You are on page 1of 8

Gait & Posture 32 (2010) 500507

Contents lists available at ScienceDirect

Gait & Posture

journal homepage:

Three-dimensional reaching tasks: Effect of reaching height and width on upper

limb kinematics and muscle activity
Annelies Vandenberghe a,*, Oron Levin a, Joris De Schutter b, Stephan Swinnen a, Ilse Jonkers a

Department of Biomedical Kinesiology, Katholieke Universiteit Leuven, Tervuursevest 101, B-3001 Heverlee, Belgium
Division of PMA, Department of Mechanical Engineering, Katholieke Universiteit Leuven, Celestijnenlaan 300B, B-3001 Heverlee, Belgium



Article history:
Received 20 January 2010
Received in revised form 8 July 2010
Accepted 13 July 2010

In previous studies, upper limb coordination was usually analyzed during two-dimensional (2D) arm
movements. Based on joint kinematics and muscle activity, it has been demonstrated that the shoulder
joint controls the multi-joint movement. This study focused on three-dimensional (3D) reaching tasks
and examined if the coordination strategies previously described in 2D can be transferred to 3D
movements and if reaching to different locations in space has an effect on kinematic and upper limb
muscle strategies.
Ten healthy subjects reached to nine different targets in 3D space placed at arm length. Kinematic
data of the shoulder and elbow and electrical activity of 10 upper limb muscles were registered.
Differences in kinematics and EMG were compared between different reaching conditions.
Activity of shoulder muscles increased earlier than elbow muscles inducing shoulder elevation prior
to elbow extension. Reaching at different widths only inuenced shoulder kinematics, whereas reaching
at different heights inuenced both shoulder and elbow joints. Modulation of reaching height induced an
immediate adaptation of elbow exion followed by an adaptation of shoulder elevation.
As previously described in 2D, the shoulder joint leads the movement during 3D reaching tasks.
Changing the 3D nature of a reaching task inuenced the interaction between shoulder and elbow joint,
with reaching height primarily affecting the elbow coordination strategy.
2010 Elsevier B.V. All rights reserved.

Upper limb control
3D reaching task
Upper limb kinematics
Upper limb electromyography

1. Introduction
In daily life, many activities such as reaching, grasping and
lifting require task-specic coordination of a large number of
degrees of freedom (DOF) of the upper limb [14], with underlying
coordinated activation of multiple muscles [5,6].
The control of a multi-joint system such as the upper limb with
multiple DOFs in the shoulder and elbow joints relies on a exible
movement pattern organization such that motion and endpoint
position can be achieved through a complex interaction of variable
excursion of the individual DOFs of each joint. During movement
tasks conned to one plane (2D), the shoulder is responsible for the
movement generation, providing acceleration and deceleration of
shoulder and elbow joint [711]. The elbow regulates the
kinematics of the other upper limb joints in order to stabilize
and ne-tune end-point trajectory [7,8,12,13]. Specic kinematic
requirements of the motion (with the ratio between shoulder
elbow excursion and the path of the end-effector as key
parameters) emphasize the role of the elbow motion [9].

* Corresponding author. Tel.: +32 16329100; fax: +32 16329196.

E-mail address: (A. Vandenberghe).
0966-6362/$ see front matter 2010 Elsevier B.V. All rights reserved.

The kinematic requirements of 2D multi-joint movements in

the horizontal plane induce a specic pattern of alternating bursts
of agonist and antagonist activity controlling elbow [1316] and
shoulder [17]. The interaction between shoulder and elbow
kinematics referenced above [713], is reected in the muscle
activity with initial EMG bursts at the shoulder preceding those of
the elbow [8,17]. At the elbow joint, the antagonistic muscles are
often coactivated and modulation of muscle activity is less
pronounced [7,10,11].
Most studies on upper limb coordination have been conducted
in a 2D plane. Little is known about the control of threedimensional (3D) reaching tasks in terms of kinematics and
muscle activity. Studies evaluating 3D motions such as reaching
[18] and throwing [19], report high correlations for all joint angles
implicated in the movement, i.e. upper arm rotation, forearm
rotation, shoulder elevation and elevation angle, but not for elbow
exion. With respect to muscle activity, Hirashima et al. [20]
conrmed the role of the shoulder muscles in accelerating the
shoulder and in assisting elbow motion during ball throwing.
The current study explored the extent to which established
characteristics of kinematics and muscle activity during multijoint coordination in 2D can be extended to 3D reaching
movements. By manipulating target height and width, we


