The role of action prediction and inhibitory control for joint action coordination in toddlers

Journal of Experimental Child Psychology 139 (2015) 203–220

Contents lists available at ScienceDirect

Journal of Experimental Child Psychology journal homepage: www.elsevier.com/locate/jecp

The role of action prediction and inhibitory control for joint action coordination in toddlers M. Meyer ⇑, H. Bekkering, R. Haartsen, J.C. Stapel, S. Hunnius Donders Institute for Brain, Cognition, and Behaviour, Radboud University Nijmegen, 6500 HE Nijmegen, The Netherlands

a r t i c l e

i n f o

Article history: Received 28 January 2015 Revised 8 June 2015 Available online 3 July 2015 Keywords: Joint action development Action prediction Inhibitory control Early childhood Social interaction Eye-tracking

a b s t r a c t From early in life, young children eagerly engage in social interactions. Yet, they still have difficulties in performing well-coordinated joint actions with others. Adult literature suggests that two processes are important for smooth joint action coordination: action prediction and inhibitory control. The aim of the current study was to disentangle the potential role of these processes in the early development of joint action coordination. Using a simple turn-taking game, we assessed 2½-year-old toddlers’ joint action coordination, focusing on timing variability and turn-taking accuracy. In two additional tasks, we examined their action prediction capabilities with an eye-tracking paradigm and examined their inhibitory control capabilities with a classic executive functioning task (gift delay task). We found that individual differences in action prediction and inhibitory action control were distinctly related to the two aspects of joint action coordination. Toddlers who showed more precision in their action predictions were less variable in their action timing during the joint play. Furthermore, toddlers who showed more inhibitory control in an individual context were more accurate in their turn-taking performance during the joint action. On the other hand, no relation between timing variability and inhibitory control or between turn-taking accuracy and action prediction was found. The current results highlight the distinct role of action prediction and inhibitory action control for the quality of joint action coordination in

⇑ Corresponding author. E-mail address: [email protected] (M. Meyer). http://dx.doi.org/10.1016/j.jecp.2015.06.005 0022-0965/Ó 2015 Elsevier Inc. All rights reserved.

204

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

toddlers. Underlying neurocognitive mechanisms and implications for processes involved in joint action coordination in general are discussed. Ó 2015 Elsevier Inc. All rights reserved.

Introduction Social interactions shape our everyday life from birth. Central to social interactions is the way we coordinate our actions with others. Accordingly, action coordination is an essential element for joint actions that are defined as ‘‘any form of social interaction whereby two or more individuals coordinate their actions in space and time to bring about a change in the environment’’ (Sebanz, Bekkering, & Knoblich, 2006, p. 70). The quality of joint action coordination (i.e., interpersonal coordination of actions in the context in which the joint action occurs) largely determines the success of a joint action. Moreover, fluent coordination between people is associated with a feeling of connectedness and liking between the interaction partners (Marsh, Richardson, & Schmidt, 2009). Thus, learning to coordinate one’s own actions skillfully with others enhances multiple aspects of joint actions such as success and affiliation. There are many examples of daily joint actions that illustrate how naturally and smoothly adults coordinate their actions with others, including setting a table together, passing a cup to another person, and playing an instrument together. Common joint actions like these have been the subject of recent behavioral and neuroimaging studies (Egetemeir, Stenneken, Koehler, Fallgatter, & Herrmann, 2011; Loehr, Large, & Palmer, 2011; Ray & Welsh, 2011). Many of these joint actions involve taking turns with another person, for example, tossing a ball back and forth, beating a drum in turns (e.g., Warneken, Chen, & Tomasello, 2006), and having a conversation with another person. Coordinating actions jointly, however, is nontrivial because it requires integrating another person’s actions into one’s own action processing. Whereas adults display refined and proficient joint action coordination, young children still have considerable difficulties in coordinating their actions with others (for a review, see Brownell, 2011). In the current study, we examine these difficulties in children’s joint action coordination and aim to provide insights into possible underlying processes of joint action coordination during early childhood. Development of joint action coordination During their first 3 years of life, young children rapidly improve in coordinating their actions with others (Brownell, 2011). Although infants are sensitive to violations in simple turn-taking interactions with adults from 2 months (Adamson & Frick, 2003), they show hardly any form of fluent coordination with others until their first birthday unless they are scaffolded by an adult (e.g., Mendive, Bornstein, & Sebastian, 2013). Around the age of 18 to 24 months, toddlers become more successful in simple coordination tasks with adults (Warneken et al., 2006) and peers (Brownell & Carriger, 1990; Brownell, Ramani, & Zerwas, 2006). They can, for instance, simultaneously act on toys to achieve a goal together with others by the age of 2 years (Brownell & Carriger, 1990). Although there are some indications that toddlers can coordinate actions with others by 2 years, successful performance seems to be mainly restricted to tasks with low coordination demands (i.e., tasks that do not require continuous action coordination). When continuous action coordination with another person is needed (i.e., coordination exceeding a single incident of coordination to more frequent instances following each other during ongoing coordination), it takes yet another year before toddlers reach a level of joint action coordination that is similar to the level they achieve in individual coordination (e.g., Meyer, Bekkering, Paulus, & Hunnius, 2010). Whereas in individual coordination actions are coordinated intrapersonally (e.g., coordinating left and right hands), joint action coordination involves coordination between people. In a recent behavioral study, 2½- and 3-year-old children were tested in a simple coordination game

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

205

during which individual and joint action coordination were compared (Meyer et al., 2010). This game required alternately pushing two buttons—either bimanually or jointly in turns with an adult partner. To determine the coordination quality, it was measured how stable the timing of children’s button presses was (timing variability) and how often children erroneously pushed a button more than once in a row (turn-taking accuracy). The findings indicated that 2½- and 3-year-olds were equally stable in their timing and equally accurate in their button pushing when they were acting on their own. During joint play, however, it became apparent that the 2½-year-olds still had difficulties in coordinating their actions with another person. The 3-year-olds, on the other hand, reached comparable levels of action coordination for individual and joint play. These findings indicate a significant discrepancy in performance that was specifically evident for joint action coordination. The question arises which processes underlie the 2½-year-olds’ difficulties in coordinating their actions with another person. Establishing smooth joint action coordination: The role of action prediction and inhibitory action control Findings from previous research (cf. Brownell, 2011; Moriguchi, 2014; Sebanz & Knoblich, 2009) suggest that two processes might be especially important for coordinating actions with another person: the ability to predict others’ actions and the ability to control one’s own actions in concord with others’ actions. Predicting the action partner’s next action is necessary to adjust one’s own action planning and execution flexibly (see Sebanz & Knoblich, 2009, for a review). For instance, when passing a ball back and forth, one can readily prepare to catch the ball by predicting the other’s next throw. Apart from action prediction, controlling one’s own actions is essential for joint action coordination. This is particularly evident in turn-taking actions, when inhibitory control over one’s own actions is needed to refrain from acting when it is the other’s turn (Sebanz et al., 2006). To our knowledge, both processes and their respective contribution to smooth joint action coordination have been studied only separately (see, e.g., Sebanz & Knoblich, 2009, for action prediction). Although both processes are relevant in the context of joint action (Brownell, 2011; Sebanz & Knoblich, 2009), their unique role for one and the same joint action has not been investigated to date. Combining measurements of both processes to predict their distinct role for joint action coordination promises to be informative for understanding the development of joint action coordination during early childhood. Besides this, combined measurements offer insight into the contribution of the two processes for joint action coordination in general. Distinct neurocognitive mechanisms have been associated with action prediction and inhibitory action control. Different accounts propose that predictions of others’ actions rely on forward models in the fronto-parietal action network of the brain (Kilner, Friston, & Frith, 2007; Miall, 2003; Wolpert, Doya, & Kawato, 2003). By matching the observed actions of another person onto one’s own motor system, predictions with respect to upcoming actions can be derived (Kilner et al., 2007; Miall, 2003; Wolpert et al., 2003). Thus, analoguous to involvement of one’s own motor system during observation of another person’s actions (see, e.g., Rizzolatti & Craighero, 2004), the motor system is involved in the prediction of another person’s next actions. Neuroimaging studies have investigated the predictive function of the action system and found supporting evidence in infants (Southgate, Johnson, Karoui, & Csibra, 2010; Stapel, Hunnius, van Elk, & Bekkering, 2010) and adults (Kilner, Vargas, Duval, Blakemore, & Sirigu, 2004). In a recent electroencephalography (EEG) study, Kourtis, Sebanz, and Knoblich (2013) expanded previous findings to a joint action context. They demonstrated that predictive activation of one’s own neural motor system for an upcoming action of a partner was closely related to the proficiency of the joint action performance. Yet, it is unclear what role the ability to predict others’ actions plays in the development of joint action coordination with others. In particular, it is an open question whether individual differences in action prediction more generally (thus, beyond a particular action context) play a role in early joint action coordination. Besides this, it is not yet understood which role inhibitory action control plays for joint action coordination in toddlers. A large body of research investigated the ability to inhibit responses at different ages (e.g., Diamond, 1990; Durston et al., 2002; Jones, Rothbart, & Posner, 2003; Munakata et al., 2011). Among others, Go/No-Go tasks have been used to investigate inhibitory control (e.g., Cragg, Fox, Nation, Reid, & Anderson, 2009). Converging findings highlight the structural and functional aspects of prefrontal brain regions playing a role in this type of inhibitory control of actions (Casey et al., 1997;

