Clinical Biomechanics 14 (1999) 653±660

www.elsevier.com/locate/clinbiomech

Passive dynamics of the knee joint in healthy children and children

a€ected by spastic paresis
Maria K. Lebiedowska *, John R. Fisk
Southern Illinois University, School of Medicine, Division of Orthopedics and Rehabilitation, P.O. Box 19652, Spring®eld, IL 62794-9652, USA
Received 30 December 1997; accepted 19 March 1999

Abstract
Objective. The purposes of this study were (1) to evaluate how changes in biomechanical parameters a€ect segment dynamics in
children and (2) to determine whether the biomechanical parameters were changed in children with spastic paresis.
Design. In vivo measurements were collected of knee viscoelastic properties.
Background. It is unknown if the inertial and viscoelastic properties of a human growing limb should be considered in motor
performance. Also unclear are whether changes in passive dynamics might be responsible for abnormal control in human spastic
paresis.
Method. Small oscillation techniques were used to measure moment of inertia of lower leg, sti€ness and viscous damping of the
knee joint. Eighty seven healthy children and 32 children with spastic paresis participated.
Results. Moment of inertia, sti€ness and the damping changed with the ®fth power of child's height. Dynamic equation of motion
parameters were a constant, independent of the child's height. Passive viscoelastic parameters were not changed in spastic patients.
Conclusions. Inertial and viscoelastic properties of a growing limb segment should not be considered in motor performance.
Passive viscoelastic properties were not changed in patients with spastic paresis and, therefore, cannot be responsible for abnormal
control in human spastic paresis.

Relevance
There is no need to adapt control patterns in children (ages 6±18). Passive viscoelastic parameters cannot be used as a descriptor
of spasticity. Ó 1999 Elsevier Science Ltd. All rights reserved.

