Acta Psychologica 153 (2014) 153–159

Contents lists available at ScienceDirect

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

Knowledge of response location alone is not sufficient to generate social inhibition of return Timothy N. Welsh a,⁎, Joseph Manzone b, Laura McDougall c a b c

Centre for Motor Control, Faculty of Kinesiology and Physical Education, University of Toronto, Canada Faculty of Kinesiology and Physical Education, University of Toronto, Canada Faculty of Kinesiology, University of Calgary, Canada

a r t i c l e

i n f o

Article history: Received 24 March 2014 Received in revised form 13 October 2014 Accepted 20 October 2014 Available online 8 November 2014 PsycINFO classification: 2330 Motor Processes 2346 Attention Keywords: Ideomotor coding Joint action Social inhibition of return

a b s t r a c t Previous research has revealed that the inhibition of return (IOR) effect emerges when individuals respond to a target at the same location as their own previous response or the previous response of a co-actor. The latter social IOR effect is thought to occur because the observation of co-actor's response evokes a representation of that action in the observer and that the observation-evoked response code subsequently activates the inhibitory mechanisms underlying IOR. The present study was conducted to determine if knowledge of the co-actor's response alone is sufficient to evoke social IOR. Pairs of participants completed responses to targets that appeared at different button locations. Button contact generated location-contingent auditory stimuli (high and low tones in Experiment 1 and colour words in Experiment 2). In the Full condition, the observer saw the response and heard the auditory stimuli. In the Auditory Only condition, the observer did not see the co-actor's response, but heard the auditory stimuli generated via button contact to indicate response endpoint. It was found that, although significant individual and social IOR effects emerged in the Full conditions, there were no social IOR effects in the Auditory Only conditions. These findings suggest that knowledge of the co-actor's response alone via auditory information is not sufficient to activate the inhibitory processes leading to IOR. The activation of the mechanisms that lead to social IOR seems to be dependent on processing channels that code the spatial characteristics of action. © 2014 Elsevier B.V. All rights reserved.

1. Introduction When an individual completes a series of responses to targets that appear randomly at several locations, reaction times (RTs) tend to be longer for responses to targets presented at the same location as a previous target than for responses to targets presented at a different location (e.g., Maylor & Hockey, 1985). This increase in RT for repeated relative to different targets is consistent with the inhibition of return (IOR) effect observed in cue–target paradigms (Posner & Cohen, 1984). The IOR effect is thought to reflect the activation of an inhibitory mechanism that facilitates efficient search patterns by hindering the return of attention to the previously responded-to location and/or the reactivation of a recently executed response used to search a given location in space (see Klein, 2000 for a review). Although IOR has most commonly been studied in individuals acting alone (i.e., individual or iIOR), a series of studies has revealed that IOR is ⁎ Corresponding author at: Faculty of Kinesiology & Physical Education, University of Toronto, 55 Harbord Street, Toronto, ON M5S 2W6, Canada. Tel.: + 1 416 946 3303; fax: +1 416 946 5310. E-mail address: [email protected] (T.N. Welsh).

http://dx.doi.org/10.1016/j.actpsy.2014.10.007 0001-6918/© 2014 Elsevier B.V. All rights reserved.

also present in social action contexts (e.g., Cole, Skarratt, & Billing, 2012; Skarratt, Cole, & Kingstone, 2010; Welsh, McDougall, & Weeks, 2009; Welsh et al., 2005, 2007). In these studies, pairs of individuals execute responses to a common set of target locations. The participants take turns responding to randomly presented targets. RTs for trials on which co-actor A responds to a target presented at the same location as co-actor B's previous response are compared to RTs for trials on which co-actor A responds to a target presented at a different location from co-actor B's previous response. The results of these comparisons consistently reveal that RTs for repeated target trials are longer than those on different trials — a pattern of RTs consistent with the iIOR effect (e.g., Maylor & Hockey, 1985; Welsh & Pratt, 2006). This social IOR (sIOR) effect is thought to be caused by the same set of mechanisms that lead to iIOR (Welsh et al., 2005). In support of the hypothesis that the same inhibitory mechanisms lead to iIOR and sIOR, Welsh et al. (2009) found that the magnitude of the IOR effect on trials in which the individual followed their own response correlated with the magnitude of the IOR effect when they followed their partner's response. The currently held view regarding the processes that lead to the activation of the inhibitory mechanisms leading to sIOR is that the

