See

discussions, stats, and author profiles for this publication at:

https://www.researchgate.net/publication/21395379

Spatial topological constraints in

a bimanual task

Article in Acta Psychologica · October 1991

DOI: 10.1016/0001-6918(91)90028-X · Source: PubMed

CITATIONS READS

201 186

3 authors, including:

Liz Franz
University of Otago
90 PUBLICATIONS 1,820 CITATIONS

SEE PROFILE

Howard Zelaznik
Purdue University
113 PUBLICATIONS 4,110
CITATIONS

SEE PROFILE

Available from: Howard Zelaznik

Retrieved on: 18 September 2016
Acta Psychologica 77 (1991) 137-151 137
North-Holland

Spatial topological constraints

in a bimanual task *

Elizabeth A. Franz, Howard N. Zelaznik and George McCabe

Piudue Uniuersify, West Lafayette, USA

Accepted March 1991

Previous research has shown that the concurrent performance of two manual tasks results in a
tight temporal coupling of the limbs. The intent of the present experiment was to investigate
whether a similar coupling exists in the spatial domain. Subjects produced continuous drawing of
circles and lines, one task at a time or bimanually, for a 20 s trial. In bimanual conditions in which
subjects produced the circle task with one hand and the line task with the other, there was a clear
tendency for the movement path of the circle task to become more line-like and the movement
path of the line task to become more circle-like, i.e., a spatial magnet effect. A bimanual circle
task and a bimanual line task did not exhibit changes in the movement path when compared to
single-hand controls, In all bimanual conditions, the hands were tightly temporally locked. The
evidence of temporal coupling and concomitant accommodation in the movement path for the
conditions in which the hands were producing different shapes suggests that spatial constraints
play a role in the governance of bimanual coordinated actions.

A recent approach toward examining the nature of constraints in

human motor actions has been to identify the characteristics that
appear to be controlled jointly for the two limbs, i.e. global aspects,
and those which are controlled separately for each limb, i.e. local
aspects (Heuer 1985). Identifying the global aspects of coordination

* Special thanks to J. Hultsman for many helpful comments on earlier drafts of this paper and to
S. Wootton for help in debugging the algorithm used to quantify the results. We are also indebted
to the members of R. Melara’s lab for their contributions and helpful questions.
This project was a continuation of master’s thesis work done by E. Franz under the advisement
of H. Zelaznik. During the initial stages of the project E. Franz was funded by a David Ross XR
Grant awarded to H.N. Zelamik and a NIH Grant No. 510-1353-2578 awarded to A. Smith, H.N.
Zelaznik and C. McGillum.
Requests for reprints should be sent to E.A. Franz, c/o H. Zelaznik, Motor Behavior
Laboratory, PEHRS, Purdue University, West Lafayette, IN 47907, USA.

OOOl-6918/91/$03.50 0 1991 - Elsevier Science Publishers B.V. All rights reserved

138 E.A. Franz et al, / Spatial constraints

provides insights into the central nervous system constraints that

govern movement control. An abundance of empirical support has
indicated that the timing properties of movements are powerful global
constraints. Studies involving bimanual tapping (Klapp 1979; Peters
1977; Yamanishi et al. 1980) and timing tasks of anatomically distinct
systems (Chang and Hammond, 1987; Klapp 1981; Muzii et al. 1984;
Smith et al. 1986) have demonstrated the difficulty people encounter
when attempting to produce two or more timing patterns that are not
related by a simple harmonic. Movements that require advance plan-
ning also have shown strong evidence of temporal constraints (Heuer
1985; Konzem 1987, cited in Schmidt 1988). Performance of both
discrete (Konzem 1987, cited in Schmidt 1988) and continuous (Heuer
1985) bimanual movement tasks is difficult when the production of
different movements is required by the two hands. From these findings
it is inferred that the spatio-temporal form of movements is a con-
straint controlled jointly for the two hands.
The treatment of dual-task motor constraints can be conceptualized
according to two prevailing views - the motor program and dynamic
pattern perspectives. Motor program theorists assert that the global
aspects of an action (relative motion patterns) are abstractly repre-
sented in a program, and the details of the action are prescribed as
mutable parameters. Accordingly, a topological representation guides
the movement but a number of variations in the details of the program
may occur. The assumption based on traditional bottleneck theories of
response execution (Keele 1973; Welford 1967) is that two different
motor programs cannot be executed concurrently.
On the other hand (no pun intended), proponents of the dynamic
pattern perspective assert that actions result from self-organizing prop-
erties of biological systems. This view, put forth by Bernstein (1967)
and rekindled by Turvey (1977) is consistent with the notion that the
temporal and topological characteristics of movements are emergent
properties of the dynamical structure of organisms. The notion of
oscillators, as extended by the seminal work of von Holst (cited in
Gallistel 1980) provides a basis for understanding this perspective. Out
of von Holst’s work on the fin movements of fish emerged concepts
such as the magnet effect and relative coordination. The former refers
to the effect one oscillator imposes on another which results in the
sustainment of a mutual phase relationship. The latter occurs when the
slowing down or speeding up of one oscillator influences the magnitude
 E.A. Franz et al. / Spatial constraints 139

