Fingertip forces during
object manipulation in
children with
hemiplegic cerebral
palsy. I: Anticipatory
scaling
Andrew M Gordon* PhD;
Susan V Duff MA PT OT; Department of Biobehavioral
Sciences, Box 199, Teachers College, Columbia University,
525 West 120th Street, New York, NY 10027, USA.
*Correspondence to first author at Department of
Biobehavioral Sciences, Box 199, Teachers College,
Columbia University, 525 West 120th Street, New York, NY
10027, USA. E-mail: [email protected]
Previous studies of grasping and object manipulation in
children with cerebral palsy (CP) have suggested a dichotomy
in the ability to use anticipatory control (planning) of the
fingertip force output, depending on the type of sensory
information (tactile or proprioceptive) on which it is based.
The present study further explores this issue by testing the
ability of 15 children with hemiplegic CP aged between 8 and
14 years to scale the fingertip force output in advance during
the lifting of small objects whose weight and surface texture
are varied. The results indicate that children with hemiplegia
can use anticipatory control based on both the weight and
texture of the object, but require a greater number of trials
than age-matched children without CP (control children)
before they can do so. We suggest that the initial lack of
anticipatory control results from an indistinct internal
representation of the object’s physical properties due to
disturbed sensory mechanisms, which may have direct
implications for therapeutic intervention.
166
Developmental Medicine & Child Neurology 1999, 41: 166–175
The integrity of the motor cortex and corticospinal pathways
are often compromised in individuals with hemiplegic cerebral
palsy (CP) (e.g. Uvebrant 1988, Yokochi et al. 1992). As a result,
skilled independent finger movements do not develop normally (e.g. Lawrence and Kuypers 1968). During tasks requiring
fine manipulation, children with CP often employ several fingers, making movements slow and clumsy (Brown et al. 1987).
There is often abnormal muscle tone, with posturing into an
upper extremity ‘flexion synergy’ (Brown et al. 1987, Yokochi et
al. 1992), reduced distal strength, and tactile and proprioceptive disturbances (Brown et al. 1987, Uvebrant 1988, Lesný et al.
1993, Yekutiel et al. 1994), which may further compromise fine
motor skills (Moberg 1962, Brown et al. 1987).
Both tactile information (signaled by slow [SA I] and fast
[FA I] adapting afferents, see Westling and Johansson 1987)
and weight-related information (signaled by muscle spindles
and tactile afferents) are used during grasping and object
manipulation to adapt the fingertip forces to the object’s
weight and texture (Johansson 1996). However, as there are
delays in the CNS and as object weight cannot be determined
until after the object is lifted, sensory information signalling
the object’s physical characteristics is not immediately available. If the initial forces are increased too quickly, excessive
force may result, risking damage to fragile objects. Conversely,
if they are developed too slowly, heavy or slippery objects
would require an inordinate amount of time to lift, and there
may be an increased risk of dropping them. To avoid these
risks, the isometric force must be scaled (planned) before
the initiation of the movement to match the object’s expected weight and texture based on internal representations of
the object gained during previous manipulatory experience
(Johansson and Westling 1987, 1988; Gordon et al. 1993).
Such ‘anticipatory control’ of the force output is characterized by continuous grip (pinch) and load (vertical lift) force
increase, with the rate of force increase scaled from the onset
towards the target load force (i.e. faster rates for heavier or
more slippery objects).
The control of fingertip forces during object manipulation in developing children without CP (control children)
generally approximates adult coordination between the ages
of 6 and 8 years (Forssberg et al. 1991, 1992, 1995; Gordon et
al. 1992, 1994; Eliasson et al. 1995a). In contrast, 6- to 8-yearold children with hemiplegic or diplegic CP often have a
force coordination resembling that of 1-year-old children
(Eliasson et al. 1991, 1992, 1995b; Gordon and Forssberg
1997). They have prolonged delays between movement
phases and sequential generation of grip and load force.
While most children with CP are capable of adjusting their
grip forces to the object’s weight and texture, their forces are
often excessive and variable, with less adaptation than agematched control children (Eliasson et al. 1992, 1995).
One of the most severe limitations in the ability of children
with CP to manipulate objects skillfully may be their impaired
anticipatory control. Children with CP are unable to scale the
rate of grip- and load-force increase based on the weight of an
object (Eliasson et al. 1992). However, they may demonstrate
some anticipatory control based on the object’s texture (i.e. a
higher rate of force increase for more slippery objects) if they
have successive lifts in which the contact surfaces are predictable (Eliasson et al. 1995b). This may suggest a dichotomy in
the ability to use anticipatory control, depending on the source
of sensory information. However, the latter study by Eliasson et
al. examined a greater number of lifts with each surface (N=10)
than did their former study with each weight (N=5).
The present study further examines the capability of children with hemiplegic CP to plan the force output based on
the object’s weight and texture, using a greater number of trials for each condition (N=25). An initial lack of anticipatory
force scaling based upon both the object’s weight and texture, but subsequent emergence after a number of lifts
would suggest that children with hemiplegic CP are capable
of anticipatory control but may need extended practice.
