Elsevier

Human Movement Science

Volume 28, Issue 4, August 2009, Pages 437-449
Human Movement Science

Visual and tactile action effects determine bimanual coordination performance

https://doi.org/10.1016/j.humov.2009.02.006Get rights and content

Abstract

Effect-based models of motor control assign a crucial role to anticipated perceptual feedback in action planning. Two experiments were conducted to test the validity of this proposal for discrete bimanual key press responses. The results revealed that the normally observed performance advantage for the preparation of two responses with homologous rather than non-homologous fingers becomes inverted when homologous fingers produce non-identical visual effects, and non-homologous fingers produce identical visual effects. In the second experiment the finger homology effect was strongly reduced when homologous fingers produced non-identical tactile feedback. The results show that representations of to-be-produced visual and tactile action effects both contribute to action planning, though possibly to a varying degree. Implications of these results for effect-based models of motor control are considered.

Introduction

The human body is to a large extent characterized by symmetry. This applies to the outer appearance as reflected in the existence of two pairs of limbs, fingers, or toes, but also to the anatomical and functional symmetry of neuronal control structures in the central nervous system. To master many of our everyday tasks, such as when lifting a heavy object, tying a tie, playing piano, or rowing a boat, these body parts have to be used in a coordinated fashion. Experimental research on bimanual coordination has often demonstrated a bias towards symmetric movements or movements involving similar parameter values for amplitude or direction (e.g., Cohen, 1971, Kelso, 1984, Spijkers et al., 1997). For example, symmetric index-finger oscillations (i.e., simultaneous inward and outward movements of the left and right index-finger) are more stable than asymmetric index-finger oscillations (i.e., one index-finger moves inward while the other moves outward): While performing these movements at low speed, both patterns are stable. However, with increasing movement speed, distinct switches from asymmetric to symmetric oscillation patterns occur, but not vice versa (Kelso, 1984).

Traditionally this superiority of symmetric movements was attributed to a tendency towards the activation of homologous muscles, which are more strongly involved in symmetric than in asymmetric body movements (Carson et al., 2000, Cohen, 1971, Swinnen et al., 1998). However, recently this view was challenged by experiments demonstrating that performance in such tasks appears to be dependent on the symmetry of the perceptual movement effects rather than on the homology of the involved effectors (Mechsner et al., 2001, Mechsner and Knoblich, 2004). For example, in the study by Mechsner and Knoblich (2004) different fingers from both hands (e.g., index- and middle-finger of the left hand, and middle- and ring-finger of the right hand) were combined to a bimanual-tapping task. Still, the spatially symmetrical pattern (i.e., left index-finger and right middle-finger vs. left middle-finger and right ring-finger in alternation) was more stable than the parallel pattern (i.e., left and right middle-fingers vs. left index-finger and right ring-finger in alternation). When considering the traditional account, this is most surprising, since in this case parallel patterns involve relatively more homologous muscle activations compared to symmetric patterns.

An explanation of these findings has been derived from ideo-motor theories of action control (e.g., Hommel, Müsseler, Aschersleben, & Prinz, 2001; for a historical overview see Stock & Stock, 2004). The crucial assumption of this approach is that motor actions are cognitively represented by their sensory effects; that is, by codes of the perceptual effects that contingently follow certain motor actions. As a consequence, a motor action can only be accessed by recollecting codes of the sensory consequences that normally accompany this action, and serve to mentally represent it. Stated differently, there is no other way to generate a motor action than by anticipating its sensory consequences. The important implication of this approach is that all the constraints of motor control we know off, such as complexity effects (Henry & Rogers, 1960), stimulus–response compatibility (Simon, 1969), limitations in dual-task performance (Welford, 1952), or symmetry tendencies in bimanual coordination (Kelso, 1984), do not arise because of constraints inherent in the ‘hardware’ of the motor system, but because of constraints in the representation of the perceptual re-afferences of to-be-produced motor actions.

By now, there exist a few studies that demonstrate the bearing capacity of this idea. For example, the initiation of long responses (such as a long key press) takes usually more time than the initiation of short responses (such as a short key press). Yet, these differences can be altered by changing the duration of the contingent sensory effects following such actions. What seems to influence the initiation of the responses most is the duration of the anticipated feedback of the forthcoming action rather than the duration of the action as such (Kunde, 2003). To give another example, key presses are executed more quickly if the desired responses match a response-affording stimulus in a certain respect (such as the faster responding to stimuli that match the required response in terms of spatial location). It has been shown that not the spatial correspondence of the stimulus and response per se matters when producing key presses, but the spatial correspondence of the stimulus and the to-be-produced visual feedback of the response (Hommel, 1993).