A. Vandenberghe et al. / Gait & Posture 32 (2010) 500507


Fig. 1. The cameras were placed 4 m away from subject, one at 908 at the right side and one at 458 at the right/front side of the subject. Targets were placed at three different
widths: at center line (medial), 458 lateral from center line (lateral), intermediate position (neutral), and three different heights: at shoulder height (middle), 458 above
shoulder height (high) and 458 below shoulder height (low). This resulted into nine reaching conditions: medial and low (ML), medial and middle (MM), medial and high
(MH), neutral and low (NL), neutral and middle (NM) the reference reaching task, neutral and high (NH), lateral and low (LL), lateral and middle (LM), and lateral and high
(LH). 26 active infrared LEDs were placed on upper arm, lower arm and trunk consisting of ve clusters xed on lower arm, upper arm, clavicle, back and thorax and six LEDs
attached to anatomical body points (radial styloid, ulnar styloid, lateral epicondyle, medial epicondyle, acromion, and processus xyphoideus). One LED was attached to the
back of the hand, and three LEDs were placed on the head.

examined the effect of reaching trajectory on the kinematics and

muscle activity of shoulder and elbow. The aim of this study was to
investigate the role of shoulder and elbow during the reaching task.
In addition, we hypothesized that the interaction between
shoulder and elbow strongly depended on the 3D target position
with a consequent effect on the underlying muscle control. This
study contributed to the understanding of joint-specic adaptation
of control strategies during reaching in 3D space.
2. Methods
2.1. Subjects
Ten healthy adolescents (six males, four females) participated in this study after
giving informed consent conforming to the institutions ethical advisory board. All
subjects were right-handed and without any history of musculoskeletal disorders of
the upper limb.
2.2. Experimental task
While sitting on a chair without backrest (knees 908 exion), subjects performed
3D reaching movements to nine different targets positioned at arm length, holding a
handle with pointer. These targets were placed in front of the subject at three
different widths and three different heights. This resulted into nine reaching
conditions shown in Fig. 1.
Each reaching cycle started with the hand on the thigh (08 shoulder elevation and
908 elbow exion) and ended at the target. The reaching duration was standardized
to 2 s and controlled by a metronome. A new cycle was initiated after 2 s of rest.
Subjects had to reach ve times to each target using a smooth motion and without
moving their trunk. To restrict wrist motion, an orthoplastic splint was applied.
2.3. Instrumentation
Kinematic data were collected at 100 Hz using a two-camera 3D motion capture
system (Krypton, Metris, Leuven). Camera position and placement protocol of 26
active infrared LEDs are shown in Fig. 1.

Bipolar surface EMG of 10 muscles was recorded (Zerowire, Aurion, Italy) at

1000 Hz conform the Seniam placement protocol [21]: deltoid anterior (DELT1),
deltoid medius (DELT2), deltoid posterior (DELT3), infraspinatus (INFRA), pectoralis
major (PMAJ), latissimus dorsi (LAT), triceps longus (TRIlong), triceps lateral
(TRIlat), biceps brachii (BIC) and brachioradialis (BRD).
2.4. Data analysis
A musculoskeletal model modied from the upper extremity model of Holzbaur
et al. [22] was used for the kinematic analysis. It denes three DOFs at the shoulder
according to the ISB standards of Wu et al. [23] and two DOFs at the elbow. The
generic model was scaled to accommodate the individual anthropometry.
Based on the three-dimensional marker trajectories, an inverse kinematics
procedure was conducted in SIMM (Software for Interactive Musculoskeletal
modeling, Motion Analysis Inc.). We analyzed shoulder elevation, elevation angle,
shoulder rotation, elbow exion and forearm rotation, shown in Fig. 2.
Prior to post hoc analysis, the kinematics were low-pass ltered (10 Hz) using a
fourth-order Butterworth lter implemented in Matlab.
Raw EMG signals were band-pass ltered (1050 Hz) using a fourth-order
Butterworth lter. RMS data (moving window: 0.05 s) were resampled to the
sampling frequency of the kinematics.
Kinematic data and RMS values of the different reaching cycles were normalized
as a function of reaching distance and expressed as a percentage of the reaching
The main effect of reaching height and width, as well as the interaction between
both factors on kinematic and RMS data, was evaluated statistically using a
repeated measure ANOVA. The effect between two targets was investigated using a
post hoc Tukey test. For both analyses, results reaching signicance level p  0.05
are reported.
To further analyze the time dependency of these ndings, the differences in
kinematics and RMS values were statistically compared over the entire reaching
cycle using a Wilcoxon signed-rank test (signicance level p  0.05).
To analyze the effect of reaching width, kinematics and muscle activity of lateral
and medial width were compared with neutral width for the three different heights.
To analyze the effect of reaching height, kinematics and muscle activity of
high and low height were compared with middle height for the three different