206

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

Munakata et al., 2011). Often, joint action situations share the same characteristics with Go/No-Go tasks; for instance, turn-taking interactions require refraining from acting when it is the action partner’s turn. In terms of development, many findings illustrate immaturity in inhibitory control early in life and a general improvement during childhood and up to adolescence (Jones et al., 2003; Williams, Ponesse, Schachar, Logan, & Tannock, 1999). Therefore, the question arises how the immaturity of inhibitory action control might affect toddlers’ coordination with an interaction partner. Studying children during toddlerhood, when an immaturity of inhibitory control can be expected, offers a unique possibility to learn more about the role of inhibitory control for joint action coordination in general. The current study The aim of the current study was to disentangle the role of action prediction and inhibitory action control for young children’s joint action coordination. For this purpose, we assessed the joint action coordination performance of 2½-year-old children by measuring timing variability and turn-taking accuracy during joint play similar to the previous study (Meyer et al., 2010). Furthermore, we measured action prediction and inhibitory action control in two separate tasks. To assess toddlers’ precision in predicting others’ actions, we measured their eye movements in an action observation task. The reason for measuring toddlers’ prediction skills in a separate task from the joint action was to allow for broader implications of prediction skills (i.e., beyond the scope of task-specific prediction skills) for early joint action coordination. Recently, eye-tracking measures of anticipatory gaze have been shown to reflect prediction processes of the motor cortex (Elsner, D’Ausilio, Gredebäck, Falck-Ytter, & Fadiga, 2013). Furthermore, we tested toddlers’ inhibitory action control by using an inhibitory task (gift delay [bow]; Carlson, 2005; Kochanska, Murray, Jacques, Koenig, & Vandegeest, 1996). Because predicting others’ actions and controlling one’s own actions are crucial processes for smooth joint action coordination in adults, we hypothesized that both might play a role in the limited coordination quality described in 2½-year-old toddlers (Meyer et al., 2010). More precisely, we investigated whether individual differences in action prediction and inhibitory action control were distinctly related to the two measures of joint action coordination (timing variability and turn-taking accuracy). We expected that action prediction abilities would be beneficial for predicting the partner’s actions during the joint action and would in turn allow timing one’s own actions in a more stable way with respect to the partner (cf. Kourtis et al., 2013; Vesper, van der Wel, Knoblich, & Sebanz, 2011). More specifically, predicting the other’s next action precisely in time allows preparing one’s own next action accordingly. Preparing and executing one’s own next action in a stable manner leads to reduced variability in timing, which in turn makes oneself more predictable for the interaction partner (see also Vesper et al., 2011). Individuals who precisely predict other people’s actions, thus, are likely to be more stable in timing their own actions in coordination with another person. Therefore, we expected that toddlers who are better at predicting others’ actions would also show less timing variability in their joint action coordination. In line with previous findings (Carlson, Mandell, & Williams, 2004), we further assumed that toddlers’ inhibitory control would relate to their turn-taking accuracy during the joint action (i.e., refraining from pushing when it is not their turn). To inhibit a prepotent response in interaction with others (e.g., not to speak before others are finished talking), it requires control over one’s own actions. Acting when it is another person’s turn due to immature inhibitory control at best impedes the joint action coordination and might even prevent a joint action from being successful entirely. Therefore, we expected children with better inhibitory control to make fewer errors in turn taking during joint play. Method Participants In total, 25 2½-year-old toddlers (18 boys) with a mean age of 29.9 months (SD = 10 days) were included in the final sample. The toddlers were recruited from a database of families who signed up for participation in child research at the Baby Research Center in Nijmegen, The Netherlands. Parents accompanied their children to the testing session and provided written informed consent

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

207

for their participation in the study. Another 20 toddlers were tested but excluded from the final data set because they did not provide sufficient data on all of the three behavioral tasks (n = 18; for details on inclusion criteria, see ‘‘Data processing’’ section) or due to parental interference (n = 1) or technical failure (n = 1) during testing. The data sample is part of a larger study as reported by Stapel, Hunnius, Meyer, and Bekkering (submitted for publication). Design We implemented three behavioral tasks in a within-participants design: a joint action coordination task, an action observation task, and an inhibitory control task. In the joint action coordination task, we recorded the toddlers’ actions during a simple turn-taking game as used in earlier joint action experiments with toddlers (Meyer et al., 2010; Meyer, Hunnius, van Elk, van Ede, & Bekkering, 2011). In accordance with previous measures of joint action coordination, we used timing variability and turn-taking accuracy to quantify the quality of the joint action coordination. By adopting the definition of joint action by Sebanz and colleagues (2006), we use joint action as a more broadly defined concept that does not entail the understanding of a joint goal or intention as a necessary criterion. Therefore, no conclusions can be drawn regarding the toddlers’ joint commitment or shared intentions during the joint action. In the action observation task, we assessed predictive and reactive eye movements with respect to the tracking of a moving actor who disappeared and then reappeared behind an occluder (cf. Elsner et al., 2013). Previous research has shown that infants as young as 6 months represent an occluded object and are able to predict movements of shortly occluded objects accurately in time (Gredebäck & von Hofsten, 2004).Timing of the eye movements was used as a measure of the toddlers’ temporal precision in predicting the observed actions. In the inhibitory control task, the classic gift delay [bow] task was adopted to assess how long the toddlers could resist looking into a gift box (see Carlson, 2005, for a review). Success in this inhibitory control task was used as dependent measure of the children’s inhibitory action control. The purpose of having the three behavioral tasks in one testing session was to ensure that the data were as comparable as possible and to avoid variability in separate measurement sessions induced by other coincidental factors (e.g., mood of the child) or developmental changes that might occur within days or weeks between measurements. Materials and stimuli Joint action coordination task The computerized setup of the joint action game consisted of a cartoon figure that needed to be moved up a ladder on a screen by pushing two buttons in alternation (see Fig. 1). The two buttons were differently colored (black and red) and connected via a tilt mechanism such that only one button at a time could be pushed down. They were placed in front of a wide screen (1200  1920 pixels) that was rotated by 90 degrees to maximize the length of the displayed ladder and, thereby, the trial length. The displayed ladder was leading up to a cartoon pig on a cloud in the upper right corner. On the foreground, a cartoon frog was shown and could be moved up the ladder toward the goal (i.e., the pig on the cloud) by pushing the two buttons alternately. Button presses elicited unique visual and auditory effects; pressing the left button moved the left leg of the frog to the next higher ladder step, whereas pressing the right button moved the right leg. Button presses triggered an auditory clapping tone unique for each button (duration = 60 ms). The ladder consisted of 42 steps such that 21 alternating presses per button were required to reach the goal. When the top was reached, a picture of both cartoon figures together on the cloud indicated that the goal was achieved. The software used to program the computerized game was Python (Version 2.5, Python Software Foundation, http:// www.python.org). During the joint action game, the timing of the button presses was registered for both buttons separately. Action observation task As stimulus material, we used short movie clips displaying infant actors moving across a room from one side to the other (i.e., from left to right and from right to left). In the movie clips, a part of the scene was visually blocked with a black occluder area (290  396 pixels) such that the moving actors