154

T.N. Welsh et al. / Acta Psychologica 153 (2014) 153–159

observation of a co-actor's response generates a representation of that action in the central nervous system of the observer. This observationevoked response representation is subsequently accessed by other systems to shape future behaviour. This hypothesis is consistent with the wealth of behavioural and neurophysiological evidence for an action observation system that enables an individual to represent the actions and the sensory consequences (effects) of the actions for use in numerous social cognitive processes (for relevant reviews see Rizzolatti & Craighero, 2004, and van der Wel, Sebanz, & Knoblich, 2013). In the case of sIOR, it is hypothesized that the observation-evoked representations generate a simulation of the performance in the observer and that this simulation of performance then activates the inhibitory mechanisms that lead to IOR. There are three sets of results that suggest that knowledge of the spatial characteristics of the response is critical to the generation of the sIOR effect. First, there are a pair of papers reporting that vision of the onset of the target and/or of the contact with the target location is not necessary for generation of sIOR because sIOR was present even when observers were only permitted to see the initial (~25%) of their partner's response trajectory (Skarratt et al., 2010; Welsh et al., 2007). Second, in a study in which participants sat next to each other (instead of across from one another), RTs were longer for movements in the same direction as the co-actor's previous response even though the response terminated on a different location (Welsh et al., 2009). Finally, there is evidence that the goal of the final action executed at the target location (e.g., write vs. erase with a pencil) does not influence sIOR (Cole et al., 2012). Thus, knowledge of the spatial characteristics of the co-actor's response, not just of the interaction with an object at a specific location in space, seems to be critical for the generation of the sIOR effect. The purpose of the present study was to determine if knowledge of the response endpoint alone, in the absence of directly witnessing some spatial components of the response, is sufficient to activate the inhibitory mechanisms leading to the sIOR effect. To this end, participants completed a sIOR task under two conditions. The Full condition was consistent with typical sIOR protocols in that participant's vision of the environment was not manipulated and they were permitted to witness the entirety of the response. The Full condition was not critical to addressing the main purpose of the study, but was included to ensure consistency with previous sIOR. The key condition for determining the role of knowledge of endpoint was the new Auditory Only condition. In the new Auditory Only condition, participants were prevented from witnessing the response, but were informed of the endpoint of the response via distinct location-contingent auditory information. In Experiment 1, the auditory information was a high or low pitched tone that was presented only when a specific button was contacted. Presentation of the effect tone alone was hypothesized to be sufficient for activating knowledge of the response, and likely the response codes that generate the tone, because there is behavioural (e.g., Elsner & Hommel, 2001) and neurophysiological (e.g., Kohler et al., 2002; Melcher, Weidema, Eenshuistra, Hommel, & Gruber, 2008, Melcher et al., 2013) evidence indicating that response representations can be activated following the perception of similar response-contingent auditory effect information when individual tones are consistently presented following a specific response. In Experiment 2, green and blue coloured paper was placed around the base of the target locations and the auditory information was the spoken words “green” or “blue” presented when the button surrounded by the green or blue paper was contacted, respectively. Thus, by perceiving the auditory information, participants knew which button was contacted by their co-actor, but did not observe the action that generated the effect tone. Hence, if knowledge of the endpoint of the partner's response alone can activate the inhibitory mechanisms leading to sIOR, then sIOR should emerge in the Auditory Only condition. On the other hand, if witnessing some spatial components of the task is necessary for the generation of sIOR, then sIOR will not be observed in the Auditory Only condition.

2. Experiment 1 2.1. Methods 2.1.1. Participants Five pairs of individuals (6 male; aged 19–24 years) were recruited to complete the study. All participants had a right-hand preference (self-report), had normal or corrected-to-normal vision, and were naïve to the purpose of the study. The procedures of the present study complied with the codes of the Declaration of Helsinki and were approved by the University of Calgary Research Ethics Board. Each participant provided written informed consent prior to data collection. 2.1.2. Apparatus, task and procedures Participants sat on opposite sides of a table. A black metal board containing four red buttons (2 cm diameter) was placed on the table in between the participants. The four buttons were arranged in a cross with each button located 14 cm from the centre of the board. One starting location button was directly in front of each participant and oriented along their midlines. The two target buttons were located on either side of an imaginary line connecting the two home buttons. Participants were asked to fixate a 2 cm by 2 cm cross located at the intersection of the imaginary line between the two target locations and the imaginary line between the two home positions throughout the blocks of trials. There were a total of 24 blocks of trials in the study. The 24 blocks were divided into 2 sets of 12 blocks. Each block of trials consisted of 33 trials. Optional breaks were provided between each block to combat mental and physical fatigue. In the each set of 12 blocks, one participant completed the Auditory Only condition while the other participant completed the Full condition. Participants switched conditions at the end of the first 12 blocks. The same sequence of auditory stimuli was presented on every trial whether the participant was in the Full or Auditory Only condition. The target for a given trial was indicated by the 80 ms illumination of a light-emitting diode (LED) under one of the potential target locations. Participants were instructed to start each trial with the index finger of their right hand depressing their starting button and then to move as quickly as possible to and touch the button that had illuminated. These movements were executed in a paired-alternating order (e.g., AABBAABB…) such that Participant A completed two responses followed by Participant B completing two responses, and so on. This trial arrangement enabled the examination of the effects of response repetition (IOR) on trials on which each individual participant followed their own response (i.e., AA and BB trials — iIOR) and when the participant followed the response of their partner (i.e., AB and BA trials — sIOR). Target location was pseudo-random on each trial with the constraints that 1) each trial combination occurred equally often within a block; and, 2) no location could be the target location on more than 4 trials in a row. Throughout the entire study, both participants wore a pair of sound attenuating headphones that reduced environmental noise and presented the auditory stimuli. The headphones worn by each participant were linked to a common output such that both participants received the same set of white noise masks and tones on each trial. The white noise was presented during the movement time interval to prevent the participants from obtaining spatial information of target contact for a trial by hearing the contact and release of the target button. The white noise mask started 50 ms after the home button was released and continued until one of the buttons was pressed. Immediately after one of the two target locations was pressed, the associated tone was presented for 200 ms. A high-pitched tone (800 Hz) was presented when one button was pressed and a low-pitched tone (200 Hz) was presented when the other button was pressed. Prior to data collection, the participants were told about and given a demonstration of the button/tone mapping. The absolute button/tone mapping was kept