A. Vandenberghe et al. / Gait & Posture 32 (2010) 500507

Fig. 2. Average segment angles (8) for the nine conditions ([TD$INLE]
LH) during one reaching cycle (%). The timing of the statistical signicant differences (Wilcoxon) between specic targets is indicated at the bottom of each
graph. Different segment angles are explained at the bottom of the gure: (1) elevation angle i.e. angle of the humerus in the transverse plane of the thorax with the rotation axis
coinciding with the longitudinal axis of the thorax (with 08 indicating shoulder abduction and 908 indicating shoulder elevation), (2) shoulder rotation i.e. angle between forearm
and sagittal plane, when shoulder is at neutral elevation angle and elbow exed 908, (3) shoulder elevation i.e. angle between longitudinal axis of humerus and longitudinal axis
of body, (4) elbow exion i.e. angle between longitudinal axis of forearm and humerus and (5) forearm rotation i.e. rotation of forearm relative to its longitudinal axis (with 08
indicating palm in neutral stand).

3. Results
3.1. Reference reaching task (NM)
Over the reaching cycle, shoulder elevation increased
from 168 (48) to 818 (118) shoulder elevation and shoulder

rotation increased from 98 (138) internal rotation to 58 (148)

external rotation (Fig. 2). The elevation angle increased from
208 (348) to 608 (148). The elbow joint extended from
928 (118) exion to 188 (88) exion at the end of the reaching
cycle. The forearm rotated from 148 (108) to 288 (108)



A. Vandenberghe et al. / Gait & Posture 32 (2010) 500507


Fig. 3. Average RMS values of the EMG (V) of 10 muscles for the nine conditions ([TD$INLE]
LH) during one reaching cycle (%). The timing of the statistical signicant differences (Wilcoxon) between specic targets is indicated at the
bottom of each graph.



A. Vandenberghe et al. / Gait & Posture 32 (2010) 500507

Fig. 4. Range of motion (8) of the different segment angles for the nine conditions. Main signicant difference in height or width (ANOVA) is indicated above or left of the
tables. Signicant differences between specic targets (Tukey) are indicated in the tables. $ p < 0.01, * 0.01 < p < 0.05.

Elevation angle and forearm supination presented the largest

changes during the rst half of the cycle, whereas elbow extension
occurred mainly during the second half of the cycle. Shoulder
elevation and rotation changed continuously.
An initial activation of DELT1, INFRA and PMAJ was present
during the rst half of the reaching cycle (060% of the cycle,
Fig. 3). In contrast, DELT2, DELT3, TRIlong, TRIlat and BIC increased
muscle activity during the second half of the reaching cycle (50
100% of the cycle). The muscle activity of LAT increased

continuously. No changes in BRD activity were found over the

3.2. Effect of reaching widths
A signicant effect of reaching width was found for the range of
motion (ROM) of elevation angle and shoulder rotation (Fig. 4).
Medial reaching increased elevation angle and resulted in a
constant shoulder internal rotation. Lateral reaching decreased


A. Vandenberghe et al. / Gait & Posture 32 (2010) 500507


Fig. 5. Average RMS value of the EMG (V) of 10 muscles for the nine conditions. Main signicant difference in height or width (ANOVA) is indicated above or left of the tables.
Signicant differences between specic targets (Tukey) are indicated in the tables. $ p < 0.01, * 0.01 < p < 0.05.

elevation angle and introduced shoulder external rotation. The

changes in elevation angle occurred during the rst half of the
reaching cycle for both conditions. Changes in shoulder rotation
occurred during the rst half of the reaching cycle for medial
reaching, whereas these changes shifted towards the middle of the
reaching cycle for lateral reaching (Fig. 2). Reaching width did not
affect elbow kinematics and forearm rotation. The observed
changes were not signicantly affected by reaching height.
A signicant effect of reaching width was found for the muscle
activity of DELT1, DELT2, BIC, TRIlong, PMAJ and LAT (Fig. 5).

