility was not significant, F (1, 15) = .002, p = .96, MSE = 216, accompanied with a significant main effect of compatibility, F (1, 15) = 22.82, p b .001, MSE = 255, η2p = .60. In Day 3, the two-way interaction between mask relevance and compatibility was also null, F (1, 15) = .97, p = .34, MSE = 238, again, it accompanied with a significant main effect of compatibility, F (1, 15) = 20.08, p b .001, MSE = 354, η2p = .60. Looking at error rates, we found a significant main effect among three days, F (1.36, 20.37)g = 8.01, p = .006, MSE = .0008, η2p = .35, showing that participants made fewer errors in Day 3 (M = 2.00%) than in Day 2 (M = 2.72%), and fewer errors in Day 2 (M = 2.72%) than in Day 1 (M = 3.94%). Furthermore, a significant main effect of compatibility, F (1, 15) = 17.91, p = .001, MSE = .0006, η2p = .54, showing that participants made fewer errors in incompatible (M = 2.15%) than in compatible (M = 3.63%) trials. However, main effect of mask relevance was not significant, F (1, 15) = .02, p = .88, MSE = .0008. Additionally, the three-way interaction was not significant, F (1.93, 28.98)g = .56, p = .58, MSE = .0003, and the two-way interactions were also not significant between test day and compatibility, F (1.85, 27.69)g = 1.07, p = .35, MSE = .0008, as well as between mask relevance and compatibility, F (1, 15) = .25, p = .63, MSE = .0010. However, the two-way interaction between test day and mask relevance was significant, F (1.76, 26.42)g = 4.12, p = .03, MSE = .0006, η2p = .22. Although the interaction between mask relevance and compatibility was not significant in error rates, the similar result pattern was observed between mean error rates and mean RTs in Fig. 3.

3.2. Results

3.3. Discussion

Assessment performance in the FC task confirmed that neither the relevant nor the irrelevant mask allowed the primes differing significantly from chance: in the relevant mask condition, t (15) = 1.55, p = .14, as well as in the irrelevant mask condition, t (15) = .47, p = .65. In the primary task, RTs of correct responses as well as errors were analyzed. Furthermore, we excluded outside of plus and minus three standard deviation from RTs in each participant. Two 3 × 2 × 2 repeated measures ANOVAs involved, respectively, test day (Day 1, Day 2, Day 3), mask relevance (relevant, irrelevant) and compatibility (compatible, incompatible). With respect to RTs, no significant difference was observed among three days, F (1.26, 18.94)g = 2.77, p = .11, MSE = 7078 (the “g” superscript indicates the use of Greenhouse–Geisser adjustments to the degrees of freedom). However, a significant main effect of mask relevance, F (1, 15) = 15.33, p = .001, MSE = 1382, η2p = .51, showed that the mean RT for relevant mask trials (M = 441 ms) was faster than that for irrelevant mask trials (M = 462 ms). This might be because the targets in the relevant mask condition (“1” and “3”) were easier to be discriminated than the other (“2” and “4”). A significant main effect of compatibility, F (1, 15) = 17.24, p = .001, MSE = 699, η2p = .54, showed that the mean RT for incompatible trials (M = 444 ms) was faster than that for compatible trials (M = 460 ms). More importantly, the analysis revealed a significant three-way interaction, F (1.59, 23.81)g = 4.55, p = .03, MSE = 132, η2p = .23. Furthermore,

The results from the forced choice (FC) task indicated that FC performance did not differ significantly from chance in either of the two mask conditions. This confirmed that there was equivalent masking effectiveness in both mask relevance conditions. In the primary task, the significant main effect in the three successive days gauged by error rates showed that the participants became familiar with the experiment through practice. Our hypothesis was tested by the two-way interaction between mask relevance and compatibility spread over three days. Specifically, we predicted that absence of an NCE in the condition of irrelevant mask (containing only NCE-NP process) would indicate that the initial S–R links were too weak to trigger inhibition in Day 1; also, the significant NCE in the condition of relevant mask (containing both NCE-P and NCE-NP processes) was predicted to be confined to the NCE-P. Results supported the hypothesis that objectupdating made a marked contribution to the NCE during early practice (i.e., Day 1). Furthermore, in Days 2 and 3, significant NCEs were predicted in both mask conditions. Results were also consistent with the prediction that when the strength of the S–R links is reinforced by practice occurring during the later days, the prime triggered activation should be strong enough to motivate the inhibitory mechanisms, and that this inhibition will automatically affect the NCE processing. However, the lack of an interaction between mask relevance and compatibility showed that the NCE's size was close to equal between the two mask conditions in Days 2 and 3. This result violated our prediction (i.e., the

