Neuropsychologia 44 (2006) 2296–2307

Deviations in upper-limb function of the less-affectedside in congenital hemiparesis

Bert Steenbergen ∗, Ruud G.J. MeulenbroekNijmegen Institute for Cognition and Information, Radboud University Nijmegen, P.O. Box 9104, 6500 HE Nijmegen, The Netherlands

Received 11 January 2006; received in revised form 8 May 2006; accepted 9 May 2006Available online 23 June 2006

Abstract

In the present study we examined upper-limb function of the less-affected side in young adolescents with congenital hemiparesis (cerebral palsy:CP). Five participants with hemiparetic CP and five control participants performed a cyclical reach-and-grasp task with the less-affected handtowards targets placed at 60%, 100%, and 140% of the participant’s arm-length. Trunk involvement, end-effector kinematics and activation of thebiceps and triceps were examined together with several clinical measures. Movements at the less-affected side were slower and peak velocity wasreached later in the experimental group. Even though total trunk involvement was identical in both groups, it was selectively limited to forwardbprCi©

K

1

babtkcsm((A2

cl

0d

ending in participants with CP. Elbow amplitudes of these participants were smaller for the 60% and 100% arm-length target distances. Additionally,articipants with CP showed weak positive correlations between agonist (triceps) activity and elbow amplitude, suggesting that deficient agonistather than antagonist innervation was responsible for the decreased elbow involvement. Especially the more severely affected participants withP proved to compensate their relatively small elbow amplitudes by increased forward bending. Collectively, the findings demonstrate deviations

n upper-limb control of the less-affected body side in congenital hemiparesis.2006 Elsevier Ltd. All rights reserved.

eywords: Cerebral palsy; Prehension; Ipsilateral side; Trunk; EMG

. Introduction

Understanding the nature of movement deviations followingrain damage is crucial for the development and systematicpplication of rehabilitation therapies. In the case of lateralizedrain damage, movement deficits are most evident on the con-ralesional side of the hemispheric lesion, leading to a disordernown as hemiparesis. Research into upper-limb control of theontralesional side in patients with hemiparesis, either fromtroke or cerebral palsy (CP), has revealed several characteristicovement deficits such as weakness of specific muscles

Bourbonnais & Vandennoven, 1989), abnormal muscle toneLance, 1980), increased levels of co-contraction (Brouwer &hsby, 1991; Damiano, Martellotta, Sullivan, Granata, & Abel,000; Lamontagne, Richards, & Malouin, 2000; but see Van

Abbreviations: CP, cerebral palsy; EMG, electromyography; IQ, intelligenceoefficient; WISC II, Wechsler intelligence scale for children; IRED, infraredight emitting diode; ADL, activities of daily living∗ Corresponding author. Tel.: +31 24 36 12642; fax: +31 24 36 12592.

E-mail address: steenbergen@nici.ru.nl (B. Steenbergen).

Roon, Steenbergen, & Meulenbroek, 2005), decreased involve-ment of the shoulder and elbow combined with increased trunkinvolvement (Cirstea & Levin, 2000; Steenbergen, Van Thiel,Hulstijn, & Meulenbroek, 2000; Van Thiel & Steenbergen,2001), less fluent movements (e.g. Trombly, 1992, 1993) and,generally, slower movements (e.g. Utley & Sugden, 1998; Utley& Steenbergen, 2006; Utley, Steenbergen, & Sugden, 2004).

Relatively few experimental studies have examined the move-ment capabilities of the less-affected extremity. However, con-servation of upper-limb function of the less-affected side ishighly important for individuals with hemiparesis, because thisside is often employed as a compensatory ‘tool’ in perform-ing activities of daily living (ADL). Neuropsychological test-ing of the less-affected limb indeed revealed subtle deficits(e.g. Dellatolas, Filho, Souza, Nunes, & Braga, 2005), thatwere also shown for reaching and grasping (Carey, Baxter, &DiFabio, 1998; Debaere, Van Assche, Kiekens, Verschueren, &Swinnen, 2001; Desrosiers, Bourbonnais, Bravo, & Roy, 1996;Hermesdorfer, Laimgruber, Kerkhoff, Mai, & Goldenberg,1999; Sunderland, Bowers, Sluman, Wilcock, & Ardron, 1999;Yarosh, Hoffman, & Strick, 2004).

028-3932/$ – see front matter © 2006 Elsevier Ltd. All rights reserved.oi:10.1016/j.neuropsychologia.2006.05.016

B. Steenbergen, R.G.J. Meulenbroek / Neuropsychologia 44 (2006) 2296–2307 2297

Despite these insights, there have been no systematic studiesof the functional loss in upper-limb function of the less-affectedside following congenital unilateral brain damage that simul-taneously address processes at the level of muscle activation,upper-limb movement kinematics, and body-segment coordi-nation. Such an approach is, in our view, important since theincidence of CP in developed countries is relatively high atabout 2–2.5/1000 live births (Lin, 2003), but at the same timethe understanding of the neuropathophysiology and motor sys-tems dysfunctions of CP remains limited (Steenbergen & Utley,2005). Consequently, the present study was set up to examinethe functional loss in upper-limb movements of the less-affectedside in young adolescents with hemiparetic CP. To that aim,muscle activation patterns, upper-limb and trunk movementkinematics, and arm and body-segment coordination patternswere examined.

A recurrent finding in upper-limb tasks performed with thecontralesional limb in hemiparesis (either as a consequence ofstroke or CP) is the excessive use of the trunk, even when theobject to be picked up is placed well within the limits of thestretched arm (e.g. Cirstea & Levin, 2000; Levin, Michaelsen,Cirstea, & Roby-Brami, 2002; Van Roon, Steenbergen, &Meulenbroek, 2004). In individuals without neurological dis-orders, the trunk is naturally recruited when movement distanceexceeds a distance of 90% of the length of the arm, the so-called‘critical boundary’ (Dean, Shephard, & Adams, 1999; Kaminski,BMawtl1&cti(o(Ltetm

wedwbdbtSao

Next to movement recording, we monitored EMG activity ofthe biceps–triceps muscles pair (prime movers) to examine thealleged role of upper-limb musculature in relation to the involve-ment of the trunk.

In sum, to uncover the deviations in control at the less-affected side we analysed trunk involvement and segmentalcontribution of the shoulder and elbow as participants with hemi-paretic CP performed cyclical grasping movements with theless-affected hand to targets at three distances. Since an anal-ysis of trunk involvement along the X-, Y-, and Z-axis has notbeen performed before in hemiparetic CP, it is unavoidable thatsuch an analysis is in part descriptive. To capture the possiblecauses of the pattern of trunk involvement we performed corre-lational analyses between elbow and trunk involvement. Basedon previous research on the affected side in hemiparetic CP (e.g.Van Roon et al., 2004; Steenbergen et al., 2000), we hypothe-sised that increased trunk involvement may compensate for thealtered (possibly limited) segmental contribution of the shoulderand elbow joints. We also examined muscle activation patternsof the prime movers to test the hypothesis that increased antag-onist activity is related to trunk involvement.