Medial reaching increased the muscle activity of DELT1, BIC,

TRIlong, PMAJ and LAT and decreased the activity of DELT2. In
contrast, lateral reaching resulted in decreased activity of DELT1,
BIC, TRIlong, PMAJ and LAT and increased muscle activity of DELT2.
3.3. Effect of reaching height
A signicant effect of reaching height was found for the ROM of
shoulder elevation and forearm rotation, but not for the ROM of
elbow exion (Fig. 4). High reaching increased shoulder elevation


A. Vandenberghe et al. / Gait & Posture 32 (2010) 500507

and forearm supination. In contrast, low reaching reduced

shoulder elevation and forearm supination. However, signicant
timing differences were present in elbow extension with earlier
changes in elbow extension than in shoulder elevation and forearm
rotation. During high reaching the changes in shoulder elevation
occurred later in the reaching cycle compared to low reaching. The
timing of the changes in forearm rotation was more variable
(Fig. 2). The observed changes were not affected by reaching width.
A signicant effect of reaching height was found for the muscle
activity of DELT1, DELT2, DELT3, BIC, TRIlat, LAT and BRD (Fig. 5).
Higher reaching increased the muscle activity of DELT2, DELT3,
TRIlat and LAT but decreased the activity of DELT1, BIC and BRD. In
contrast, lower reaching resulted in decreased muscle activity of
DELT2, DELT3, TRIlat and LAT, but increased activity of BRD. The
activity of DELT1 and BIC decreased during lower reaching.

4. Discussion
Different coordination strategies have been identied during
2D movements of the upper limb, dening specic relations
between muscle activity and segmental kinematics [7,10,11,13
17]. However, little is known on how these control strategies
transfer to reaching tasks in 3D space. Therefore, we present a
descriptive analysis of the kinematics and muscle activity of
shoulder and elbow of 10 normal subjects during 3D reaching
movements at three different heights and widths.
During reaching at the reference target (NM), the motion
depended mostly on increased elevation angle, elbow extension
and shoulder elevation (ROMs 808, 748 and 658 respectively).
Rotational DOFs (i.e. shoulder external rotation and forearm
supination) were the smallest with 148. Analyzing the sequence of
the kinematics, shoulder elevation and external rotation increased
continuously throughout the reaching cycle. Changes in elevation
angle and forearm supination occurred at the beginning of the
cycle, whereas elbow extension was initiated in the second half of
the cycle. These ndings are consistent with the observations of
previous 2D and 3D studies, which conrm that the shoulder
initiates the movement. We identied shoulder elevation and
rotation to control the end-point trajectory throughout the
reaching cycle, whereas the initial positioning with respect to
the reaching target was mainly controlled through the elevation
angle. Elbow extension contributed to end-point control only
during the second half of the cycle [712,18,19]. Although, the
magnitude of the rotational DOFs was small, their contribution was
important in the initial phase (pronation) and throughout the
reaching motion (shoulder rotation).
The burst pattern in muscle activity reported in previous 2D
studies [7,1417] of horizontal plane movements, clearly differs
from the activity patterns observed during 3D reaching in the
present study, where most shoulder and elbow muscles presented
a gradual increase of muscle activity over the reaching cycle. The
sustained gravitational force on the arm and the 3D nature of the
movement contributed to this difference. During the reference
reaching task, DELT1 was identied as prime contributor to
movement, presenting an immediate increase in muscle activity at
the start of the reaching movement. During the second half of the
reaching cycle, DELT2 and DELT3 took over this role. Elbow
extension during the second part of the reaching cycle was
controlled by co-activation of the elbow muscles (BIC, TRIlong and
TRIlat). Stabilization at the shoulder was achieved through coactivation of INFRA and PMAJ coinciding with DELT1 action during
the beginning of reaching, and through the continuously increasing
activity of LAT. The observed earlier onset of EMG activity of the
shoulder muscles compared to the activity of the elbow muscles is
in agreement with the ndings of previous studies [8,17,20].