Conversely, anticipatory force scaling based on an object’s
frictional characteristics, but not its weight, would suggest
that the deficit in force scaling may depend on the type of
sensory information on which it is based.
Method
SUBJECTS
Fifteen children (10 males and five females) with hemiplegic
CP and 15 age-matched children (five males and 10 females)
without CP (control children) participated in the study. All
children were recruited from schools and clinics in the New
York city metropolitan area and were between 8 and 14
years of age. This age range was chosen because the grasping behavior in developing children with CP approximates
that of adults by 8 years of age (Gordon and Forssberg
1997). Each child with CP was individually screened for eligibility before participation. Inclusion criteria were as follows: the ability to grasp and lift a small object between the
fingertips, an attention span of at least one hour, the ability
to follow verbal instructions, attendance of a mainstream
school, normal cognitive abilities according to the Kaufman
Brief Intelligence Test (Kaufman and Kaufman 1990).
Spasticity in the involved extremity of the children with
hemiplegia ranged from zero to moderate (0 to 2) according
to the Modified Ashworth Scale (0 to 4) (Bohannon and
Smith 1987). Informed consent was obtained from all children and their parents.
with hemiplegia often used additional fingers (‘three-digit
pinch’) and/or lifted the instrument between the thumb and
lateral portion of their index finger (‘lateral pinch’). However,
while the use of lateral pinch may result in some subjects taking longer and using slightly higher grip forces (due to the
lower density of mechanoreceptors innervating the lateral
skin), it generally does not influence coordination of forces
or anticipatory control (see Figs 7 and 8). Four of the children
with hemiplegia tended to posture the wrist into flexion.
After a demonstration, children performed a series of 25
lifts with the object’s weight adjusted to 200 g. This was followed by 25 lifts with the weight adjusted to 400g. The sandpaper grip surfaces were used for these two lifting series.
Finally, the children performed 25 lifts with the rayon contact
surfaces and the weight adjusted to 200 g. The children’s eyes
were open, and they were aware when the object’s properties
were altered. The object was held in the air for 5 seconds during each trial. Rest periods (15 to 20 minutes) were provided
between each lifting series. One child with CP performed the
weight and texture series during separate sessions.
After these trials, subjects were asked to hold the object in
the air and gradually release the grip until the object slipped
from the digits so the slip grip force (the minimal grip force
required) could be measured. This was repeated for each
weight and texture.
DATA ACQUISITION AND ANALYSIS
Data were sampled at 500 Hz, digitized with 12-bit resolution
and stored in a flexible data acquisition/analysis system
(SC/ZOOM, Umeå University, Sweden). The grip-force rate
(dGF/dt) and load-force rate (dLF/dt) were calculated within
Load force
APPARATUS
th)
(
rce
The grip instrument (Fig. 1) had exchangeable contact surfaces
(35 × 35mm, 20 mm apart) covered with either fine (200 grit)
sandpaper or rayon. The object’s weight was adjusted to either
200 g or 400 g using an exchangeable mass. The grip force at
each contact surface and the total load force were measured
with strain-gauge transducers (dc 160Hz). The instrument
contained an infrared light-emitting diode, which projected to
a position-sensing camera 50cm away, and an accelerometer to
detect slips between the skin and contact surfaces.
ip
Gr
d)
All children washed their hands before the experiment to
remove sweat and excessive oil from the skin. They sat on a
chair in front of a table, which was adjusted to position the
forearm horizontal to the table when the object was grasped.
Children with hemiplegia lifted the object with their
involved hand. As our earlier work suggested that force coordination is similar in the dominant and non-dominant hands
of individuals without CP (Gordon et al. 1994), the control
children lifted the object with their dominant hand for comparison. Children were instructed to grasp the object
between the thumb and index finger (precision grip) and lift
it so that it was adjacent to a marker 6cm high. The children
c
a
c
ip
Gr
PROCEDURE
in
e(
fo
for
b
Figure 1: Schematic drawing of the grip instrument. (a)
Exchangeable grip surfaces (sandpaper or rayon) covering
strain-gauge force transducers at the thumb (th) and index
finger (ind), (b) exchangeable mass (200 g or 400 g), and
(c) infrared light-emitting diode projecting to a
photoelectric position-sensing camera (not shown).
Object Manipulation in Children with CP Andrew M Gordon and Susan V Duff 167
a ±10 ms window. A graphics terminal was used to define the
time, events, and force amplitudes.
The grip force at the initiation of a positive load force was
measured to determine the extent to which grip and load
force are produced simultaneously. The peak rate of development of grip and load forces (i.e. force rates) were measured to determine whether they are higher for heavier or
more slippery objects, indicating anticipatory force scaling.
The grip force at lift-off was also measured to determine
whether higher forces are achieved at this point for the heavier or more slippery object. The durations between contact
of each digit (finger contact) and load-force onset (preloadphase), and between positive load-force onset and movement onset (load phase) were measured.