of response of another oscillator, even though a 1 : 1 frequency ratio is

not necessarily maintained. In other words, certain effecters may speed
up and slow down because of the influence of other effecters.
The work of Kelso et al. (1979) has been interpreted in light of the
dynamic pattern perspective. It was demonstrated in a bimanual Fitts
aiming task that each hand did not obey Fitts’ Law when a high index
of difficulty movement assigned to one hand was combined with a low
index of difficulty movement assigned to the other. The low index of
difficulty hand produced a larger movement time than would be
expected based on its single hand movement time, in an attempt to
accommodate the movement time of the high difficulty hand (but see
Corcos (1984) and Marteniuk et al. (1984) for slightly different inter-
pretations).
It is apparent from the abundance of literature on timing that the
examination of temporal constraints has been the dominant approach
used to investigate coordinated movements. An investigation of the
contributions of spatial topological constraints in the control of such
tasks, therefore, seems worthwhile. Recent evidence has, in fact, sug-
gested that spatial constraints may play a significant role in coordina-
tion.
In a bimanual Fitts aiming task, Kelso et al. (1983) placed an
obstacle in the movement path of one limb but not the other. The limb
that did not go over the obstacle produced a movement trajectory
similar in form to that produced by the limb that ‘hurdled’ over the
real obstacle. Kelso et al. (1983) claimed that the limb that jumped over
the phantom obstacle produced this trajectory in an attempt to main-
tain temporal synchrony. An equally valid interpretation of these
findings would be that the limbs were constrained spatially to move
with similar trajectories.
Additional evidence (Swinnen et al. 1988; 1990) suggests that devia-
tions from temporal synchrony occur when the required limb move-
ments are dissimilar in their spatio-temporal trajectories. In these
experiments, one limb was required to produce an elbow flexion
(unidirectional) movement, and the other limb was required to produce
an elbow flexion-extension-flexion (reversal) movement. Initial perfor-
mance of the bimanual task revealed a strong tendency for the reversal
movement to impose its spatio-temporal pattern on the movement of
the limb producing the unidirectional movement.
The extant literature includes numerous examples of tasks whose
140 E.A. Franz et al. / Spatial constramts

temporal characteristics have been emphasized over the spatial char-

acteristics. In tasks in which the spatio-temporal patterns of the move-
ments were complex (e.g. Swinnen et al. 1988) the precise timing of the
flexion-extension movements was required for their successful execu-
tion. Thus, it is clear that the timing control has been the dominant
feature of these tasks. Most theorists in motor control postulate that
the spatio-temporal form, i.e. pattern, of movements is the determinant
of control in coordinated activity despite the lack of knowledge of how
spatial constraints contribute to such actions.
The present experiment is our initial attempt to examine the spatial
rules of coordination. Because of the distinction that has been made in
the neuromotor science literature concerning the fundamental nature of
straight line versus curved line trajectory formation (Hollerbach 1981;
Hollerbach and Flash 1982; Morass0 1981) we chose to employ tasks
of these two forms, lines and circles. Specifically, we assumed, a priori,
that a line movement is of a different spatial form than a circle
movement .
To apply a spatial analog of the temporally-based notions extended
by von Holst, we attempted to determine whether absolute or relative
spatial coordination would be observed in the concurrent production of
circle movements by one hand and line movements by the other hand.
One might think of absolute spatial coordination as the effect of one
spatial form taking on the form of the other, i.e. a line becomes a circle.
The term relative spatial coordination might be applied if the shapes of
both tasks take on characteristics that are common, yet the line remains
line-like and the circle circle-like. Continuous circle drawing move-
ments and line drawing movements were performed in three types of
experimental conditions: (1) producing circles or lines with one hand
alone (single), (2) producing circles or lines with both hands (dual-same),
or (3) producing circles with one hand and lines with the other
(dual-different).