T.N. Welsh et al. / Acta Psychologica 153 (2014) 153–159

consistent throughout the study, but the relative (i.e., to the right or left of a given participant) was counterbalanced across participants as a natural consequence of the absolute mapping. A second white noise mask was presented for 50 ms after the tone was presented to prevent the participant from hearing the release of the target button. There was a further 1170 ms from the offset of the white noise mask until the onset of the next target. Hence, there was a total of 1220 ms from the offset of the auditory tone until the onset of the subsequent target to allow for the participant to process the auditory tone and to return the hand to the home position and prepare for the subsequent trial. During the Full condition, the participant had an unrestricted view of the movement environment during both movement execution (when the participant responded) and observation (when the co-actor responded) trials. Likewise, during movement execution trials in the Auditory Only condition, the participant had an unrestricted view of the environment. Thus, participants heard the tone and witnessed the response that generated the tone on all movement execution trials and on observation trials in the Full condition. On the observation trials in the Auditory Only condition, however, the participant was prevented from seeing the onset of the target stimulus and any part of the response of their co-actor. The participants in the Auditory Only condition wore a pair of goggles with liquid crystal lenses (Translucent Technologies, Toronto, ON). The lenses in these goggles can change from a transparent (open) to a translucent (closed) state in approximately 4 ms. To prevent the participant from seeing the onset of the targets and the partner's subsequent responses during observation trials, the lenses of the goggles closed 20 ms prior to the onset of the target and remained closed while the partner executed their pair of responses. The lenses of the goggles did not open again until 170 ms prior to the onset of the target for the first of the participant's own pair of movement execution trials. Note that participants received the same auditory stimuli on all trials in both Auditory Only and Full conditions. Thus, on observation trials in the Auditory Only condition, the participant heard the auditory tone generated by the partner's action, but did not witness target onset or the movement that generated the tone. Prior to data collection, participants completed a familiarization phase in which they were first instructed about the task and the specific button-tone mapping. Following these instructions, they completed a short practice sequence of a block of 33 trials to familiarize themselves with the paired alternating sequence of responses and to experience the button-tone mapping. Immediately after the familiarization trials were completed, the test trials were completed. A custom E-Prime (V 1.0) program controlled the stimulus events and recorded the timing and location of button presses. 2.1.3. Data reduction and analysis For each trial, the time and identity of home button release and target button contact was recorded. RT was defined as the time between target onset and movement onset (home button release). Movement time (MT) was defined as the time between movement onset and target button contact. Temporal and location data for trials with RT values shorter than 100 ms, or with RTs or MTs over 1000 ms were removed from the data set as anticipation and inattention errors, respectively. Subsequently, data that were greater than two standard deviations above the mean for a given condition were deleted as outliers. Data were also deleted from trials on which participants moved in the incorrect order. The data for the trial immediately after each order error was also removed because the analysis is dependent on an accurately completed movement sequence for the preceding response. Finally, the first trial of each block was excluded because there was no response preceding it and therefore belonged to no experimental condition. There were no movements to the incorrect target location. These criteria resulted in exclusion of 0.5–2.9% of trials per participant; an average of 2.1% of total trials. After the data for the errors had been removed from the data set, mean RTs and MTs were calculated and submitted to a set of statistical