2. Methods

2.1. Participants

aaaSoppcw

ciat&Pc(flsnd

aaatnsatwh

mR

ock, & Gentile, 1995; Mark et al., 1997; Saling, Stelmach,escheriakov, & Berger, 1996). Such a preferred critical bound-

ry may correspond to an arm configuration for grasping inhich relative comfort is attained that may be associated with

he orientation of the hand or the avoidance of extreme angu-ar positions (Gentilucci, Deprati, Gangitano, Saetti, & Toni,997; Kamper & Rymer, 1999; Roby-Brami, Bennis, Mokhtari,

Baraduc, 2000). In individuals with hemiparetic stroke, thisritical boundary is dramatically reduced to 50% of the length ofhe arm (Levin et al., 2002). This may be due to a decreased abil-ty to fully extend the arm due to weakness of agonist musclese.g. anterior deltoid and triceps, Colebatch & Gandieva, 1989)r to an excessive antagonist muscle activation, or co-contractionWing, Lough, Turton, Fraser, & Jenner, 1990). According toevin et al. (2002) the poor cooperation between the antagonis-

ic muscles pairs of the upper arm (increased co-activation) andxcessive stretch reflexes lead to a limited ability to fully stretchhe arm. As a consequence, the trunk is recruited to reach the

ovement goal.To examine in detail the role of co-activation of the trunk

ith arm movements in individuals with hemiparetic CP, wevaluated the pattern of recruitment of the less-affected shoul-er, elbow and wrist concurrent with the contribution of the trunkhen natural reaching movements were made to targets placedoth within and beyond the reach of the arm at three differentistances. While most studies have examined trunk involvementy looking at trunk displacement in the sagittal movement direc-ion only (e.g. Michaelsen, Jacobs, Roby-Brami, & Levin, 2004;teenbergen et al., 2000; Van Roon et al., 2004), we performedmore detailed analysis by examining the three components

f trunk involvement along the X-, Y-, and Z-axis separately.

Five participants with hemiparetic CP and five controls participated onvoluntary basis in the study. The hemiparetic participants were students

t a school for special education called ‘Werkenrode’ where they followedn adapted educational program (all male, mean age = 16.3 years/months,.D. = 1.1 years/months). Selection of the hemiparetic participants was basedn information in school records made available to the experimenters with fullermission of the participants and their tutors. Due to the fact that the hemi-aretic participants were students at a school rather than patients at a medicallinic, the information laid down in the files about the individual neuropathologyas limited.

To provide a good clinical picture, each participant underwent a series oflinical assessments administered by a trained physiotherapist. This was donen the weeks following the experiment. Hand function of the affected and less-ffected side was established through the administration of the Purdue–Pegboardest (Tiffin, 1968) and the Box-and-Block test (Mathiowetz, Volland, Kashman,

Weber, 1985) according to the instructions in the test protocols. While theurdue–Pegboard is a test of fine manipulative skills, gross dexterity is typi-ally measured by the Box-and-Block test. In addition, IQ scores were obtainedWISC II), both verbal and performal. Spasticity levels at the wrist and elbowexors and extensors of the affected and less-affected arm were assessed viacores on the Ashworth Scale of Spasticity (Bohannon & Smith, 1987). We wereot able to assess the level of spasticity of one participant as this participant hadeparted from the school after the experiment was done.

All participants selected were able to understand the task instruction. Inddition, only participants were selected who were diagnosed with hemiparesiss a consequence of CP. For three participants the affected side was the left sidend for two it was the right side. Likewise, participants without functional sit-ing balance or lacking the cognitive capacities to perform the experiment wereot included. Due to the relatively long duration of the complete experimentalession (approximately 2 h) only participants were selected who had sufficientttentional capacities to concentrate for such a long time. All hemiparetic par-icipants had undergone extensive rehabilitation programs and their situationsere described as non-progressive, or stable. Participants with receptive aphasia,emi-neglect or apraxia were not included in the study.

The control group (two males and three females, mean age = 22.6 years/onths, S.D. = 1.5 years/months) consisted of psychology students from theadboud University Nijmegen who participated as part of a college research

2298 B. Steenbergen, R.G.J. Meulenbroek / Neuropsychologia 44 (2006) 2296–2307

Table 1Clinical data of the participants with hemiparetic cerebral palsy

Participant

1 2 3 4 5

Age (years/months) 14.8 16.1 17.7 16.8 16.1Gender m m m m mAffected side Left Left Right Left RightFunctional reaching distance (less-affected side) 45.5 45 44 49.9 50

DexterityPurdue–Pegboard (affected side) n.p. n.p. n.p. n.p. 30Purdue–Pegboard (less-affected side) 43 42 19 40 31Box-and-Block test (affected side) 17 8 7 26 48Box-and-Block test (less-affected side) 62 50 24 63 49

IQ-scores (WISC-II)IQ-verbal 103 59 71 77 71IQ-performal 62 69 50 103 64

Spasticity (Ashworth scale: 0 [no spasticity]–5 [severe spasticity])Wirst palmar flexion: affected (less-affected) n.a. 0 (0) 0 (0) 0 (0) 0 (0)Wrist palmar extension: affected (less-affected) n.a. 0 (0) 2 (0) 0 (0) 0 (0)Elbow flexion: affected (less-affected) n.a. 1 (0) 2 (0) 2 (0) 0 (0)Elbow extension: affected (less-affected) n.a. 1 (0) 2 (0) 1 (1) 1 (0)

Note: n.p., not possible; n.a., not available.

credit requirement. The control subjects had no musculoskeletal problems affect-ing their arm or trunk movements and had no history of neurological disorders.Participants of the control group served as baseline measure to signify ‘adult’coordination patterns. As previous research indicated that coordination patternsin CP can be considered ‘adult-like’ at the age of 16 (e.g. Steenbergen et al.,2000) there was no specific need to exactly match both groups on age. Allparticipants signed an informed consent form. The study was approved by thelocal ethics committee and has been performed in accordance with the ethicalstandards laid down in the 1964 Declaration of Helsinki.

Table 1 displays the results of the clinical assessments. For both, thePurdue–Pegboard test and the Box-and-Block test, standardised norms scoresfor different age groups are available in the test manuals. The norm score of thePurdue–Pegboard test for individuals between 16 and 17 years of age is 49.5when using their preferred hand. Standardised norm scores for the Box-and-Block test are not available for individuals younger than 20 years. Therefore,we used the norm scores for the youngest age group available, individualsbetween 20 and 24, which is 88.2 when using their preferred hand. All par-ticipants with hemiparetic CP attained lower test scores than the norm value onthe Purdue–Pegboard test, even when using their less-affected hand. The lowestscore was 19 (participant 3) and the highest was 43 (participant 1). A similartrend was found for the test scores on the Box-and-Block test. Again, participant3 attained the lowest score (24), and the highest score was 63 (participant 4). Insum, these clinical assessment scores indicate that dexterity of the less-affectedside of the participants with hemiparetic CP was consistently worse comparedto the norm scores.

As shown in Table 1, for the affected side spasticity ranged from absent(Ashworth 0: no resistance to passive movement) to mild (Ashworth 2: increasedresistance to passive movement during the complete range of motion). For theless-affected side, only one participant (participant 4) had minor spasticity asobserved during passive elbow extension (Ashworth 1). For the other participantsno spasticity was observed at their less-affected side.

lb

2

w

in approximately 90◦ flexion if placed on the table. Hemiparetic participantswere instructed to perform the experimental task with their less-affected handand the control group were instructed to use their preferred (right) hand. Speedof moving was not explicitly stressed. At the start of each trial, subjects had toposition their hand on a start button. The task participants were asked to performconsisted of: (1) grasping a ball located at one of three distances in front of theirbody midline; (2) lift and transport the ball and release it in a wide gutter placedabout 10 cm further away along the body midline; (3) return their hand to thestart button. The gutter was equipped with a small rail that faced the participantand that was placed on a small slope. Once the ball was released in the widegutter it rolled back through the rail (see Fig. 1). Four balls were present in therail, such that a ball was always present for the participant to be picked up.