208

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

joint turn-taking game

Inhibitory Control Task

Fig. 1. The experimental setup consisted of three tasks tested in a within-participants design. Left: In the joint action coordination task, toddlers pushed buttons in alternation with an adult action partner to move up a cartoon figure displayed on the screen. Right top: In the action observation task, toddlers were presented with short movie clips in which an actor moved through the scene. A black area occluded the moving actor for part of the presentation. Children’s eye movements were recorded during the movie presentation. Right bottom: In the inhibitory control task, the classic gift delay task (Kochanska et al., 1996) was adopted. In this task, toddlers received a gift box containing a gift. They were not allowed to look inside the gift box for 3 min. The picture displays a failed attempt.

disappeared behind the occluder (see Fig. 1). The time from stimulus onset to full occlusion was fixed (55 frames). Occlusion duration ranged between 290 and 720 ms and was varied to ensure that predictions could not be based on fixed duration between stimulus onset and reappearance of the actor. The total duration of each movie was approximately 4 s (73–112 frames, 25 frames per second). The displayed infant actors were moving by either walking or crawling through the scene. Eight unique movie clips (with different actors) per locomotion type (walking and crawling) were presented. These stimuli belonged to a subset of movie clips created for an eye-tracking study on action prediction (see Stapel et al., submitted for publication).1 The stimulus set also comprised movie clips displaying objects moving with constant speed from one side of the screen to the other side. Analogous to the human movement clips, the moving objects disappeared and reappeared behind an occluder. The different durations were matched across conditions. Although the main focus of this study was on the observation and prediction of human actions (walking and crawling), gaze behavior in response to moving objects was explored in additional analyses (see ‘‘Data processing’’ and ‘‘Statistical analysis’’ sections). In total, children were presented with two randomized blocks of 24 movie clips (8  walking, 8  crawling, 8  object) interleaved with short clips (2 s) of moving pictures to keep the children attending to the screen. The stimuli were presented on a Tobii 1750 eye-tracker (Tobii Technology, Stockholm, Sweden) that was used to record the toddlers’ eye movements.

Inhibitory control task For the inhibitory control task, we used a small gift (a teddy bear) hidden in a gift box (width = 21.5 cm, height = 15 cm, depth = 15 cm) decorated with wrapping paper (see Fig. 1). The lids of the gift box could be closed tightly with a bow. When no bow was present, the lids were only loosely closed but were still hiding the gift from sight. The gift box containing the gift (but initially lacking the bow) was placed in front of the child, who was sitting on a chair at the table. With the excuse to get the missing bow, the experimenter left the room for 3 min (see ‘‘Procedure’’ section). 1 Stapel and colleagues (submitted for publication) compared prediction performance of differently aged participants to examine the role of motor experience for action prediction. The participants in the current study were engaged only in the eye-tracking paradigm and in no other measurements by Stapel and others.

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

209

Video recordings of the child’s behavior during the joint action coordination task and the inhibitory control task were made using a digital video camera (Sony Handycam, DCR-SR190E). Procedure A brief warm-up period (10 min) preceded the actual testing session. During this warm-up period, two experimenters (E1 and E2) were present. The parent was informed about the experimental procedures and instructed not to interfere during the testing session. Next, E1 accompanied the child and the parent to the testing rooms. The study was conducted in two separate testing rooms and had a fixed order of tasks (first: action observation task; second: joint action coordination task; third: inhibitory control task). The experimental session started with the action observation task in a testing room with the eye-tracker. The reason for starting the experimental session with the action observation task was two-fold. It allowed the toddlers to get acquainted with the new environment and the experimenter, who would later become their joint action partner. In addition, this task required the toddlers to sit and watch movie clips for approximately 4 min. By presenting this task first, we opted for an order in which children were most likely to finish the full set of trials (i.e., 4 min of movie clips). After the action observation task, children and parents were accompanied to the adjacent testing room in which the joint action coordination task, and subsequently the inhibitory control task, took place. The inhibitory control task was chosen to end the session because receiving a gift for the participation in the study naturally finalizes the testing session. In total, the testing session took approximately 40 min. Action observation task During the action observation task, the child was asked to sit on the lap of the parent in front of the eye-tracker. The child and parent were seated at a distance of approximately 60 cm from the screen. A calibration phase preceded the actual measurement phase. The toddler was presented with nine contracting and expanding calibration points arranged in a 3  3 grid. Subsequently, the toddler was told that he or she would see a set of short movie clips and was asked to watch them carefully. Joint action coordination task In the second testing room, the joint action game setup was installed on a table (see Fig. 1). Again, the child was asked to sit on the lap of the parent in front of the screen. E1 sat to the left of the child such that they were seated centrally in front of two buttons on the table. To introduce the goal of the joint action game, E1 explained that the moving character (frog) at the bottom of the scene wanted to climb up the ladder to reach the other character (pig) at the top of the screen (see Fig. 1). The joint action coordination task started with a demonstration phase in which E1 demonstrated the joint action game together with E2. Then, the actual testing phase was initiated by E1, who invited the child to play jointly together with her by saying, ‘‘Shall we now play the game together?’’ During play, E1 received metronome sounds via a hidden earphone to standardize her actions to the child’s response. The metronome sound consisted of three consecutive beep tones locked to the child’s button press. The last beep was presented exactly 1 s after the child’s button press. Thus, E1 could time her press with respect to the last beep tone and, thereby, could standardize her actions to the child’s presses. Moreover, E1 was instructed to encourage the child during the game by saying ‘‘good’’ and ‘‘well done’’ and to remind the child to use his or her own right hand on the right button whenever the child would not do so. In addition, the parent was asked to refrain from engaging during the game so as not to influence the child. For each round of the game (i.e., to bring the cartoon figure up to the top), E1 and the child needed to continuously coordinate their actions for a total of 42 button presses. E1 strived to play the joint action game with the child for at least two rounds. Inhibitory control task After the joint action game, E1 initiated the inhibitory control task by asking the parent to sit at the opposite end of the room with E2 to fill out a form. The child remained seated at the table. The parent was instructed not to interact with the child during this period. Throughout the inhibitory control task, the parent and E2 stayed in the room with their backs turned to the child. While the parent

