Exp Brain Res (2002) 146:205–212 DOI 10.1007/s00221-002-1179-5
RESEARCH ARTICLE
Timothy D. Lee · Quincy J. Almeida · Romeo Chua
Spatial constraints in bimanual coordination: influences of effector orientation Received: 3 December 2001 / Accepted: 28 May 2002 / Published online: 26 July 2002 4 Springer-Verlag 2002
Abstract Two experiments are reported that examined the influence of spatial orientation of the upper limbs in bimanual coordination. In both experiments, the upper limbs were oriented in either parallel, orthogonal, or obtuse spatial configurations and participants were asked to move their limbs continuously in temporal (1:1) synchrony, prepared in either in-phase or anti-phase modes of coordination. Bimanual coordination trials in Experiment 1 were paced by a metronome at one of four frequencies (1.0, 1.5, 2.0 or 2.5 Hz). Measures of relative phase accuracy and stability both revealed that, as metronome frequency increased, in-phase coordination dominated for the parallel spatial orientation, anti-phase coordination dominated for the orthogonal spatial orientation, and neither pattern dominated for the obtuse spatial orientations. In Experiment 2, an intentional switch method replicated and extended these influences of spatial orientation. The time to voluntarily switch from an anti-phase pattern to an in-phase pattern was faster than an in-phase to anti-phase switch (confirming support for the dominance of the in-phase pattern), but this was true only for the parallel spatial orientation. The reverse was true for the orthogonal spatial orientation (i.e., faster from in-phase to anti-phase), and no difference in switch times was observed for an obtuse spatial orientation. These findings support and extend previous research regarding the influence of spatial orientation in bimanual coordination and may be attributed to the role of, and potential interactions between, egocentric, allocentric, and mechanical constraints during action. The present research was supported by operating grants from the Natural Sciences and Engineering Research Council of Canada awarded to T.D.L. and R.C. T.D. Lee ()) · Q.J. Almeida Department of Kinesiology, McMaster University, 1280 Main St. West, Hamilton, Ontario L8S 4K1, Canada e-mail:
[email protected] Fax: +1-905-5236011 R. Chua University of British Columbia, British Columbia, Canada
Keywords Motor · Bimanual · Coordination · Spatial constraints · Egocentric allocentric
Introduction Coordinating the hands to perform everyday actions is such a natural and common activity that we rarely give it much thought. Yet, we still manage to carry out the actions with few errors. For example, when tossing a salad the upper limbs operate in spatial and temporal congruence. Coordinated movements of this type have an egocentric frame of reference, so named because their actions are centered with respect to the midline of the body (e.g., Swinnen et al. 1997). In such egocentric actions the upper limbs move toward and away from the midline reference point. This type of bimanual coordination is often described as moving in mirror-image symmetry, or simply in-phase. These in-phase movements are inherently easy to perform and show minimal improvement with additional practice (e.g., reviews of early and contemporary research can be found in Kelso 1995; Schmidt and Lee 1999, chap. 8; Turvey 1990). Considerable research during the past 2 decades has been directed at the performance of continuous bimanual coordination of the upper limbs, most often by contrasting the accuracy and variability of in-phase coordination with anti-phase movements – coordinated actions in which the effectors move in the same direction, in spatial and temporal asymmetry with respect to an egocentric reference. In general, anti-phase movements are coordinated as accurately and as consistently as in-phase movements when performed at a preferred or relatively slow frequency (Kelso 1984). However, the performance of anti-phase movements deteriorates when the cycling frequency is increased. The anti-phase pattern loses stability and, at a critical frequency, often switches to and remains in an in-phase coordination mode unless resisted by intention (e.g., Kelso 1984). Haken et al. (1985) have captured these features of bimanual coordination (e.g., differential stability, pattern transitions) in a
206
model based on concepts and principles from a general framework of dynamical pattern theory (e.g., Kelso 1995). Briefly, the HKB model proposed by Haken et al. (1985) maps stable patterns onto attractors of a dynamical system’s potential landscape. The relative stability and strength of attractors are determined by the depth and slope of an attractor’s basin within the potential landscape. Changes to the landscape, and to attractor basins, can lead to changes in pattern stability and transitions to new patterns. Recent work by Swinnen and colleagues has raised the issue of how spatial constraints modify or otherwise interact with typical bimanual coordination performance characteristics (Serrien et al. 1999; Swinnen et al. 1997, 1998). For example, in an important study by Swinnen et al. (1998), participants moved hand-held styli over two digitizing tablets in time to an auditory metronome (1.33 Hz). Of interest to us were the two spatial directions of motion and their mode of coordination. Movements of the left or right hand were performed in either a left to right motion in the frontal plane (along the X-axis) or toward and away from the body (along the Y-axis). Bimanual coordination was examined when both limbs moved in spatial symmetry (e.g., both limbs moving along the X-axis) and when the limbs moved in spatial asymmetry (e.g., one limb along the X-axis and one limb along the Y-axis). Regardless of the spatial orientation of the limbs, coordination was considered to