Two factors were manipulated, the distance of the ball to the participant andthe size of the ball. Distance of the ball was related to the functional reachingdistance of each individual participant and determined as follows. Participantswere asked to stretch out their arm without using their trunk, and touch the table.

Fm(p

In sum, the results of the clinical assessment revealed that dexterity of theess-affected body side of the participants with hemiparetic CP was impaired,ut that the level of spasticity at this body side was negligible.

.2. Apparatus, task, experimental procedure, and instructions

Participants were seated comfortably at a table on a height-adjustable chairith a flat seat. Seat height was adjusted so that the participant’s elbow was

ig. 1. Top view of the experimental set-up, with a definition of the three move-ent axes. Seated participants start the trial with their hand placed on the button

depicted by the dashed ‘arm’), pick up the ball (depicted by solid ‘arm’), andlace it in the gutter.

B. Steenbergen, R.G.J. Meulenbroek / Neuropsychologia 44 (2006) 2296–2307 2299

The distance from the acromion process of the shoulder to the metacarpopha-langeal joint of the index finger was determined and defined as the functionalreaching distance (46.9 and 47.6 cm for participants with hemiparesis and con-trols, respectively). Functional reaching distances were not significantly differentamong groups as assessed by a t-test. If the ball was placed at the functionalreaching distance, movement of the trunk would not be necessary in order to pickup the ball. This distance was set as 100%. Two other distances were derivedfrom this, one at 60% and one at 140%. If the ball was located at the latterdistance it could not be reached without involvement of the trunk, while notrunk involvement was needed at the 60% distance. Two ball sizes were used,measuring either 1.5 or 2.5 cm in diameter.

Combining these manipulations yielded six unique conditions, defined byBall Size (small, large) and Ball Distance (small, medium, large). In each con-dition five trials were performed. One trial consisted of 10 repetitions of the taskin immediate succession. Thus, participants had to perform a total of 300 move-ments (6 conditions × 5 trials × 10 repetitions). Trials were blocked accordingto the factor condition and the conditions were randomized across participants.

Two precalibrated Optotrak 3020 3D motion-tracking systems (spatial accu-racy better than 0.1 mm in X, Y and Z) were used for movement recording.Infrared light emitting diodes (IREDs) were placed on the wrist and elbow ofthe moving arm, and on both shoulders (see Fig. 1). A rigid body placed onthe sternum was used to record trunk involvement along three rotation axes(Cartesian X-, Y-, and Z-axis, corresponding to trunk forward bending, trunkinclination, and trunk torsion, respectively, see Fig. 1). Positions of the IREDswere sampled in raw data mode at a frequency of 100 Hz. All data were analysedoff-line.

EMG activity of two muscles was recorded using surface electromyography.Adhesive, disposable pre-gelled Ag/AgCl electrodes (diameter, 9 mm, inter-

electrode distance 2 cm) were placed in a bipolar arrangement parallel to thefibers at the bellies of the agonist–antagonist muscle pair of the prime moversM. Biceps Brachii (B) and M. Triceps (T) (see Delagi & Perotto, 1981 for elec-trode placement). The reference electrode was placed on C5. Prior to electrodeplacement the skin was cleaned with ethyl alcohol, abraded with scrub and againcleaned with ethyl alcohol until skin resistance was below 10 k�. EMG signalswere registered via the ODAU channel of the Optotrak at a sampling rate of1024 Hz. The experiment was videotaped, for verification purposes afterwards.

The total experiment took approximately two to two and a half hours. Stan-dard breaks were scheduled in the experiment, but participants could alwaysindicate when they needed extra rest.

2.3. Data analysis

The raw position data were filtered by means of a third-order, dual-pass But-terworth filter with a cut-off frequency of 10 Hz. Subsequently, each movementcycle of the series of ten cycles generated per trial (see Fig. 2) was split up inthree phases on the basis of the tangential velocity profile of the wrist marker.The first phase, denoted reach phase in what follows, started at the momentthe local velocity minimum occurred when the hand was at the start button andended at the moment the local velocity minimum occurred when the hand hadgrasped the ball. The second phase, denoted transport phase in what follows,was the phase starting at the end of the reach phase and ending at the momentthe local velocity minimum occurred when the hand released the ball. The thirdphase, return phase in what follows, started at the end of the transport phaseand ended at the moment the local velocity minimum occurred when the handhad returned to the start button. For the present purpose, we only report data of

Fab

ig. 2. Example of raw data collected in one trial in which a ball was picked up 10re the displacement of the hand (X-, Y-, and Z-axis), angular excursion of the elbowiceps and triceps, respectively. Displayed are the data from a control participant.

times at a distance of 30 cm and released at approximately 40 cm. Displayedand shoulder joint, trunk rotation along its three axes, and EMG-values of the

2300 B. Steenbergen, R.G.J. Meulenbroek / Neuropsychologia 44 (2006) 2296–2307

the reach phase, to enable comparison with existing studies on reaching. On thebasis of the position data of the individual IREDs we calculated shoulder andelbow angles. The elbow angle was computed by taking the 3D vector betweenthe wrist IRED and elbow IRED, and that between the elbow IRED and shoulderIRED, and consequently computing the enclosed angle between them. For theshoulder angle the same procedure was followed, but here the vector betweenthe elbow IRED and shoulder IRED, and that between the shoulder IRED andthe IRED on the contralateral shoulder, were used. With respect to trunk motion,we analysed total trunk involvement and trunk rotation around its constituentaxes, the X-axis (trunk forward bending), the Y-axis (trunk inclination) and Z-axis (trunk torsion, see Fig. 1 for definition of the axes). Total trunk involvementwas calculated by squaring each of the three constituent axes (forward bending,inclination and torsion, respectively), calculating their sum and then taking thesquare root of that value. As we were primarily interested in angular changes,instead of the absolute values of the different joint angles, we reset all angularvalues at 0◦ at the start of the reach phase.

Raw EMG signals were preprocessed as follows. A root mean square filterwith a time constant of t = 0.02 s was applied to the raw EMG data yielding afiltered surface EMG signal. To synchronize the EMG signal with the positionrecordings, a constant time shift of 50 ms was applied to compensate for the timediscrepancy that exists between the recorded EMG signal and the detectablemotor response, the so-called electromechanical delay. This delay is caused bythe elastic properties of the musculo-tendinous unit. For the muscles studiedhere, 50 ms has been indicated to be good estimator of this delay (Bloemsaat,Meulenbroek, & Van Galen, 2005). Finally, a third order zero phase-lag, low-pass Butterworth filter with a cut-off frequency of 10 Hz was applied to focusthe analyses on variations in the surface EMG measurements that correspondedwith the hand displacements. The preprocessed EMG signals were expressed aspercentages of the median EMG in each condition. Muscular co-contraction forthe biceps–triceps pair was determined as follows. The percentage of movementtime during which normalized EMG-activity for both muscles of a muscle paired

foffbt

3

pitcch

3

ieatBgf

m

Fig. 3. Final hand position (upper panel), movement time (middle panel), andpeak hand speed (lower panel) for the three target distances, and for controlsand participants with hemiparesis.

effect on movement duration, F(1,8) = 69.83, p < 0.001 (Fig. 3,middle panel). In addition, for both groups movement dura-tion increased linearly with movement distance, F(2,16) = 6.09,p < 0.05. Although there existed a trend for peak hand speedto be smaller for the hemiparetic participants, this trend failedto reach the conventional level of significance, F(1,8) = 3.95,p = 0.082. On the other hand, as expected movement distance sig-nificantly affected peak hand speed, F(2,16) = 342.24, p < 0.001,which was larger when a larger distance needed to be cov-ered. Planned comparisons of the Group × Distance interaction(F(2,16) = 8.52, p < 0.01) revealed that for control participants

xceeded 100% was calculated to assess the duration of phasic co-contractionuring a movement (cf. Lamontagne et al., 2000; Van Roon et al., 2005).