210

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

started filling out a form, E1 announced that she had a present for the child. E1 then placed the gift box on the table in front of the child and said that she had forgotten to put the bow on the gift box. Before leaving the room to get the missing bow, E1 instructed the child not to look inside the gift box. Then, E1 left the room for 3 min. The child was video-recorded during the absence of E1. After 3 min, E1 returned with the bow and fixed it to the gift box (if the present was still inside). Finally, the child was allowed to open the gift box to retrieve the toy. The testing session finished with a debriefing of the parent with regard to the purpose of the study. Data processing To answer our current research question, we analyzed four dependent measures: the timing variability and turn-taking accuracy in the joint action coordination task, the timing of eye movements closest to the reappearance of the observed actor behind the visual occluder in the action observation task, and the latency to peek at the gift in the inhibitory control task. Joint action coordination task Based on previous measures of coordination quality in toddlers (Meyer et al., 2010), we assessed the timing variability and turn-taking accuracy from the button press data and the video recordings made during the joint action game. The data processing was identical to the processing applied previously (Meyer et al., 2010; note that turn-taking accuracy was previously termed performance accuracy). The first two button presses of each trial were excluded from the data to prevent outliers due to initiating a trial. Furthermore, video recordings were analyzed to exclude data in which a button press was caused by rule violations (e.g., the child using the parent’s hand to push the button) other than the registered errors (i.e., the child pushing the button more than once in a row). Button presses during which the parent interfered actively (e.g., by leading the child’s hand) were also excluded from further analysis. Together, for all participants (i.e., including participants who are not in the final data), 27% of all trials were excluded as violating the task rule based on the video coding. The interrater reliability was found to be excellent (Cohen’s kappa = .97). To investigate timing variability and turn-taking accuracy, only button presses in which children pressed the right button with their right hand were included. This was done to be consistent with previous research (Meyer et al., 2010), in which button presses of the right button performed with the right hand were compared between a unimanual joint action and a bimanual individual action coordination task (left hand and bimanual button presses on the right button are part of the 27% excluded trials). For all children, the first two rounds of playing were analyzed as we strived for finalizing at least two rounds during testing (see ‘‘Procedure’’ section). Turn-taking accuracy. The correct execution of the joint action game consisted of E1 and the children pressing the left and right buttons in turns. Consequently, children pressing the right button more than once in a row indicated an action coordination error. The mean percentage of these errors was calculated per child. Timing variability. We also calculated the children’s variability in pressing their button with respect to the button press of their joint action partner (E1). Invalid button presses (i.e., 27% of all trials) and registered errors (i.e., 14% of the remaining trials) were excluded before determining the timing variability. Even though this resulted in a relatively high exclusion of trials, children still provided on average 37 trials (SD = 8) for the analysis of timing variability. The basis of the timing variability measure was the interval between the last left button press by E1 and the right button press of the children (right minus left). An average time interval per child was calculated. The variability in timing was then computed using the coefficient of variation (COV), which is calculated by dividing the standard deviation by the average time interval. To prevent a potential bias caused by differences in children’s average time interval, we used the COV instead of the standard deviation as a measure of variability. Higher values of the COV indicate more variability; for instance, a value of 1 indicates that the standard deviation equals the mean. In smooth joint action coordination between adults, it was found that individuals perform in a temporally stable manner to make themselves more predictable for the interaction

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

211

partner (Vesper et al., 2011). The COV reflects children’s temporal stability in coordinating their actions with respect to their partner’s actions.

Action observation task We defined the part of the stimulus movie where the actor or object reappeared from behind the occluder as the spatial area of interest (AoI) in which fixations (minimum required gaze period of 40 ms and maximum allowed distance of 30 pixels radius) were considered for further analysis. Data from the first 200 ms in each trial were discarded to exclude potential gaze that might still be lingering at a location determined by the previous stimulus movie. Any fixation in the AoI after the first 200 ms until the end of the movie presentation was taken into account (this differs from Stapel et al., submitted for publication, who cut off their data window for their comparisons between conditions). From these fixations, those with the smallest difference with respect to the moment of reappearance of the actor or object were selected for each trial per participant. Negative values indicate the closest fixation to be before the actual reappearance of the actor or object, whereas positive values represent the closest fixation to be after the actual reappearance. The average time of the selected fixations over trials was calculated per participant and collapsed over human actions (walking and crawling) to be entered as an action prediction measure in the comparison between tasks.2 Additional analyses were used to investigate the gaze behavior to human actions and to nonbiological object movements in their relation to joint action coordination (see ‘‘Statistical analysis’’ section for more details). The fixation closest to the actual reappearance was taken as the index of temporal precision of predicting the moving actor or object to assess how precise gaze arrives at around the time of reappearance (see Bennett & Barnes, 2006, for information about optimal ways to track observed movements).3 The closer in time a fixation to the AoI was made in relation to the actual reappearance, the more precise the prediction was. In addition, fixations were considered predictive if they were before and up to 200 ms after the actual reappearance of the actor or object (values < 200 ms) or were considered reactive when they occurred later (values > 200 ms) due to the processing lag of the oculomotor system (see Gredebäck, Johnson, & von Hofsten, 2009). The eye-tracking analysis was identical to the analysis performed by Stapel and colleagues (submitted for publication) unless explicitly mentioned.

Inhibitory control task Analogous to Kochanska, Murray, and Harlan (2000), we analyzed the time children managed to refrain from peeking into the gift box during E1’s absence. Based on the video recordings, the start of E1’s absence was defined as the moment E1 closed the door behind her. From the start of E1’s absence, the analysis period lasted for 3 min. Peeking at the gift within this analysis period was defined as the moment in time children looked into the gift box. At least one of two criteria needed to be fulfilled to ensure that the children looked into the gift box: (a) the position of the children with respect to the gift box needed to allow them to be able to look inside the box (i.e., usually achieved by the children climbing on the chair) or (b) the children needed to make an effort by stretching out and/or by touching and tilting the box to peek at it. If the first moment in time fulfilling these criteria was within the first 3 min (i.e., the toddlers peeked into the gift box), it was coded as failing the task. If the children did not fulfill these criteria within 3 min or never looked into the gift box at all, this was coded as succeeding. Success and failure in the inhibitory control task were used as measures of inhibitory action control (see online supplementary material for a time-resolved measure of the inhibitory action control task). 2 Given the strong positive relationship of processing walking and crawling actions (r = .503, p = .01) and the associated risk of multicollinearity between the two predictors (violating the assumptions of multiple regression models), values for gaze behavior to walking and crawling were collapsed for future statistical comparisons. 3 Often eye-tracking studies investigating predictive gaze use the first fixation to a specific AoI as a measure of prediction (e.g. Falck-Ytter et al., 2006; Gredebäck & Melinder, 2010). Although the first gaze to a specific location gives a good indication of a target prediction in space, it typically requires a target to be present at all times in the visual scene (e.g., a bucket in which objects are being placed; Falck-Ytter et al., 2006), which was not the case here where the focus was on timing accuracy rather than the speed of predictions.

212

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

Statistical analysis To investigate whether individual differences in action prediction and inhibitory action control measures were related distinctly to the two joint action coordination measures, two multiple regression models were computed separately. Only datasets providing data for all four measures were included in the final data set (data points exceeding 2 standard deviations in any measure were excluded; n = 2). To test the unique contribution of each of the individual processes (action prediction and inhibitory action control) to predicting the joint action coordination measures, we entered the variables simultaneously (i.e., forced entry) as predictors in a multiple regression model with timing variability and turn-taking accuracy as outcome variables. Besides this main analysis, we tested whether the role of predictions for joint action timing is specific to predicting human actions or whether it holds for predictions in general (e.g., in response to moving objects). For this purpose, we investigated the toddlers’ predictions of human actions compared with nonbiological movements and to what extent those related to children’s timing variability during the joint play. We hypothesized a specific contribution of action prediction for toddlers’ timing variability during joint play. To test this, we used a multiple regression model to investigate the unique contribution of human actions and nonbiologically moving objects for the timing variability during joint action coordination. For completeness, correlations between all individual variables are provided in the supplementary material. All calculations and statistical tests were carried out using MATLAB (Version 7.8, The MathWorks) and SPSS 19.0.