70

P. Liu, Y. Wang / Acta Psychologica 153 (2014) 66–73

Day1

Error Rates(%)

RT(ms) 520 500 480 460 440 420

8 400

4 0

380 Relevant

Irrelevant

effect size of relevant mask condition should be larger than that in the irrelevant mask condition) and seemed to indicate that the strengthened S–R links dominated the perceptual influence (NCE-P) that occurs with a weaker association, and then the primary contribution to the NCE should originate from automatic motor inhibition at the response level (NCE-NP) in the later days. Nevertheless, the object-updating would still occur in the later days. Supporting evidence came from the differences of the NCE's size in Day 3, which was shown in Fig. 4. In Day 3, the effect size of the irrelevant mask condition was approximately 17 ms, but it increased to approximately 25 ms in the relevant mask condition. This difference could only be due to contribution from the object-updating. Note that the difference in effect sizes in the relevant mask condition between Day 2 (19 ms) and Day 3 (25 ms) is not discussed here because this difference was not significant. This was calculated as the interaction between Day 2 and Day 3 and compatible and incompatible trials in the relevant mask condition, F (1, 15) = 1.23, p = .29, MSE = 104.

Mask Relevance 4. Experiment 2

Day2

Error Rates(%)

RT(ms) 520 500 480 460 440 420

In Experiment 1 the level of mask relevance was blocked in order to accelerate acquisition of the S–R links. However, blocking might carry a risk from expectancy effects, due to repetitive mask relevance condition within each block, could interfere with the size of NCE. That is, the updating might become routine throughout one block but not the other. To rule out such expectancy effects, Experiment 2 manipulated the same factors, except that the mask relevance was randomized and equiprobable within the two blocks. If the pattern of results for Experiment 2 is similar to that of Experiment 1, it should demonstrate that results from Experiment 1 are reliable and unaffected by expectancy effects.

8 400

4

380

4.1. Method

0 Relevant

Irrelevant

Mask Relevance

Day3 RT(ms)

Error Rates(%)

520 500

4.1.1. Participants A separate set of sixteen paid college students (5 males in contrast to 6 in Experiment 1) ranging in age from 18 to 22 years (mean age = 18.75 years) took part in the experiment. Two participants were left-handed and the others were right-handed, all the participants had normal or corrected-to-normal vision.

4.1.2. Stimuli and apparatus The experimental stimuli and apparatus were the same as those described in Experiment 1.

480 460

35

440

30

420

25

400

4 0

380 Relevant Irrelevant Mask Relevance

Effect Size (ms)

8

20

Mask Relevance

15

Relevant

10

Irrelevant

5 0 -5 -10

Fig. 3. Results of Experiment 1. Mean response times (RTs: line graphs) and mean error rates (bar graphs) obtained in the condition of relevant and irrelevant masks for compatible and incompatible trials spread over three successive days in Experiment 1. Error bar represents one standard error of the mean (SE).

Day1

Day2

Day3

Fig. 4. Comparison of the effect size between the relevant and irrelevant mask conditions over the three successive days. The effect size is calculated by mean RTs in the compatible trials minus that in the incompatible trials. Error bar represents one standard error of the mean (SE).

P. Liu, Y. Wang / Acta Psychologica 153 (2014) 66–73

4.1.3. Procedure The experimental procedure was the same as described in Experiment 1, with the following exceptions: in the primary task, both compatibility and mask relevance were randomized and equiprobable within each of the two blocks, this entailed a 4:2 mapping of stimuli to responses in each block (in half of the participants, the stimuli “1” and “2” mapped to left-hand response, “3” and “4” mapped to righthand response. This response role was counterbalanced in the other half of participants). Similar randomizations were applied to each of the two blocks of trials in the FC task where all types of prime were equiprobable. Again, the FC task was conducted after the primary task in the third day with each participant.