be in-phase when the limbs arrived at their most proximal referent point simultaneously (and, conversely, when the limbs arrived simultaneously at the point furthest from egocentric reference). Coordination was considered to be antiphase when the opposite phasing occurred (e.g., when one limb was egocentric, the other limb was at its furthest point from being egocentric). Typical of research in bimanual coordination, Swinnen et al. (1998) found that in-phase performance was better than anti-phase performance when the spatial orientations of the two limbs were symmetric (e.g., both moving along the X-axis). This finding was true for coordination accuracy (average mean error in relative phase) and consistency (standard deviation about the mean relative phase). In contrast, under conditions of spatial asymmetry, the anti-phase coordination pattern was either equal to or better than the in-phase pattern. Spatial orientation appeared, therefore, to be a critical factor or constraint in determining the performance characteristics of in-phase and anti-phase spatial-temporal coordination patterns. The Swinnen et al. (1998) study, together with other studies from Swinnen’s laboratory (e.g., Bogaerts and Swinnen 2001; Swinnen et al. 2002) and by others (e.g., Carson et al. 2000), suggest that modifications to the original HKB model of bimanual coordination (Haken et al. 1985) are necessary in order to account for spatial symmetry influences. Indeed, in a recent paper, Fuchs and Jirsa (2000) have proposed an extension to the HKB model to account for symmetry in the dynamics of coordination. The modifications to the Haken et al. equation come in the form of a symmetry parameter,
Fig. 1a–d Illustration of the experimental conditions, depicting the two coordination patterns under different spatial orientations (Experiment 1)
which captures the degree of symmetry breaking in the dynamics, and influences the stability of intrinsic coordination states (Fuchs and Jirsa 2000). The present studies further investigate the role of spatial symmetry and its influence on bimanual coordination. Two studies are reported here which provide additional experimental evidence that spatial orientation influences spatial-temporal bimanual coordination.
Experiment 1 In the present study we included two of the spatial orientation-movement conditions used by Swinnen et al. (1998) – their “XX” (spatially symmetric) and their “XY” (spatially asymmetric) conditions – together with two intermediate conditions in which the angle of spatial orientation between the limbs was obtuse (between 90L and 180L). The four spatial orientations used in this study are illustrated in Fig. 1 (note that Fig. 1a, spatial angle = 90L, is similar to the asymmetric condition used by Swinnen et al. 1998; Fig. 1d, spatial angle = 180L, is similar to the symmetric condition in Swinnen et al.). In addition, similar to previous studies using movement frequency as a control parameter, we required participants to perform trials at frequencies matching a driving
207
metronome frequency that varied between 1.0 and 2.5 Hz. Thus, the goal of the present study was to examine the performance of in-phase and anti-phase patterns, at different movement frequencies, under conditions of varying degrees of spatial symmetry/asymmetry. Our predictions were: (a) that in-phase coordination would be the dominant pattern under conditions of spatial symmetry (Fig. 1d), (b) that anti-phase coordination would be the dominant pattern under conditions of spatial asymmetry (Fig. 1a), (c) that neither pattern would be dominant when the spatial orientation was obtuse (Fig. 1b, c), and (d) that pattern dominance would be smaller under lower frequency metronome conditions (1 Hz) than under higher frequency conditions (2.5 Hz). Participants The research reported in this experiment (and in Experiment 2) was approved by the McMaster University Research Ethics Board and complies with the ethical standards of the 1964 Declaration of Helsinki regarding the treatment of human participants in research. Written, informed consent was obtained from all subjects prior to the start of their participation. Forty undergraduate students (age range 20–23 years) from McMaster University volunteered to participate in the present study. The participants were assigned at random to one of four experimental groups (corresponding to different spatial orientations) with the restriction that each group be represented by an equal number of males and females and also balanced for handedness. All participants were naOve to the purpose of the experiment. Apparatus The apparatus used for this experiment is illustrated in Fig. 1. It consisted of two identical linear sliding devices, each attached to linear potentiometers to encode displacement. Each linear slide consisted of a 10-cm plastic handle that was attached vertically to a metal slide carriage (9P13 cm). Four ball bearing units were located under the carriage, which rode atop two metal rods and permitted low friction, linear movements. The rods were attached to a metal base, which was fixed to the tabletop. The spatial orientation of the two linear slides was a critical feature of the present experiment. As illustrated in Fig. 1a, the spatial orientation for the Orthogonal group was 90L (i.e., the left limb moved to the left and right in front of the body; the right limb moved toward and away from the body). For the Parallel group both limbs moved to the left and right in front of the body (resulting in a 180L spatial orientation; see Fig. 1d). For the other two groups, the spatial angle separating the right limb in relation to the left limb was either 120L (the Obtuse 120 group; see Fig. 1b) or 150L (the Obtuse 150 group; see Fig. 1c). The reversal points for left and right limb movements were marked on the base of the apparatus.