155

analyses. Although the key analysis for evaluating the hypotheses was between RTs on repeated and new target trials (the assessment of the (non-)presence of sIOR) in the Auditory Only condition, the data were initially submitted to omnibus 2 (Response Information: Full, Auditory Only) by 2 (Person: Same, Different) by 2 (Target: Repeated, New) 3-factor repeated measures ANOVAs. The assumption of sphericity was not violated in the analysis of RT and MT. Subsequent to the omnibus analysis, post hoc analysis in which the data for repeated and new target trials in each condition was completed using Tukey's HSD to determine if IOR emerged in each condition. Alpha was set at 0.05 for all analyses. 2.2. Results and discussion The only significant effect to emerge from the analysis of MTs was a significant effect of Person, F(1, 9) = 10.25, p b .05, η2p = 0.53. Participants had shorter MTs when they followed their partner (mean = 196 ms, SD = 11.9) than when they followed their own responses (mean = 200 ms, SD = 11.7). No other effect approached significance, Fs b 1.0. MT data for each condition are presented in Table 1. The analysis of RTs revealed a main effect for Person, F(1, 9) =23.02, p b .001, η2p = 0.72. RTs were shorter when the participants followed their own response (mean = 252 ms, SD = 32.6) than when they followed the response of their partner (mean = 261 ms, SD = 31.1). There was also a main effect for Target, F(1, 9) = 28.16, p b .001, η2p = 0.76, revealing an overall IOR effect (Repeated Target RTs — mean = 262 ms, SD = 32.4; New Target RTs — mean = 251 ms, SD = 31.3). Although there was no main effect for Response Information, F(1, 9) = 0.03, p N .87, η2p = 0.003, Response Information interacted with Target, F(1, 9) = 29.40, p b .001, η2p = 0.77. Analysis of this interaction revealed that the repeated/new target differences in the Full condition was larger (mean = 17.0 ms, SD = 8.4) than in the Auditory Only condition (mean = 4.63 ms, SD = 6.2). Repeated/new target differences were also smaller when the participant followed their partner (mean = 5.3 ms, SD = 10.8) than when they followed their own response (mean = 16.4 ms, SD = 9.4); Person by Target interaction, F(1, 9) = 4.99, p = .05, η2p = 0.357. Thus, repeated/new target differences were smaller overall in the Auditory Only than in the Full condition, and when participants followed the response of their partner (sIOR) than when they followed their own response (iIOR). Although these interactions are informative, they do not specifically address the main a priori predictions of whether or not RTs on repeated target and new target trials (i.e., IOR) were statistically different in any one specific condition and, of greatest importance, whether or not a sIOR effect emerged in the Auditory Only condition. Thus, despite the absence of a significant 3-way interaction, F(1, 9) b 1.0, a post hoc analysis comparing RTs for repeated and new target locations was conducted to determine if IOR was present in each condition. The results of this analysis revealed significant iIOR effects in both Auditory Only and Full conditions (Fig. 1). Likewise, sIOR was present when the participant followed their partner's movement in the Full condition. Critically, a small and non-significant facilitation effect (i.e., no sIOR) was present when the participant followed their partner in the Auditory Only condition. In sum, the above results indicate that the knowledge of a co-actor's response endpoint alone is not sufficient to activate the processes Table 1 Mean (and standard deviation) movement times in milliseconds as a function of response information, person, and target for Experiment 1. Response information

Full Auditory only

Same person

Different person

New target

Repeated target

New target

Repeated target

200 (39.8) 201 (38.4)

201 (40.2) 200 (33.9)

195 (40.3) 196 (36.7)

196 (39.7) 197 (37.9)

156

T.N. Welsh et al. / Acta Psychologica 153 (2014) 153–159

New Target

Reaction Time (ms)

340

Repeated Target

Auditory Only condition. If some visuospatial information is necessary to drive the effect then no sIOR will occur in the Auditory Only condition.

*

3. Experiment 2

320 300

*

ns

*

3.1. Methods

280 260 240 220 Same Person

Different Person

Auditory Only

Same Person

Different Person

Full

Response Information Fig. 1. Mean reaction times (ms) as a function of Response Information, Person, and Target for Experiment 1 (tones). Open bars represent reaction times for different target trials and filled bars represent reaction times for repeated target trials. * denotes significant differences between RTs on repeated and different target trials. SEM bars are shown.

leading to IOR. This conclusion is based on the observation that an IOR effect was present in each condition except the critical Auditory Only condition in which participants responded after their partner, but were prevented from observing any visuospatial information of their partner's action. In this Auditory Only condition, participants only gained knowledge about the location of their partner's response via the presentation of either a high or low tone that corresponded to each target. The absence of an IOR effect in this condition suggests that some visuospatial information is needed to activate the processes leading to sIOR and, relatedly, the mere knowledge of a co-actor's response endpoint is not sufficient in activating these processes. Although the results lead to such a conclusion, it is possible that ambiguity in the auditory tone and location mapping may be the reason sIOR did not occur in this condition. That is, even though the specific tones were only presented when one of the buttons was contacted, the auditory tones that were presented may have been too ambiguous to sufficiently generate the knowledge of the co-actor's response to a certain location. Further, it is also possible that the sIOR effect did not emerge because, even though motor codes are thought to be efficiently activated by the presentation of the associated auditory effects, the time it took to convert the auditory information into the knowledge of partner's response location was longer than the time it took for the observer to plan, initiate, and execute the subsequent response. If the process of converting the auditory information into the knowledge of partner's response location took longer than planning and executing the subsequent action, then it is unlikely that this knowledge could affect the efficiency of the process of planning and initiating that action. Given these possibilities, a second experiment was conducted in which the auditory information that provided the knowledge of the coactor's response endpoint more directly (and thereby more efficiently) identified the target location. In Experiment 2, the location-contingent auditory tones were replaced with a male voice saying words that identified the target that was pressed. Directional words like “left” and “right” were avoided because of: 1) potential confusion over who's “left” and “right” to which the word referred; and, 2) the words themselves contain spatial information, a characteristic that we wanted to avoid. Instead, the target locations were surrounded by colours (one target was surrounded with blue and the other with green) and the auditory information that was presented was the colour word “blue” and “green” after the related button was contacted. The assumptions and predictions from Experiment 1 remain the same for Experiment 2; if knowledge of the co-actor's response is sufficient to activate the mechanisms that lead to IOR, then a sIOR effect will emerge in the