Method

Subjects

The subjects were four male and four female right-handed graduate and under-
graduate volunteers from Purdue University. All subjects were naive to the purpose of
the study.
 E.A. Franz er al. / Spatial constraints 141

The apparatus consisted of a standard rectangular desk (90.0 cm long, 65.0 cm wide,
72.0 cm high) covered by black posterboard. Two white sheets of paper (28.0 cm x 36.0
cm) were placed on the desk. An audio amplifier was interfaced via a Scientific
Solutions 12 bit Digital to Analog converter, with a Zenith ZW-248 computer. A
mallory Sonalert 1800 Hz buzzer also was interfaced with the Zenith computer via a
digital output board.
Infrared light emitting diodes (IREDS) were mounted just above the writing tip of
each of two 0.9 mm mechanical pencils. Ireds were also placed at each of three joints -
shoulder, elbow, and wrist - of each limb. ’ All ireds were sampled at 250 Hz via a
Watsmart infrared recording system which was interfaced with a Compaq 386/16
computer. The static calibrations for each session ranged from 2.1-3.0 mm of RMSE.

Task

Two tasks were performed: a line task and a circle task. The line task consisted of
drawing reciprocal lines in rhythm for a 20 s trial along the Y-dimension. The circle
task consisted of drawing circles in rhythm (the direction of motion was not specified).
Each task was to be drawn the approximate size indicated by templates which were
presented to the subject at the beginning of the testing session. The line template was
27.0 cm long and the circle template was 54.0 cm in circumference. Both tasks were to
be performed with a ‘stiff wrist and only the point of the pencil was to touch the
drawing surface. Tasks were to be performed in pace with a metronome (600 ms per
cycle). The metronome pace was terminated midway through each 20 s trial and the
subject was to self-pace the remainder at the 600 ms cycle duration.

Conditions and design

Subjects performed four single-hand conditions (each hand performed lines or

circles alone) and four dual-hand conditions (two tasks by two hands). The dual task
conditions will be referred to as dual-same (both hands attempted to produce the same
shape) or dual-different (one hand attempted to draw circles and the other attempted
to draw lines). There were eight trials per condition. Subjects performed single-hand
conditions first (with those conditions randomized across subjects) followed by dual-
hand conditions (also randomized).

Procedure

Upon entering the laboratory, the task requirements were explained after which
each subject was asked to read and sign an informed document. A template of a line,

i The purpose of this paper was not concerned with the joint angle kinematics. Before each testing
session, we measured the subject’s limbs in order to use anthropometric measures in our analyses
at a later date. We are fully aware of the importance in analyzing these data but at the present
time our software programs are in their preparatory stages of development.
142 E.A. Franz et al. / Spatial constraints

27.0 cm in length, oriented along the Y-axis and another template of a circle, 17.2 cm in
diameter (54.0 cm in circumference), were placed on the desk to indicate to the subject
the shape and size for each task. The subject was informed that the templates were to
be used only to indicate the approximate size task to draw, and that the shape of the
movement was more important than the exact size.
A trial began by a verbal ‘ready’ signal from the experimenter. The metronome then
began and the subject performed the assigned task(s) in pace with the metronome.
After 10 seconds the metronome terminated and the subject was instructed to continue
performing the task(s) for another 10 seconds when a signal to stop occurred.