Reaching height and width affected the kinematics of the

shoulder and elbow as well as the muscle activity. Depending on
the different widths or heights to be reached, different DOFs were
inuenced. However, no statistical interaction was found between
reaching height and width for kinematics and muscle activity,
indicating that both parameters independently affect the shoulderelbow strategy.
Specically, reaching at different widths affected the kinematics of the shoulder, leading to changes in the elevation angle and
shoulder rotation, but did not inuence the elbow joint and
forearm. Additionally, the sequence of the changes in shoulder
kinematics was affected by the reaching width: medial reaching
resulted in an initial change in shoulder rotation followed by a
secondary change in elevation angle, whereas during lateral
reaching the elevation angle changed rst, followed by changes in
shoulder rotation.
The effect on shoulder rotation and elevation angle was related
to the observed changes in muscle activity. The earlier change in
elevation angle compared to shoulder external rotation during
lateral reaching was also reected in the earlier activity of DELT2
when reaching lateral, which is in line with the shoulder abduction
function of DELT2. The increased co-activation of the elbow
muscles (TRIlong and BIC) observed during medial reaching did not
affect the elbow kinematics.
Reaching at different heights inuenced shoulder elevation and
forearm rotation. Compared to altering reaching width, where the
rotation occurred mainly at shoulder level, altering reaching height
inuenced forearm rotation. We can conclude that altering the 3D
nature of the reaching task dictates whether the rotational
component is shifted from shoulder to elbow joint. As all targets
were placed at arm length, reaching height had no effect on elbow
extension. The timing of the observed adaptations was jointspecic: changing reaching height resulted in an immediate
modulation of elbow extension followed by a subsequent
modulation of shoulder elevation and forearm rotation. This
nding indicates that reaching to higher targets was still initiated
by the shoulder joint, but the adaptations compared to lower
reaching occurred primarily at the elbow level. This meant that the
interaction and timing between shoulder and elbow during 3D
reaching differed from 2D reaching.
We expected the activity of DELT1, BIC and BRD to increase due
to the increase in shoulder elevation and supination when
reaching higher, but only the muscle activity of DELT2, DELT3,
TRIlat and LAT increased. These ndings showed that the
individual contribution of the muscles during the reaching task
changed when reaching to a higher target. The signicant changes
in DELT1, DELT2, DELT3 and LAT activity lasted longer and had
larger amplitude when they reached lower. This showed that
there were more differences in muscle activity between middle
low reaching than between highmiddle reaching. The early
elbow extension when reaching higher could not be conrmed at
the level of the underlying muscle activation pattern. The specic
requirement of 3D reaching to move the limb against gravity
potentially obscured this isolated effect on the elbow muscles.
Furthermore, the majority of the muscles acting at the elbow had a
biarticular role and therefore contributed to this antigravity
The current study focused on the coordination of shoulder and
elbow during 3D reaching with the subject sitting without back
support. Although subjects were instructed not to use the trunk,
some limited movement may have been present and could
inuence our results. Whilst changes in forearm rotation were
described in this study, this could not be attributed to changes in
activity of the forearm rotator muscles (pronator teres, pronator
quadratus and supinator) as these muscles cannot be reliably
measured with surface EMG.

A. Vandenberghe et al. / Gait & Posture 32 (2010) 500507

In conclusion, this study conrms that during 3D reaching, the

control strategy as reected in kinematics and muscle activity was
affected by the width and height of the target. Gravity further
modulated the observed muscle activity. Changes in reaching
width had a clear effect on kinematics and muscle activity of the
shoulder. Changes in reaching height affected the kinematics and
muscle function of the shoulder and elbow. Both manipulations
resulted in a unique pattern of temporal changes in kinematics and
muscle activity. In line with the observations from previous 2D
movements, our study indicates that the shoulder plays the role of
primary contributor to movement in a 3D reaching task. In
contrast, depending on the position in space of the reaching task,
the amplitude and timing of the change differed for the three
heights and widths. A change in reaching width had no effect on
the elbow kinematics, whereas a change in reaching height
affected initially elbow exion and subsequently shoulder elevation. It can therefore be concluded that changing the 3D nature of a
reaching task inuences the interaction between the elbow and
shoulder joint, with reaching height mainly affecting the elbow
coordination strategy. Therefore, when conducting upper limb
evaluation in subjects with pathology who cannot perform an
imposed reaching trajectory, altered 3D spatial orientation should
be considered together with the pathology as a potential
contributor to the observed changes in kinematics and muscle
activation during reaching.
Scientic responsibility for this work remains with the authors.
Conict of interest statement

