Abstract Previous research has shown that the repro-
duction of a criterion distance is biased towards previ-
ously coded endpoints. The purpose of this research was
to illustrate that, in addition to the retention of endpoint
information, the presence of conflicting sources of spa-
tial information within a trial causes systematic response
biases in distance reproduction. Three experiments were
conducted in which participants performed rapid aiming
movements on a digitising tablet that translated to move-
ment of a cursor on a computer monitor. The required
movement amplitude in all three experiments was 20 cm.
In experiment 1, the location of the home and target
positions on the monitor was fixed, but the initial position
of the hand was varied randomly from trial to trial. In
experiment 2, the change in position of the limb was
matched by a corresponding change in the location of the
monitor display. In experiment 3, the initial position of
the limb was fixed, but the location of the display on the
monitor varied from trial to trial. The results of experi-
ments 1 and 2 showed that error varied as a function of
the initial position of the limb. However, this effect was
greater in experiment 1, where the mapping between the
location of the monitor display and limb position varied
from trial to trial. There was also an effect of varying the
location of the monitor display in experiment 3, but this
was smaller than varying initial limb position in experi-
ment 1. These findings suggest that both the retrieval of
previously specified endpoints and conflicts in the coding
of spatial information contributed to the observed response
biases in distance reproduction.
Keywords Aiming movements · Distance coding ·
Position coding · Human
Introduction
In numerous everyday activities, people perform rapid
movements to specific locations in space (e.g. pointing,
reaching and grasping an object, moving a cursor on a
computer monitor). In order to perform these goal-directed
movements, an individual must code and translate spatial
information into the appropriate motor commands needed
to achieve the task goal (Abrams et al. 1994; Bock and
Eckmiller 1986). There has been considerable debate
concerning which features are coded in sensory space
and transformed into corresponding parameters in motor
space. According to one viewpoint, the distance from the
initial position of the limb to the target is specified by
programming the timing and amplitude of force pulses
(i.e. distance control; Schmidt et al. 1979; Wallace
1981). Alternatively, it has been argued that the final
position of the limb is coded by specifying an equilibrium
point based upon the length-tension relationships of the
agonist and antagonist muscles (i.e. position control;
Feldman 1986; Polit and Bizzi 1978).
In order to investigate the nature of the control signal
that is specified in performing rapid aiming movements,
researchers have typically employed variations of either
distance- or position-reproduction tasks. In both tasks,
participants are required to produce movements from
randomly varying initial positions. The goal in distance-
reproduction tasks is to produce movements having a
constant amplitude, while the requirement in position-
reproduction tasks is to terminate movements at a fixed
location. It is expected that if participants code the
distance from the initial position of the limb to the target,
reproducing the same amplitude from varying starting
M.A. Khan (✉) · A. Fourkas
School of Sport, Health and Exercise Sciences,
University of Wales, Bangor, George Building, Bangor,
Gwynedd, LL57 2PX, Wales, UK
e-mail: m.khan@bangor.ac.uk
Tel.: +44-1248-388275, Fax: +44-1248-371053
I.M. Franks
School of Human Kinetics,
University of British Columbia, Canada
E. Buckolz
School of Kinesiology, University of Western Ontario, Canada
L. Hardy
School of Sport, Health and Exercise Sciences,
University of Wales, Bangor, UK
Exp Brain Res (2002) 145:231–238
DOI 10.1007/s00221-002-1116-7
R E S E A R C H A RT I C L E
Michael A. Khan · Alissa Fourkas · Ian M. Franks
Eric Buckolz · Lew Hardy
Conflicting sources of spatial information
in a distance-reproduction task
Received: 20 September 2001 / Accepted: 19 March 2002 / Published online: 17 May 2002
© Springer-Verlag 2002

positions would be relatively easy. However, coding
movement distance would cause error to vary systemati-
cally when participants are required to produce the same
endpoint from different initial positions. Participants
would overshoot the target when the initial position is
shifted closer to the desired endpoint, while they would
undershoot when the starting location is moved further
from the target. On the other hand, if participants code
movement endpoints, it is expected that performance
will be accurate in position-reproduction tasks. However,
in distance-reproduction tasks, the coding of endpoints
would cause participants to undershoot the desired target
when the initial position is shifted towards the endpoint
of the previous trial, whereas the target would be over-
shot when the initial position is moved away from the
previous endpoint.