Results

Data reduction

The kinematic displacement data were filtered digitally, forward and backward, at a
cut-off frequency of 10 Hz. A three-point difference technique was utilized to generate
velocity measures. Numerical algorithms were used to ascertain kinematic landmarks.
The first and last seconds of each trial were omitted, resulting in 3 equal time intervals
each lasting 6 s.

Spatial effects

The line trajectory of a typical dual-different trial is depicted in fig. 1. The

trajectory of lines became elliptical when the other hand produced circles concurrently

Lines

Single-hand dual-hand dual-hand

SQrne different

Fig. 1. Typical displacement in mm in the X and Y dimensions of the line task performed alone
(single), when the other hand also produced lines (dual-same) and when the other hand produced
circles (dual-different), for a 20 s trial.
 E.A. Franz et al. / Spatial constraints 143

Circles
I

tingle-hand dual-hcnd dual-hand

same different

Fig. 2. Typical displacement in mm in the X and Y dimensions of circles when performed alone
(single), when the other hand also produced circles (dual-same) and when the other hand
produced lines (dual-different), for a 20 s trial.

(dual-different) but not when the other hand produced a line (dual-same) or in the
single hand condition (single).
When the circle task was combined with the line task the circle topologies tended to
take one of two forms. Either they became elliptical or they produced lines intermit-
tently with the circles. The majority of cases were like those depicted in fig. 2 which
illustrates that the path of the pencil became less circular and more elliptical when
paired with the lines. Very little spatial disruption was observed when two circles were
paired (dual-same) or when circles were performed alone (single).
In order to quantify the shape of the two tasks, an index of circularity was
computed1 for each period of these repetitive movements. If a subject produced a
perfect circle, then the ratio of the two diameters should be one. On the other hand, if
the X axis diameter is divided by the Y axis diameter of a movement that is
approaching a perfect line, then the ratio should be equal to zero. This ratio measure
was devised as an index for the shape of the circle and line tasks produced.
The index of circularity measure was based on four points from the Y-dimension
velocity profile that comprised one cycle of movement. The four points were the first
zero cross, peak positive velocity, the second zero cross, and peak negative velocity. The
X and Y position coordinates at each of these velocity landmarks were used to
compute distances joining the points at the zero crossings and the points at the peak
velocities. The ratio of these two distances was computed with the larger of the two
numbers as the denominator. Thus, the index of circularity was within the range of 0
(most linear) to 1 (most circular).
Fig. 3 depicts the index of circularity for the three conditions, (1) single-hand, (2)
dual-same, and (3) dual-different, for the line and circle tasks across the 8 trials. As
144 E.A. Franz et al. / Spatial constraints

Circles

x
.E 0.7 -
D
0.6 - ,* .* .*. .* .* ” a ” a
z
.$ t’
0.5 -
x 0.4 -
v
D 0.3 -
& Lines
0.2- 0..
‘D Et 0 43 .D .a
0.1 - ~+-a-_&q-+,

0.0 ’ B c ’ a c ’ ’ s
1 2345678

Trial
Fig. 3. Mean index of circularity for lines and circles under the 3 conditions, single (dashed),
dual-same (solid) and dual-different (dotted) for the 8 trials. Data are collapsed across left and
right hands. Circle data are depicted as filled dots and line data are depicted as open squares. A
perfect circle would be an index of 1.0 and a perfect line would be an index of 0.0.

depicted in the figure, these measures for lines and circles do not overlap, F(1,7) =
762.68, p < 0.001, which indicates that circle tasks could be distinguished from line
tasks even under the conditions that are presumably most difficult (dual-different). As
seen in fig. 3, the magnitude of change in the index of circularity for all circles was
greater than that for all lines.
The interaction of condition and task (circle or line) was significant, F(2,14) = 26.74,
p < 0.001. Planned comparisons of the mean index of circularity produced in the single
and dual-same conditions versus those produced in the dual-different condition were
significant for both the line and the circle tasks, respectively, F(l,14) = 17,13 and
F(1,14) = 60.66, p’s < 0.001 in both cases. This is depicted in fig. 3 by the decrease in
the ratio metric for circles in the dual-different conditions and the increase in the ratio
metric for lines in the dual-different conditions. ’
The main effect of hand was not significant, F(1,7) = 5.13, p > 0.05, which suggests
that the disruption in spatial trajectories was not markedly different for the two hands.
A trial b,y task interaction, F(7,49) = 5.58, p < 0.001, and a three-way interaction
between trial, condition and task, F(14,98) = 5.52, p < 0.001, also were found. As
indicated in fig. 3, the index of circularity for the first trial most likely produced these
relatively small interactions (see the Appendix for a table of means and standard
deviations for the spatial data).
The main findings from the spatial analysis indicated that the index of circularity of
the line task increased and the index of circularity of the circle task decreased in