Preliminary analysis of the data revealed no main effects or interactionsor the factor Ball Size. Therefore, data were evaluated using an ANOVA withne between-subject (Group: hemiparetic CP, control) and one within-subjectactor (Ball Distance: small, medium, large) with repeated measures on the lastactor. The critical p-level was set at α = 0.05. Pearson correlation coefficientsetween selected experimental variables were subsequently calculated to assesheir specific relationships.

. Results

First, the kinematic characteristics of the cyclical hand dis-lacements are described, followed by data on the relativenvolvement of the trunk, shoulder joint and elbow joint. Third,he results of the EMG measurements are presented. Fourth,orrelations between the various measurements are presented toapture possible causes of the pattern of trunk involvement inemiparetic CP.

.1. Kinematic characteristics of the end effector

In order to verify that participants had complied with the tasknstruction, we first analysed the final hand position of the endffector. As expected, the repeated measures ANOVA revealedmain effect of Distance, F(2,16) = 617.9, p < 0.001, showing

hat, as instructed, final hand position increased linearly withall Distance (see upper panel of Fig. 3). This was true for bothroups as no Group effect or interaction with any other factoror final hand position was found.

In general, hemiparetic participants made slower reachingovements than controls as revealed by a significant Group

B. Steenbergen, R.G.J. Meulenbroek / Neuropsychologia 44 (2006) 2296–2307 2301

Fig. 4. Total trunk involvement for the three target distances, and for controlsand participants with hemiparesis.

peak hand speed increased with an increase in distance (Fig. 3,lower panel). For the hemiparetic group, peak hand speed wassignificantly different between the 60% and 140% distance con-dition but it was not when separately comparing the 60% and100% or the 100% and 140% distance conditions. Similar topeak hand speed, an effect of Distance on the moment of peakhand speed was found, F(2,16) = 19.64, p < 0.001. When dis-tance increased, peak hand speed was attained earlier, leadingto a prolonged deceleration phase. In addition, the hemipareticparticipants reached peak hand speed later than the control par-ticipants, F(1,8) = 6.87, p < 0.05.

Taken collectively, it is clear that at the level of the endeffector, movement performance of the less-affected side of thehemiparetic participants differred from that of controls. Move-ments made by the hemiparetic participants were slower andpeak velocity was reached later. Hence, corroborating the clin-ical findings, our analysis of the hand displacement indicatesthat the kinematics of arm movements performed by the less-affected side of the hemiparetic participants are also affected tosome extent.

3.2. Segmental involvement

Total trunk involvement increased with increasing distance,F(2,16) = 178.01, p < 0.001 (see Fig. 4). As shown in Fig. 4,tcsrireiwlwm

Fig. 5. Trunk forward bending (top), trunk inclination (middle) and trunk torsionamplitude (lower) for the three target distances, and for controls and participantswith hemiparesis.

effects of Distance (p’s < 0.01) and the Group × Distance inter-actions (p’s < 0.05) were significant. In addition, a main effect ofGroup on trunk inclination (F(1,8) = 5.42, p < 0.05) was found.As shown in Fig. 5, both trunk inclination and torsion scaledwith movement distance in controls. However, in hemipareticparticipants, rotations in these dimensions were small and notscaled to movement distance.

otal trunk involvement in hemiparetic participants was largerompared to controls at 60% and 100%, but this trend was notignificant. When disentangling total trunk involvement into itsotational components around the three basic movement axes,.e. forward bending, inclination, and torsion, the followingesults were found. Similar to total trunk involvement, the onlyffect found for trunk forward bending was an increase with anncrease in distance, F(2,16) = 48.03, p < 0.001, but again for-ard bending in the 60% and 100% condition was somewhat

arger in the hemiparetic group (see Fig. 5, top panel). The effectsere different for trunk inclination and trunk torsion see Fig. 5,iddle and lower panels). For both rotation dimensions, main

2302 B. Steenbergen, R.G.J. Meulenbroek / Neuropsychologia 44 (2006) 2296–2307

Fig. 6. Elbow amplitude for the three target distances, and for controls andparticipants with hemiparesis, respectively.

For shoulder amplitude no main Distance or Group effectsor interactions were found. This was different for elbow ampli-tude. Although elbow amplitude increased with increasing dis-tance, F(2,16) = 280.40, p < 0.001, the Group × Distance inter-action, F(2,16) = 4.46, p < 0.05 revealed that elbow amplitudewas smaller for the hemiparetic group in the 60% and 100% dis-tance conditions as compared to controls. At the farthest distance(viz., 140%) elbow amplitude was similar in both participantgroups (see Fig. 6). With respect to the timing of segmentalinvolvement, no effects were found for moment of maximumtrunk speed, moment of maximum shoulder speed, or momentof maximum elbow speed, respectively.

Taken collectively, most of the trunk involvement in thehemiparetic participants was due to forward bending. Trunkinclination and torsion were neglible in this group. Comparedto controls, elbow amplitude was significantly smaller in thehemiparetic participants at 60% and 100% movement distanceswhile trunk involvement (both, total involvement and forwardbending) was somewhat enlarged for hemiparetic participants atthese movement distances. In the 140% movement distance con-dition, both total trunk involvement and elbow extension werenot different among both participant groups.

3.3. Muscle activation patterns

bdd

tttbatv

Table 2Correlations per participant and per target distance between trunk forward bend-ing on the one hand and elbow excursion, shoulder excursion, biceps activity,triceps activity, and co-contraction on the other

Participant Distance

1 2 3

ElbowC1 n.s. n.s. −0.34C2 0.66 0.33 0.47C3 0.63 0.39 0.21C4 0.36 −0.21 0.2C5 0.25 n.s. -0.3H1 n.s. −0.43 n.s.H2 −0.4 −0.31 −0.5H3 −0.35 n.s. −0.5H4 n.s. n.s. n.s.H5 −0.27 n.s. n.s.

ShoulderC1 −0.36 −0.73 0.69C2 n.s. −0.44 n.s.C3 0.28 −0.5 n.s.C4 n.s. −0.64 0.73C5 −0.22 n.s. n.s.H1 n.s. 0.55 0.73H2 −0.24 n.s. −0.59H3 0.52 0.43 0.82H4 −0.51 −0.25 −0.39H5 −0.29 n.s. n.s.

BicepsC1 0.22 −0.29 −0.51C2 0.42 −0.23 n.s.C3 0.32 n.s. n.s.C4 n.s. −0.25 −0.35C5 −0.45 −0.43 −0.51H1 −0.73 −0.5 −0.43H2 −0.28 n.s. n.s.H3 n.s. −0.43 −0.43H4 −0.43 0.27 n.s.H5 −0.3 −0.37 −0.35

TricepsC1 n.s. n.s. −0.39C2 n.s. n.s. 0.23C3 −0.33 0.25 n.s.C4 n.s. −0.24 n.s.C5 n.s. n.s. n.s.H1 n.s. n.s. n.s.H2 n.s. n.s. n.s.H3 n.s. 0.46 0.29H4 −0.37 −0.26 n.s.H5 0.21 n.s. n.s.