Results During the joint action game, toddlers made on average 6.4% turn-taking errors (SD = 4.1) and their COV was on average .53 (SD = .21), which indicates that the standard deviation in toddlers’ timing of button presses was approximately half their mean button pressing time. In the action observation task, toddlers looked at the area in which the actors reappeared on average 122 ms after the actual reappearance of the actors. Based on the processing lag of the oculomotor system (i.e., 200 ms; see Gredebäck et al., 2009), 17 toddlers can be regarded as showing on average predictive gaze patterns, and 8 toddlers as showing reactive gaze patterns, to the observed actions. In the inhibitory control task, 10 toddlers failed at inhibiting their actions (i.e., peeked into the gift box within the first 3 min) and 15 toddlers succeeded in the task (i.e., never peeked into the gift box). To address the main hypotheses, we applied two regression models, one for each of the two joint action coordination measures, with as predictors the eye-tracking measure of the toddlers’ temporal precision in action prediction and the behavioral measure of inhibitory action control. Fig. 2 illustrates the results. In both models (see Tables 1 and 2), the linear combination of action prediction and inhibitory control measures was significantly related to the respective outcome variables (i.e., timing variability and turn-taking accuracy) (both ps < .05). As shown in Table 1, action prediction significantly predicted toddlers’ timing variability in the joint play (b = .506, p < .05), thereby contributing significantly to the model. No indications were found for a unique contribution of inhibitory control to explain toddlers’ timing variability (b = .154, p > .05). Thus, the results show a strong positive relationship between predictive gaze to human actions and timing of the toddlers’ actions during joint play. The opposite data pattern was found for the model of toddlers’ turn-taking accuracy during joint play (see Table 2). Action prediction did not significantly contribute to the model of toddlers’ turn-taking accuracy (b = .249, p > .05). In contrast, inhibitory control significantly predicted toddlers’ turn-taking accuracy during joint play and, thus, uniquely contributed to the model (b = –.458, p < .05). Toddlers who failed to inhibit their actions made more errors during the joint action coordination. In addition to the main analysis, we used a multiple regression model to investigate the unique contribution of human actions and nonbiologically moving objects for the timing variability during toddlers’ joint action coordination. The linear combination of the gaze measures was significantly related to the timing variability of joint action coordination, F(2, 24) = 3.8, p = .038. As shown in Table 3, the gaze behavior during human actions significantly predicted the timing variability in the

213

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

Inhibitory Control Task [inhibitory control measure] (A)

(B) r = .204

r = -.434*

16

14

14

12

12

Errors (%)

Errors (%)

16

10 8 6

10 8 6

4

4

2

2 0

0 −0.2

−0.1

0

0.1

0.2

0.3

0.4

0.5

Failed

Succeeded

Success in gi delay task

Fixaons to AoI closest in me to reappearance (s)

(D)

(C)

r = -.104

r = .491*

(c) 1

1

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0

−0.2

−0.1

0

0.1

0.2

0.3

0.4

0.5

0 Failed

Fixaons to AoI closest in me to reappearance (s)

Succeeded

Success in gi delay task

Fig. 2. The figure illustrates the toddlers’ individual differences on the four measures. The joint action coordination measures (coefficient of variation and percentage errors) are represented on the y-axes, and the action prediction measure (left panel) and inhibitory control measure (right panel) are represented on the x-axes. Thus, scatter plots represent the relations between percentage turn-taking errors and action prediction measure (i.e., fixation closest to reappearance) (A), percentage turn-taking errors and inhibitory control measures (i.e., success at the gift delay task) (B), the coefficient of variation and the action prediction measures (C), and the coefficient of variation and the inhibitory control measure (D).

Table 1 Linear regression model testing the unique contribution of action prediction and inhibitory control for the outcome variable timing variability during joint action coordination.

Constant Success in gift delay task Fixation closest in time to reappearance (walking and crawling)

B

SE B

b

0.500 –0.065 0.636

0.065 0.078 0.231

–.154 .506*

Note: R2 = .265 (p < .05). * p < .05.

joint play (b = .432, p < .05), thereby contributing significantly to the model. No indications were found for a unique contribution of gaze behavior to moving objects to toddlers’ timing variability (b = –.138, p > .05). In sum, individual differences in the two measures of the joint action coordination were distinctly related to action prediction and inhibitory action control. Individual differences in the toddlers’ timing

214

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

Table 2 Linear regression model testing the unique contribution of action prediction and inhibitory control for the outcome variable turntaking errors during joint action coordination.

Constant Success in gift delay task Fixation closest in time to reappearance (walking and crawling)

B

SE B

b

7.976 –3.802 6.102

1.275 1.540 4.547

–.458* .249

Note: R2 = .250 (p < .05). * p < .05.

Table 3 Linear regression model testing the unique contribution of gaze behavior to human actions and nonbiologically moving objects for the outcome variable timing variability during joint action coordination.

Constant Fixation closest in time to reappearance (walking and crawling) Fixation closest in time to reappearance (moving objects)

B

SE B

b

0.444 0.543 –0.071

0.055 0.255 0.105

.432* –.138

Note: R2 = .257 (p < .05). * p < .05.

variability were associated with their precision in predicting others’ actions. Individual differences in the toddlers’ turn-taking accuracy specifically related to their inhibitory control. Moreover, the strong relationship between gaze behavior in response to human actions, but not nonbiological movement and joint action timing, shows a unique contribution of processing human actions for joint action coordination. Discussion Although eager to engage in joint actions, toddlers still have difficulties in coordinating their actions with another person (Brownell, 2011). In the current study, we investigated the role of action prediction and inhibitory action control for different aspects of joint action coordination during early childhood. The following distinct patterns were observed. The precision of toddlers’ action prediction was correlated with timing variability of their actions during joint play, whereas the degree of toddlers’ inhibitory action control was correlated with their turn-taking accuracy in the joint action. More specifically, those toddlers who were more precise in their prediction of another person’s actions were more stable in their action timing when playing in turns with an adult action partner. In addition, the more inhibitory action control toddlers displayed in resisting a gift, the less they were inclined to act when it was not their turn during the joint action. No relation between timing variability and inhibitory action control or between turn-taking accuracy and action prediction was found. Thus, distinctive roles for action prediction and inhibitory action control were observed for separable aspects of joint action coordination. In the following, we discuss the role of action prediction and inhibitory action control for joint action coordination and its development. Role of action prediction for joint action coordination during early childhood Eye-tracking and electrophysiological studies have shown that already during their first year of life, infants make predictions about other people’s actions (Falck-Ytter, Gredebäck, & von Hofsten, 2006; Southgate et al., 2010; Southgate, Johnson, Osborne, & Csibra, 2009; Stapel et al., 2010). Many studies employing eye-tracking measures during action observation found that infants and young children make eye movements in anticipation of the goal of an action (Cannon & Woodward, 2012; Falck-Ytter et al., 2006; Hunnius & Bekkering, 2010; Kochukhova & Gredebäck, 2010). Toddlers in the current study observed moving actors disappearing and reappearing from behind an occluder.