2 All interactions that occurred across both tasks together will not be reported because, due to the
directional nature of the index of circularity, circle ratios get smaller and line ratios get larger as
more accommodation in the tasks occurs.
 E.A. Franz et al. / Spatial constraints 145

Table 1
Means and standard deviations of half period and VE for lines and circles across the three
conditions.

Condition
Single-hand Dual-hand Dual-hand
same different
Lines
M 284 (33) 296 (63) 292 (09)
VE 43 (18) 38 (21) 35 (10)
Circles
M 287 (21) 290 (10) 289 (24)
VE 34 (11) 33 (16) 28 (08)

Note: Means and (standard deviations) are represented in ms.

conditions in which line tasks were combined with circle tasks. These results provide
evidence for relative spatial coordination, i.e. a spatial magnet effect.

Temporal effects

The temporal measures of interest were the average time between successive velocity
peaks, i.e. the half period, and its within-subject standard deviation, commonly called
variable error, VE. Half periods were measured as the time between two successive
peaks in the velocity profile (scalar absolute values) in the Y-dimension of motion.
With respect to the dependent measure, half period, a main effect of time interval,
F(2,14) = 8.88, p -C 0.01, indicated that the first 6 seconds of all trials averaged
together were slightly faster than the subsequent 12 seconds. The mean cycle times for
intervals 1, 2, and 3, were 285 ms, 291 ms, and 292 ms, respectively which indicates that
subjects slowed down only 7 ms on average during each trial, across all conditions. An
interaction between condition and interval, F(4,28) = 4.91, p < 0.01, indicated that
single and dual-same conditions slowed down by approximately 10 ms and 5 ms,
respectively, while dual-different conditions maintained the same average speed
throughout the trial.
Of primary importance with respect to the temporal data was the lack of an
interaction between task and condition, F(2,14) < 1, which indicates that all tasks were
produced at the same average speed under single- and dual-hand conditions (see table
1). The absence of such an interaction suggests that our temporal manipulation
(metronome pace) was successful and that subjects were able to maintain the same
average temporal metric even after the metronome was turned off.
For VE in half period a main effect of interval was found, F(2,14) = 12.89,
p < 0.01, indicating a larger VE for the first 6 s of the trial than for the remaining 12 s
(38 ms compared to 35 ms). A significant interaction of hand and trial also was found,
F(7,49) = 4.19, p < 0.01, wherein the VE became slightly smaller for the left hand
across trials while the VE for the right hand remained the same.
146 E.A. Franz et al. / Spatial constramts

In sum, the lack of robust, and in our minds meaningful, temporal effects for the
interaction of task and condition holds considerable importance. The lack of such
effects indicates that the hands remained tightly temporally locked regardless of
whether they produced the same tasks or two different tasks, circles and lines.