The same basic idea seems to apply to bimanual coordination as well. As already noted above, Mechsner and Knoblich (2004) found that finger tapping was more stable when spatially symmetric rather than asymmetric fingers were involved – independent of finger homology. A similar observation was reported by Kunde and Weigelt (2005). They asked participants to turn two wooden blocks from a horizontal starting position into a specific vertical goal position in each trial. Importantly, these blocks were marked with blue on one end and the goal positions were either congruent (i.e., the color marks on both blocks were either upward or downward) or incongruent (i.e., the color mark of one block was upward and of the other one was downward). It turned out that the congruency of the intended goal positions determined the participants’ performance, irrespective of whether this required symmetric rotations (i.e., homologous muscle innervations) or asymmetric rotations (i.e., non-homologous muscle innervations) of the two hands. In both cases responses were initiated and executed faster with congruent goal positions than with incongruent goal positions. Hence, bimanual coordination is facilitated if both actions are carried out to produce similar goal states, rather than by symmetry constraints inherent in the neuromuscular-skeletal system.

Up to now, the impact of action effects in bimanual coordination has been studied with repetitive actions such as finger oscillation or tapping (Mechsner and Knoblich, 2004, Mechsner et al., 2001) or non-repetitive but continuous actions such as object manipulations (Kunde et al., 2009, Kunde and Weigelt, 2005, Weigelt et al., 2006). The problem with these tasks is that planning and execution aspects of these actions are hard to disentangle. With repetitive actions the planning of the movement normally occurs while another movement is still executed. The problem with continuous actions is that they are temporally extended. For example, the object manipulations used by Kunde and Weigelt (2005) took about 800 ms to be completed. Therefore, and in accordance with previous observations on deferred programming effects in bimanual coordination (Spijkers et al., 1997), some of the action specification processes are concluded before the start of the action, while others influence action execution. This renders it somewhat ambiguous to which extent programming processes are reflected in reaction times (from stimulus to response onset; RT) or movement times (from start of the movement to its end; MT) or both. To remove this problem of distributed action specification before and after action onset, we used discrete bimanual responses in the present study.

To this end, participants were asked to press two response buttons simultaneously (with one finger from each hand) according to a response signal (e.g., the left and right index-finger, or the left index-and right middle-finger). Importantly, these responses required almost no time for their execution (the button is either pressed or not) and thus, they could be entirely planned in advance. Therefore, RT is the main dependent variable here, and it entails all the planning processes necessary to execute the action. Consequently, RTs in discrete bimanual tasks presumably reflect a purer measure of action planning than available from repetitive and continuous actions. The first purpose of the present study was thus to test if the impact of action effects, that has been previously observed for repetitive and continuous actions, extends to discrete bimanual responses as well.

The second purpose was to study the impact of action effects of varying remoteness. Regarding this issue already James (1890, 1981) made an interesting distinction into ‘resident’ and ‘remote’ action effects. By resident effects he denotes those body-related sensory consequences that accompany every physical activity of the body, such as the tactile and proprioceptive changes that we feel in the index-finger when bending it. These effects are resident in that they occur whenever we move a part of our body (except in rare pathological cases). In contrast to this, he referred to remote effects as those action consequences that occur outside the body. These are consequences of physical activity taken in through exteroceptive channels, such as the lightening of a lamp after pressing a light switch, or a tone after pressing a piano key. We want to pick up these intuitive plausible terms although on closer inspection it is clear that resident and remote effects differ in several respects, such as their physical distance to the body, the perceptual channel through which they are perceived, and the certainty with which they follow a certain motor output (see below). We also want to make clear that the terms resident and remote should not be confused with the terms ‘proximal’ and ‘distal’ as used in the theory of event coding (TEC; Hommel et al., 2001). According to TEC a code is distal irrespective of the sensory channel informing it if it refers to events in the world, where the body is in some sense part of the world as well. In contrast, a code is proximal irrespective of the sensory channel informing it if it refers to the specifics of the sensory transduction, that is, of how our system responds to external events. Accordingly, it simply does not matter whether fingers are perceived visually or proprioceptively: if the code refers to a finger movement as an event in the world this code is distal, it codes for a distal aspect of the event. Hence, resident and remote action effects in James’ terminology would be both considered as being distal.1

Ideo-motor inspired research on motor control has mainly focused on remote perceptual effects, presumably because such effects can be easily controlled in the psychological lab. However, one should assume that the findings relative to motor coordination and remote effects generalize to resident effects as well. Yet, there may be good reasons to doubt this assumption. As noted above, resident and remote action effects differ in certain respects and might serve different roles in action planning. Consider, for example, that the same motor output, such as the bending of the right index-finger, can be linked to several exchangeable remote effects, such as the lightening of a desk lamp in one context, or the sound of a piano tone in another context. In contrast, resident effects are less exchangeable. Although the proprioceptive feedback of bending a finger might differ from time to time, such variability has obviously limitations. For example, it is hard to imagine that bending the finger feels like moving the leg. At least the body-related location of proprioceptive feedback from a moving effector is constant. These properties might assign different roles to remote and resident action effects in action generation. The anticipation of remote effects might activate several body movements or classes of body movements that are in principle instrumental to obtain these remote effects, whereas the anticipation of resident effects might activate very specific motor patterns. This speculation is not meant to say that we know already the different roles of resident and remote effects. We do not. It should simply illustrate that we need experimental approaches to clarify potential differences between them, and the present study is a first attempt to do so.