Keywords: Passive dynamics; Children; Spasticity

1. Introduction inertia, sti€ness and viscous damping of the joint ele-

The central nervous system (CNS) monitors biome-
chanical variables such as force, position and velocity.
The manner in which the CNS maintains these variables
constant, stable and reproducible in motor tasks is un-
clear. Hill [1] made an assumption that the muscle sys-
tem consists of active and passive elastic structures. The
three basic components of the viscoelastic properties of
a limb segment system are commonly evaluated at the
level of an intact joint. The passive, intrinsic and re¯ex
properties are related to the state of segment's muscle
activity (relaxation, voluntary and re¯ex, respectively).
The full dynamic description (Appendix A Eq. (A.1)) of
an individual segment (including segmental moment of
*
Corresponding author. E-mail: mlebiedowska@siumed.edu
0268-0033/99/$ - see front matter Ó 1999 Elsevier Science Ltd. All rights reserved.
PII: S 0 2 6 8 - 0 0 3 3 ( 9 9 ) 0 0 0 2 1 - 2
ments) is available for the foot [2], forearm [3,4] and
lower leg [5,6]. The equation of motion is usually re-
duced by a measurement design. This reduction relates
to the biomechanical degrees of freedom, activation of
muscles (relaxation or steady e€ort) and gravitational
forces (unloading or constant loading condition).
Methods for identifying the parameters of a reduced
equation depend on the choice of the structural ele-
ments, the complexity of the model, the identi®cation
procedures and the technical methods of force or motion
analysis being utilized. The identi®cation procedures are
based on a given system response to the di€erent input
functions and controllable initial conditions. The
methods for the moment of inertia measurements of
di€erent human body segments based on the geometrical
distribution of a segment's mass are limited to its inertial
properties [7,8] and have been discussed elsewhere [9].
654 M.K. Lebiedowska, J.R. Fisk / Clinical Biomechanics 14 (1999) 653±660
Static sti€ness, de®ned as a coecient between the
change of force and displacement, was described by
Wilkie [10] and has been measured in the elbow [11],
knee [12,13], ankle [14], metacarpophalangeal [15] and
hip [5] joints. An apparatus supplying constant velocity
displacement (the inertial properties in the equation of
motion negligible) and measuring force (or vice versa)
has been used in ankle joint sti€ness measurements [16±
18]. The resistance to sinusoidal movement or the
movement as a result of sinusoidal forces was used to
measure interphalangeal [19], elbow [20], wrist [21], an-
kle [2], hip [22] and knee muscle [12] sti€ness. Force and
displacement step and pulse input functions were used to
identify the viscoelastic properties of interphalangeal
[23], wrist [24], elbow [25], ®nger [26] and ankle [27,28]
joints.
A popular method involves driving the limb system
with sinusoidal force of increasing frequency to deter-
mine its natural frequency. Dynamic sti€ness is calcu-
lated from the free frequency with an estimated or
measured moment of inertia. This method was used to
identify the sti€ness of muscles acting on the ankle
[2,29,30], wrist [31], elbow [3,31,32], knee [6,33] and in-
terphalangeal [19,34] joints. The random perturbation
sequences (nonlinear identi®cation technique) was ap-
plied to identify the biomechanical properties of the
knee joint [35]. The biomechanical properties of di€erent
body segments were measured during muscle relaxation
and during activation by di€erent sources (voluntary
and re¯ex).
There are no data, however, describing the changes in
viscoelastic properties in children populations. It is un-
certain to what extent the inertial and viscoelastic
properties of a growing limb segment should be con-
sidered in motor performance. Do adaptive control
mechanisms automatically recalibrate the motor system
[36]? There were several studies investigating the changes
in passive sti€ness and viscous damping in human
spastic paresis. The ®ndings for adult spastic patients
are inconsistent [5,30,33], and there are no data for
children with spastic paresis.
The objectives of this study are twofold. The ®rst
objective is to evaluate the in¯uence of changes in bio-
mechanical parameters on segment dynamics in a chil-
dren population. The second objective is to determine
whether the passive segment dynamics are a€ected by
human spastic paresis and can be used as a quantitative
measure of spasticity in children.
2. Methods
Eighty seven healthy school children (33 girls and 54
boys) were evaluated. They ranged in age from 5 to 18
yrs. Their body height ranged from 113 to 189 cm, and
their body mass ranged from 19 to 81 kg. Children were
examined and determined to be healthy prior to the
experiments. Thirty two patients with spasticity partic-
ipated in the study. Twenty three children had spastic
hemiplegia (Table 1). Nine children exhibited spasticity
in both legs (Table 2).
All children were patients of the Rehabilitation De-
partment, Child Health Centre, Warsaw, Poland, and
were referred for evaluation by a physician. All subjects
and their parents gave oral consent to be included in the
study.
We incorporated the free oscillation technique origi-
nally used by Hill [1] to measure the forearm moment of
inertia. This method was described for the moment of
inertia elsewhere [9]. The shin in the free hanging posi-
tion was attached to a strain-gauge beam at the ankle
joint (Fig. 1). The strain gauge beam was used to mea-
sure transitional position.
The system was perturbed by manually hitting (rep-
resented by d(t) in Appendix A Eq. (A.8)) the leg at the
ankle joint perpendicularly to the long axis of the seg-
ment to elicit small oscillations. The signal from the
strain gauge bridge (CMT-83) was ampli®ed (Tektronix
AM502) and fed to an electromyograph, Disa 1500.
Surface electrodes were attached to the skin over the
rectus femoris and hamstring muscles to monitor EMG
activity and to verify when the muscles were relaxed (no
signs of EMG activity > 50 lV). A one degree of free-
dom model of the lower leg and the apparatus were used
(Appendix A Eq. (A.8)). The reduced mass (0.005 kg)
and the damping of the apparatus (0.007 N s/m) were
negligible compared to other parameters, and thus, were
not in the model. It was found that the in¯uence of the
viscous damping on the period of free oscillations of the
system was less than 0.28%, thus the system was reduced
to an undamped second order system. The standard
solution of Eq. (A.8) (Appendix A) relates these pa-
rameters, which can be calculated from the periods of
free system oscillations. The periods were measured with
an external mass loading me of 0, 0.5, 1 and 2 kg. The
average values for up to ten oscillations for each loading
were calculated. Next the square of free system oscilla-
tion x2 was calculated, and the function m ˆ f(x2 ) was
plotted (Fig. 2A).
The function value at 0 was assigned as the reduced
mass of the whole system (mr ). The moment of inertia
was calculated according to the formula: I ˆ mr2 , where
r is the radius of reduction, equaling the distance from
the knee axis of rotation to the point of attachment. The
error of the moment of inertia calculations depended on
the correlation coecient (which was > 0.95) and was in
the range of 2±10%. The slope of the function was
considered as the sum of joint sti€ness (reduced to the
attachment point), apparatus sti€ness and the gravita-
tional component (Eq. (A.10)). The mass of the lower
leg was estimated as 0.061 of total body mass (37). The
distance of center of mass from the knee joint was
 M.K. Lebiedowska, J.R. Fisk / Clinical Biomechanics 14 (1999) 653±660 655

Table 1
Group of children with spasticity of di€erent origin with hemiplegia a

No Sex Age Body height Body mass Type of Diagnosis Years after
(years) (m) (kg) paresis incident