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

215

To quantify exactly how well toddlers were able to predict another person’s actions, we focused on the precision in timing eye movements to the moment of reappearance. The current results illustrate that whereas some toddlers showed more predictive gaze with respect to the actor’s reappearance, others tended to show more reactive gaze. As hypothesized, we found that toddlers who were lagging behind more in their gaze (i.e., less precise in their predictions) were more variable in timing their actions during joint play. Toddlers who displayed more predictive gaze patterns were more stable in their action timing during the joint play. These findings suggest a direct link between the ability to predict actions observed in others and to coordinate actions in time with others, thereby connecting the developmental findings on predictive gaze and the adult findings on timing variability during joint actions (e.g., Vesper et al., 2011). The results further suggest that timing variability is specifically linked to the precision in predicting actions given that it was not related to inhibitory action control. Precisely predicting others’ actions potentially allows for preparing and executing one’s own next action in a stable manner, which in turn helps to increase stability in one’s action timing. As a consequence, acting in a stable manner makes oneself more predictable for the joint action partner, thereby establishing smooth joint action coordination (cf. Vesper et al., 2011). The tight link between the ability to predict actions and to coordinate actions jointly is also in line with electrophysiological findings in adults (Kourtis et al., 2013). Kourtis and colleagues (2013) found anticipatory motor activity for an action partner to be related to the proficiency of subsequent joint action performance. In sum, the ability to predict others’ actions in time seems to be crucially connected to the optimal timing of one’s own actions in a joint action context. Our additional analysis demonstrates that predictions of human actions, but not nonbiological movement, uniquely contributed to the model explaining variability in joint action timing. Thus, individual differences in predictions specific to human actions seem to play a role for joint action timing in toddlers. The question arises where individual differences in predicting others’ actions stem from. Previous research indicates that, like in adults, children’s own motor system is crucially involved in generating and timing predictions about others (Elsner et al., 2013; Gredebäck & Kochukhova, 2010; Kanakogi & Itakura, 2011; Southgate et al., 2010; Stapel et al., 2010). Consequently, one possible explanation for the observed individual differences might be developmental differences in the maturation and shaping of the neural motor system during early childhood. Alternatively, differences might reflect stable individual differences in the level to which individuals integrate another person into their own motor system. Findings from neuroimaging studies have suggested that the general level with which we activate our own action system when observing (and predicting) another person’s actions differs between individuals (e.g., Gazzola, Aziz-Zadeh, & Keysers, 2006). In some cases, this was associated with disorders defined by social malfunction (e.g., autism spectrum disorder; Oberman et al., 2005). Future studies are required to disentangle developmental and stable individual differences in the ability to predict others’ actions and to determine potential consequences for joint action coordination. Immature skills in predicting others’ actions during early childhood and associated variability in action timing likely affect not only the quality but possibly also the success and social affiliation established during joint action. A recent study on peer coordination in 19-month-olds investigated the relation between the ability to predict another person’s actions based on inferring the actor’s intentions and toddlers’ affiliative behavior during peer interaction (Hunnius, Bekkering, & Cillessen, 2009). Hunnius and colleagues (2009) found that toddlers who were more proficient in predicting an actor’s goal showed more positive affiliative behaviors during the interaction with a peer. Role of inhibitory action control for joint action coordination during early childhood Besides action prediction, we investigated the role of inhibitory action control for joint action coordination during early childhood. We found that the longer toddlers were able to refrain from peeking at a gift, the less likely they were to act when it was not their turn during the joint action game. This result is consistent with previous findings reporting an association between the performance in classic gift delay tasks and turn-taking tasks in 3-year-old children (Carlson et al., 2004). In contrast to previous studies (using the Tower Building task: Kochanska et al., 1996, 2000), the current study measured turn-taking violations by the young children directly instead of children’s reactions to turn-taking violations of an adult action partner (Kochanska et al., 1996). Still, both tasks—the joint

216

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

action coordination task and the previously used Tower Building task (Kochanska et al., 1996)—provide converging evidence for a close relationship between inhibitory action control in an individual context (gift delay task) and performance in a joint context. In addition, a considerable number of studies have shown associations between inhibitory control and higher social cognitive abilities (e.g., reflected in theory of mind; Carlson et al., 2004; Hughes, 1998). Not only is inhibitory action control in an individual context associated with the toddlers’ performance in a joint context, but this link also seems to be specific for the toddlers’ ability to maintain the turn-taking structure. For instance, it is not related to other qualities of the action coordination (e.g., toddlers’ timing within the joint turn taking). By illustrating the specificity of inhibitory action control for joint action coordination in toddlers, the current findings extend the existing literature on the link between social cognition and inhibitory control. Still, although the current joint turn-taking task requires elements of inhibitory control, it cannot simply be reduced to an individual inhibitory control task considering previous literature on individual and joint action coordination (Kirschner & Tomasello, 2009) and aspects of the joint action coordination unrelated to the inhibition character but relevant for the interaction with the social partner such as children’s action timing. Several toddlers in the current experiment showed poor inhibitory action skills. Related behavioral and neuroimaging research suggests that individual differences in inhibitory action control can reflect both stable individual differences measurable over decades (Casey et al., 2011) and developmental differences (Harnishfeger & Bjorklund, 1994; Jonkman, Lansbergen, & Stauder, 2003). Developmental differences in inhibitory action control have mainly been attributed to maturation of (pre)frontal cortical brain regions (e.g., Cuevas, Swingler, Bell, Marcovitch, & Calkins, 2012). In accordance with this, structural differences in the prefrontal cortex were related to stable behavioral differences in inhibitory control between individuals (Casey et al., 1997). Prefrontal areas show protracted brain development up to adolescence and even early adulthood (Gogtay et al., 2004). Thus, immaturity of the prefrontal cortex resulting in immaturity of inhibitory control closely relates to difficulties in coordinating actions jointly during early childhood—specifically suggesting effects on turn-taking accuracy during joint turn-taking actions. Action prediction and inhibitory action control: Distinct but interconnected processes of joint action coordination Taken together, the current data suggest that action prediction and inhibitory action control are associated with the quality of joint action coordination in toddlers. These two processes are distinctly related to two aspects of joint action coordination: timing stability and turn-taking accuracy. Hence, different processes are underlying and affecting certain aspects of the quality of joint action coordination during early childhood. This differentiated pattern is in accord with prior neurocognitive research that suggests distinct neural mechanisms for generating predictions on others’ actions (Kilner et al., 2004) and for inhibitory control over one’s own actions (e.g., Casey et al., 1997, 2011; Durston et al., 2002). The sensorimotor cortex was shown to play a key role in action prediction (e.g., Kilner et al., 2004; Southgate et al., 2009), whereas the prefrontal cortex has been associated with inhibitory action control (e.g., Casey et al., 1997, 2011; Durston et al., 2002). In terms of structural brain development, sensorimotor regions mature earlier than prefrontal regions, which show a protracted maturation up to adolescence (Casey, Tottenham, Liston, & Durston, 2005). This is indicative of a diverging developmental time course for the ability to predict actions and to inhibit one’s own actions. If both processes indeed diverge in their development, we would also expect their role for the quality of joint action coordination to deviate as children grow older. For instance, during late childhood, the immaturity of inhibitory action control processing might still negatively affect joint action coordination, whereas this would not be expected of action prediction processing that supposedly is more mature at that time. Because the relation between structural brain development and the current joint action processes is likely interdependent, experience in acting jointly with others likely also shapes the underlying neural structures in turn (Casey et al., 2005; Holtmaat & Svoboda, 2009). Although distinct in their behavioral and neural underpinnings (Elsner et al., 2013; Rubia et al., 2001), action prediction and inhibitory action control (including their dissociable relation to joint action coordination) likely do not reflect isolated processes but rather processes that are

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

217

interconnected in a larger network. Indications for this connection are provided by findings from a recent electrophysiological study that was conducted with 3-year-old toddlers (Meyer et al., 2011). The study compared to what extent children made use of their own neural motor system when observing another person acting—either when they were engaged in a joint action with this person or when they were not engaged in a common joint action. Stronger activation of sensorimotor areas for another person’s action was observed when children were engaged in the joint action. More important, this distinction in motor system activation during action observation was significantly correlated to children’s turn-taking accuracy during the joint action. This potentially links processing of sensorimotor areas (associated with action prediction and observation) with processing of prefrontal areas (associated with inhibitory control of one’s own actions). One might speculate that the sensitivity of neural motor areas to being involved in a joint action is signaled forward to prefrontal areas, in turn leading to enhanced inhibitory control and, thus, turn-taking accuracy in a joint action. In other words, selectively incorporating the joint partner’s actions into one’s own motor system potentially emphasizes this action in the processing stream. Crucially, the partner’s action also constitutes the event during which children need to inhibit their actions. If enhanced processing of this event is passed on to prefrontal areas, it might lead to enhanced inhibitory activation and, thereby, the integration of the partner’s action into one’s own action sequence. The adjacent anatomical location and connectivity among primary motor, premotor, and prefrontal cortical areas (e.g., Haggard, 2008) further support the assumption that distinct processes such as action prediction and inhibitory action control are interconnected in a larger network. Thus, the distinct relation between action prediction and inhibitory action control for different aspects of toddlers’ joint action coordination does not exclude interplay between these processes.