Discussion

The present study examined the possible influence of spatial con-

straints in the coordination and control of bimanual movements. Sub-
jects when asked to produce two movements of different spatial forms,
circles and lines, tended to exhibit spatial accommodation in the
performances of both tasks. In most cases this was manifest in a
tendency for each task to look more like the task being performed by
the other hand. These effects in the spatial domain occurred despite a
tight temporal coupling of the limbs which suggests that a simple
temporal explanation would not apply. It appears that a spatial magnet
effect may be operating and it is the nature of that presumed effect that
remains to be explained.
It seems a parallel can be drawn between previous findings in the
temporal domain of bimanual movement tasks and our present find-
ings in the spatial domain. The majority of previous work has required
subjects to perform two different movements as quickly as possible
(i.e., Kelso et al. 1979; 1983; Konzem 1987, cited in Schmidt 1988), or
two different movements in a discrete fashion where the precise timing
of flexions and extensions of the limbs formed the basis of the
temporal structure of the movement (Swinnen et al. 1988). These
movement tasks all involved strict temporal requirements and minimal
spatial demands. In such tasks the typical finding is a tendency of one
hand to influence the temporal structure of the other hand’s movement.
Simultaneity in the initiation and termination of movements has been
reported (Kelso et al. 1979) albeit departures from simultaneity appear
to be the rule rather than the exception (Corcos 1984; Martenuik et al.
1984; Swinnen et al. 1988). These findings provide strong evidence of
relative coordination in timing, i.e. a magnet effect.
In the present approach the opportunity for spatial deviations in the
trajectories was maximized. Unlike previous work, any number of
possible trajectories could have been produced because the movements
were virtually unaffected by external physical constraints (other than
 E.A. Franz et al. / Spatial constraints 147

the desktop surface). A relaxed temporal metric was employed so that

the average speed of movement production would be maintained
throughout the different task conditions. Our findings indicate that the
two tasks did not become identical, rather the spatial topology of one
task tended to influence that of the other. These findings provide
evidence of relative spatial coordination, i.e. a spatial magnet effect. As
in the temporal domain of movements, the effect of coupling in space is
not absolute but relative (see Smith and Zelaznik (1991) for a review of
coupling effects). It is reasonable to assume that the same processes are
operating whether the basis of the observed magnet effect is temporal
or spatial.
One possible explanation for the results of the present experiment
might be that movement organization is based on the minimization of
one or several control parameters. The optimization of certain dynamic
variables has been postulated as a possible mechanism in the formation
of smooth single limb trajectories (Abend et al. 1982; Flash and Hogan
1985). The minimum jerk model explains trajectory formation by
positing that a smooth trajectory is a desired one. Smoothness is
maximized by minimizing the sum of the mean square jerk, i.e., the
transient changes in acceleration. If a minimum jerk model could be
applied to bimanual movements it would be expected that the two
limbs operate by minimizing total jerk. Because the smallest magnitude
of jerk results in a straight line trajectory, the prediction for the tasks
employed in the present study would be that circles become more
line-like while lines remain as straight lines. The observation that lines
in the present study became more circle-like suggests that a strict
version of this minimization model would not apply.
An important feature of the minimum jerk model, relevant to our
understanding of trajectory formation, is that predictions are based
solely on hand trajectories without regard to specific neuromuscular
variables. However, while it is generally believed that movement repre-
sentation can occur either in joint coordinate space (Abend et al. 1982;
Hollerbach and Atkeson 1986; Soechting and Lacquaniti 1981; Viviani
and Terzuolo 1982) or in endpoint trajectory space (Hollerbach and
Flash 1982; Merton 1972; Morass0 1981; Viviani and Terzuolo 1980),
the rejection of a strong version of the minimum jerk model does not
necessarily exclude either of these possibilities. Although the minimum
jerk model is based on the kinematics of the hand movement in
extracorporeal space, it is possible that the smoothness of hand trajec-
148 E.A. Franz et al. / Spatial constraints

