Elsevier

Acta Psychologica

Volume 115, Issues 2–3, February–March 2004, Pages 185-209
Acta Psychologica

Development of the acquisition and control of action–effect associations

https://doi.org/10.1016/j.actpsy.2003.12.006Get rights and content

Abstract

Voluntary action is anticipatory and, hence, must depend on associations between actions and their perceivable effects. We studied the acquisition of action–effect associations in 4–5-vs. 7-year-old children. Children carried out key-pressing actions that were arranged to produce particular auditory effects. In a subsequent test phase, children were to press keys in response to the previous effect sounds, with the sound–key mapping being either consistent or inconsistent with previous key–sound practice. As the processes underlying voluntary action controls are known to significantly improve between 4 and 7 years of age, it was expected that younger children were more prone to automatic effects of acquired sound–key associations. This hypothesis was confirmed, but reaction times and accuracy measures showed different and dissociable patterns. Four-year-olds but not 7-year-olds were more likely to commit an error––i.e., to perform a sound-compatible rather than the correct action––if the sound–key mapping was inconsistent with previous practice. This effect strongly depended on previous practice, suggesting that it reflects long-term learning. In contrast, reaction time effects of mapping consistency did not depend on previous experience but only on the consistency between stimulus and action effect in the present task. Taken altogether, the results suggest that children acquire response–effect associations automatically and that younger children are more likely to suffer from frequent goal neglect; i.e., they tend to forget the current action goal, so that their behavior is dominated by automatic, stimulus-triggered response tendencies.

Introduction

The success of achieving our goals depends on the success of our actions. Hence, when planning an action we must have some anticipatory notice about the impact that action is likely to have on our environment. Indeed, it seems plausible to assume that we plan our actions on the basis of the knowledge we have about their consequences or effects (Hommel, 1998). An example may clarify this line of thought. Imagine that you are driving a new spacecraft and want to slow down its speed. Assume further that for some unfortunate reason nobody had told you where the brakes of this vehicle are and how they work, and you forgot to ask. To reach your goal, you will probably try pushing and switching all buttons and pedals you come across to see what effect that may have until, as we hope, you have found the one that does the job. Odds are very high that next time you are in the same situation you will recall which pedal or button needs to be moved in which way, which will enhance your action repertoire and speed up your performance. Thus, storing information about the effects or consequences of particular actions provides a helpful basis for the control and organization of your actions.

This logic inspired Elsner and Hommel (2001) to develop a two-stage model of action control, which takes up ideas from the ideomotor approach to voluntary action (Harless, 1861; James, 1890; Lotze, 1852). In this model action control is attributed to the automatic integration of movements and their sensory consequences (for an extended overview of the model, see Elsner & Hommel, 2001).

In the first stage the model claims, contingencies between actions and their effects are acquired. That is, when an event consistently follows a particular movement its representation becomes associated with the representation of that movement. Indeed, it can be demonstrated that when two responses are consistently followed by a low- and high-pitched tone, respectively, the tones become associated with the accompanying response, even in tasks where the tones are irrelevant and non-informative (Elsner & Hommel, 2001; Hommel, 1996). These tasks often comprise two parts or phases. For instance, in Elsner and Hommel’s (2001) study, subjects first carried out a number of freely chosen responses to a neutral trigger stimulus, with each response consistently producing a particular tone. Then, in a following test phase, subjects again performed a free-choice task but this time the trigger stimulus could be one of the tones that in the acquisition phase served as action (-produced) effects. As expected, subjects frequently carried out the response that previously had produced the current trigger tone, that is, if in the acquisition phase Response 1 produced Tone A and Response 2 produced Tone B, the trigger Tone A was more likely to induce selection of Response 1 than Response 2, and vice versa. Apparently, then, (even irrelevant) action effects are integrated with the action producing them.