Limitations and future directions We investigated joint action according to the definition by Sebanz and colleagues (2006), which does not presuppose joint intentions as necessary for joint actions, in contrast to other conceptual views (Bratman, 1992; Carpenter, 2009). Whether joint intentions are present in the current interaction or whether this type of joint action coordination may precede the emergence of joint intentions (see also Brownell, 2011), therefore, remains an open question. Regardless, the current findings provide insights into how children differed in their joint action coordination and potentially into which processes play a role for smooth coordination between joint action partners. To thoroughly study the relationship between joint action coordination and the potential underlying processes (i.e., action prediction and inhibitory control), data were required for all three of the tasks. Although this ensures optimal interpretability of the findings because each participant contributed to each measure, it led to a relatively high attrition rate (44.5%). Various factors possibly contributed to the high attrition rate, for instance, technical challenges (e.g., insufficient calibration data) and individual factors (e.g., attentional capacities because some toddlers did not play the joint game long enough, current condition because some toddlers got tired at the end of the session). The excluded group, therefore, is very heterogeneous. Still, it remains an open question whether systematic differences might underlie the included and excluded sample of participants. Future studies could now, based on the current findings of distinct relationships, direct their focus separately at either of the discovered relations. This would allow a possibly reduced attrition rate for further investigations, for instance, studying how the two processes might diverge in development. Finally, the current findings have potential implications beyond neurocognitive development mentioned above. For instance, the findings highlight the influence of inhibitory control on social interaction. Inhibitory control also relates to academic skills such as arithmetic (Passolunghi & Siegel, 2001) and literacy (McClelland et al., 2007). Although individual differences in inhibitory control are relatively stable across age (Casey et al., 1997, 2011), it has been suggested that plasticity especially during preschool years allows for potential interventions (Zelazo & Carlson, 2012). Interventions that improve inhibitory control, thus, might not only affect academic performance but also have a positive effect on children’s social interaction skills. Future research is required to explore to what extent social development might be affected by interventions targeting inhibitory control.

218

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

Conclusions In a nutshell, we examined the difficulties that 2½-year-old toddlers display when coordinating their actions jointly with another person. In particular, toddlers face difficulties in timing their actions with respect to their joint action partner and in refraining from acting when it is not their turn in a joint action. We found that two processes, action prediction and inhibitory action control, were distinctly related to these two aspects of joint action coordination. Toddlers who showed more precision in their action predictions were more stable in timing their actions with an action partner. Furthermore, toddlers who showed more inhibitory control in an individual context were more likely to refrain from acting when it was not their turn in the joint action. The current results highlight the role of action prediction and inhibitory action control for the quality of joint action coordination in young children. Moreover, the results bridge previous findings on children and adults, extending our knowledge about joint action coordination in general. Acknowledgments We thank the children and parents for participating in this study. Moreover, we thank our laboratory managers, Angela Khadar and Margret van Beuningen, for their help. We also acknowledge technical support from the Technical Support Group at Radboud University Nijmegen. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi. org/10.1016/j.jecp.2015.06.005. References Adamson, L. B., & Frick, J. E. (2003). The still face: A history of a shared experimental paradigm. Infancy, 4, 451–473. Bennett, S., & Barnes, G. (2006). Combined smooth and saccadic ocular pursuit during the transient occlusion of a moving visual object. Experimental Brain Research, 168, 313–321. Bratman, M. E. (1992). Shared cooperative activity. The philosophical review, 101(2), 327–341. Brownell, C. A. (2011). Early developments in joint action. Review of Philosophy and Psychology, 2, 193–211. Brownell, C. A., & Carriger, M. S. (1990). Changes in cooperation and self-other differentiation during the second year. Child Development, 61, 1164–1174. Brownell, C. A., Ramani, G. B., & Zerwas, S. (2006). Becoming a social partner with peers: Cooperation and social understanding in one- and two-year-olds. Child Development, 77, 803–821. Cannon, E. N., & Woodward, A. L. (2012). Infants generate goal-based action predictions. Developmental Science, 15, 292–298. Carlson, S. A. (2005). Developmentally sensitive measures of executive function in preschool children. Developmental Neuropsychology, 28, 595–616. Carlson, S. M., Mandell, D. J., & Williams, L. (2004). Executive function and theory of mind: Stability and prediction from ages 2 to 3. Developmental Psychology, 40, 1105–1122. Carpenter, M. (2009). Just how joint is joint action in infancy? Topics in Cognitive Science, 1(2), 380–392. Casey, B. J., Somerville, L. H., Gotlib, I. H., Ayduk, O., Franklin, N. T., Askren, M. K., et al (2011). Behavioral and neural correlates of delay of gratification 40 years later. Proceedings of the National Academy of Sciences of the United States of America, 108, 14998–15003. Casey, B. J., Tottenham, N., Liston, C., & Durston, S. (2005). Imaging the developing brain: What have we learned about cognitive development? Trends in Cognitive Sciences, 9, 104–110. Casey, B. J., Trainor, R. J., Orendi, J. L., Schubert, A. B., Nystrom, L. E., Giedd, J. N., et al (1997). A developmental functional MRI study of prefrontal activation during performance of a Go–No-Go task. Journal of Cognitive Neuroscience, 9, 835–847. Cragg, L., Fox, A., Nation, K., Reid, C., & Anderson, M. (2009). Neural correlates of successful and partial inhibitions in children: An ERP study. Developmental Psychobiology, 51, 533–543. Cuevas, K., Swingler, M. M., Bell, M. A., Marcovitch, S., & Calkins, S. D. (2012). Measures of frontal functioning and the emergence of inhibitory control processes at 10 months of age. Developmental Cognitive Neuroscience, 2, 235–243. Diamond, A. (1990). Developmental time course in human infants and infant monkeys, and the neural bases of, inhibitory control in reaching. Annals of the New York Academy of Sciences, 608, 637–676. Durston, S., Thomas, K. M., Yang, Y. H., Ulug, A. M., Zimmerman, R. D., & Casey, B. J. (2002). A neural basis for the development of inhibitory control. Developmental Science, 5(4), F9–F16. Egetemeir, J., Stenneken, P., Koehler, S., Fallgatter, A. J., & Herrmann, M. J. (2011). Exploring the neural basis of real-life joint action: Measuring brain activation during joint table setting with functional near-infrared spectroscopy. Frontiers in Human Neuroscience, 5. http://dx.doi.org/10.3389/fnhum.2011.00095. Elsner, C., D’Ausilio, A., Gredebäck, G., Falck-Ytter, T., & Fadiga, L. (2013). The motor cortex is causally related to predictive eye movements during action observation. Neuropsychologia, 51, 488–492.