Biceps–triceps co-contractionC1 0.23 n.s. −0.24C2 0.38 −0.24 n.s.C3 0.3 n.s. n.s.C4 n.s. n.s. −0.28C5 −0.41 n.s. −0.27H1 −0.72 −0.32 −0.24H2 n.s. n.s. n.s.H3 n.s. −0.45 −0.33H4 −0.32 0.29 n.s.H5 −0.27 n.s. n.s.

Note: n.s., not significant; italics values are significant at p < 0.05; bold valuesare significant at p < 0.01.

No main Group effects were found for the mean normalisediceps activity. However, mean normalised triceps activityecreased with increasing distance (F(2,16) = 7.89, p < 0.01), asid co-contraction of biceps–triceps (F(2,16) = 8.71, p < 0.01).

We also examined the relation between elbow amplitude onhe one hand, and muscle activation of the biceps and triceps, andheir co-contraction on the other hand. To that aim, we calculatedhe Pearson correlation coefficients (one-tailed testing, p < 0.05)etween elbow amplitude, and biceps activity, triceps activitiy,nd co-contraction of biceps–triceps, for each movement dis-ance separately and for each participant separately, yielding 15alues (three distances × five participants correlation pairs, with

B. Steenbergen, R.G.J. Meulenbroek / Neuropsychologia 44 (2006) 2296–2307 2303

the number of comparisons within each pair ranging from 89 to100). If reduced elbow amplitudes were associated with deficitsin agonist (triceps) activity, a significant positive correlation wasexpected to be found. The results showed that four values werepositive and significant in controls and eight values in partici-pants with CP. For both groups, and for antagonist activity and

co-contraction, significant negative values were expected (e.g.more antagonist activity resulting in smaller elbow amplitudes).For antagonist activity, only three values were significant andnegative in controls and only one value in participants with CP.Similar results were found for co-contraction and elbow involve-ment. Here, only four values were significant and negative for

Fsm0[Si

ig. 7. Scatter plots of the relationship between elbow amplitude (Y-axis) and trunkignificant correlations at p < 0.05, n.s.: not significant). Displayed are a control particild to severe based on the scores of the affected side on the Box-and-Block test (see

.33* [medium], and 0.47* [large]). (b) Shows the data of participant 3 (ranked 1 in smedium], and −0.50* [large]). (c) Shows the data of participant 2 (ranked 2 in severhows the data of participant 4 (ranked 4 in severity: correlations are −0.11 [small],

n severity: correlations of −0.27* [small], −0.17 [medium], and −0.17 [large]).

forward bending (X-axis) for each target distance separately (values denoteipant (a) and four participants with hemiparesis (b–e) ordered in severity fromTable 1). (a) Shows the data of control subject 2 (correlations of 0.66* [small],

everity: 1 [most severe]–5 [least severe], correlations of −0.35* [small], −0.03ity: correlations of −0.40* [small], −0.31* [medium], and −0.50* [large]). (d)0.17 [medium], and 0.07 [large]). (e) Shows the data of participant 5 (ranked 5

2304 B. Steenbergen, R.G.J. Meulenbroek / Neuropsychologia 44 (2006) 2296–2307

Fig. 7. (Continued ).

controls and two values for participants with CP. Thus, for par-ticipants with CP there appears to be a relation between elbowamplitude and agonist activity, with deficiencies in latter activ-ity leading to smaller elbow excursions. The suspected relationsbetween increases in antagonist activity or co-contraction andelbow amplitude were absent in this group.

3.4. Relationship between trunk involvement and (1) elbowinvolvement, (2) shoulder involvement, and (3) muscleactivation patterns

Several causes for trunk involvement have been put forward(see Section 1). With respect to the upper-limb movement sys-tem these include an inability to fully extend the arm (decreasedshoulder or elbow extension), weakness of agonist muscles (tri-ceps activity), excessive antagonist activation (biceps activity),and increased co-contraction. These assumptions can be verifiedby examining the sign and value of the correlation coefficients.Therefore, for each movement distance separately, and for eachparticipant separately, we calculated the Pearson correlationcoefficients between, on the one hand, trunk-involvement relatedvariables (in this case the trunk forward bending, as this was themain dimension of trunk involvement) and possible causes forthe pattern of trunk involvement on the other hand (changes inelbow involvement, shoulder involvement, agonist and antago-nist activation patterns and the co-contraction of both). Table 2s

acc(inot

with decreased elbow excursion in participants with hemipare-sis. It thus appears that trunk forward bending did to some degreecompensate for smaller angular excursions in the elbow joint inparticipants with hemiparesis, but not in controls. This inferencewas further scrutinised via examination of the individual scatterplots of the trunk forward bending–elbow excursion relation. InFig. 7 these are displayed for one control participant and fourparticipants with CP. Severity of the hemiparetic condition forthe four participants with CP is assessed on the basis of the scoresof the affected side on the Box-and-Block test (see Table 1).

From Fig. 7 it can be inferred that the two most severe partici-pants with hemiparesis (Fig. 7b and c) five out of six correlationswere significant and negative. For the two least severe par-ticipants (Fig. 7d and e) only one out of six correlations wassignificant and negative, suggesting that the degree to which thetrunk forward bending compensated for the decreased elbowexcursion was related to the severity of the impairment. For thecontrol participant (Fig. 7a) all three correlations were signifi-cant and positive.

Less unequivocal results were found for the correlationsbetween trunk forward bending and shoulder involvement. Incontrast to elbow involvement, the pattern of results was sim-ilar among both participant groups and suggest that shoulderinvolvement was not related to trunk forward bending (seeTable 2). Finally, with respect to muscle activation patterns ofthe upper-limb (agonist triceps activity, antagonist biceps activ-iamobcotfbttsrHntt

4

sltnmctwH

hows these coefficients.With respect to the correlation between elbow excursion

nd trunk forward bending an interesting difference betweenontrols and the hemiparetic group was found. Whereas mostorrelations in the control group were significant and positivenine out of 15), no significant positive correlations were foundn participants with CP. In the latter, correlations were either sig-ificant and negative (seven out of 15) or not significant (eightut of 15). Thus, increased trunk forward bending was relatedo more elbow excursion in controls, while it was associated

ty, and co-contraction) the pattern of results was again similarmong both participant groups. For the agonist activity (triceps)ost correlations were not significant in both groups (10 out

f 15), indicating a lack of relationship between trunk forwardending and triceps activity. For the antagonist activity mostorrelations were negative and significant in both groups (eightut of 15 in controls and 10 out of 15 in hemiparesis) suggestinghat excessive biceps activity did not cause an increased trunkorward bending. Thus, the sign and value of the correlationsetween agonist and antagonist activity on the one hand andrunk forward bending on the other were contrary to our expecta-ions. Finally, the co-contraction correlations were generally notignificant (seven in both groups), which may suggest a lack ofelationship between trunk forward bending and co-contraction.owever, in hemiparetic participants seven correlations wereegative and significant. Again, contrary to our expectation,here is a weak relation between increased co-contraction andrunk forward bending in hemiparetic participants.