tories would result from the internal mechanics of the neuromuscular

system.
An intimate relationship exists between torque and joint angle
movement because torque must be generated by the musculature to
produce changes in joint angles (Hogan 1984). Thus, another likely
candidate for minimization may be the rate of change in torque.
Accordingly, the minimum torque change model, proposed by Kawato
and colleagues (Kawato et al. 1990; Uno et al. 1989), purports that
limb movements find a unique trajectory by minimizing torque change.
Unlike the minimum jerk model, this model depends on the dynamics
of the musculoskeletal system. The model involves minimizing the sum
of the square of the first derivative of torque. Because the torque
produced by the components of the musculoskeletal system and the
objective amount of force produced by the movement are not isomor-
phic, exact predictions based on this model involve elaborate measure-
ments of neuromuscular variables. Thus, although the model would be
difficult to implement for bimanual movements, it seems reasonable to
conjecture that the change in torque is a plausible control variable
minimized for the two limbs combined.
We are suggesting that the way in which the bimanual movement
system circumvents some of the problems that arise in controlling its
vast number of degrees of freedom is by a single minimization princi-
ple. We believe that this attempt toward minimization forms the basis
of the spatial magnet effect like that observed in the present work.
Specifically, we are postulating that the minimization of relevant con-
trol parameters provides a constraint that operates by the two limbs
combined in order to reduce the many degrees of freedom in producing
coordinated movements.
The present findings suggest that spatial effects hold considerable
importance in bimanual movement control. At least two main issues
for future work remain. One is concerned with representation, and the
other is concerned with spatio-temporal constraints in the neuromuscu-
lar system. The minimization models put forth as possible mechanisms
of control in bimanual movements address each of these issues. Whether
the spatio-temporal constraints are emergent characteristics of the
neuromuscular dynamics or whether they are prescribed by higher-order
cognitive mechanisms, are issues that are difficult to resolve.
Insight regarding the relative contributions of higher-order processes
and lower levels of constraint and the way the two levels interact, might
 E.A. Franz et al. / Spatial constraints 149

be gained if the appropriate manipulations are applied. It is possible,

for example, given an appropriate higher-order organizational frame-
work, that one might be able to produce circles and lines concurrently
without producing spatial accommodation effects. Such evidence would
suggest that higher-order constraints may be able to override the
spatio-temporal constraints that occur at the lower levels of the system.
This possibility is currently under investigation in our laboratory. Our
laboratory is also investigating the effects of practice in these tasks to
ascertain whether the inherent constraints can be altered through the
development of new coordination strategies.
In sum, the present study provides strong evidence that spatial
constraints play an important role in the coordination of bimanual
movement control. Circles, when performed concurrently with lines,
became more line-like and lines became more circle-like. These spatial
effects occurred despite a tight temporal coupling. Such findings may
be interpreted as support for a spatial magnet effect which occurs in
the coordination of bimanual movements that require different spatial
patterns. The nature of this accommodation may be in a minimization
principle that operates for the combined efforts of the two limbs.

Appendix

Table A.1
Index of circularity means and standard deviations for lines and circles for the three conditions
collapsed across trials.

Condition
Single-hand Dual-hand same Dual-hand different
Lines
M 0.104 0.086 0.164
SD 0.072 0.028 0.078
Circles
M 0.852 0.819 0.598
SD 0.072 0.117 0.129
150 E.A. Franz et al. / Spatial constratnts