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

219

Falck-Ytter, T., Gredebäck, G., & von Hofsten, C. (2006). Infants predict other people’s action goals. Nature Neuroscience, 9, 878–879. Gazzola, V., Aziz-Zadeh, L., & Keysers, C. (2006). Empathy and the somatotopic auditory mirror system in humans. Current Biology, 16, 1824–1829. Gogtay, N., Giedd, J. N., Lusk, L., Hayashi, K. M., Greenstein, D., Vaituzis, A. C., et al (2004). Dynamic mapping of human cortical development during childhood through early adulthood. Proceedings of the National Academy of Sciences of the United States of America, 101, 8174–8179. Gredebäck, G., Johnson, S., & von Hofsten, C. (2009). Eye tracking in infancy research. Developmental Neuropsychology, 35, 1–19. Gredebäck, G., & Kochukhova, O. (2010). Goal anticipation during action observation is influenced by synonymous action capabilities: A puzzling developmental study. Experimental Brain Research, 202, 493–497. Gredebäck, G., & Melinder, A. (2010). Infants’ understanding of everyday social interactions: A dual process account. Cognition, 114, 197–206. Gredebäck, G., & von Hofsten, C. (2004). Infants’ evolving representations of object motion during occlusion: A longitudinal study of 6- to 12-month-old infants. Infancy, 6, 165–184. Haggard, P. (2008). Human volition: Towards a neuroscience of will. Nature Reviews Neuroscience, 9, 934–946. Harnishfeger, K. K., & Bjorklund, D. F. (1994). A developmental perspective on individual-differences in inhibition. Learning and Individual Differences, 6, 331–355. Holtmaat, A., & Svoboda, K. (2009). Experience-dependent structural synaptic plasticity in the mammalian brain. Nature Reviews Neuroscience, 10, 647–658. Hughes, C. (1998). Executive function in preschoolers: Links with theory of mind and verbal ability. British Journal of Developmental Psychology, 16, 233–253. Hunnius, S., & Bekkering, H. (2010). The early development of object knowledge: A study of infants’ visual anticipations during action observation. Developmental Psychology, 46, 446–454. Hunnius, S., Bekkering, H., & Cillessen, A. H. N. (2009). The association between intention understanding and peer cooperation in toddlers. European Journal of Developmental Science, 3, 368–388. Jones, L. B., Rothbart, M. K., & Posner, M. I. (2003). Development of executive attention in preschool children. Developmental Science, 6, 498–504. Jonkman, L. M., Lansbergen, M., & Stauder, J. E. A. (2003). Developmental differences in behavioral and event-related brain responses associated with response preparation and inhibition in a go/no-go task. Psychophysiology, 40, 752–761. Kanakogi, Y., & Itakura, S. (2011). Developmental correspondence between action prediction and motor ability in early infancy. Nature Communications, 2. http://dx.doi.org/10.1038/ncomms1342. Kilner, J. M., Friston, K. J., & Frith, C. D. (2007). Predictive coding: An account of the mirror neuron system. Cognitive Processing, 8, 159–166. Kilner, J. M., Vargas, C., Duval, S., Blakemore, S. J., & Sirigu, A. (2004). Motor activation prior to observation of a predicted movement. Nature Neuroscience, 7, 1299–1301. Kirschner, S., & Tomasello, M. (2009). Joint drumming: Social context facilitates synchronization in preschool children. Journal of Experimental Child Psychology, 102, 299–314. Kochanska, G., Murray, K. T., & Harlan, E. T. (2000). Effortful control in early childhood: Continuity and change, antecedents, and implications for social development. Developmental Psychology, 36, 220–232. Kochanska, G., Murray, K., Jacques, T. Y., Koenig, A. L., & Vandegeest, K. A. (1996). Inhibitory control in young children and its role in emerging internalization. Child Development, 67, 490–507. Kochukhova, O., & Gredebäck, G. (2010). Preverbal infants anticipate that food will be brought to the mouth: An eye tracking study of manual feeding and flying spoons. Child Development, 81, 1729–1738. Kourtis, D., Sebanz, N., & Knoblich, G. (2013). Predictive representation of other people’s actions in joint action planning: An EEG study. Social Neuroscience, 8, 31–42. Loehr, J. D., Large, E. W., & Palmer, C. (2011). Temporal coordination and adaptation to rate change in music performance. Journal of Experimental Psychology: Human Perception and Performance, 37, 1292–1309. Marsh, K. L., Richardson, M. J., & Schmidt, R. C. (2009). Social connection through joint action and interpersonal coordination. Topics in Cognitive Science, 1, 320–339. McClelland, M. M., Cameron, C. E., Connor, C. M., Farris, C. L., Jewkes, A. M., & Morrison, F. J. (2007). Links between behavioral regulation and preschoolers’ literacy, vocabulary, and math skills. Developmental Psychology, 43, 947–959. Mendive, S., Bornstein, M. H., & Sebastian, C. (2013). The role of maternal attention-directing strategies in 9-month-old infants attaining joint engagement. Infant Behavior and Development, 36, 115–123. Meyer, M., Bekkering, H., Paulus, M., & Hunnius, S. (2010). Joint action coordination in 2½- and 3-year-old children. Frontiers in Human Neuroscience, 4. http://dx.doi.org/10.3389/fnhum.2010.00220. Meyer, M., Hunnius, S., van Elk, M., van Ede, F., & Bekkering, H. (2011). Joint action modulates motor system involvement during action observation in 3-year-olds. Experimental Brain Research, 211, 581–592. Miall, R. C. (2003). Connecting mirror neurons and forward models. NeuroReport, 14, 2135–2137. Moriguchi, Y. (2014). The early development of executive function and its relation to social interaction: A brief review. Frontiers in Psychology, 5. http://dx.doi.org/10.3389/fpsyg.2014.00388. Munakata, Y., Herd, S. A., Chatham, C. H., Depue, B. E., Banich, M. T., & O’Reilly, R. C. (2011). A unified framework for inhibitory control. Trends in Cognitive Sciences, 15, 453–459. Oberman, L. M., Hubbard, E. M., McCleery, J. P., Altschuler, E. L., Ramachandran, V. S., & Pineda, J. A. (2005). EEG evidence for mirror neuron dysfunction in autism spectrum disorders. Cognitive Brain Research, 24, 190–198. Passolunghi, M. C., & Siegel, L. S. (2001). Short-term memory, working memory, and inhibitory control in children with difficulties in arithmetic problem solving. Journal of Experimental Child Psychology, 80, 44–57. Ray, M., & Welsh, T. N. (2011). Response selection during a joint action task. Journal of Motor Behavior, 43, 329–332. Rizzolatti, G., & Craighero, L. (2004). The mirror–neuron system. Annual Review of Neuroscience, 27, 169–192. Rubia, K., Russell, T., Overmeyer, S., Brammer, M. J., Bullmore, E. T., Sharma, T., et al (2001). Mapping motor inhibition: Conjunctive brain activations across different versions of go/no-go and stop tasks. NeuroImage, 13, 250–261.

220

M. Meyer et al. / Journal of Experimental Child Psychology 139 (2015) 203–220

Sebanz, N., Bekkering, H., & Knoblich, G. (2006). Joint action: Bodies and minds moving together. Trends in Cognitive Sciences, 10, 70–76. Sebanz, N., & Knoblich, G. (2009). Prediction in joint action: What, when, and where. Topics in Cognitive Science, 1, 353–367. Southgate, V., Johnson, M. H., Karoui, I. E., & Csibra, G. (2010). Motor system activation reveals infants’ on-line prediction of others’ goals. Psychological Science, 21, 355–359. Southgate, V., Johnson, M. H., Osborne, T., & Csibra, G. (2009). Predictive motor activation during action observation in human infants. Biology Letters, 5, 769–772. Stapel, J. C., Hunnius, S., Meyer, M., & Bekkering, H. (submitted for publication). Motor system contribution to social cognition: Action prediction depends on motor experience. Stapel, J. C., Hunnius, S., van Elk, M., & Bekkering, H. (2010). Motor activation during observation of unusual versus ordinary actions in infancy. Social Neuroscience, 5, 451–460. Vesper, C., van der Wel, R. P. R. D., Knoblich, G., & Sebanz, N. (2011). Making oneself predictable: Reduced temporal variability facilitates joint action coordination. Experimental Brain Research, 211, 517–530. Warneken, F., Chen, F., & Tomasello, M. (2006). Cooperative activities in young children and chimpanzees. Child Development, 77, 640–663. Williams, B. R., Ponesse, J. S., Schachar, R. J., Logan, G. D., & Tannock, R. (1999). Development of inhibitory control across the life span. Developmental Psychology, 35, 205–213. Wolpert, D. M., Doya, K., & Kawato, M. (2003). A unifying computational framework for motor control and social interaction. Philosophical Transactions of the Royal Society B: Biological Sciences, 358, 593–602. Zelazo, P. D., & Carlson, S. M. (2012). Hot and cool executive function in childhood and adolescence: Development and plasticity. Child Development Perspectives, 6, 354–360.