3.1.1. Participants Seven pairs of individuals (all female; aged 18–25 years) were recruited from the University of Toronto community to complete the study. All participants had self-reported right hand preference and normal, or corrected-to-normal, vision. All participants were naïve to the purpose of the experiment. All procedures and protocols taken were approved by the University of Toronto Research Ethics Board. Each participant provided written informed consent prior to data collection. 3.1.2. Apparatus, task, and procedures The apparatus, task, and procedures of Experiment 2 were essentially identical to those of Experiment 1. There were two important differences. First, the black metal response board that was used in Experiment 1 was slightly modified in that 6 cm × 6 cm squares of blue and green construction paper were cut and pasted on the board to surround the target buttons. A hole in the centre of the squares allowed the buttons to emerge unobstructed by the construction paper. The second important modification was the auditory stimuli presented to the participants after the button contact. Specifically, in Experiment 2 participants were not presented with a high or low tone when a target was depressed (as in Experiment 1), but were presented with a male voice saying the words “Blue” or “Green”. These auditory stimuli acted as the effect tones and were presented when the locations surrounded by the blue and green paper, respectively. This word/ location mapping was constant throughout the study to ensure the words acted as an absolute reference throughout the entire experiment. These sounds were presented through noise cancelling headphones. The headphones were used to present the colour words and the white noise masks while reducing environmental noise. The timing of the presentation of the target stimuli and the auditory stimuli matched that of Experiment 1. The headphones were linked to a common output such that the same set of auditory stimuli was presented to each participant on each trial throughout the entire experiment. Prior to data collection participants were notified of the response/ word mapping and were given practice trials to familiarize themselves with the task. Practice trials were given to the participants until they understood the effect tone mapping and the task. The experimental conditions immediately followed. A custom program in E-Prime (Version 2.0) was used to present all stimuli and record the dependent variables. 3.1.3. Data reduction and analysis RT and MT definitions and exclusion criteria matched that of Experiment 1. Although there were no incorrect button presses (i.e., the participant always contacted the correct target location), there were some trials in which the contact was not correctly recorded. Hence, data for trials were also excluded if the target button was not depressed properly. In these cases, the data for the next trial in the sequence was not excluded because the auditory information was still presented on each trial. An average of 16.16% of trials was excluded per participant for the above reasons. From the error free data, outliers were considered greater than the mean plus two standard deviations. The outlier analysis allowed for an average of 7.3% of trials to be excluded per participant. Statistical analysis matched that of Experiment 1. 3.2. Results and discussion The MT analysis revealed a main effect of Person, F(1, 13) = 8.57, p b .05, η2p = 0.397. Participants had shorter MTs when they followed

T.N. Welsh et al. / Acta Psychologica 153 (2014) 153–159

Table 2 Mean (and standard deviation) movement times in milliseconds as a function of response information, person, and target for Experiment 2. Response information

Full Auditory only

Same person

Different person

New target

Repeated target

New target

Repeated target

273 (60.4) 278 (73.5)

272 (59.4) 281 (77.5)

273 (60.1) 291 (74.0)

277 (61.3) 287 (75.6)

New Target

340

Reaction Time (ms)

