Elsevier

Acta Psychologica

Volume 136, Issue 3, March 2011, Pages 311-320
Acta Psychologica

The learning of 90° continuous relative phase with and without Lissajous feedback: External and internally generated bimanual coordination

https://doi.org/10.1016/j.actpsy.2010.12.004Get rights and content

Abstract

Results from recent experiments (e.g., Kovacs, Buchanan, & Shea, 2009a–b, 2010a,b) suggest that when salient visual information is presented using Lissajous plots bimanual coordination patterns typically thought to be very difficult to perform without extensive practice can be performed with remarkably low relative phase error and variability with 5 min or less of practice. However, when this feedback is removed, performance deteriorates. The purpose of the present experiment was to determine if reducing the frequency of feedback presentation will decrease the participant's reliance on the feedback and will facilitate the development of an internal representation capable of sustaining performance when the Lissajous feedback is withdrawn. The results demonstrated that reduced frequency Lissajous feedback results in very effective bimanual coordination performance on tests with Lissajous feedback available and when feedback is withdrawn. Taken together the present experiments add to the growing literature that supports the notion that salient perceptual information can override some aspects of the system's intrinsic dynamics typically linked to motor output control. Additionally, the present results suggest that the learning of both externally and internally driven bimanual coordination is facilitated by providing reduced frequency Lissajous feedback.

Introduction

Bimanual movements have been used extensively in the study of interlimb coordination, especially with respect to developing theories of temporal order. This research, which has employed a variety of bimanual coordination tasks, has revealed remarkable temporal constraints (Kelso & DeGuzman, 1988) that limit spontaneous bimanual coordination to two relatively stable patterns. Coordination patterns of in-phase (relative phase between oscillating components, Φ = 0°) and anti-phase (Φ = 180°) have repeatedly been found to be inherently more stable and easy to perform without extended practice. Other relative phase coordination patterns (e.g. Φ = 30°–150°) are quite difficult to perform and require substantial practice in order to achieve relatively stable performance (Yamanishi et al., 1980, Zanone and Kelso, 1992). In other words, relative phase patterns other than in-phase and anti-phase are not inherently stable. When trying to perform coordination patterns with other phase relations, the motor system shows a bias towards what has been labeled the intrinsic dynamics of in-phase and anti-phase coordination (Schöner & Kelso, 1988). The difficulty in performing other phase relationships has traditionally been attributed to the attraction toward the intrinsic dynamics of in-phase and anti-phase (Schöner & Kelso, 1988), the instability associated with the activation of homologous and non-homologous muscles via crossed and uncrossed corticospinal pathways (Cattaert et al., 1999, Kagerer et al., 2003, Kennerley et al., 2002) as well as the time delays associated with the integration of the afferent and efferent signals due to anatomical connectivity (Banerjee and Jirsa, 2006, Peper and Beek, 1999).

More recently perceptual information and attentional factors have also been found to play an important role in the coordination between the limbs (e.g., Bingham, 2004a, Bingham, 2004b, Bingham et al., 1999, Kovacs et al., 2009a, Kovacs et al., 2009b, Kovacs et al., 2010a, Kovacs et al., 2010b, Kovacs and Shea, 2010, Mechsner et al., 2001, Shea et al., 2009). Recent experiments have demonstrated that changes in the perceptual information available in the testing environment and the attention demands associated with monitoring the movement of the two limbs can greatly influence bimanual coordination, e.g., stabilize anti-phase and destabilize in-phase coordination patterns by changing the orientation of the hand (Mechsner et al., 2001). Moreover, recent experiments by Kovacs et al., 2009a, Kovacs et al., 2009bhave demonstrated that when attention splitting features of a task are reduced by not using a visual or auditory metronome, vision of the moving limbs is restricted, and salient integrated visual feedback is provided (Lissajous plot and template), the perceptual information in the Lissajous feedback appears to free the perception–action system from constraints that typically limit it to the bi-stable regime of the intrinsic dynamics. For example, participants in the Kovacs et al. (2009a) experiment were asked to perform a 1:1 bimanual coordination task with a 90° relative phase while receiving concurrent Lissajous feedback. The Lissajous feedback was in the form of a cursor representing the joint position of the two limbs overlaid on a Lissajous template depicting the required phase relation between the limbs. Note that movement of the right limb moved the cursor horizontally and movement of the left limb moved the cursor vertically. The results indicated that participants were quite effective (relative phase errors and variability < 10°) in performing the required coordination pattern after only 5 min of practice when an auditory metronome was not used and vision of the limbs was not permitted. This level of relative phase errors and variability was lower than typically observed after multiple days of practice. When a metronome was used relative phase errors and variability increased substantially (> 25°). More importantly for the purpose of the present paper, a test without Lissajous feedback resulted in greatly increased bimanual coordination errors and variability suggesting that participants had developed a dependency on the feedback. The dramatic increase in relative phase errors and variability when Lissajous feedback was withdrawn suggests that participants had not developed an internal representation of the coordination pattern. While the extremely small relative phase errors and variability resulting from the use of the Lissajous information were quite shocking especially following minimal practice, it was also clear that participants could not continue to perform the coordination pattern in the absence of the feedback. Similarly, Kovacs et al. (2009b) have demonstrated that when participants are provided Lissajous feedback, they can very effectively perform various 1:1 continuous bimanual coordination patterns between 30° and 150° relative phase after only 4 min of practice at each relative phase. Indeed, the patterns of relative phase error and variability for each relative phase condition was reduced to levels typically observed only after multiple days of practice, and then only for the relative phase pattern that was practiced (Zanone & Kelso, 1992). In contrast, participants that were paced by a visual metronome and allowed vision of their limbs exhibited much poorer performance.