Error Rates(%)

500 480 460 440 420 400

In the FC task, identification performance was not above chance in either the relevant mask condition, t (15) = 1.12, p = .28, or in the irrelevant mask condition, t (15) = .19, p = .86. In the primary task, data analysis was similar to that of Experiment 1. With respect to RTs, a significant main effect among three days, F (1.73, 26)g = 17.88, p b .001, MSE = 1412, η2p = .54, showed that responses were faster in Day 2 (M = 469 ms) and Day 3 (M = 463 ms) than in Day 1 (M = 498 ms). Furthermore, a significant main effect of mask relevance, F (1, 15) = 9.55, p = .007, MSE = 501, η2p = .39, showed that the mean RT for relevant mask trials (M = 472 ms) was faster than that for irrelevant mask trials (M = 482 ms), and a main effect of compatibility, F (1, 15) = 12.64, p = .003, MSE = 373, η2p = .46, showed that the mean RT for incompatible trials (M = 472 ms) was faster than that for compatible trials (M = 482 ms). However, the analysis revealed no significant three-way interaction, F (1.98, 29.62)g = .79, p = .46, MSE = 120, and no significant two-way interactions, all ps N .21. Although both the three-way and two-way interactions were not significant in this experiment, in order to compare the results between the two experiments, we also looked separately at the interaction between mask relevance and compatibility by day, as shown in Fig. 5. As can be seen, the mean RT was faster for incompatible than that for compatible trials in the relevant mask condition in Days 1 and 2, but the mean RT was close to equal between incompatible and compatible trials in the irrelevant mask condition in the first two days. These observations were substantiated by paired t-tests, which confirmed significant NCEs in the relevant mask condition in Day 1, t (15) = 2.31, p = .035, SE = 4.82, and Day 2, t (15) = 2.94, p = .01, SE = 4.43; but no significant compatibility effect in the irrelevant mask condition in Days 1 and 2, all ps N .25. However, in Day 3, both significant NCEs can be observed in the relevant, t (15) = 2.61, p = .02, SE = 4.23, as well as in the irrelevant mask condition, t (15) = 2.35, p = .03, SE = 5.48. With respect to error rates, a significant main effect among three days, F (1.52, 22.79)g = 3.94, p = .04, MSE = .0010, η2p = .21, showing that participants made fewer errors in Day 2 (M = 2.94%) and Day 3 (M = 2.63%) than in Day 1 (M = 3.91%). However, both main effects of compatibility and mask relevance were not significant, all ps N .12. Additionally, the three-way interaction was significant, F (1.97, 29.51)g = 4.93, p = .015, MSE = .0004, η2p = .25, but the two-way interactions were not significant between either two of them, all ps N .30.

380

As found in Experiment 1, the results from the FC task in this experiment indicated that the inference from the primary task was effective under equivalent masking effectiveness. Results from the primary task should be compared to those in Experiment 1. A significant NCE in the condition of relevant mask and no compatibility effect in the condition of irrelevant mask were observed in Day 1. Furthermore, significant NCEs were observed in both mask relevance conditions in Day 3. These results resemble the findings of

Day1

RT(ms) 520

4.2. Results

4.3. Discussion

71

8 4 0

Relevant Irrelevant Mask Relevance RT(ms) 520

Day2

Error Rates(%)

500 480 460 440 420 8 4 0

400 380 Relevant Irrelevant Mask Relevance

Day3 RT(ms) RT 520

Error Rates(%)

500 480 460 440 420 8 4 0

400 380

Relevant Irrelevant Mask Relevance

Fig. 5. Results of Experiment 2. Mean response times (RTs: line graphs) and mean error rates (bar graphs) obtained in the condition of relevant and irrelevant masks for compatible and incompatible trials spread over three successive days in Experiment 2. Error bar represents one standard error of the mean (SE).