The grip force during the static phase (when the object was
held in the air) was analyzed by taking the mean and SD of all
samples for 2 seconds, starting 1.5 seconds after the peak grip
force. The coefficient of variation (SD divided by the mean)
First block (control nr 15)
was used to quantify the variability in grasping forces. The
slip at the end of each measured trial was signaled by the
accelerometer and a sudden drop in the load force. The
mean safety margin was calculated by subtracting the grip
force at slip from the mean grip force.
Preliminary analysis of each subject’s data indicated skewness (for example, the skewness for grip and load-force rates
averaged approximately 1.3, and was as high as 3.5 for some
individuals). The median value was thus chosen as a representative measure to ensure robustness of our statistical
tests. To assess the influence of the object’s weight and texture on the measured parameters at various points during
the lifting series, each series of 25 lifts was divided into five
blocks of five trials. We performed both non-parametric
(Wilcoxon sign-rank and Mann–Whitney U tests) and parametric (weight [2] × group [2] × trial block [5] ANOVAs) on
each measure at the P<0.05 level. While the results were
qualitatively similar, independent of the statistical methods
Fifth block (control nr 15)
5N
Grip force
20 N/s
Grip-force rate
4N
Load force
20 N/s
Load-force rate
6 cm
Position
First block (CP nr 11)
Fifth block (CP nr 11)
5N
Grip force
20 N/s
Grip-force rate
4N
Load force
20 N/s
Load-force rate
6 cm
Position
0.2 s
Figure 2: Grip force, grip-force rate, load force, load-force rate and position from a representative control subject and child
with hemiplegic CP during one lift with 200 g (- - - - - - - -) and one lift with 400 g (––––––) during the first and fifth trial blocks.
The traces are aligned at the onset of vertical movement and the grip- and load-force rates are shown using a ±20-point
numerical differentiation. Note the higher force rates for the 400 g object during both the first and fifth blocks in the control
child, but only during the fifth block for the child with hemiplegia.
168
Developmental Medicine & Child Neurology 1999, 41: 166–175
employed, we report the results of the ANOVA due to the
large number of tests associated with comparisons of measures during each trial block. Newman–Keuls post-hoc tests
were subsequently performed where appropriate.
Results
ANTICIPATORY CONTROL BASED ON OBJECT WEIGHT
In contrast to earlier findings (Eliasson et al. 1992), anticipatory control according to the object’s weight was observed during the 25 lifts. Figure 2 shows recordings from a representative
control subject and child with hemiplegia for lifts with each
weight during the first trial block (trials 1 to 5) and fifth trial
block (trials 21 to 25). A higher rate of load-force increase can
already be seen for the 400g object during the first block for the
control child, which typifies intact anticipatory control. This
was maintained during the fifth block. In contrast, the child
with hemiplegic CP showed little difference in the force rates
during the first block, but the force rates were considerably
higher for the heavier object by the fifth block.
These results were representative of the children in the
control and hemiplegic groups (Fig. 3). The load-force rates
were higher for lifts with the 400 g object for both the control
children and children with hemiplegic CP (F=77.1, P<0.001).
While the employed grip-force rates were slightly higher for
the 400g object for both the control children and children
with hemiplegia, surprisingly the differences did not reach significance (P<0.1) in either case. Overall, the load-force rates
(but not grip-force rates) were higher in the control children
than in the children with hemiplegia (F=5.33, P<0.05). For
the control children, there were differences in the load-force
rates already during the first block with only minor changes
during practice. In contrast, the children with hemiplegia
exhibited only small differences between the load-force rates
employed for each weight during the initial blocks, with the
difference increasing across blocks. Despite a weight × block
interaction (F=3.51, P<0.01), the weight × block × group interaction failed to reach significance, presumably because of the
variability between the CP and control groups. In order to
evaluate this issue further, we performed a separate weight ×
block ANOVA on each group. The results indicated no interaction for the control group (P>0.05), but a significant interaction for the CP group (F=3.19, P<0.05). Post-hoc analysis
indicated that the differences in force rates did not reach significance in the children with hemiplegia until the third trial
block. Thus, anticipatory control of the load force according to
the object’s weight was achieved only after extended practice
with the object in the children with hemiplegia.
Fourteen of the 15 children with hemiplegia exhibited
Children with CP
Blocks
Blocks
Grip-force rate (N/s)
Load-force rate (N/s)
Control children
Figure 3: Mean ± SEM load-force rates and grip-force rates employed for the 200 g (- - - - - - -) and 400 g (––––––) object
(sandpaper contact surfaces) during each block of trials across all 15 children in the control and hemiplegic CP group. Note
the large differences in the load-force rates for the control children during all blocks and small initial differences which
increase across blocks for the children with hemiplegia.
Object Manipulation in Children with CP Andrew M Gordon and Susan V Duff 169
higher load-force rates and 12 of 15 exhibited higher gripforce rates for the heavier object by the last trial block, indicating that most children were capable of achieving anticipatory
control. Interestingly, the relative change in load-force rate
was significantly correlated with the relative change in gripforce rate from the 200 g to the 400 g object for the children
with hemiplegia (r=0.60, P<0.05), but not for the control
children (r=0.03), suggesting proximal and distal anticipatory control tend to be affected in a similar manner in children
with hemiplegia. Neither the relative change in grip-force rate
nor the load-force rate significantly correlated with the spasticity ratings (r=–0.04 and r=–0.34, respectively).