. Discussion

In the present study we showed that congenital hemipare-is has subtle but systematic consequences for control with theess-affected side. Not only were the scores on the dexterityests lower in participants with hemiparetic CP compared toorm scores, but also basic kinematic measures, such as move-ent duration and time to peak hand speed differed among

ontrols and participants with hemiparesis. Similar effects athe less-affected side have been shown previously in patientsith unilateral stroke (e.g. Fisk & Goodale, 1988; Haaland &arrington, 1989; Hermesdorfer, Blankenfeld, & Goldenberg,

B. Steenbergen, R.G.J. Meulenbroek / Neuropsychologia 44 (2006) 2296–2307 2305

2003; Kim, Pohl, Luchies, Stylianou, & Won, 2003; Wetter,Poole, & Haaland, 2005; Yarosh et al., 2004) but studies on theeffects of congenital hemiparesis on control of the less-affectedside are relatively scarce (see, however, Steenbergen & Van derKamp, 2004).

The presently demonstrated restrictive use of the trunk incli-nation and torsion in participants with hemiparesis may pointto difficulties controlling the many degrees of freedom of thetrunk segment (cf. Bernstein, 1967). According to Bernstein,the first stage in the acquisition of new motor skills is the ‘freez-ing’ of some of the degrees of freedom of the motor system inorder to obtain a controllable system (for examples in individualswithout neurological disorders, see Steenbergen, Marteniuk, &Kalbfleisch, 1995; Vereijken, Van Emmerik, Whiting, & Newell,1992). Progress in skill acquisition is associated with the grad-ual release of the formerly ‘frozen’ degrees of freedom and theincorporation of these degrees of freedom into a dynamic sys-tem, which is represented by so-called synergies, or coordinativestructures (Kelso, Buchanan, DeGuzman, & Ding, 1993; Turvey,Shaw, & Mace, 1978). The trunk is an example of a redundanteffector system because, even when oversimplified representedas a single rigid body, it can move along three dimensions.Exploitation of the full capacity of the trunk demands coor-dination among the movements in these three dimensions. Theresults from the present study showed that controls were ableto fully exploit the capacity of the trunk as movement ampli-ttceTpcTmmrtbccoTAdsadtsm

aapis

as a compensation for the decreased elbow amplitude. At thesame time, these findings suggest a functional, task-specific syn-ergy between elbow and trunk. While a similar synergy has beenshown at the affected side of individuals with hemiparesis as aconsequence from stroke (Levin, 1996), or CP (Steenbergen etal., 2000), its existence at the less-affected side has not beenreported before in CP. Furthermore, the present results showthat the pattern of trunk forward bending was similar amongboth groups. Thus, when only considering this dimension oftrunk involvement there appears to be no abnormality in control.Rather, it is the absence of any task related (or, distance related)variations in trunk inclination and torsion in the hemiparetic par-ticipants which was deviant from controls. Our detailed analysisof trunk involvement in its three constituent rotation axes hastherefore uncovered trunk involvement patterns that were notshown before, potentially because previous studies on upper-limb control of the affected side in individuals with brain damagehave primarily focussed on trunk involvement in the primarymovement direction (e.g. Cirstea & Levin, 2000; Levin, 1996;Steenbergen et al., 2000).

These findings beg the question as to the origin of thedecreased elbow amplitude at the 60% and 100% movement dis-tances in participants with CP. Three possible causes have beenproposed, viz., increases in antagonistic (biceps) muscle activ-ity, deficiencies in agonist (triceps) muscle activity (Colebatch &Gandieva, 1989), and increases in biceps–triceps co-contraction(acsoiprbncBttlaccmtpmtot

edbwi

udes in all three trunk dimensions were considerable and scaledo movement distance. Hence, in terms of Bernstein’s classifi-ation, this represents a highly skilled movement system thatxploits its full potential in a dynamic, task-related manner.he pattern of trunk involvement in the participants with hemi-aresis suggest the opposite. Here, the only dimension whichhanged in a task-related manner was trunk forward bending.runk inclination and torsion did not change with variations inovement distance, and their amplitudes were restricted. Thisay indicate that these degrees of freedom were not incorpo-

ated in the trunk synergy. Or, to use Bernstein’s terminology,hese degrees of freedom are ‘frozen’, leaving the system at theeginning of the skill-acquisition stage. These findings are inlose agreement with previous reports that revealed that posturalontrol in children with CP is worse than in typically devel-ping children (Hadders-Algra, Van der Fits, Stremmelaar, &ouwen, 1999; Liao, Yang, Hsu, Chan, & Wei, 2003). Hadders-lgra et al. argued that the main problem in CP was related to aeficient capacity to modulate the postural adjustments to task-pecific constraints, which is similar to the present findings oflack of scaling of trunk inclination and torsion to movementistance. Roncevalles, Woollacott, and Burtner (2002) showedhat a possible neural factor underlying this impairment may beought in the insufficient level of contraction of agonist posturaluscles.Our correlation analysis on trunk forward bending and elbow

mplitude as a function of movement distance yielded a remark-ble finding. In the most severely affected hemiparetic partici-ants (as measured by dexterity of the affected hand), the major-ty of the correlation values were significant and negative lendingupport to the notion that increased trunk forward bending served

e.g. Levin et al., 2002). Our analysis of the mean normalisedctivity of both muscles did not reveal any differences betweenontrols and participants with hemiparesis. Neither was our mea-ure of co-contraction different among both groups. On the basisf these findings we cannot conclude that deficiencies in musclennervation per se caused the decreased elbow involvement inarticipants with hemiparesis. However, we also examined theelationship between these three factors and elbow involvementy means of correlation analysis. This analysis suggested thateither increased antagonist (biceps) activity, nor increased co-ontraction was responsible for the decrease in elbow amplitude.oth factors are major determinants of spasticity, and therefore

his finding corroborates our clinical assessment of spasticity athe less-affected side (Ashworth) which was shown to be neg-igible. In contrast, the results suggest that the smaller elbowmplitudes were probably caused by deficiencies in agonist (tri-eps) activation (see also, Colebatch & Gandieva, 1989). Thisorrelation analysis-based suggestion warrants further experi-entation, particularly as we only tested five participants in

he present study. In addition, to put this inference into propererspective we have to bear in mind that, although elbow involve-ent is primarily accomplished by activation of the biceps and

riceps muscles, it may still be that pathological synergies withther muscles involved (e.g. shoulder) may also have contributedo the restricted elbow range of motion.

In conclusion, our results extend the growing body of lit-rature in unilateral stroke of the cerebral motor areas whichemonstrates that the less-affected (often denoted ‘unaffected’)ody side exhibits deficits as well. In the present study, thisas shown in congenital hemiparesis (CP) in which primar-

ly the motor cortices and the corticospinal tract are affected.

2306 B. Steenbergen, R.G.J. Meulenbroek / Neuropsychologia 44 (2006) 2296–2307

As our study was primarily aimed at examining in detail theinvolvement of the less-affected side in congenital hemipare-sis, rather than to search for neurophysiological mechanismswe did not perform any neuroimaging and, consequently, hadno detailed lesion data. However, neuroimaging may informus on the central control system related to ipsilateral or bilat-eral control. Concerning bilateral control, it is important to notethat 10–30% of fibers in the lateral corticospinal tract remainuncrossed at the motor decussation (Nathan & Smith, 1973) andthere is an occurrence of bilateral lesions in approximately one-third of individuals with hemiparetic CP (Okumura, Kato, Kuno,Hayakawa, & Watanabe, 1997) which may promote the impair-ment of fine motor control of each body side. With respect toipsilateral control, it is known that the neonatal brain undergoessignificant reorganisation following early damage, for instancethrough the development or persistence of ipsilateral motor path-ways (e.g. Nezu, Kimura, Takeshita, & Tanaka, 1999; Staudt etal., 2004). Especially the role of these ipsilesional pathways forcontrol of the less-affected side in congenital hemiparesis war-rants further investigation, because its contribution to functionalrecovery later in development may be negative as shown in strokepatients (Werhahn, Conforto, Kadom, Hallet, & Cohen, 2003).The lack of any lesion data in the present data only allows us tospeculate on the central control processes for our findings. Ourfinding of decreased elbow involvement, and its co-occurrencewith agonist deficiency, may suggest that there exists bilaterali

icbrtfnp

A

acTtd

R

B

B

B

B

Brouwer, B., & Ahsby, P. (1991). Altered corticospinal projections to lowerlimb motoneurons in subjects with cerebral palsy. Brain, 114, 1395–1407.