The majority of evidence has indicated that, in both
distance and position-reproduction tasks, there is a sys-
tematic pattern of undershooting and overshooting the
target, depending on the initial position of the limb
(Bock and Eckmiller 1986; Ilic et al. 1996; Imanaka and
Abernethy 1992a; Jaric et al. 1992, 1994). However, the
effect of varying starting location on endpoint accuracy
has been shown to be less in position- than distance-
reproduction tasks. For example, in a study by Jaric et al.
(1994), two groups of participants practised aiming
movements consisting of horizontal, right arm elbow
flexion movements and were given visual feedback
throughout acquisition. One group of participants practised
a distance control task in which they were required to
flex their elbow 36° from different initial elbow angles
(113–137° in 4° increments, where 180° was full exten-
sion). A second group practised a location control task,
in which they were required to end their movements at a
criterion angle of 89° from different initial elbow angles
(113–137° in 4° increments, where 180° was full exten-
sion). Following acquisition, both groups were first
tested on the task they had practised and then on the task
the other group had practised but without visual feed-
back. Results showed that both groups reproduced move-
ment endpoints relatively well when tested on the location
task. However, when participants were tested on the
distance task, the distance travelled increased as initial
positions shifted to the right (i.e. elbow angles closer to
full extension) and decreased when initial positions shifted
towards the left (i.e. elbow angles closer to full flexion).
This was the case even for participants who practised
the distance-reproduction task. Therefore, regardless of
which task was practised, it appeared that participants
coded final positions, and the tendency to reproduce
these endpoints caused the distance travelled to vary as a
function of the initial position of the limb.
Of interest to researchers has been whether the inter-
ference caused from endpoint coding is due to local
neuromuscular factors as outlined in equilibrium-point
hypotheses or originates at a more central level (see
Imanaka et al. 1998). Equilibrium-point models are
based primarily on neuromuscular factors such as the
specification of tonic stretch reflex thresholds of the
muscles (e.g. λ model; Feldman 1986). However, findings
from studies on motor short-term memory have revealed
that systematic response biases are present when the
criterion and reproduction movements are performed
with separate limbs (Imanaka and Abernethy 1992a). On
this basis, it is argued that the retrieval of abstract
memory codes is the primary source of interference,
rather than limb-specific proprioceptive information.
Along these lines, Imanaka and Abernethy (1992b) have
illustrated that cognitive strategies had a significant
impact on the degree to which error varied as a function
of initial limb position. They have shown that distance
reproduction is improved when participants are instructed
to attend to the initial and end locations of their move-
ments and to take into account changes in start positions
from one trial to the next. According to Imanaka and
Abernethy, the interference caused by location coding
may be a consequence of automatic processing of irre-
levant endpoint information. Hence, bringing this infor-
mation to conscious awareness through selective atten-
tion may enable participants to reduce the interference
caused from its automatic processing. It is also quite
possible that, by simply attending to the changes in the
start locations from trial to trial, participants are better
able to calibrate target positions in space, which then
facilitate the reproduction of the required movement
distance from the different initial positions.
Although neuromuscular and cognitive explanations
differ in terms of the locus of interference effects, both
are based on the premise that systematic response biases
observed in the production of a criterion distance arise
from the tendency to reproduce previously coded end-
points. Thus, changing the initial position of the limb
results in a conflict between the required movement
distance and the endpoint retained from the previous
trial. The aim of the present research was to illustrate
that, in addition to the retention of endpoint information,
the presence of conflicting sources of spatial information
within a trial contributes to the interference from end-
point coding in a distance-reproduction task. These two
accounts are not necessarily mutually exclusive, as the
retention of endpoint information may be the precursor
to conflicts in the coding of spatial parameters. In previ-
ous investigations of the effects of endpoint coding on
distance reproduction, the task was performed without
vision of the limb and target (Imanaka and Abernethy
1992a, 1992b; Jaric et al. 1992, 1994). Therefore, partic-
ipants had to rely on visual and/or proprioceptive repre-
sentations of spatial parameters such as the initial posi-
tion of the limb and the target. It is possible that errors in
the representation of these spatial features that arise as a
result of randomly varying the initial position of the limb
could be the cause of systematic biases in movement
error.
It is well known that the processing of visual informa-
tion dominates other feedback sources (e.g. propriocep-
tive) and the retention of motor commands (Posner et al.
1976). Hence, the precise manner in which visuo-spatial
information is coded would be expected to have a signifi-
232