It should be noted that Lissajous feedback has been used in a number of previous bimanual coordination experiments dating at least back to the early 1990s. One of the first reports of the use of Lissajous feedback was presented by Swinnen, De Pooter, and Delrue (1991) with a number of subsequent experiments describing the role of perceptual, attentional, and conceptual factors on bimanual coordination (e.g., Lee, Swinnen, & Verschueren, 1995, 1996; Swinnen, Dounskaia, et al., 1997, Swinnen, Lee, et al., 1997; see Swinnen & Wenderoth, 2004 for a review). These earlier results provided the initial hints as to the potential power of Lissajous feedback. The experiments by Kovacs and colleagues demonstrated how factors like metronomes used to direct or pace movements, vision of the limbs, and Lissajous templates of the desired coordination pattern influence the ability of participants to fully utilize the Lissajous feedback provided. That is, even with concurrent Lissajous information provided, 1:1 bimanual coordination patterns with 90° relative phase typically required several days of practice for this coordination pattern to be produced with relatively small relative phase errors and variability (e.g., Hurley and Lee, 2006, Lee et al., 1995, Swinnen, Dounskaia, et al., 1997, Swinnen, Lee, et al., 1997) when vision of the limbs and/or metronomes were used. What the Kovacs et al., 2009a, Kovacs et al., 2009b experiments have demonstrated is that when a Lissajous template is provided and attention splitting features of a task are reduced in the testing environment by eliminating vision of the limbs and metronomes the effect of the Lissajous feedback is very powerful allowing participants to essentially tune their movements to the various templates provided. Note that the earlier use of Lissajous feedback by Swinnen and colleagues (e.g., Lee et al., 1995, Lee et al, 1996; Swinnen et al., 1991, Swinnen, Dounskaia, et al., 1997, Swinnen, Lee, et al., 1997) often did not include a presentation of a Lissajous template with the online Lissajous feedback or only provided the template as terminal feedback, allowed vision of the limbs, and/or used a metronome to pace the movement production. These factors appear to play a role in limiting the utility of this form of feedback.

A critical feature of the Kovacs et al. experiments, however, was that participants became highly dependent on the information provided by the Lissajous feedback. Indeed, as noted earlier Kovacs et al. (2009a) found that levels of continuous relative phase error and variability where remarkably low when tested after 5 min of practice with concurrent Lissajous feedback available but not when the Lissajous feedback was withdrawn. This finding suggests that the Lissajous plot with cursor and template provided participants a mean by which they were able to detect their coordination errors and perform the necessary corrections. This would be considered a short-term adaptation using online feedback control. On a subsequent test when the Lissajous feedback was removed values of error and variability in relative phase increased significantly. This increase indicated that participants had not learned to perform the required coordination pattern, but had learned to use the concurrent information provided to tune-in the required coordination pattern. In other words, participants were dependent on the concurrent feedback and had not developed an internal representation of the task on which they could rely in the absence of concurrent feedback. The notion of externally and internally driven bimanual response production is not new. Debaere et al., 2001, Debaere et al., 2003 has provided evidence that suggest the existence of distinct cortico-cortical and subcortico-cortical neural pathways for externally (augmented feedback) and internally guided cyclical bimanual movements (e.g., Debaere et al., 2001, Debaere et al., 2003). Presumably, the externally guided movements are dependent on the feedback available in the testing environment, while internally guided movements are not dependent on this type of extrinsic feedback.