The second stage of Elsner and Hommel’s (2001) model addresses the selection of actions. As the experience of action–effect sequences is assumed to result in the formation of bidirectional links between action and accompanying effect, actions can be primed, retrieved, and launched by activating representations of their effects––be it by “thinking of” the intended consequences of an action (the topic of ideomotor theory) or, more accidentally, by stimuli that happen to share features with action effects. In our example, thinking of your goal to slow down the spacecraft will prime actions that previously have been produced the wanted effect. If you happen to have already a successful experience with decelerating that particular spacecraft, the corresponding action pattern is likely to be activated most strongly, leading to the most efficient performance. If you lack that experience, however, other action patterns associated with the same or with similar goals (and with similar situative contexts) will be activated most strongly instead, which will make you look for knobs, buttons, and pedals that turned out to be useful in braking cars, bicycles, and other vehicles you may be familiar with. However, as goal representations are no (much) more than representations of previously perceived (and now wanted) action effects, any action–effect stimulus has the potency of activating the associated actions to at least some degree. Hence, the braking action may also be primed by perceiving stimuli that this action had produced in the past or by stimuli that are associated with such effect stimuli (e.g., red traffic lights or a child crossing the street). In other words, associating actions and effect stimuli renders the latter (as well as stimuli similar to them) effective primes of those actions––a particularity the present study was thought to capitalize on.

Indeed, apart from the auditory stimuli used by Elsner and Hommel (2001), there is considerable evidence that all kinds of action–effect stimuli can become effective action primes, suggesting that the integration of actions and their effects, and the thereby implied transformation of action effects into action primes, are a general phenomenon. For instance, action–effect learning was established with tones of varying location (Hommel, 1996), tones of different pitch (Elsner & Hommel, 2001; Elsner et al., 2002; Hazeltine, 2002; Hoffmann, Sebald, & Stöcker, 2001; Hommel, 1996; Hommel & Elsner, 2000; Kunde, Hoffmann, & Zellman, 2002), visual stimuli of varying location (Ansorge, 2002; Hommel, 1993), visual letters (Ziessler, 1998; Ziessler and Nattkemper, 2001, Ziessler and Nattkemper, 2002), visual stimuli of different affective valence (Van der Goten, Caessens, Lammertyn, De Vooght, & Hommel, submitted for publication), words (Hommel, Alonso, & Fuentes, 2003), Stroop stimuli (Hommel, in press), and with electrocutaneous stimuli (Beckers, De Houwer, & Eelen, 2002).

To summarize, there is evidence that actions can be primed by both anticipations of intended action effects and by any stimulus that looks like or has been experienced to be an effect of the respective action. For us, the important implication of this observation is that we can experimentally set action priming via intentional anticipation (making use of the instructed stimulus–response mapping) in opposition to action priming via external triggering by previous action effects, and study the interaction of the resulting voluntary and involuntary response tendencies. The outcome of this competition should depend on the relative strength of the respective contribution: the smaller the contribution from intentionally controlled processes, the more should a response reflect the impact of an irrelevant action–effect stimulus. Studying this interplay between the automatic stimulus-induced processes and the intentionally controlled processes should increase our understanding of human goal-directed behavior, especially if we compare subjects who differ in the efficiency of action-control processes. In the present study this was accomplished by studying the acquisition and use of action–effect associations in children from different age groups. In particular, we compared age groups that are likely to differ in the efficiency of action control processes, that is, in the balance between reflexive and goal-directed behavior.

Control processes that guide behavior gradually improve during childhood. One major growth spurt in the development of action control seems to occur at about 5–6 years of age. In this period, reflexive behavior becomes less frequent and the ability to inhibit prepotent responses and perseverative behavior in favor of the production of intentionally guided movements improves substantially. This finding is supported by many studies investigating efficiency of cognitive control on a variety of inhibition tasks.