The distance between the centermost and outermost positions was 16 cm. The linear potentiometers were connected to an 80486 microprocessor, which sampled data at a frequency of 200 Hz. Customized software controlled the timing of events for each trial and delivered a metronome pulse to a tone generator. Data were stored for later analyses. Procedure Participants were assigned at random to one of four separate groups, each group corresponding to a different spatial orientation (parallel, obtuse 150, obtuse 120, orthogonal). Each participant completed a total of 40 trials, comprising 5 blocks of 8 trials per block. Each block consisted of four in-phase and four anti-phase trials, paced by the auditory metronome at frequencies of 1.0, 1.5, 2.0, and 2.5 Hz. The ordering of trials was not random. Rather, a sequence of four trials, in ascending order from 1.0 to 2.5 Hz, was completed for one coordination pattern (e.g., four in-phase trials) before a series of four trials was completed for the other pattern (e.g., four anti-phase trials). The sequence of patterns was counterbalanced in ABBA order. Trial duration was 15 s. Instructions to the participants emphasized two procedural matters. First, they were encouraged to try to maintain their movement speed along with the metronome frequency, completing one full cycle of movement per metronome beat. Second, participants were told to try and maintain the initial coordination pattern throughout the trial, with the deliberate intention to regain the pattern if it was destabilized during a trial (cf. Lee et al. 1996). Regardless of spatial orientation, movement patterns were defined in terms of egocentric coordinates (similar to Swinnen et al. 1998). The goal of the in-phase pattern was to move both limbs such that they arrived at their closest, centermost position (with respect to the body midline) at the same time and arrived at the extreme excursion points at the same time (as illustrated on the left side of Fig. 1a–d). The primary joint involved in the production of these movements was the elbow. In all cases, simultaneous flexion (or simultaneous extension) of both elbow joints was required to perform in-phase movements. In contrast, anti-phase movements were described as actions in which one limb moved toward (i.e., elbow flexion) the closest, centermost position at the same time that the other limb moved away from (i.e., elbow extension) the closest, centermost position (as illustrated on the right side of Fig. 1a–d). All participants were asked to move as steadily and continuously as possible. The timing of one limb within its cycle relative to the timing of the other limb (the phase angle) within its cycle was the primary measure of coordination performance, using the formula suggested by Scholz and Kelso (1989, p. 129). These measures of relative phase were determined at the reversal points of the left limb, providing approximately 30–75 data points per trial. Two measures
208
of performance were then calculated from each set of data points per trial. First, the mean relative phase was the average of the data points collected per trial and was expected to be near 0L for in-phase trials and near 180L for anti-phase trials. In order to compare these more directly, each mean relative phase score was recast as an error score by determining the absolute difference between the observed mean score and the intended score (i.e., 0L or 180L). These mean error scores were determined on a trial-by-trial basis. Second, the standard deviation was used as a measure of within-trial coordination consistency – the average discrepancy by which each relative phase data point deviated from the trial mean. All participants performed five replications in each condition. Therefore, for each individual, the median mean error and the median standard deviation for each condition were determined and used in further statistical analyses. Mixed design, three-factor ANOVAs were used to analyze each dependent measure: a 4 (spatial orientation group) P 2 (coordination pattern) P 4 (metronome frequency) design, with repeated measures on the pattern and frequency factors. Differences between means involved in significant ANOVA effects were tested for significance using Tukey’s HSD procedure. We used the 0.05 level for statistical significance in all tests.
Fig. 2 Mean relative phase error as a function of movement frequency and spatial orientation in Experiment 1
Results and discussion The two measures of coordination performance revealed similar results, which are illustrated in Figs. 2 and 3. Mean error and standard deviation results corresponded to the following trend. There were no differences between the performance of the in-phase and anti-phase patterns at the 1.0 and 1.5 Hz metronome frequencies, regardless of the spatial orientation of the limbs. For the parallel group (180L spatial orientation), the in-phase pattern was performed with low error at all frequencies, whereas the performance of the anti-phase pattern deteriorated significantly with increasing speed. The opposite trend in results occurred for the orthogonal group (90L spatial orientation): the anti-phase pattern was performed with low error at all frequencies, whereas a significant deterioration in performance occurred for the in-phase pattern with increasing speed. Continuing with this trend, the groups that were spatially oriented “between” the parallel and orthogonal groups produced a set of results that were between the two sets of extremes established by the parallel and orthogonal groups. The only exception to this trend was the in-phase, standard deviation performance for the Obtuse 150 group, which was maintained across metronome frequencies. Statistical analysis of the mean error data resulted in significant main effects and two-way interactions for all factors (minimally at P