their own response (mean = 276 ms, SD = 66.3) than when they followed their partner (mean = 282 ms, SD = 66.1). This main effect was superseded by a significant 3-way interaction between Response Information, Person, and Target, F(1, 13) = 5.06, p b .05, η2p = 0.28. Although this 3-way interaction emerged, post hoc analysis did not reveal any theoretically-relevant (i.e., repeated vs. new target) significant differences. That is, there were no differences between repeated and new targets when responding after their own response or their partner in either the Full or the Auditory Only condition. MT data for each condition are presented in Table 2. Analysis of the RT data revealed main effects for Response Information, F(1, 13) = 7.66, p b .05, η2p = 0.371, Person, F(1, 13) = 9.83, p b .05, η2p = 0.431, and Target, F(1, 13) = 36.45, p b .05, η2p = 0.737. The main effect of Response Information revealed that RTs were shorter when participants were in the Auditory Only condition (mean =262 ms SD = 39.3) compared to the Full condition (mean = 281 ms, SD = 39.6). The main effect of Person revealed that participants had shorter RTs when following their own response (mean = 266 ms, SD = 34.7) compared to when following the response of their partner (mean = 278 ms, SD = 41.3). Lastly, an overall IOR effect was revealed by the main effect of Target (Repeated Target RTs — mean = 281 ms, SD = 38.9; New Target RTs — mean = 262 ms, SD = 36.9). In addition to these main effects, there were significant Response Information by Person, F(1, 13) = 17.77, p b .05, η2p = 0.578, and Response Information by Target, F(1, 13) = 15.09, p b .05, η2p = 0.537, interactions. Post hoc analysis of the Response Information by Person interaction revealed that the facilitation effect (participants had shorter RTs following their own response) only appeared in the Full condition (mean = 32 ms, SD = 25.4); an opposite effect occurred in the Auditory Only condition where participants had shorter RTs following their partner's response (mean = − 8 ms, SD = 20.5). Further, post hoc analysis of the Response Information by Target interaction revealed that there were larger repeated/new target differences for the Full condition (mean = 27 ms, SD = 14.8) compared to the Auditory Only condition (mean = 12 ms, SD = 12.9). As in Experiment 1, however, these interactions do not directly address the main a priori predictions of whether or not RTs for repeated target and new target trials (i.e., IOR) were different within a given condition. For this reason, a closer examination of the 3-way interaction is needed. Although the 3-way interaction approached but did not surpass conventional levels of significance, F(1, 13) = 3.83, p N .05, η2p = 0.228, a post hoc analysis was completed in which RTs for repeated and new target locations were compared in each condition to determine if IOR was present in any of the conditions. Consistent with Experiment 1, significant iIOR effects were present in both Full and Auditory Only conditions. Critically, significant sIOR was only present in the Full condition. A small non-significant IOR effect was present when the participant followed their partner in the Auditory Only condition. This interaction is displayed in Fig. 2. In sum, the results of Experiment 2 are consistent with those of Experiment 1 and indicate that an IOR effect occurred in each condition except for when the participant did not witness the action and only heard the response location contingent auditory information (the Auditory Only condition). Based on the repeated pattern of results, it can be concluded that knowledge of the response location alone is not sufficient to activate the mechanisms of IOR. Conversely, it appears

157

320

Repeated Target

*

ns

*

*

Same Person

Different Person

Same Person

Different Person

300 280 260 240 220

Auditory Only

Full

Response Information Fig. 2. Mean reaction times (ms) as a function of Response Information, Person, and Target in Experiment 2 (colour words). Open bars represent reaction times for different target trials and filled bars represent reaction times for repeated target trials. * denotes significant differences between RTs on repeated and different target trials. SEM bars are shown.

that some visuospatial information about the partner's response is necessary for the activation of the mechanisms leading to sIOR. A more extensive treatment of these conclusions is provided in the following section. 4. General discussion The purpose of the present study was to determine if knowledge of a co-actor's response, generated via associated auditory information presented in the absence of any visual information, is sufficient to activate the inhibitory processes leading to sIOR. The results of the present study reveal that knowledge of the response alone is not sufficient to lead to sIOR. In both experiments, sIOR was present in the Full condition when participants saw the response and heard the auditory effect of their partner, but sIOR was not present in the critical Auditory Only condition in which participants did not witness their co-actor's response but received information about which button was contacted via a distinct auditory information. The data thus reveal that knowledge of the partner's response alone is not sufficient to generate sIOR and that some visuospatial information about the co-actor's action is necessary to activate the processes leading to sIOR. To recap, in Experiment 1, it was predicted that if the perception of the auditory effect can activate the representation of the response leading to that effect via common coding mechanisms (Elsner & Hommel, 2001), then sIOR might emerge when the observer heard the tone associated with a specific button press of their partner, even though they did not witness the action. The same predictions were made for Experiment 2 in which more direct information about the location of the response was provided to the participants because contact with a button generated the presentation of a location-specific colour word. It was predicted that sIOR might emerge under these circumstances because the response representation generated via perception of the auditory information could have been used to simulate the performance of the co-actor in the same way an observation-evoked representation would (as in the Full condition in the present experiments; see also Skarratt et al., 2010; Welsh et al., 2005, 2007, 2009). The data did not support these predictions. Thus, it appears as though perception of location-specific auditory after-effect information alone is not sufficient to generate the mechanisms of sIOR. There are at least three reasons that may explain why sIOR effects did not emerge in the Auditory Only conditions. First, it could be that action–effect binding did not occur in Experiment 1. Second, it could have been that the auditory effect associated with a co-actor's response did not generate knowledge of the response location or activate the