ANTICIPATORY CONTROL BASED ON OBJECT TEXTURE
Figure 4 shows recordings from a representative control
child and a child with hemiplegia for a lift with the non-slippery (sandpaper) and slippery (rayon) grip surfaces during
the first and fifth trial blocks. Neither subject displayed large
differences in the grip-force rates for the two grip surfaces
during the first block, while the grip-force rate is clearly higher for the rayon surface by the fifth block in both cases.
These results are representative of both groups of subjects
(Fig. 5). Overall, the children with hemiplegia and control
First block (control nr 2)
children showed a slight influence of object texture on the
grip-force rates (higher force rates for the rayon surfaces)
across the 25 trials with both the rayon and sandpaper grip
surfaces, though surprisingly the differences were not significant. Neither the control children nor the children with hemiplegia exhibited clear differences between the grip-force
rates employed for the sandpaper or rayon surfaces during
the initial blocks, but by the last trial block, some differences
can be seen. This was true in 11 of 15 children with hemiplegia. Despite a texture × block interaction (F=4.83, P<0.001),
the texture × block × group interaction failed to reach significance. When a separate texture × block ANOVA was performed on each group, a significant interaction (F=3.02,
P<0.05) was seen for the children with hemiplegia, with the
post-hoc analysis yielding significant differences by the fifth
trial block (but, F=2.30, P<0.07 for the control children).
Thus weak anticipatory control appears to be present, but
only after considerable practice with each grip surface.
The general lack of anticipatory control of the grip force in
the control children and children with hemiplegia contradicts the results of Eliasson et al. (1995b). To explore this
issue further, we measured the grip force at slip for each texture and weight (see horizontal lines in Fig. 6g) to determine
Fifth block (control nr 2)
Grip force
5N
Grip-force rate
40 N/s
Load force
2N
6 cm
Position
First block (CP nr 10)
Fifth block (CP nr 10)
5N
Grip force
40 N/s
Grip-force rate
2N
Load force
6 cm
Position
0.2 s
Figure 4: Grip force, grip-force rate, load force and position from a representative control subject and child with hemiplegic
CP during one lift with the sandpaper (- - - - - - - ) and one lift with rayon (–––––––) grip surfaces during the first and fifth trial
blocks. The traces are aligned at the onset of vertical movement and the grip-force rates are shown using a ±20-point
numerical differentiation. Note that the grip-force rates only differ in the fifth block for both children.
170
Developmental Medicine & Child Neurology 1999, 41: 166–175
the minimum grip force required to maintain contact with
the object. Overall, the grip force at slip did not differ
between the children with hemiplegia and control children.
However, it was not as high for the rayon surfaces (1.7 N for
both groups of children) as it was for the 400 g weight (2.0 N
and 1.9 N for the children with hemiplegia and control children, respectively), i.e. the differences between coefficients
of friction of the two surfaces may not have been large
enough to necessitate entirely anticipatory control.
Interestingly, the relative change in grip-force rate from
the sandpaper to the rayon surfaces was significantly correlated with the relative change from the 200 g to the 400 g
object in the previous experiment (r=0.71 for the children
with CP and r=0.68 for the control children, P<0.01 in both
cases). This suggests that these object properties influence
the scaling of the grip force output in a similar manner. As in
the weight experiment, the relative change in grip-force rate
between textures did not significantly correlate with the
spasticity ratings (r=–0.04).
POSSIBLE COMPENSATORY STRATEGIES
Across all weights and textures, the children with CP contacted the object first with the thumb on 55% of the trials, whereas
the control children contacted the object first with the index
finger on 56% of the trials. The contact order did not vary
appreciably according to the object’s weight or texture.
However, the duration between contact of each digit was considerably longer in the children with hemiplegia compared
with the control children (F=16.45, P<0.001 and F=16.54,
P<0.001) (Fig. 6a) for the weight and texture experiments,
respectively. This was also true of the preload-phase duration
(F=48.36, P<0.001 and F=54.54, P<0.001) (Fig. 6b) and
loading-phase duration (F=35.29, P<0.001 and F=38.76,
P<0.001) (Fig. 6c) as described previously (Eliasson et al.
1991). Furthermore, due to the longer preload phase (during
which the children with hemiplegia pushed the object down
against its support, creating a negative load force compared
with the control children [F=19.52, P<0.001 and F=30.69,
P<0.001 in each experiment, respectively]) (Fig. 6d), the grip
forces at the onset of load force (Fig. 6e) were higher for the
children with hemiplegia (nearly 3N) than control children
(F=8.73, P<0.01 and F=6.35, P<0.05 in each experiment,
respectively). In fact, for the 200 g object, the children with
hemiplegia had already achieved 65% of the force level
employed during the static phase by the onset of load force,
compared with 27% for the control children.
Despite the longer time to develop grip force before loadforce onset, the preload phase, negative load force and grip
force at load-force onset did not vary according to the object’s
weight or texture (P>0.05 in all cases), suggesting that if
these behaviors are compensatory, they are not dependent on
the specific physical properties of the object. The longer loading phases in the children with hemiplegia (particularly for
the 400g object) resulted in a higher grip force at lift-off (Fig.