In the motor learning literature information provided by extrinsic feedback regarding the progress and outcome of movements acquired through concurrent and terminal feedback is perceived as one of the most important variables in the learning of motor skills (Adams, 1971, Bilodeau and Bilodeau, 1958). Indeed, the feedback literature is replete with examples of benefits of providing salient information to the learner. However, this literature also chronicles the strong tendency for participants to develop dependencies on various types of feedback. In an attempt to characterize the research on feedback and to offer potential strategies to minimize problems associated with feedback withdrawal Salmoni, Schmidt, and Walter (1984) proposed the guidance hypothesis. The guidance hypothesis postulates that although frequent feedback provided during practice is very powerful in guiding the learner toward the correct response, it also results in a dependency on that feedback and blocks the processing of other important sources of information (Bjork, 1988, Schmidt, 1991, Schmidt and Wulf, 1997) that are necessary to develop an internal representation of the movement task that is capable of producing the movement when feedback is withdrawn. One of the methods used to reduce participants' dependencies on extrinsic feedback is to reduce the percent of trials with feedback. For example, Winstein and Schmidt (1990) reduced the frequency (50% of the trials) with which one group was provided feedback while providing feedback on every trial for another group. In this experiment feedback was reduced in what was called a faded schedule whereby feedback was provided on all trials at the beginning of practice with the number of trials with extrinsic feedback gradually reduced until relatively little feedback was provided at the end of practice. They found that the reduced frequency group was able to maintain performance during practice at the same level as the group provided feedback on every trial. However, on no feedback tests administered after 5 min and 24 h the reduced frequency group performed substantially better than the group that had been provided feedback on every trial. That is, participants were able to reduce their dependency on the feedback which improved their ability to produce the required movement pattern when feedback was withdrawn. Additional experiments have also demonstrated the beneficial influence of reduced frequency feedback (e.g., Lee et al., 1990, Sparrow and Summers, 1992, Weeks et al., 1993). Thus, it is possible that by reducing the amount of time that Lissajous feedback is provided during practice, test performance can be maintained at levels observed when Lissajous feedback was continually provided. According to the guidance hypothesis the reduced feedback will result in the participant processing other sources of intrinsic information that will aid the development an internal representation capable of sustaining performance when feedback is withdrawn. If these hypotheses prove true, we would expect participants to be able to greatly reduce the time required to develop an internal representation from days to a few minutes.

In sum, there is compelling experimental evidence indicating that concurrent Lissajous feedback is very effective in enhancing performance on a wide variety of bimanual coordination tasks (e.g., Kovacs et al., 2010a, Kovacs et al., 2010b, Swinnen, Lee, et al., 1997), however extensive practice still appears to be required for participants to be able to perform the task with relatively low error and variability in the absence of Lissajous feedback. That is, participants develop a dependence on the extrinsic information provided by the Lissajous feedback (Kovacs et al., 2009a) such that withdrawing the Lissajous information results in substantial increases in relative phase errors and variability. Therefore, the primary purpose of this experiment was to determine if reducing the frequency with which Lissajous feedback is provided will facilitate the development of an internal representation of the coordination pattern that will allow participants to effectively perform the bimanual coordination task in the absence of Lissajous feedback. It is hypothesized that presenting extrinsic feedback (termed knowledge of results or KR) on some but not all practice trials would improve performance on a subsequent test when extrinsic feedback is removed. One method of presenting reduced frequency feedback involves providing extrinsic feedback relatively often during the initial stages of practice and then gradually withholding the presentation of feedback more and more toward the end of practice (Winstein and Schmidt, 1990, Wulf and Schmidt, 1989). A secondary purpose of the present experiment was to determine if by simply increasing the amount of time that participants practice with Lissajous feedback on all trials will also facilitate the internalization of a representation of the required coordination pattern. We feel this is unlikely because previous research indicates that participants become increasingly dependent on extrinsic feedback when it provides salient information (e.g., Blandin, Toussaint, & Shea, 2008). We do hypothesize that participants provided reduced frequency Lissajous feedback will performs the bimanual coordination task as well as a group provided Lissajous feedback throughout practice when Lissajous feedback is provided on a Lissajous test and will be able to maintain performance when feedback is withdrawn on a no Lissajous feedback test. Simply stated, the present experiment is aimed at determining a practical method whereby salient perceptual information, which is capable of supporting the production of difficult coordination patterns when present, cam be used to develop an internal representation which sustains performance when the extrinsic information is withdrawn.