[1] Buckley MA, Yardley A, Johnson GR, Carus DA. Dynamics of the upper limb
during performance of the tasks of everyday living a review of the current
knowledge base. Proc Inst Mech Eng [H] 1996;210(4):2417.
[2] Magermans DJ, Chadwick EK, Veeger HEJ, van der Helm FCT. Requirements for
upper extremity motions during activities of daily living. Clin Biomech


[3] Petuskey K, Bagley A, Abdala E, James MA, Rab G. Upper extremity kinematics
during functional activities: three-dimensional studies in a normal pediatric
population. Gait Posture 2007;25(4):5739.
[4] Van Andel CJ, Wolterbeek N, Doorenbosch CA, Veeger HEJ, Harlaar J. Complete 3D
kinematics of upper extremity functional tasks. Gait Posture 2008;27(1):1207.
[5] Kronberg M, Nemeth G, Brostrom LA. Muscle activity and coordination in the
normal shoulder. An electromyographic study. Clin Orthop 1990;257:7685.
[6] Veeger HEJ, van der Helm FCT. Shoulder function: the perfect compromise
between mobility and stability. J Biomech 2007;40(10):211929.
[7] Dounskaia N, Ketcham CJ, Stelmach GE. Commonalities and differences in
control of a large set of drawing movements. Exp Brain Res 2002;146:1125.
[8] Dounskaia N. The internal model and the leading joint hypothesis: implications for control of multi-joint movements. Exp Brain Res 2005;166:116.
[9] Galloway JC, Koshland GF. General coordination of shoulder, elbow and wrist
dynamics during multijoint arm movements. Exp Brain Res 2002;142:16380.
[10] Levin O, Ouamer M, Steyvers M, Swinnen SP. Directional tuning effects during
cyclical two-joint arm movements in the horizontal plane. Exp Brain Res
[11] Levin O, Forner-Cordero A, Li Y, Ouamer M, Swinnen SP. Evidence for adaptive
shoulderelbow control in cyclical movements with different amplitudes,
frequencies, and orientations. J Motor Behav 2008;40(6):499515.
[12] Buchanan JJ, Kelso SA, de Guzman GC. Self-organization of trajectory formation. I. Experimental evidence. Biol Cybern 1997;76:25773.
[13] Dounskaia N, Swinnen SP, Walter CB, Spaepen AJ, Verschueren SMP. Hierarchical control of different elbowwrist coordination patterns. Exp Brain Res
[14] Berardelli A, Hallett M, Rothwell JC, Agostino R, Manfredi M, Thompson PD,
et al. Single-joint rapid arm movements in normal subjects and in patients
with motor disorders. Brain 1996;119:66174.
[15] Gottlieb GL. Muscle activation patterns during two types of voluntary singlejoint movements. J Neurophysiol 1998;80:18607.
[16] Hallett M, Shahani BT, Young RR. EMG analysis of stereotyped voluntary
movements in man. J Neurol Neurosurg Psychiatry 1975;38:115462.
[17] Karst GM, Hasan Z. Timing and magnitude of electromyographic activity for twojoint arm movements in different directions. J Neurophysiol 1991;66(5):1594
[18] Desmurget M, Prablanc C, Rossetti Y, Arzi M, Paulignan Y, Urquizar C, et al.
Postural and synergic control for three-dimensional movements of reaching
and grasping. J Neurophysiol 1995;74(2):90510.
[19] Hirashima M, Kudo K, Ohtsuki T. Utilization and compensation torques during
ball-throwing movements. J Neurophysiol 2003;89:178496.
[20] Hirashima M, Kadota H, Sakurai S, Kudo K, Ohtsuki T. Sequential muscle
activity and its functional role in the upper extremity and trunk during
overarm throwing. J Sports Sci 2002;20:30110.
[21] SEN (Surface EMG for Non-Invasive Assessment of Muscles) Research Group.
European recommendations for surface electromyography results of the
SENIAM project. Enschede, The Netherlands: Roessingh Research and Development; 1999.
[22] Holzbaur KRS, Murray WM, Delp SL. A model of the upper extremity for
simulating musculoskeletal surgery and analyzing neuromuscular control.
Ann Biomed Eng 2005;33(6):82940.
[23] Wu G, van der Helm FCT, Veeger HEJ, Makhsous M, Van Roy P, Anglin C, et al.
ISB recommendation on denitions of joint coordinate systems of various
joints for the reporting of human joint motion part II: shoulder, elbow, wrist
and hand. J Biomech 2005;38:98192.