6f ) for the heavier (F=6.74, P<0.05) and more slippery (but
P<0.09) objects, which were maintained during the subsequent static phase (F=9.17, P<0.01 and P<0.14, respectively) (Fig. 6g). In contrast to previous results (Eliasson et al.
1991, 1992), the children with hemiplegia did not have significantly higher grip forces in the static phase or safety margins
(area above the horizontal lines, Fig. 6g) than the control children. Their coefficient of variation (Fig. 6h) was greater during the weight (F=5.55, P<0.05) but not texture (P<0.15)
experiment compared with control children.
The above temporal and force measures suggest that
despite the general lack of anticipatory control based on the
object’s weight and texture during the initial trial blocks, the
children with hemiplegia achieved higher forces for the
heavier or more slippery objects mainly by developing the
grip and load force sequentially, prolonging the durations of
force increase and by adjusting their grip force using sensory
feedback. The longer phase durations and higher forces
would reduce the need to modulate the rate of force increase
during the subsequent loading phase.
INFLUENCE OF GRASPING POSTURE ON FINGERTIP COORDINATION
In order to investigate whether the ‘three-digit’ or ‘lateral’
pinch grips frequently employed by the children with hemiplegia influenced the force coordination and anticipatory
control, six of the control subjects were asked to grasp and
Children with CP
Blocks
Blocks
Grip-force rate (N/s)
Control children
Figure 5: Mean ± SEM grip-force rates employed for the sandpaper (- - - - - - - -) and rayon (––––––) grip surfaces during each
block of trials across all 15 children in the control and hemiplegic CP group. Note the small initial differences in grip-force
rates which increase across blocks for both the controls and the children with hemiplegia.
Object Manipulation in Children with CP Andrew M Gordon and Susan V Duff 171
tive grasping patterns in children with hemiplegia may contribute to the higher and more variable forces and longer
phase durations, but do not account for the apparent initial
lack of anticipatory control in children with CP.
lift the object five times with each weight using these grasp
patterns. Representative recordings from one of these subjects are shown in Figure 7. Overall, the coordination is not
greatly influenced by the different grips used by the control
subject. In general, the grip and load force continue to
increase in parallel and the grip- and load-force rates, when
plotted on a graph, are largely bell-shaped as previously
described for precision grip (Gordon and Forssberg 1997).
Despite the general similarities in force coordination,
some differences in the temporal and force measures occur
during each of the three grip patterns. The difference in finger contact (Fig. 8a) was longer for the lateral pinch (F=7.49,
P<0.05). Similarly, the preload phase (but F=5.10, P<0.07)
(Fig. 8b) and loading phase (F=9.49, P<0.05) (Fig. 8c) were
also longer for the lateral pinch. The mean static grip force
(Fig. 8g) increased by 72 and 122% (F=9.29 P<0.05) for the
200 g and 400 g weights, respectively, and the static grip force
coefficient of variation (Fig. 8h) increased by 42 and 70%
(F=10.38, P<0.05), respectively, compared with the precision grip. In contrast, except for an increased coefficient of
variation in static grip force (F=7.56, P<0.05) (Fig. 8h),
none of the measured parameters was different between the
three-digit pinch and precision grip.
Despite the slight differences between the temporal and
force parameters seen during precision grip and lateral
pinch, these six control subjects still exhibited anticipatory
control during lifts with each grip. This can be seen in Figure
8e and 8f, which show that the load-force rates (but not gripforce rates) were higher for the 400g object during all three
grips (P<0.05 in all cases). Thus, the employment of alterna-
Discussion
The results indicate that the ability of children with hemiplegic CP to use anticipatory control of fingertip forces is not
dependent on the type of sensory information on which it is
based. Rather, they suggest that children with hemiplegia
have a general deficit in the scaling of fingertip forces in
advance, but may develop anticipatory control based on the
object’s weight or texture if provided with enough practice.
POSSIBLE MECHANISMS OF IMPAIRED ANTICIPATORY CONTROL
The initially impaired capability of children with hemiplegic
CP to scale the rate of force increase could be a result of the
general impairment in motor output in CP (e.g. Leonard et al.
Neilson et al. 1990, Eliasson et al. 1991). During the grasping
and lifting of an object between the fingertips, children with
CP have an impaired ability to coordinate the grip force and
load force simultaneously (Eliasson et al. 1991); they typically first increase their grip force (often excessively) in conjunction with a negative load force (pushing the object down
against its support). Instead of the continuous force increase
typically associated with anticipatory control, they show a
discontinuous force increase. In addition, the phases comprising the grip-lift movement are greatly prolonged.
Similarly, voluntary control of isometric contractions is
impaired in children with CP (Neilson et al. 1990).