We report two experiments here. In both experiments one group of participants produced identical and non-identical perceptual effects with homologous and non-homologous finger presses, respectively (congruent response–effect (R–E) mapping), while a second group produced identical effects with non-homologous and non-identical effects with homologous finger presses (incongruent R–E mapping). The crucial question was whether the normally observed advantage in RT with homologous compared to non-homologous fingers is affected (or possibly even reversed) when homologous fingers produce non-identical effects and non-homologous fingers produce identical effects (i.e., with an incongruent R–E mapping). Experiment 1 used remote (visual) effects and Experiment 2 used resident (tactile) effects.

Section snippets

Experiment 1

In Experiment 1 participants were to respond with one finger (either the index- or the middle-finger) of both hands simultaneously to the onset of a visually presented stimulus. In each trial, pressing a response button produced a visual effect. These visual effects were growing and shrinking columns that were directly located above the associated response buttons, thus representing salient remote effects (cf. Fig. 1). Importantly, these columns grew either high or low and this characteristic

Experiment 2

In Experiment 1 we successfully demonstrated the impact of (remote) visual effects in the bimanual-tapping task. Performance was superior when identical rather than non-identical effects were produced, independent of whether homologous or non-homologous fingers produced them. Yet, the advantage of identical over non-identical effects was numerically larger when identical effects were produced by homologous rather than by non-homologous fingers. This suggests that finger homology still played a

General discussion

Two experiments investigated the impact of action effects on response planning and execution in the bimanual finger-tapping task. The often observed advantage of symmetrical response patterns has frequently been ascribed to the involvement of homologous muscle portions (Carson et al., 2000, Cohen, 1971, Swinnen et al., 1998). However, recent experimentation casts doubt on this assertion by demonstrating the impact of perceptual effects rather than homology of involved effectors on the stability

References (24)

  • J. Adam et al.

    Preparing for perception and action. I: The role of grouping in the response-cuing paradigm

    Cognitive Psychology

    (2003)
  • W. Spijkers et al.

    Preparation of bimanual movements with same and different amplitudes: Specification interference as revealed by reaction time

    Acta Psychologica

    (1997)
  • R.G. Carson et al.

    Neuromuscular-skeletal constraints upon the dynamics of unimanual and bimanual coordination

    Experimental Brain Research

    (2000)
  • L. Cohen

    Synchronous bimanual movements performed by homologous and non-homologous muscles

    Perceptual and Motor Skills

    (1971)
  • F.M. Henry et al.

    Increased response latency for complicated movements and a “memory drum” theory of neuromotor reaction

    Research Quarterly of the American Association for Health, Physical Education, and Recreation

    (1960)
  • B. Hommel

    Inverting the Simon effect by intention: Determinants of direction and extent of effects of irrelevant spatial information

    Psychological Research

    (1993)
  • B. Hommel et al.

    The theory of event coding (TEC): A framework for perception and action

    Behavioral and Brain Sciences

    (2001)
  • W. James

    The principles of psychology (orig. 1890)

    (1890, 1981)
  • J.A.S. Kelso

    Phase transitions and critical behavior in human bimanual coordination

    American Journal of Physiology: Regulatory, Integrative, and Comparative Physiology

    (1984)
  • W. Kunde

    Temporal response–effect compatibility

    Psychological Research

    (2003)
  • W. Kunde et al.

    Goal congruency in bimanual object manipulation

    Journal of Experimental Psychology: Human Perception and Performance

    (2005)
  • W. Kunde et al.

    Goal-congruency without stimulus congruency in bimanual coordination

    Psychological Research

    (2009)
  • Cited by (39)

    • Resource limitations in bimanual pointing

      2022, Human Movement Science
      Citation Excerpt :

      In fact, action effects have also been shown to affect and determine the efficacy of bimanual movements. As one example, the RT advantage of homologous over non-homologous finger presses can be reversed if the latter lead to similar and the former to dissimilar action effects (Janczyk, Skirde, Weigelt, & Kunde, 2009; see also Janczyk & Kunde, 2014; Kunde, Krauss, & Weigelt, 2009; Kunde & Weigelt, 2005; Mechsner, Kerzel, Knoblich, & Prinz, 2001). One objection to the present study and its rationale is the perhaps overly simplifying distinction between cognitive and motor processes.

    • Towards a Unitary Approach to Human Action Control

      2017, Trends in Cognitive Sciences
    • Thinking with portals: Revisiting kinematic cues to intention

      2014, Cognition
      Citation Excerpt :

      As a reliable finding across these studies, responses are initiated faster when responses and their to-be-produced effects are compatible rather than incompatible. Because the effects only occur after action initiation, these findings clearly indicate that effect anticipations are at work during action selection and planning (see also Janczyk, Pfister, Hommel, & Kunde, 2014, and Janczyk, Skirde, Weigelt, & Kunde, 2009; for converging results from other paradigms). Whether or not such anticipations also affect the trajectories of a manual movement has not been studied to date, though.

    View all citing articles on Scopus

    The research reported here was supported by a grant from the German Research Foundation (DFG) awarded to WK (Ku 1964/2-1). We thank Bernhard Hommel and an anonymous reviewer for helpful comments on an earlier draft of this manuscript.

    View full text