For instance, Levy (1980), who studied stopping behavior in a go/no-go task, reported a rapid rise in the speed of responding and an even more rapid decline in errors between the ages 3 and 7 (from 68% in 3-year-olds to 0.02% in 7-year-olds). Other studies replicated these findings but suggest an even earlier improvement of action control: children 3–4 years of age often fail to inhibit their responses in no-go trials while 5–6 year-olds perform very well (Bell & Livesey, 1985; Dowsett & Livesey, 2000; Livesey & Morgan, 1991). In another frequently used inhibition task, the antisaccade task (in which a prepotent eye movement toward a stimulus has to be inhibited and an intentionally guided eye movement in the opposite direction has to be generated), performance improves dramatically from the age of 6 years on (Klein & Foerster, 2001). Examination of developmental changes on the Wisconsin Card Sorting Task (WCST), in which a switch to a new sorting dimension requires the inhibition of a previously relevant dimension, shows that performance improves most rapidly between the ages 6 and 7 (Chelune & Baer, 1986). In a simplified and adjusted version of the WCST, children of 2.5–3 years often fail and show perseverance of the old sorting rule––even though they have no difficulty remembering and verbalizing the new, correct rule (Zelazo, Frye, & Rapus, 1996; Zelazo, Reznick, & Piñon, 1995; Zelazo, Craik, & Booth, 2004).

Comparable patterns of performance are found in tasks in which actions are guided by rules that require acting contrary to intuitive responding. In a Stroop-like task, the day–night task (in which children have to say “day” to black/moon cards and say “night” to white/sun cards), it was found that children younger than 5 years of age show very poor performance (long reaction times and accuracy 70% or lower) while performance is very good by the age of 7 (accuracy more than 90%; e.g., Diamond, Kirkham, & Amso, 2002; Gerstad, Hong, & Diamond, 1996). Similar results in children 3–4 years of age and 6 years of age were found on the tapping task (Diamond & Taylor, 1996; Luria, 1966), in which subjects are instructed to tap once or twice if the experimenter taps twice or once, respectively. Furthermore, children 3–4 years of age, but not 5–6-year-old, have serious difficulties in delay-of-gratification paradigms (Mischel & Mischel, 1983), in which they are to wait for a more preferred or bigger reward in the presence of a less preferred or smaller, but immediately available reward.

In sum, many developmental studies reveal a transition in the efficiency of goal directed behavior around the age of 5–6 years. Accordingly, we thought that investigating the resistance to stimulus-induced actions in children before and after this critical period in action-control development would be particularly diagnostic in unveiling the basis of how goal-directed behavior is controlled.

The purpose of the present study was to investigate how children acquire and use associations between actions and their effects. The underlying idea was that developmental changes in the trade-off between automatic stimulus-induced action processes and controlled action processes would provide more insight into the role of effect-based learning in action control.

The four experiments of this study followed the logic underlying the experiments of Elsner and Hommel (2001). The first three experiments were divided into two parts. The purpose of the first part, the acquisition phase, was to provide children with the opportunity to experience the sequence of two motor actions (M1 and M2) and the two auditory events following them (E1 and E2; M1E1 and M2E2), which should lead to bidirectional associations between the cognitive representations of actions and effects (m1e1 and m2e2). Children performed a free-choice response task, in which two different responses were contingently followed by one of two different sounds. If children would in fact form bidirectional action–effect associations, presenting effect stimuli should prime the action they accompanied (i.e., if m1e1, then E1M1).

This prediction was tested in the second part of the experiment, the test phase. The higher likelihood (in Experiment 1) and/or the faster speed of performing an action m1 after the presence of an action–effect stimulus e1 (in Experiments 2–4)––i.e., the degree of effect-induced action priming––served as evidence that bidirectional associations between actions and effects had been formed. Note that the relation between actions and effects was irrelevant in both the acquisition phase and the test phase, so that the degree to which actions were primed by perceiving their effects represents an automatic, stimulus-induced and, in a sense, reflexive impact on action control. The stronger this external impact, we reasoned, the weaker must be the internal, intentional control of action. Accordingly, less action priming in the older children as compared to the younger children would point to an age-related increase of internal action control.