200 g
400 g
Rayon
Controls
CP
CP
500
Controls
(d)
CP
Controls
300
100
0
(e)
0
100
Ð0.6
Ð0.4
Ð0.2
0
0
(f)
(g)
4
2
0
(h)
6
Static GF (N)
2
200
6
GF at lift-off (N)
4
300
0
6
Negative LF (N)
40
200
400
4
2
0
Static GF CV (%)
80
CP
Ð0.8
Loading (ms)
120
Preload (ms)
Finger contact (ms)
Controls
GF and LF onset (N)
(c)
(b)
(a)
12
8
4
0
Figure 6: Mean ± SEM temporal and force measures associated with lifts with the 200 g and 400 g weights, and the rayon
surface: (a) duration between contact of the thumb and index finger, (b) preload-phase duration, (c) loading-phase
duration, (d) negative load force (LF), (e) grip force (GF) at positive load-force (LF) onset, (f) grip force (GF) at lift-off, (g) grip
force (GF) during static phase (horizontal lines represent grip force at slip, and area above line represents employed safety
margin), and (h) coefficient of variation of static grip force (GF CV). Note the longer latencies and higher forces before lift-off
for the children with hemiplegia.
172
Developmental Medicine & Child Neurology 1999, 41: 166–175
Nevertheless, the amount and coordination of forces did not
change during the repeated lifts with a given object in the present study (Neilson et al. 1990). Other than the ability to scale
the rate of force increase, the measured temporal or force
parameters remained the same for both groups. Thus,
impaired motor output does not preclude anticipatory force
scaling, and its emergence after practice was not a result of a
general improvement in motor output.
The impaired anticipatory control in children with CP
could be a result of the spasticity often associated with the
movement disorder. However, for several reasons this is
unlikely. Firstly, considerable training of force production
reduces the spasticity, but has little influence on the efficiency of motor commands (Neilson et al. 1990). Secondly, there
is no conclusive evidence that pharmacological treatment of
spasticity improves motor performance (Nathan 1969,
McLellan 1977). Finally, the children we tested had spasticity
ratings ranging from normal to moderate. Nevertheless,
even the children without spasticity initially exhibited
impaired anticipatory control, and we did not find a relation
between anticipatory control and spasticity.
Precision grip (control nr 9)
The use of the lateral pinch grasping pattern also could
have contributed to the impaired force scaling. For example,
Woollacott et al. (1996) showed that when in a crouched
position (approximating the posture of children with hemiplegia) and subjected to standing postural perturbations,
control children exhibit temporal disorganization of muscle
activation patterns as well as increased antagonistic muscle
activity, which is similar to that of children with CP. Similarly,
when control children are asked to walk with their knees
bent and trunk inclined forward, they also exhibit abnormalities in muscle activation which resemble those in CP
(Sienko-Thomas et al. 1996). Thus, it is conceivable that musculoskeletal constraints or altered biomechanical patterns
may underlie the movement impairments in CP and contribute to impaired anticipatory control. Overall, the use of
the lateral pinch resulted in higher grip forces and small
increases in some temporal parameters in control children,
most likely due to the decreased density of mechanoreceptors located in the lateral skin (Johansson and Vallbo 1983).
However, their overall coordination was not greatly affected
and anticipatory force scaling persisted in these children. Yet
Three-digit pinch (control nr 2)
3N
Grip force
20 N/s
Grip-force rate
2N
Load force
20 N/s
Load-force rate
30 cm
Position
0.2 s
Lateral pinch (control nr 9)
0.2 s
Precision grip (CP nr 2)
3N
Grip force
20 N/s
Grip-force rate
2N
Load force
20 N/s
Load-force rate
Position
30 cm
0.2 s
0.4 s
Figure 7: Grip force, grip-force rate, load force, load-force rate and position from a control subject lifting the 200 g object
using the precision grip, three-digit pinch and lateral pinch, compared with a lift with the precision grip from one child with
hemiplegic CP. The traces are aligned at the onset of vertical movement and the grip and load-force rates are shown using a
±20-point numerical differentiation. Note the different time scale for the child with CP and overall similarity in
coordination for the three grips in the control.
Object Manipulation in Children with CP Andrew M Gordon and Susan V Duff 173
erties (i.e. their initial representations may be distorted). When
the texture of the object is presented in a random order, control children scale the fingertip forces according to the previous lift, and rapidly adapt their grip force to the new texture
when it is erroneously scaled toward an unexpected surface
(Forssberg et al. 1995). In contrast, children with hemiplegia or
diplegia do not alter their grip force according to the previous
texture, nor do they adapt their motor output to the actual texture when it is inappropriate (Eliasson et al. 1995b). Similarly,
6- to 8-year-old control children are capable of scaling their
forces according to an object’s weight after one or two lifts
(Forssberg et al. 1992), whereas children with CP were not,
even after five lifts (Eliasson et al. 1992). Children with CP
appear to have the capacity to scale the grip force when the
same weights or textures are presented repeatedly (cf.