72

P. Liu, Y. Wang / Acta Psychologica 153 (2014) 66–73

Experiment 1, except that in Experiment 2, there was a significant NCE in the relevant mask condition but no compatibility effect in the irrelevant mask condition in Day 2. This might have been caused by the randomized design of the two mask relevance conditions, which made the S–R links more difficult to acquire in this experiment than in the first experiment. Specifically, two alternative stimuli were used as primes and targets in each block in Experiment 1, i.e., “1” and “3” in one block, “2” and “4” in the other block, but four alternative stimuli were used as primes and targets in each block in Experiment 2. Accordingly, objectupdating rather than self-inhibition made a greater contribution to the NCE in Days 1 and 2 under the circumstance of invalid motor inhibition, and when the strength of the S–R links was reinforced by practice in Day 3, self-inhibition made a greater contribution to the NCE. These essentially similar results pattern between the two experiments indicates that the results from Experiment 1 are reliable and unaffected by the expectancy effects. Furthermore, mean RTs in Experiment 2 significantly increased relative to those in Experiment 1, as evident in the main effect of experiment F (1, 15) = 29.19, p b .001, MSE = 2022, η2p = .66. This is likely to be the result of increased difficulty in acquiring the S–R links in Experiment 2. 5. General discussion 5.1. Overview of the study The purpose of the current study was to investigate contributions to the NCE from object-updating at the perceptual level and self-inhibition at the response level while avoiding the potentially influential factors found in the previous studies. To this end, in two experiments, mask relevance was manipulated together with prime/target compatibility in tasks that extended over three successive days. Experiment 1 was divided into two blocks, with different mask conditions in each block, in order to accelerate participants' acquisition of the S–R links. This aimed to ensure that the prime's triggering of motor activation met requirements for exceeding the inhibitory threshold as early as possible. In this case, the NCE-NP should take effect in the NCE processing. In Experiment 2, the two levels of mask relevance were presented randomly (with equal probability) over trials in both of the two blocks (again with equal probability) in order to confirm that the results from Experiment 1 are reliable and unaffected by the expectancy effects. Results revealed that under equivalent masking effectiveness, early emergence of the NCE was only observed in the relevant rather than irrelevant mask condition. Specifically, in Experiment 1, early emergence of an NCE was evident in Day 1, and in Experiment 2, it was evident in Days 1 and 2. This implied that the inhibitory threshold was not exceeded when the S–R links were weak in the early days. Consequently, object-updating at the perceptual level, rather than self-inhibition at the response level, had a greater impact on the NCE in the early days. However, as the strength of S–R links increased over three days of practice and exceeded the inhibitory threshold, then nearly equivalent NCE's size emerged in the two mask conditions. These results indicated that in the later days, self-inhibition at the response level had a greater impact on the NCE. However, although automatic motor inhibition at the response level primarily contributed to the NCE in the later days, objectupdating at the perceptual level would still take effect. 5.2. Processing levels of the traditional NCE The traditional NCE paradigm has relied upon arrow stimuli as primes and targets. These would activate the motor area automatically because of the natural physical characteristics of these objects. Furthermore, evidence from ERPs and fMRI using the traditional NCE paradigm showed that the motor activation and the following inhibition would take effect even without practice (Aron et al., 2003; Boy, Husain, Singh, & Sumner, 2010; D'Ostilio & Garraux, 2012; Jaśkowski, Białuńska,