The test phase differed across the four experiments. In Experiment 1 the children were to choose freely one of two possible responses after presentation of a particular sound (the previous action effects). It was predicted that subjects would be more likely to perform the response that in the acquisition phase was associated with that sound, which we will call the effect-consistent response. Experiments 2 and 3 employed a binary-choice task that required pressing one key in response to one sound and the other key in response to another sound. Subjects were divided into two groups: A consistent-mapping group, where the sound–key mapping in the test phase was consistent with that in the acquisition phase (E1M1 and E2M2), and an inconsistent-mapping group, where the sound–key mapping in the test phase was inconsistent with that in the acquisition phase (E1M2 and E2M1). If producing sounds by pressing keys creates a bidirectional association between the action and its effect, performance should be better with a consistent than an inconsistent mapping. The last experiment served as a control for Experiment 3 and was a replication without the acquisition phase.

Section snippets

Experiment 1

As pointed out, the purpose of the first, acquisition part of Experiment 1 was to bring about bindings between the actions performed––two key presses––and their effects––two task-irrelevant sounds that contingently followed the key presses. If such bindings would be formed, we would expect that in the second phase of the experiment, the test phase, the presence of a sound would prime the key press it had followed previously. If so, children should, when free to choose which of the two keys to

Experiment 2

The major modification we made in Experiment 2 was to change the test phase from a free-choice into a forced-choice task. Subjects were now to respond with a left key press to one, and with a right key press to the other sound. Furthermore, the test phase was split into two between-subjects conditions: subjects could be assigned to a consistent-mapping group or an inconsistent-mapping group. For those assigned to the consistent-mapping group the sound–key mapping was consistent with the

Experiment 3

Apart from some, mostly motivational improvements of our design, in Experiment 3 pressing a key triggered the corresponding sounds in both the acquisition phase and the test phase. That is, participants in the consistent-mapping group always heard the same two sounds in each test trial (e.g., right key  “uh-oh” in acquisition phase and “uh-oh”  right key  “uh-oh” in test phase), whereas subjects in the inconsistent-mapping condition always heard two different sounds (e.g., right key  “uh-oh” in

Experiment 4

In Experiment 4 subjects performed exactly the same task as in the test phase of Experiment 3 without having carried out any acquisition trial before. That is, subjects had no opportunity to acquire any mapping-consistent or -inconsistent action–effect association that could affect their performance. The consistency manipulation thus referred only to the relation between the relevant stimulus sound and the irrelevant action–effect sound, which in each trial were the same for the consistent-sound

General discussion

The present study shows that it is possible to demonstrate action–effect acquisition in children but that this is more difficult than one would expect. Why did we fail to find reliable effects of action–effect learning in the first experiment? Apart from the somewhat smaller age range than in Experiments 2–4, the main reason seems to be the free-choice nature of the task, which apparently confused the children and motivated the development of individual strategies. Such strategies are likely to

Acknowledgements

This research was supported by a grant of the Deutsche Forschungsgemeinschaft to B.H. (Priority Program on Executive Functions, HO 1430/8-1). We wish to thank Birgit Elsner and an anonymous reviewer for constructive and insightful comments and suggestions.

References (51)

  • J.A. Bell et al.

    Cue significance and response regulation in 3- to 6-year old-children’s learning of multiple choice discrimination tasks

    Developmental Psychobiology

    (1985)
  • G.J. Chelune et al.

    Developmental norms for the Wisconsin Card Sorting Test

    Journal of Clinical and Experimental Neuropsychology

    (1986)
  • J.D. Cohen et al.

    A computational approach to prefrontal cortex, cognitive control, and schizophrenia: Recent developments and current challenges

  • J.D. Cohen et al.

    On the control of automatic processes: A parallel distributed processing account of the Stroop effect

    Psychological Review

    (1990)
  • M.-P. Deiber et al.