Steenbergen et al. 1998). This indicates that they are capable of
modulating the rate of force increase to some extent. Thus,
children with CP appear to have impaired abilities to form the
internal representations unless considerable ‘blocked’ practice is provided.
the lateral pinch, while being less flexible than a precision
grip, is a grip that is routinely employed (e.g. when turning a
key). Four of the children with hemiplegia in the present
study tended to grasp and lift objects with the wrist postured
in flexion. The changed length–tension relation in the finger
flexors may weaken pinch force (cf. An et al. 1985) and influence the motor output. Nevertheless, all of them still exhibited anticipatory control by the fifth trial block. Thus, the
altered grip patterns and hand posturing were not likely to
be the cause of the impaired anticipatory control.
The impaired anticipatory control may have resulted from
sensory disturbances which are common in CP. Anticipatory
control of fingertip forces during grasping and manipulating
objects is based on an internal representation of the object’s
physical characteristics. This is achieved through tactile and
proprioceptive signals during earlier manipulatory experience conveying information about the object’s weight and
texture (Johansson and Westling 1987, 1988; Gordon et al.
1993). Normally, anticipatory control can be used after minimal manipulatory experience with an object as only one or
two trials are necessary to achieve such representations
(Gordon et al. 1993).
While many children with CP have sensory disturbances,
previous results suggest that they do not clearly relate to the
ability to adjust the fingertip forces to the object’s texture
(Eliasson et al. 1995b), which all of the children in the present
study could do. However, children with hemiplegic CP may not
be able to extract enough sensory information during single
lifts to form vivid internal representations of the object’s prop-
CLINICAL IMPLICATIONS
The results suggest that children with hemiplegic CP may be
able to acquire representations of new objects, but take
longer to do so than control children. Therefore, children
with CP could benefit if the environment is structured to
match their capabilities. For example, objects could be kept
consistent within the child’s environment, and new ones
introduced gradually. Therapists could also encourage use of
Precision grip
Three-digit pinch
Lateral pinch
(f)
400 g
400 g
GF at LF onset (N)
200 g
(h)
(g)
Static GF (N)
LF rate (N/s)
(d)
200 g
400 g
Loading (ms)
200 g
Preload (ms)
Finger contact (ms)
400 g
(e)
GF rate (N/s)
(c)
(b)
200 g
Static GF CV (%)
(a)
Figure 8: Mean ± SEM temporal and force measures associated with lifts with the precision grip, three-digit pinch and lateral
pinch with the 200 g and 400 g object for six control subjects: (a) duration between contact of the thumb and index finger, (b)
preload-phase duration, (c) loading-phase duration, (d) grip force (GF) at positive load-force (LF) onset, (e) grip-force (GF)
rate, (f) load-force (LF) rate, (g) grip force (GF) during static phase, and (h) coefficient of variation of static grip force (GF
CV). Note the temporal latencies are longer and forces higher for the lateral pinch, but the force rates are higher for the 400 g
object independent of the grip employed.
174
Developmental Medicine & Child Neurology 1999, 41: 166–175
alternative, compensatory strategies such as to grasp first the
object while simultaneously pushing it down against its support. This would provide additional sensory information
(e.g. friction) about the object, stabilize it in the hand before
lifting and reduce the need for anticipatory control, potentially reducing functional impairments. Children with CP
could also be taught to rely on visual information signalling
the object’s size, weight, and texture. These interventions,
which may minimize the requirement of anticipatory control, could potentially improve hand function in children
with CP, a possibility which requires further testing.
Accepted for publication 17th July 1998.
Acknowledgements
This project was supported by a grant from the United Cerebral Palsy
Research and Education Foundation (#R-709-96), the VIDDA
foundation and an Untenured Faculty Research Grant from Teachers
College, Columbia University to Andrew Gordon. We are grateful to
Ellen Godwin, Sheila Walsh, Lauren Robertson, Molly Roffman, Gail
Lavander, Amy Shrank and the other therapists who assisted with
recruiting subjects, and the parents and children who made this
project possible. We also thank Jeanne Charles for assistance with
data collection, and Dr Ann Gentile for helpful comments.
References
An K-N. (1985) Forces in the normal and abnormal hand. Journal of
Orthopedic Research 3: 202–11.
Bohannon RW, Smith MB. (1987) Interrater reliability of a modified
Ashworth scale of muscle spasticity. Physical Therapy 67: 206–7.
Brown JK, Rensburg van E, Walsh G, Lakie M, Wright GW. (1987) A
neurological study of hand function of hemiplegic children.
Developmental Medicine and Child Neurology 29: 287–304.
Eliasson AC, Gordon AM, Forssberg H. (1991) Basic coordination of
manipulative forces in children with cerebral palsy.
Developmental Medicine and Child Neurology 33: 661–70.
— — — (1992) Impaired anticipatory control of isometric forces
during grasping by children with cerebral palsy. Developmental
Medicine and Child Neurology 34: 216–25.
— Forssberg H, Ikuta K, Apel I, Westling G, Johansson RS. (1995a)
Development of human precision grip V: anticipatory and
triggered grip actions during sudden loading. Experimental
Brain Research 106: 425–33.
— Gordon AM, Forssberg H. (1995b) Tactile control of isometric
fingertip forces during grasping in children with cerebral palsy.
Developmental Medicine and Child Neurology 37: 72–84.
Forssberg H, Eliasson AC, Kinoshita H, Johansson RS, Westling G.