P1 M 12 1.48 30 HS S 8
P2 F 15 1.60 52 HD S 1
P3 F 10 1.42 29 HS CP ÿ
P4 M 10 1.25 24 HS SS 6/12
P5 F 13 1.62 49 HD CP 6/12
P6 F 12 1.54 56 HS BS 1
P7 M 5 0.97 15 HS S 5
P8 M 7 1.32 28 HD CP ÿ
P9 F 17 1.62 49 HD CP ÿ
P10 F 7 1.25 25 HS TBI 2
P11 M 14 1.77 47 HD TBI 3/12
P12 M 7 1.34 32 HS SI 1
P13 F 5 1.18 26 HD TBI 6/12
P14 M 9 1.31 27 HD TBI 6/12
P15 M 9 1.33 34 HS TBI 1
P16 M 12 1.53 44 HD TBI 2 6/12
P17 F 7 1.30 29 HS TBI 1
P18 F 3 0.98 22 HD TBI 1
P19 M 9 1.27 25 HD TBI 3
P20 M 9 1.29 24 HS TBI 2
P21 F 14 1.63 44 HD BS 8
P22 M 12 1.44 36 HD CP ÿ
P23 F 12 1.33 30 HS CP ÿ
a
S: stroke BS: brain surgery, CP: cerebral palsy, TBI: traumatic brain injury, SI: spinal cord injury, SS: spinal cord surgery, HS: hemiplegia of left
side of the body, HD: hemiplegia of right side of the body.

Table 2
Group of children with spasticity of di€erent origin with diplegia (D), paraplegia (P) or tetraplegia (T) a

No Sex Age Body stature Body mass Type of Diagnosis Years after
(years) (m) (kg) paresis incident

P1 M 10 1.6 56 P TBI 1
P2 F 15 1.52 48 P TBI 16/12
P3 M 14 1.43 34 T SI 0.2
P4 M 10 1.40 36 P TBI 6/12
P5 F 15 1.34 31 T TBI 2/12
P6 M 9 1.71 46 T TBI 1
P7 F 14 1.69 43 D SS 2
P8 F 12 1.58 46 P SI 7
P9 M 8 1.2 21 P CP ÿ
a
CP: cerebral palsy, TBI: traumatic brain injury, SI: spinal cord injury, SS: spinal cord surgery.

estimated as 0.606 of lower leg length (from greater

Fig. 1. The block diagram of the arrangement of biomechanical pa-
rameter measurements.
trochanter to medial malleolus) (37). The joint sti€ness
was calculated as the di€erence of the apparatus sti€ness
and gravitational component from the function slope.
The linear viscous damping of the elements acting on the
knee joint (reduced to the ankle joint) was calculated
from the decrement of amplitude (Fig.2B) (Appendix A,
Eq. (A.11)). To obtain moment of inertia, rotational
sti€ness and viscous damping around the joint axis, the
transitional parameters were multiplied by the square of
the radius of reduction. This was equal to the distance
from the knee axis of rotation to the location of the
transducer. The appropriate coecients of the dynamic
656 M.K. Lebiedowska, J.R. Fisk / Clinical Biomechanics 14 (1999) 653±660

If the re¯ex activation was elicited in the patientsÕ

muscles during the time of measurement (10±15 min) in
certain trials, then these trials were excluded from the
calculation.
Only the trials with the EMG level below 50 lV were
used in further analysis. We observed that the scatter of
the data points around the function m ˆ f(x2 ) was re-
lated to the activation of muscles below the 50 lV
threshold and the activation of other knee muscles
which were not controlled. Thus the correlation coe-
cient of the m ˆ f(x2 ) function (Fig. 2A) was used to
quantify the ¯uctuations of the sti€ness Kfl during the
period in which data were collected. Kfl in controls was
equal to 0.992 ‹ 0.008.
In 21 children with spasticity restricted to one side of
the body, the results in both limbs were compared using
the paired t-test. Next all parameters for all a€ected
limbs (n ˆ 41) were pooled together and compared with
the control values at the signi®cance level a < 0.05. The
t-test was used if the data were normally distributed;
otherwise, Wilcoxon or sign tests were applied.