Section snippets

Participants

Right-handed college students (N = 63) with no prior experience with the experimental task served as participants. Informed consent was obtained prior to participation in the experiment.

Apparatus

The apparatus consisted of two horizontal levers and a projector. The levers were affixed at the proximal ends to near frictionless vertical axles. The axles, which rotated freely in ball-bearing supports, allowed the levers to move in the horizontal plane over the table surface. Near the distal end of each lever,

Results

Mean RMSE and VE of relative phase were analyzed in a 3 (Feedback condition: 100%, 50%, 0%) × 3 (Practice time: 5, 10, 20 min) × 2 (Test: Lissajous, no-Lissajous) ANOVAs with repeated measures on Test. Mean cycle duration was analyzed in a 3 Feedback condition (100%, 50%, 0%) × 3 Practice time (5, 10, 20 min) × 2 Test (Lissajous, no-Lissajous) × 2 Limb (left, right) ANOVAs with repeated measures on Test and Limb. These measures are presented in the bimanual and unimanual performance sections, respectively.

Discussion

The present experiment was designed to examine whether reduced frequency Lissajous feedback presented in a fading schedule would facilitate the development of an internal representation of a 1:1 bimanual coordination pattern with a 90° relative phase. Note that recent research using 100% Lissajous feedback found remarkably low levels of relative phase error and variability while performing various difficult bimanual coordination patterns including 90° relative phase after only a few minutes of

References (52)

  • D.M. Wolpert et al.

    Internal model in the cerebellum

    Trends in Cognitive Sciences

    (1998)
  • J.A. Adams

    A closed-loop theory of motor learning

    Journal of Motor Behavior

    (1971)
  • T.R. Armstrong

    Training for the production of memorized movement patterns

  • A. Banerjee et al.

    How do neural connectivity and time delays influence bimanual coordination?

    Biological Cybernetics

    (2006)
  • E.A. Bilodeau et al.

    Variable frequency of knowledge of results and the learning of a simple skill

    Journal of Experimental Psychology

    (1958)
  • G.P. Bingham

    A perceptually driven dynamical model of bimanual rhythmic movement (and phase perception)

    Ecological Psychology

    (2004)
  • G.P. Bingham

    Another timing variable composted of state variables: Phase perception and phase driven oscillators

  • G.P. Bingham et al.

    Visual perception of relative phasing in human limb movements

    Perception and Psychophysics

    (1999)
  • R.A. Bjork

    Retrieval practice and the maintenance of knowledge

  • Y. Blandin et al.

    Specificity of practice: Interaction between concurrent sensory information and terminal feedback

    Journal of Experimental Psychology: Learning, Memory and Cognition

    (2008)
  • D. Cattaert et al.

    Simulating a neural cross-talk model for between-hand interference during bimanual circle drawing

    Biological Cybernetics

    (1999)
  • F.A. Kagerer et al.

    Instabilities during antiphase bimanual movementes: Are ipsilateral pathways involved?

    Experimental Brain Research

    (2003)
  • J.A.S. Kelso et al.

    Order in time: how cooperation between the hands informs the design of the brain

  • S.W. Kennerley et al.

    Callosotomy patients exhibit temporal uncoupling during continuous bimanual movements

    Nature Neuroscience

    (2002)
  • E. Klaiman et al.

    Bimanual adaptation: Internal representations of bimanual rhythmic movements

    Experimental Brain Research

    (2006)
  • A.J. Kovacs et al.

    Bimanual 1:1 with 90° continuous phase: Difficult or Easy?

    Experimental Brain Research

    (2009)
  • Cited by (0)

    View full text