References

Abend, W., E. Bizzi and P. Morasso, 1982. Human arm trajectory formation. Brain 105, 331-348.
Bernstein, N.. 1967. The co-ordination and regulation of movements. Oxford: Pergamon Press.
Chang, P. and G.R. Hammond, 1987. Mutual interactions between speech and finger movements.
Journal of Motor Behavior 19, 265-274.
Corcos, D.M.. 1984. Two-handed movement control. Research Quarterly for Exercise and Sport
55, 117-122.
Flash, T. and N. Hogan, 1985. The coordination of arm movements: An experimentally confirmed
mathematical model. Journal of Neuroscience 5. 168881703.
Gallistel, C.R., 1980. The organization of action: A new synthesis. Hillsdale, NJ: Erlbaum.
Heuer, H.. 1985. Intermanual interactions during simultaneous execution and programming of
finger movements. Journal of Motor Behavior 17, 335-354.
Hogan, N., 1984. An organizing principle for a class of voluntary movements. The Journal of
Neuroscience 4, 2745-2754.
Hollerbach, J.M., 1981. An oscillation theory of handwriting. Biological Cybernetics 39. 139-156.
Hollerbach, J.M. and C.G. Atkeson, 1986. ‘Characteristics of joint-interpolated arm movements’.
In: H. Heuer and C. Fromm (eds.), Generation and modulation of action patterns. Berlin:
Springer-Verlag. pp. 41-54.
Hollerbach, J.M. and T. Flash, 1982. Dynamic interactions between limb segments during planar
arm movement. Biological Cybernetics 44, 67-77.
Kawato, M., Y. Maeda, Y. Uno and R. Suzuki, 1990. Trajectory formation of arm movement by
cascade neural network model based on minimum torque-change criterion. Biological
Cybernetics 62. 275-288.
Keele, S.W., 1973. Attention and human performance. Pacific Palisades, CA: Goodyear.
Kelso, J.A.S., CA. Putnam and D. Goodman, 1983. On the space-time structure of human
interlimb co-ordination. Quarterly Journal of Experimental Psychology 35A, 347-375.
Kelso, J.A.S., D.L. Southard and D. Goodman, 1979. On the coordination of two-handed
movements. Journal of Experimental Psychology: Human Perception and Performance 5,
2299238.
Klapp. A.T., 1979. Doing two things at once: The role of temporal compatibility. Memory and
Cognition 7. 375-381.
Klapp, A.T., 1981. Temporal compatibility in dual motor tasks II: Simultaneous articulation and
hand movements. Memory and Cognition 9, 3988401.
Marteniuk, R.G., C.L. MacKenzie and D.M. Baba, 1984. Bimanual movement control: Informa-
tion processing and interaction effects. Quarterly Journal of Experimental Psychology 36A,
335-365.
Merton, P.A., 1972. How we control the contraction of our muscles. Scientific American 226,
30-37.
Morasso, P., 1981. Spatial control of arm movements. Experimental Brain Research 42, 223-227.
Muzii, R.A., C.L. Warburg and A.M. Gentile, 1984. Coordination of the upper and lower
extremities. Human Movement Science 3, 337-354.
Peters, M., 1977. Simultaneous performance of two motor activities: The factor of taming.
Neuropsychologia 15. 461-465.
Schmidt, R.A., 1988. Motor control and learning: A behavioral emphasis. Champaign, IL: Human
Kinetics Publishers.
Smith, A. and H.N. Zelaznik, 1991. ‘Comparative investigations of speech and other neuromotor
systems’. In: P. Chang and G.R. Hammond (eds.), Cerebral control of speech and other
neuromotor systems. Amsterdam: North-Holland.
 EA. Franz et al. / Spatial constraints 151

Smith, A., D.H. McFarland and C.M. Weber, 1986. Interactions between speech and finger
movements: An exploration of the dynamic pattern perspective. Journal of Speech and
Hearing Research 29, 471-480.
Soechting, J.F. and F. Lacquantini, 1981. Invariant characteristics of pointing movement in man.
Journal of Neuroscience 1, 710-720.
Swinnen, S.P., C.B. Walter, J.M. Pauwels, P.F. Meugens and M.B. Beirinckx, 1990. The dissocia-
tion of interlimb constraints. Human Performance 3, 187-215.
Swinnen, S., C.B. Walter and D.C. Shapiro, 1988. The coordination of limb movements with
different kinematic patterns. Brain and Cognition 8, 326-347.
Turvey, M.Y., 1977. ‘Preliminaries to a theory of action with reference to vision’. In: R. Shaw and
J. Bransford (eds), Perceiving, acting and knowing: Toward an ecological psychology. Hills-
dale, NJ: Erlbaum. pp. 211-266.
Uno, Y., M. Kawato and R. Suzuki, 1989. Formation and control of optimal trajectory in human
multijoint arm movement-minimum torque-change model. Biological Cybernetics 61, 89-101.
Viviani, P. and C. Terzuolo, 1980. ‘Space-time invariance in learned motor skills’. In: G.E.
Stelmach and J. Requin (eds.), Tutorials in motor behavior. Amsterdam: North-Holland. pp.
525-533.
Viviani, P. and C. Terzuolo, 1982. Trajectory determines movement dynamics. Neuroscience 7,
431-437.
Welford, A.T., 1967. Single channel operations in the brain. Acta Psychologica 27, 5-22.
Yamanishi, J., M. Kuwato and R. Suzuki, 1980. Two coupled oscillators as a model for the
coordinated finger tapping by both hands. Biological Cybernetics 37, 219-225.