158

T.N. Welsh et al. / Acta Psychologica 153 (2014) 153–159

associated response representation. The third, and most likely, explanation is that the manner of knowledge of the response and response code activation did not provide the information necessary to generate the mechanisms underlying sIOR. That is, it is possible that the sIOR effect is dependent on the observers witnessing and representing the spatial aspects of how the action was executed. Before expounding the rationale for the preferred option, the first two possibilities will be addressed. Although there is no data in the present study to definitively indicate whether or not action–effect binding occurred in Experiment 1 or if a response representation was activated via the presentation of the auditory information, it is believed that these first two possibilities do not account for the absence of sIOR in the Auditory Only conditions for several reasons. First, for Experiment 1, it is very likely that action–effect binding occurred (a button–tone association was built) because participants were specifically instructed on the relationship between specific button contact and a tone. There is extensive research showing that such action–effect bindings occur even when participants are not specifically instructed about the relationship (i.e., an incidental action–effect binding; see for example Elsner & Hommel, 2001). Note also that Experiment 2 included an additional visual reference for each target location and the auditory effect was a word that was unambiguously related to that location. In addition, there is evidence that such action–effect associations can be developed through social settings where the observer views another person executing the response (Paulus, van Dam, Hunnius, Lindemann, & Bekkering, 2011). Finally, previous research from our laboratory has shown that action–effect associations can influence response initiation times in social action settings (joint Simon effect paradigm) when the effect information is visuospatial in nature (Kiernan, Ray, & Welsh, 2012) as predicted by ideomotor approaches to joint action (see van der Wel et al., 2013). Thus, action–effect binding can occur in individual and social settings, and the learned action–effect associations can influence response selection in joint action contexts. Based on all these considerations, it is intuited that action–effect binding occurred in the present study and that the perception of the auditory information generated by the co-actor's response activated a response representation, or at minimum knowledge of the co-actor's response location, in the observer. The third and most likely explanation for the absence of a sIOR effect in the Auditory Only condition is that spatial information regarding how the response was executed is the critical information needed to activate the processes leading to sIOR. This conclusion is based on a careful integration of the present findings with those of previous work. Recall that Skarratt et al. (2010) and Welsh et al. (2005, 2007) revealed that witnessing a visual transient (target onset) at the target location and contact with the target location is not necessary for sIOR because sIOR emerged when the observers only witnessed the initial portion of the movement. In addition, the study by Welsh et al. (2009) showed that the direction of the observed movement is important because it was found that RTs for responses in a similar direction to a previously observed response were longer than those in a different direction, even when the endpoints of the movements were different. Thus, there is converging evidence that witnessing endpoint contact maybe sufficient, but not necessary, to generate the sIOR effect. On the other hand, spatial information about how the action is completed is what seems to be the common feature necessary for the activation of the inhibitory processes underlying sIOR. In the present study, participants had knowledge about what button was contacted, but did not witness the visuospatial characteristics of the action that brought about the effect. Thus, the absence of a sIOR effect when the participant knew the end point, but did not see the action, is consistent with the previous work. The conclusion that knowledge of which button was contacted did not activate the processes leading to sIOR is also consistent with recent research indicating that the goal of the action does not influence sIOR. Specifically, Cole et al. (2012; cf. Ondobaka, de Lange, NewmanNorlund, Wiemers, & Bekkering, 2012) conducted a series of studies in

which participants reached out to one of two locations and performed either the same or a different action with the object located at that location (e.g., write or erase something with pencils located at the two locations). Cole et al. revealed that the goal of the task did not influence the emergence of magnitude of the sIOR effect. Participants in that study had longer RTs for targets at the same location as their partners' previous response, but the magnitude of the increase in RT for repeated target locations was similar when the pair of participants performed the same task (e.g., both wrote with the pencil) or different tasks (e.g., one erased and the other wrote). It is perhaps instructive to note here that the spatio-temporal characteristics of the response to the target location were also similar across the goals of the task, providing more support for the critical role of observing these response characteristics in generating the sIOR effect.1 Interestingly, a distinction has been made between how individuals may code and use the goal (higher-order) and sensorimotor (lower-order) aspects of the task in individual and social action contexts (e.g., Ondobaka et al., 2012; Pacherie, 2008). Converging evidence from the present and these recent studies suggest that the observationevoked sensorimotor representations, not the goal-evoked representations, are necessary for activating the inhibitory mechanisms leading to sIOR. The reliance of these mechanisms on sensorimotor representations with a distinct spatial component seems entirely consistent with the notion that the mechanisms underlying IOR itself are spatiallydriven and likely evolved to facilitate search patterns (e.g., Klein, 2000; Klein & MacInnes, 1999). That is, regardless of the outcome of the search of a particular location (i.e., identify and obtain a target vs. reject and ignore a non-target), it would be less efficient for an individual to search a location that they or their co-actors just searched than to search a new location. In sum, the results of the present study reveal that the presentation of the auditory effect information of an action does not activate the processes that lead to sIOR. This finding is consistent with previous work showing that sIOR is not dependent upon the observation of contact with the endpoint (e.g., Skarratt et al., 2010; Welsh et al., 2007, 2009) and that goal of the action does not influence the sIOR effect (Cole et al., 2012). These overall findings seem to fit with the notion that goal and sensorimotor aspects of a given movement are processed and utilized differently (Pacherie, 2008). In the case of sIOR, it seems that the activation of the mechanisms that lead to the effect is critically dependent on the activation of the sensorimotor processing channels that code the spatial characteristics of the action. Acknowledgements This research was supported by an Early Research Award from the Ontario Ministry of Research and Innovation (TW), a Discovery Grant (TW) and an Undergraduate Summer Research Award (JM) from the Natural Sciences and Engineering Research Council of Canada, and a Graduate Scholarship from the Canadian Institutes of Health Research (LM). References Cole, G.G., Skarratt, P.A., & Billing, R. -C. (2012). Do action goals mediate social inhibition of return? Psychological Research, 76, 736–746. Elsner, B., & Hommel, B. (2001). Effect anticipation and action control. Journal of Experimental Psychology: Human Perception and Performance, 27, 229–240. Kiernan, D., Ray, M., & Welsh, T.N. (2012). Inverting the joint Simon effect by intention. Psychonomic Bulletin & Review, 19, 914–920. Klein, R.M. (2000). Inhibition of return. Trends in Cognitive Sciences, 4, 138–147.