Tomanek, & Verleger, 2008; Sumner et al., 2007). Combined with the results from the current study, it seems clear that in the NCE paradigm, when the prime's activation in the motor cortex exceeds the inhibitory threshold, as in the later days of the current study, automatic motor inhibition at the response level has a dominate impact on the traditional NCE; meanwhile, object-updating at the perceptual level continue to exert some effect on the NCE. 5.3. Relevance of previous studies There is some merit in attempting to relate the previous divergent results about the self-inhibition hypothesis to the present findings. Often in the previous studies, central arrows were used as primes and targets, flankers (i.e., overlays of two types of primes presented on both sides from fixation) were used as distractors and presented between primes and targets (Bennett, Lleras, Oriet, & Enns, 2007; Lleras & Enns, 2006). The results obtained a significant NCE with high prime visibility. Arguably these results did not favor the self-inhibition explanation because the flankers did not remove the perceptual evidence of the prime, thus violating the theoretical precondition in triggering the NCE according to the self-inhibition hypothesis. However, according to the present findings, the NCE-NP is based on the response inhibition, which occurs when a stimulus that initially begins to activate a response readiness is no longer available. Thus, the NCE-NP should have little effect on the NCE due to high prime visibility. In this case, the NCE might primarily originate from the NCE-P. Consequently, these results would only support the object-updating hypothesis, but could not refute the self-inhibition hypothesis, for the mismatch of its precondition. 5.4. Another explanation It is noted that the present study focuses on the self-inhibition hypothesis and the object-updating hypothesis, but other main theories for the NCE, such as the attentional shift hypothesis (Cole & Kuhn, 2010) and the mask-triggered inhibition hypothesis (Jaśkowski & Przekoracka-Krawczyk, 2005), should not be ignored. Since the stimuli used in the current study seem less likely to shift one's attention, we could not evaluate the attentional shift hypothesis using the present results. However, another mask-triggered inhibition hypothesis fits some of the present results. This hypothesis is a hybrid containing features taken both from the object-updating and the selfinhibition hypothesis. It claims that the NCE occurs whenever the mask contains relevant features or searching for objects. In such situations, the inhibitory processing evoked by the mask is able to overcome the motor activation evoked by the prime. Furthermore, it also assumes that the inhibition should be more effective for relevant than that for irrelevant masks. According to this hypothesis, the magnitude of the NCE in the relevant mask conditions should be larger than in the irrelevant mask conditions in both experiments. This prediction is supported by part of the present results, specifically, in the first experiment, a significant NCE was observed in the relevant mask condition and no compatibility effect was observed in the irrelevant mask condition in Day 1. Moreover, both significant NCEs were observed in the two mask conditions in Day 3, and a larger magnitude of NCE emerged in the relevant than in the irrelevant mask condition; in the second experiment, a significant NCE was observed in the relevant mask condition and no compatibility effect was observed in the irrelevant mask condition in Days 1 and 2. However, applying the mask-triggered inhibition hypothesis to the current results becomes problematic when trying to explain why magnitudes of the NCE were close to equal between the two mask conditions in Day 2 of the first experiment and in Day 3 of the second experiment. Consequently, although the mask triggered inhibition hypothesis has explanatory merit, the conclusion of the current study presents a more reasonable explanation for the present results.

P. Liu, Y. Wang / Acta Psychologica 153 (2014) 66–73

6. Conclusion The strength of the S–R links moderates the impact of both objectupdating at the perceptual level as well as self-inhibition at the response level on the NCE under equivalent masking effectiveness. Specifically, if the strength of the links is too weak to trigger the following inhibition, the NCE primarily originates from object-updating at the perceptual level; if the strength of the links is reinforced by practice, which exceeds the inhibitory threshold, although the NCE may originate from the both processing levels, the automatic motor inhibition at the response level would be primarily responsible. Acknowledgment This research was supported by a grant from the Natural Science Foundation of China (31371026) to Yonghui Wang. We thank Prof. Kevin Miller for assistance with the manuscript. References 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(3), 713–723. Bennett, J., Lleras, A., Oriet, C., & Enns, J. (2007). A negative compatibility effect in priming of emotional faces. Psychonomic Bulletin & Review, 14(5), 908–912. Bowman, H., Schlaghecken, F., & Eimer, M. (2006). A neural network model of inhibitory processes in subliminal priming. Visual Cognition, 13(4), 401–480. Boy, F., Husain, M., Singh, K.D., & Sumner, P. (2010). Supplementary motor area activations in unconscious inhibition of voluntary action. Experimental Brain Research, 206(4), 441–448. Boy, F., & Sumner, P. (2010). Tight coupling between positive and reversed priming in the masked prime paradigm. Journal of Experimental Psychology: Human Perception and Performance, 36(4), 892–905. Cole, G., & Kuhn, G. (2010). What the experimenter's prime tells the observer's brain. Attention, Perception, & Psychophysics, 72(5), 1367–1376. D'Ostilio, K., & Garraux, G. (2012). Dissociation between unconscious motor response facilitation and conflict in medial frontal areas. The European Journal of Neuroscience, 35(2), 332–340. 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 and Performance, 24(6), 1737–1747. Eimer, M., & Schlaghecken, F. (2002). Links between conscious awareness and response inhibition: evidence from masked priming. Psychonomic Bulletin & Review, 9(3), 514–520.