(1991) Development of human precision grip I: basic
coordination of force. Experimental Brain Research 85: 451–7.
— Kinoshita H, Eliasson AC, Johansson RS, Westling G, Gordon AM.
(1992) Development of human precision grip II: anticipatory
control of isometric forces targeted for object’s weight.
Experimental Brain Research 90: 393–8.
— — — — — (1995) Development of human precision grip IV:
tactile adaptation of isometric finger forces to the frictional
condition. Experimental Brain Research 104: 323–30.
Gordon AM. (1994) Development of the reach to grasp movement.
In: Bennett KMB, Castiello U, editors. Insights into the Reach to
Grasp Movement. Amsterdam: Elsevier. p 37–56.
— Forssberg H. (1997) The development of neural control
mechanisms for grasping in children. In: Connolly KJ, Forssberg
H, editors. Neurophysiology and Psychology of Motor
Development. Clinics in Developmental Medicine No. 143/144.
London: Mac Keith Press. p 214–31.
— — Johansson RS, Eliasson AC, Westling G. (1992) Development
of human precision grip III: integration of visual size cues during
the programming of isometric forces. Experimental Brain
Research 90: 399–403.
— — Iwasaki N. (1994) Formation and lateralization of internal
representations underlying motor commands during precision
grip. Neuropsychologia 32: 555–68.
— Westling G, Cole KJ, Johansson RS. (1993) Memory
representations underlying motor commands used during
manipulation of common and novel objects. Journal of
Neurophysiology 69: 1789–96.
Johansson RS, Vallbo AB. (1983) Tactile sensory coding in the
glabrous skin of the human hand. Trends in Neuroscience: 27–31.
— Westling G. (1987) Signals in tactile afferents from the fingers
eliciting adaptive motor responses during precision grip.
Experimental Brain Research 66: 141–54.
— — (1988) Coordinated isometric muscle commands adequately
and erroneously programmed for the weight during lifting task
with precision grip. Experimental Brain Research 71: 59–71.
Johansson RS. (1996) Sensory control of dextrous manipulation in
humans. In: Wing AM, Haggard P, Flanagan JR, editors. Hand and
Brain: Neurophysiology and Psychology of Hand Movement.
New York: Academic Press.
Kaufman AS, Kaufman NL. (1990) Kaufman Brief Intelligence Test.
Circle Pines, MN: American Guidance Service.
Lawrence DG, Kuypers HGJM. (1968) The functional organization
of the motor system in the monkey. I.The effects of bilateral
pyramidal lesions. Brain 91: 1–14.
Leonard CT, Moritani T, Hirschfeld H, Forssberg H. (1990) Deficits
in reciprocal inhibition in children with cerebral palsy as revealed
by H reflex testing. Developmental Medicine and Child
Neurology 32: 974–84.
Lesný, Stehlik A, Tomasek J, Tomankova A, Havlicek I. (1993)
Sensory disorders in cerebral palsy: two-point discrimination.
Developmental Medicine and Child Neurology 35: 402–05.
McLellan DL. (1977) Co-contractions and stretch reflexes in
spasticity during treatment with baclofen. Journal of Neurology,
Neurosurgery, and Psychiatry 40: 30–8.
Moberg E. (1962) Criticism and study of methods for examining
sensibility in the hand. Neurology 12: 8–19.
Nathan PW. (1969) Treatment of spasticity with perineural injections
of phenol. Developmental Medicine and Child Neurology 11: 384.
Neilson PD, O’Dwyer NJ, Nash J. (1990) Control of isometric muscle
activity in cerebral palsy. Developmental Medicine and Child
Neurology 32: 778–88.
Sienko-Thomas S, Moore CA, Kelp-Lenane C, Norris C. (1996)
Simulated gait patterns: the resulting effects on gait parameters,
dynamic electromyography, joint moments, and physiological
cost index. Gait and Posture 4: 100–7.
Steenbergen B, Hulstijn W, Lemmens IH, Meulenbroek RG. (1998)
The timing of prehensile movements in subjects with cerebral
palsy. Developmental Medicine and Child Neurology 40: 108–14.
Uvebrant P. (1988) Hemiplegic cerebral palsy aetiology and
outcome. Acta Paediatrica Scandinavica 345: (Suppl.)
Westling G, Johansson RS. (1987) Responses in glabrous skin
mechanoreceptors during precision grip in humans.
Experimental Brain Research 66: 128–40.
Woollacott MH, Burtner P. (1996) Neural and musculoskeletal
contributions to the development of stance balance control in
typical children and in children with cerebral palsy. Acta
Paediatrica Scandinavica 416: (Suppl.) 58–62
Yekutiel M, Jariwala M, Stretch P. (1994) Sensory deficits in the hand
of children with cerebral palsy: a new look at prevalence.
Developmental Medicine and Child Neurology 36: 619–24.
Yokochi K, Hosoe A, Kodama M, Kodama K. (1992) Assessment of
upper and lower extremity movements in hemiplegic children.
Brain Development 14: 18–22.
Object Manipulation in Children with CP Andrew M Gordon and Susan V Duff 175