1 A short comment regarding the examination of the role of “goal” compatibility in the sIOR effect. Note that “goal” has been operationally defined by researchers as the final task that participants complete with the object. Although this approach is a valid and appropriate in the context of the research, the goal of the task could be defined differently by the participant such as, perhaps, simply the reaching and/or the grasping of an object at a location in space regardless of the subsequent task performed at the end.

T.N. Welsh et al. / Acta Psychologica 153 (2014) 153–159 Klein, R.M., & MacInnes, W.J. (1999). Inhibition of return is a foraging facilitator in visual search. Psychological Science, 10, 346–352. Kohler, E., Keysers, C., Umiltà, M.A., Fogassi, L., Gallese, V., & Rizzolatti, G. (2002). Hearing sounds, understanding actions: Action representation in mirror neurons. Science, 297, 846–848. Maylor, E.A., & Hockey, R. (1985). Inhibitory component of externally controlled covert orienting in visual space. Journal of Experimental Human Perception & Performance, 11, 777–787. Melcher, T., Weidema, M., Eenshuistra, R.M., Hommel, B., & Gruber, O. (2008). The neural substrate of the ideomotor principle: An event-related fMRI analysis. NeuroImage, 39, 1274–1288. Melcher, T., Winter, D., Hommel, B., Pfister, R., Dechent, P., & Gruber, O. (2013). The neural substrate of the ideomotor principle revisited: Evidence for asymmetries in action– effect learning. Neuroscience, 231, 13–27. Ondobaka, S., de Lange, F., Newman-Norlund, R., Wiemers, M., & Bekkering, H. (2012). Interplay between action and movement intentions during social interaction. Psychological Science, 23, 30–35. Pacherie, E. (2008). The phenomenology of action: A conceptual framework. Cognition, 107, 179–217. Paulus, M., van Dam, W., Hunnius, S., Lindemann, O., & Bekkering, H. (2011). Action–effect binding by observational learning. Psychonomic Bulletin & Review, 18, 1022–1028.

159

Posner, M.I., & Cohen, Y. (1984). Components of visual orienting. In H. Bouma, & D. Bowhuis (Eds.), Attention and performance (pp. 531–556). Hillsdale, New Jersey: Erlbaum. Rizzolatti, G., & Craighero, L. (2004). The mirror–neuron system. Annual Review of Neuroscience, 27, 169–192. Skarratt, P.A., Cole, G.G., & Kingstone, A. (2010). Social inhibition of return. Acta Psychologica, 134, 48–54. van der Wel, R.P.R.D., Sebanz, N., & Knoblich, G. (2013). Action perception from a common coding perspective. In K. Johnson, & M. Schiffrar (Eds.), People watching: Social, perceptual and neurophysiological studies of body perception (pp. 101–119). New York: Oxford University Press. Welsh, T.N., Elliott, D., Anson, J.G., Dhillon, V., Weeks, D.J., Lyons, J.L., et al. (2005). Does Joe influence Fred's action? Inhibition of return across different nervous systems. Neuroscience Letters, 385, 99–104. Welsh, T.N., Lyons, J., Weeks, D.J., Anson, J.G., Chua, R., Mendoza, J., et al. (2007). Withinand between-nervous-system inhibition of return: Observation is as good as performance. Psychonomic Bulletin & Review, 14, 950–956. Welsh, T.N., McDougall, L.M., & Weeks, D.J. (2009). The performance and observation of action shape future behaviour. Brain and Cognition, 71, 64–71. Welsh, T.N., & Pratt, J. (2006). Inhibition of return in cue–target and target–target tasks. Experimental Brain Research, 174, 167–175.