    Cortical areas and the selection of movement: A study with positron emission tomography

    Experimental Brain Research

    (1991)
  • R. De Jong

    Adult age differences in goal activation and goal maintenance

    European Journal of Cognitive Psychology

    (2001)
  • A. Diamond

    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

    (1990)
  • A. Diamond et al.

    Conditions under which young children can hold two rules in mind and inhibit a prepotent response

    Developmental Psychology

    (2002)
  • A. Diamond et al.

    Development of an aspect of executive control: Development of the abilities to remember what I said and to “do as I say not as I do”

    Developmental Psychobiology

    (1996)
  • S.M. Dowsett et al.

    The development of inhibitory control in preschool children: Effects of “executive skills” training

    Developmental Psychobiology

    (2000)
  • J. Duncan

    Attention, intelligence and the frontal lobes

  • Dutzi, I. B., & Hommel, B. (submitted for publication). Spontaneous but goal-dependent binding of actions and their...
  • B. Elsner et al.

    Effect anticipation and action control

    Journal of Experimental Psychology: Human Perception and Performance

    (2001)
  • B. Elsner et al.

    Linking actions and their perceivable consequences in the human brain

    Neuroimage

    (2002)
  • J.M. Fuster

    The prefrontal cortex

    (1989)
  • Cited by (37)

    • Dissociating distractor inhibition and episodic retrieval processes in children: No evidence for developmental deficits

      2018, Journal of Experimental Child Psychology
      Citation Excerpt :

      More important for the current purposes, however, is the fact that we obtained reliable episodic retrieval effects in the child sample. The current finding nicely fits with other studies that reported intact automatic binding and retrieval processes already in very young children, that is, from the age of 4 years onward (Eenshuistra, Weidema, & Hommel, 2004; Karbach, Kray, & Hommel, 2011). Therefore, we can conclude that automatic processes of stimulus–response binding and retrieval are present and functionally intact already at a young age.

    • The developing cognitive substrate of sequential action control in 9- to 12-month-olds: Evidence for concurrent activation models

      2015, Cognition
      Citation Excerpt :

      After acquisition, they test if exogenously cueing an effect cues the action that previously caused it (Elsner & Hommel, 2001; Greenwald, 1970). This approach resulted in demonstrations of bidirectional action–effect acquisition for a wide range of actions and effects in children (Eenshuistra, Weidema, & Hommel, 2004; Kray, Eenshuistra, Kerstner, Weidema, & Hommel, 2006) and adults, suggesting the mechanism responsible to be fast-acting (Dutzi & Hommel, 2009), automatic (Band, van Steenbergen, Ridderinkhof, Falkenstein, & Hommel, 2009; Elsner & Hommel, 2001), implicit (Elsner & Hommel, 2001; Verschoor, Spapé, Biro, & Hommel, 2013), and modulated by the same factors that influence instrumental learning (Elsner & Hommel, 2004) (for a review on action–effect learning see: Hommel & Elsner, 2009). Furthermore, action–effects have also been found to be important for action evaluation (Band et al., 2009; Verschoor et al., 2013).

    • Perceiving by proxy: Effect-based action control with unperceivable effects

      2014, Cognition
      Citation Excerpt :

      Matters might be entirely different for infants who may well have to rely on actual experience for setting up initial action–effect associations. Especially the acquisition of bidirectional R–E associations is a developmental challenge that is overcome only after extensive practice (Eenshuistra, Weidema, & Hommel, 2004; Hauf, Elsner, & Aschersleben, 2004; Verschoor, Weidema, Biro, & Hommel, 2010) and that is preceded by only unidirectional use of R–E associations (Verschoor, Spapé, Biro, & Hommel, 2013). The emergence of purely intentional formations of R–E associations in ontogenetic development certainly is an interesting topic for future inquiry, as are possible effects of actual experiences on R–E associations that were built by mere intention (and vice versa).

    • Binding in voluntary action control

      2010, Neuroscience and Biobehavioral Reviews
    View all citing articles on Scopus
    View full text