Brown, J. V., Schumacher, U., Rohlmann, A., Ettlinger, G., Schmidt, R. C.,& Skreczek, W. (1989). Aimed movements to visual targets in hemi-plegic and normal children: Is the “good” hand of children with infantilehemiplegia also normal? Neuropsychologia, 27, 283–302.

Carey, J. R., Baxter, T. L., & DiFabio, R. P. (1998). Tracking control in thenon-paretic hand of subjects with stroke. Archives of Physical Medicineand Rehabilitation, 79, 435–441.

Cirstea, M. C., & Levin, M. F. (2000). Compensatory strategies for reachingin stroke. Brain, 123, 940–953.

Colebatch, J. G., & Gandieva, S. C. (1989). The distribution of muscularweakness in upper motor neuron lesions affecting the arm. Brain, 112,749–763.

Damiano, D. L., Martellotta, M. S., Sullivan, D. J., Granata, K. P., & Abel,M. F. (2000). Muscle force production and functional performance inspastic cerebral palsy: Relationship of co-contraction. Archives of PhysicalMedicine and Rehabilitation, 81, 895–900.

Dean, C., Shephard, R., & Adams, R. (1999). Sitting balance I: Trunk-armco-ordination and the contribution of the lower limbs during self-pacedreaching in sitting. Gait and Posture, 10, 135–146.

Debaere, F., Van Assche, D., Kiekens, C., Verschueren, S. M. P., & Swinnen,S. P. (2001). Coordination of upper and lower limb segments: Deficits onthe ipsilesional side after unilateral stroke. Experimental Brain Research,141, 519–529.

Delagi, E. F., & Perotto, A. (1981). Anatomic guide for the electromyogra-pher: The limbs. Springfield, IL: Thomas.

Dellatolas, G., Filho, G. H., Souza, L., Nunes, L. G., & Braga, L. W. (2005).Manual skill, hand skill asymmetry, and neuropsychological test perfor-mance in schoolchildren with spastic cerebral palsy. Laterality, 10(2),161–182.

D

F

F

G

H

H

H

H

K

K

K

K

nvolvement of the affected ipsilateral hemisphere.Finally, when using neuroimaging in children with CP it is

mportant to note that different brain lesions may underly thelinical diagnosis of hemiparetic CP and the location of therain lesion is not always a good predictor of the clinical rep-esentation (Kwong, Wong, Fong, Wong, & So, 2004). Instead,he timing of the brain lesion may be more critical for residualunctional deficits (Okumura et al., 2000). Conversely, relativeormal anatomy cannot rule out functional alterations due tolasticity of the brain.

cknowledgements

The authors would like to thank the participants of this studys well as Marlies Croonen and Marijn Post for their help in dataollection. This research was supported by a grant awarded byhe Netherlands Organization for Scientific Research (NWO) to

he first author for the research project ‘Adaptation in movementisorder’.

eferences

ernstein, N. (1967). The coordination and regulation of movements. Oxford:Pergamon Press.

loemsaat, J. G., Meulenbroek, R. G. J., & Van Galen, G. P. (2005). Differ-ential effects of mental load on proximal and distal arm muscle activity.Experimental Brain Research, 167, 622–634.

ohannon, R. W., & Smith, M. B. (1987). Assessment of strength deficitsin eight paretic upper extremity muscle groups of stroke patients withhemiplegia. Physical Therapy, 67, 522–525.

ourbonnais, D., & Vandennoven, S. (1989). Weakness in patients with hemi-paresis. American Journal of Occupational Therapy, 43(5), 313–319.

esrosiers, J., Bourbonnais, D., Bravo, G., & Roy, P. M. (1996). Performanceof the “unaffected” upper extremity of elderly stroke patients. Stroke, 27,1564–1570.

illoux, F. M. (1996). Neuropathophysiology of movement disorders in cere-bral palsy. Journal of Child Neurology, 11, S5–S12.

isk, J. D., & Goodale, M. A. (1988). The effects of unilateral brain damageon visually guided reaching: Hemispheric differences in the nature of thedeficit. Experimental Brain Research, 72, 425–435.

entilucci, M., Deprati, E., Gangitano, M., Saetti, M. C., & Toni, I. (1997).On orientating the hand to reach and grasp an object. Neuroreport, 7,589–592.

aaland, K. Y., & Harrington, D. L. (1989). Hemispheric control of the initialand corrective components of aiming movements. Neuropsychologia, 27,961–969.

adders-Algra, M., Van der Fits, I. B. M., Stremmelaar, E. F., & Touwen,B. C. L. (1999). Development of postural adjustments during reaching ininfants with CP. Developmental Medicine and Child Neurology, 41(11),766–776.

ermesdorfer, J., Blankenfeld, H., & Goldenberg, G. (2003). The dependenceof ipsilesional aiming deficits on task demands, lesioned hemisphere, andapraxia. Neuropsychologia, 41, 1628–1643.

ermesdorfer, J., Laimgruber, K., Kerkhoff, G., Mai, N., & Goldenberg, G.(1999). Effects of unilateral brain damage on grip selection, coordination,and kinematics of ipsilesional prehension. Experimental Brain Research,128, 41–51.

aminski, T. R., Bock, C., & Gentile, A. M. (1995). The co-ordinationbetween trunk and arm motion during pointing movements. ExperimentalBrain Research, 106, 457–466.

amper, D. G., & Rymer, W. Z. (1999). Effects of geometrical joint con-straints on the selection of final arm posture during reaching: A simulationstudy. Experimental Brain Research, 126, 134–138.

elso, J. A. S., Buchanan, J. J., DeGuzman, G. C., & Ding, M. (1993). Spon-taneuous recruitment and annihilation of degrees of freedom in biologicalcoordination. Physics Letters A, 179, 364–371.

im, S. H., Pohl, P. S., Luchies, C. W., Stylianou, A. P., & Won, Y. (2003).Ipsilateral deficits of targeted movements after stroke. Archives of PhysicalMedicine and Rehabilitation, 84(5), 719–724.

B. Steenbergen, R.G.J. Meulenbroek / Neuropsychologia 44 (2006) 2296–2307 2307

Kwong, K. L., Wong, Y. C., Fong, C. M., Wong, S. N., & So, K. T. (2004).Magnetic resonance imaging in 122 children with spastic cerebral palsy.Pediatric Neurology, 31, 172–176.

Lamontagne, A., Richards, C. L., & Malouin, F. (2000). Coactivation duringgait as adaptive behavior after stroke. Journal of Electromyography andKinesiology, 10, 407–415.

Lance, J. W. (1980). Symposium synopsis. In R. G. Feldman, R. R. Young, &W. P. Koella (Eds.), Spasticity: Disordered motor control (pp. 485–494).Chicago: Yearbook Medical.

Levin, M. F. (1996). Interjoint coordination during pointing movements isdisrupted in spastic hemiparesis. Brain, 119, 281–293.