73

Jaśkowski, P., Białuńska, A., Tomanek, M., & Verleger, R. (2008). Mask- and distractortriggered inhibitory processes in the priming of motor responses: an EEG study. Psychophysiology, 45(1), 70–85. Jaśkowski, P., & Przekoracka-Krawczyk, A. (2005). On the role of mask structure in subliminal priming. Acta Neurobiologiae Experimentalis, 65(4), 409–417. Klapp, S.T. (2005). Two versions of the negative compatibility effect: comment on Lleras and Enns (2004). Journal of Experimental Psychology: General, 134(3), 431–435. Klapp, S.T., & Greenberg, L.A. (2009). Temporary activation of perceptual–motor associations: a stimulus–response interpretation of automaticity. Journal of Experimental Psychology: Learning, Memory, and Cognition, 35(5), 1266–1285. Klapp, S.T., & Haas, B.W. (2005). Nonconscious influence of masked stimuli on response selection is limited to concrete stimulus–response associations. Journal of Experimental Psychology: Human Perception and Performance, 31(1), 193–209. 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(2), 255–269. Lleras, A., & Enns, J.T. (2004). Negative compatibility or object updating? A cautionary tale of mask-dependent priming. Journal of Experimental Psychology: General, 133(4), 475–493. 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(3), 495–500. Lleras, A., & Moore, C.M. (2003). When the target becomes the mask: using apparent motion to isolate the object-level component of object substitution masking. Journal of Experimental Psychology: Human Perception and Performance, 29, 106–120. Neill, W.T., Terry, K.M., & Valdes, L.A. (1994). Negative priming without probe selection. Psychonomic Bulletin & Review, 1, 119–121. Schlaghecken, F., Blagrove, E., & Maylor, E.A. (2007). Incidental learning of S–R contingencies in the masked prime task. Journal of Experimental Psychology: Human Perception and Performance, 33(5), 1177–1188. Schlaghecken, F., & Eimer, M. (2002). Motor activation with and without inhibition: Evidence for a threshold mechanism in motor control. Perception & Psychophysics, 64(1), 148–162. Schlaghecken, F., & Eimer, M. (2004). Masked prime stimuli can bias “free” choices between response alternatives. Psychonomic Bulletin & Review, 11(3), 463–468. Schlaghecken, F., & Eimer, M. (2006). Active masks and active inhibition: A comment on Lleras and Enns (2004) and on Verleger, Jaskowski, Aydemir, van der Lubbe, and Groen (2004). Journal of Experimental Psychology: General, 135(3), 484–494. Sumner, P. (2008). Mask-induced priming and the negative compatibility effect. Experimental Psychology, 55(2), 133–141. 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(5), 697–711. Tipper, S.P., Weaver, B., Cameron, S., Brehaut, J.C., & Bastedo, J. (1991). Inhibitory mechanisms of attention in identification and localization tasks: Time course and disruption. Journal of Experimental Psychology: Learning, Memory, & Cognition, 17, 681–692. Verleger, R., Jaśkowski, P., Aydemir, A., van der Lubbe, R.H.J., & Groen, M. (2004). Qualitative differences between conscious and nonconscious processing? On inverse priming induced by masked arrows. Journal of Experimental Psychology: General, 133(4), 494–515.

Perceptual and motor contributions to the negative compatibility effect.

The current study focused on contributions to the negative compatibility effect (NCE) from object-updating at the perceptual level and self-inhibition...
511KB Sizes 2 Downloads 5 Views