Levin, M. F., Michaelsen, S. M., Cirstea, C. M., & Roby-Brami, A. (2002).Use of the trunk for reaching targets placed within and beyond the reachin adult hemiparesis. Experimental Brain Research, 143, 171–180.

Liao, S. F., Yang, T. F., Hsu, T. C., Chan, R. C., & Wei, T. S. (2003).Differences in seated postural control in children with spastic cerebralpalsy and children who are typically developing. American Journal ofPhysical Medicine and Rehabilitation, 82(8), 622–626.

Lin, J.-P. (2003). The cerebral palsies: A physiological approach. Journal ofNeurology Neurosurgery and Psychiatry, 74(Suppl I), i23–i29.

Mark, L. S., Nemeth, K., Gardner, D., Dainhoff, M. J., Paasche, J., Duffy, M.,et al. (1997). Postural dynamics and the preferred critical boundary forvisually guided reaching. Journal of Experimental Psychology: HumanPerception and Performance, 23, 1365–1379.

Mathiowetz, V., Volland, G., Kashman, N., & Weber, K. (1985). Adult normsfor the box and block test for manual dexterity. American Journal ofOccupational Therapy, 39, 386–391.

Michaelsen, S. M., Jacobs, S., Roby-Brami, A., & Levin, M. F. (2004). Com-pensation for distal impairments of grasping in adults with hemiparesis.Experimental Brain Research, 157, 162–173.

Nathan, P. W., & Smith, M. C. (1973). Effects of two unilateral cordotomies

N

O

O

R

R

S

S

S

Steenbergen, B., & Utley, A. (2005). Cerebral palsy: Recent insights intomovement deviations. Motor Control, 9, 353–356.

Steenbergen, B., & Van der Kamp, J. (2004). Control of prehension in hemi-paretic cerebral palsy: Similarities and differences among the ipsi- andcontra-lesional side of the body. Developmental Medicine and Child Neu-rology, 46, 325–332.

Steenbergen, B., Van Thiel, E., Hulstijn, W., & Meulenbroek, R. G. J. (2000).The coordination of reaching and grasping in spastic hemiparesis. HumanMovement Science, 19, 75–105.

Sunderland, A., Bowers, M. P., Sluman, S., Wilcock, D. J., & Ardron, M.E. (1999). Impaired dexterity of the ipsilateral hand after stroke and therelationship to cognitive deficit. Stroke, 30(5), 949–955.

Tiffin, J. (1968). Purdue–Pegboard examiner manual. Chicago: ScienceResearch Associates.

Trombly, C. A. (1992). Deficits of reaching in subjects with left hemipare-sis: A pilot study. American Journal of Occupational Therapy, 46, 887–897.

Trombly, C. A. (1993). Observations of improvement of reaching in fivesubjects with left hemiparesis. Journal of Neurology Neurosurgery andPsychiatry, 56, 40–45.

Turvey, M. T., Shaw, R. E., & Mace, W. (1978). Issues in the theory of action:Degrees of freedom, coordinative structures and coalitions. In J. Requin(Ed.), Attention and performance: vol. VII, (vol. VII, (pp. 557–595).Hillsdale (NJ): Lawrence Erlbaum.

Utley, A., & Steenbergen, B. (2006). Discrete bimanual co-ordination inchildren and young adolescents with hemiparetic cerebral palsy: Recentfindings, implications and future research directions. Pediatric Rehabili-tation, 9(2), 127–136.

Utley, A., Steenbergen, B., & Sugden, D. A. (2004). The influence of objectsize on discrete bimanual co-ordination in children with hemiparetic cere-bral palsy. Disability and Rehabilitation, 26, 603–613.

U

V

V

V

V

W

W

W

Y

on motility of lower limbs. Brain, 96, 471.ezu, A., Kimura, S., Takeshita, S., & Tanaka, M. (1999). Functional

recovery in hemiplegic cerebral palsy: Ipsilateral electromyographicresponses to focal transcranial magnetic stimulation. Brain Development,21, 162–165.

kumura, A., Kato, T., Kuno, K., Hayakawa, F., & Watanabe, K. (1997).MRI findings in patients with spastic cerebral palsy. 2. Correlation withtype of cerebral palsy. Developmental Medicine and Child Neurology, 39,369–372.

kumura, A., Watanabe, K., Negoro, T., Ishiguro, Y., Miura, K., Matsumoto,A., et al. (2000). MRI findings in patients with symptomatic localization-related spilepsies beginning in infancy and early childhood. Seizure, 9,566–571.

oby-Brami, A., Bennis, N., Mokhtari, M., & Baraduc, P. (2000). Hand orien-tation for grasping depends on the orientation of the reaching movement.Brain Research, 869, 121–129.

oncevalles, M. N., Woollacott, M. W., & Burtner, P. A. (2002). Neuralfactors underlying reduced postural adaptability in children with cerebralpalsy. Neuroreport, 13(18), 2407–2410.

aling, M., Stelmach, G. E., Mescheriakov, S., & Berger, M. (1996). Pre-hension with trunk assisted reaching. Behavioural Brain Research, 80,153–160.

taudt, M., Gerloff, C., Grodd, W., Holthausen, H., Niemann, G., & Krageloh-Mann, I. (2004). Reorganization in congenital hemiparesis acquired atdifferent gestational ages. Annals of Neurology, 56, 854–863.

teenbergen, B., Marteniuk, R. G., & Kalbfleisch, L. E. (1995). Achiev-ing coordination in prehension: Joint freezing and postural contributions.Journal of Motor Behavior, 27, 333–348.

tley, A., & Sugden, D. A. (1998). Interlimb coupling in children withhemiplegic cerebral palsy during reaching and grasping at speed. Devel-opmental Medicine and Child Neurology, 40, 396–404.

an Roon, D., Steenbergen, B., & Meulenbroek, R. G. J. (2004). Trunkrecruitment during spoon use in tetraparetic cerebral palsy. ExperimentalBrain Research, 155, 186–195.

an Roon, D., Steenbergen, B., & Meulenbroek, R. G. J. (2005). Trunkuse and co-contraction in cerebral palsy as regulatory mechanisms foraccuracy control. Neuropsychologia, 43, 497–508.

an Thiel, E., & Steenbergen, B. (2001). Shoulder and hand displacementsduring hitting, reaching, and grasping movements in hemiparetic cerebralpalsy. Motor Control, 2, 72–88.

ereijken, B., Van Emmerik, R. E. A., Whiting, H. T. A., & Newell, K.M. (1992). Free(z)ing degrees of freedom in skill acquisition. Journal ofMotor Behavior, 24, 133–142.

erhahn, K. J., Conforto, A. B., Kadom, N., Hallet, M., & Cohen, L. G.(2003). Contribution of the ipsilateral motor cortex to recovery afterchronic stroke. Annals of Neurology, 54, 464–472.

etter, S., Poole, J. L., & Haaland, K. Y. (2005). Functional implicationsof ipsilesional motor deficits after unilateral stroke. Archives of PhysicalMedicine and Rehabilitation, 86(4), 776–781.

ing, A. M., Lough, S., Turton, A., Fraser, C., & Jenner, J. R. (1990).Recovery of elbow function in voluntary positioning of the hand follow-ing hemiplegia due to stroke. Journal of Neurology Neurosurgery andPsychiatry, 53(2), 126–134.

arosh, C. A., Hoffman, D. S., & Strick, P. L. (2004). Deficits in movementsof the wrist ipsilateral to a stroke in hemiparetic subjects. Journal ofNeurophysiology, 92, 3276–3285.