Fig. 2. (A) A function m ˆ kxÿ2 which was used to calculate segmentÕs 3. Results
moment of inertia I, and passive sti€ness (k) of the knee joint.
M ˆ mr + me , k ˆ kr + ke , x2 ˆ 4p2 /Tÿ2 , I ˆ mr \r2 , where mr is the re-
The results of measurements for the moment of in-
duced mass of the lower leg to the point of attachment (kg), me the
external mass loading (kg), kr the passive transitional sti€ness of knee ertia of the lower leg (I), sti€ness (K), damping (B) and
joint reduced to the point of, attachment (N/m), ke the sti€ness of the gravitational component (G) of the elements acting on
apparatus (148 N/m), T the period of free oscillations (s), r the radius the knee joint as the functions of body height are dis-
of reduction (m). (B) Calculation of passive transitional viscous played in Fig. 3 (A, B, C and D, respectively).
damping (b) of the knee joint, (Eq. (A.11)) where Dt ˆ Ti ÿ Ti ÿ 1 , n is
the number of oscillations used in calculations, Ti the amplitude of i
Moment of inertia, sti€ness and damping changes
oscillation used in calculations. with a ®fth power of body stature. The gravitational
component changes with a fourth power of body stat-
ure. The results obtained in spastic children are pre-
equation of the segment were calculated according to
sented in Tables 3 and 4.
their de®nitions (Appendix A, Eqs. (A.3)±(A.7). The
All parameters except for sti€ness ¯uctuation were
error of the moment of inertia and sti€ness measurement
normally distributed. No di€erences were found be-
had depended on the correlation coecient values (Fig.
tween the a€ected and the non a€ected side of the body
2) and could reach two percent. The error of the dam-
(in patients with hemiplegia), except for the sti€ness
ping measurement was 25%. The best least square ®t of
¯uctuation which was larger (correlation coecient at
the values of moment of inertia, sti€ness and viscous
the range: 0.82±0.99) in a€ected than in nona€ected
damping as a function of the ®fth power of body height
limbs (range: 0.98±0.99). The signi®cance of sti€ness
was calculated. The best least square ®t of gravitational
¯uctuation di€erences was con®rmed (z ˆ 3.1) with the
component G as a function of the fourth power of body
sign test.
height was calculated.
Sti€ness (5.17 N radÿ1 mÿ4 ‹ 2.19 N radÿ1 mÿ4 ) in all
The data were normalized by dividing the individual
a€ected limbs was larger than in healthy limbs (4.87 N
results of moment of inertia, sti€ness and viscous dam-
radÿ1 mÿ4 ‹ 1.83 N radÿ1 mÿ4 ) but was not statistically
ping by the ®fth power of body stature according to an
signi®cant. Moment of inertia in all a€ected limbs (0.032
earlier established procedure [38]. The mean values and
kg mÿ3 ‹ 0.01 kg mÿ3 ) was larger than in healthy limbs
standard deviations of normalized moment of inertia
(0.0295 kg mÿ3 ‹ 0.076 kg mÿ3 ) but was not statistically
(In ), sti€ness (Kn ) and viscous damping (Bn ) were:
signi®cant. Viscous damping in all a€ected limbs (0.053
In ˆ 0:0295  0:076…k mÿ3 †; N s radÿ1 mÿ4 ‹ 0.025 N s radÿ1 mÿ4 ) was smaller than
in healthy limbs (0.06 ‹ 0.022) but was not statistically
Kn ˆ 4:87  1:83…N radÿ1 mÿ4 †; signi®cant. The ¯uctuation of sti€ness in all a€ected
limbs was larger (correlation coecient range: 0.8±0.99)
Bn ˆ 0:06  0:022…N s radÿ1 mÿ4 †: than in controls (0.96ÿ1), and the signi®cance of
 M.K. Lebiedowska, J.R. Fisk / Clinical Biomechanics 14 (1999) 653±660 657

Fig. 3. Biomechanical parameters as a function of body height. (A) lower leg moment of inertia (I), I ˆ 0.0299 H5 , r2 ˆ 0.74; (B) passive sti€ness (K)
K ˆ 4.588 H5 , r2 ˆ 0.68; (C) passive viscous damping (B), B ˆ 0.0609 H5 , r2 ˆ 0.68; (D) gravitational component, G ˆ 1.072 H4 , r2 ˆ 0.93.

di€erences was statistically con®rmed (z ˆ ÿ6.625) with
the Wilcoxon test.
4. Discussion
There are no data describing the sti€ness and viscous
damping in children. The values of moment of inertia
obtained with this study agree with other data reported
in children, which were discussed in detail elsewhere [9].
Our values of sti€ness in the oldest children (Fig. 3) were
similar to the values obtained by Tai and Robinson [6]
(90±130 N m/rad in ®ve adults) and were close to the
values obtained by Mansour [13] (6±50 N m/rad in four
adults). The sti€ness values were larger than those ob-
tained by Douglas et al. [33] and Stein et al. [5] (5±8 N
m/rad). Tai and Robinson [6] used a small range of
motion in their study as we did in ours; it could be hy-
pothesized that the short range [39] sti€ness played a
role in both studies. However, the comparability of re-
sults obtained by di€erent methods needs more study.
Our values of viscous damping obtained in the oldest
children (Fig. 3) are close to the values obtained in ten
adults (0.78 ‹ 0.53) [5]. The increase of moment of in-
ertia, sti€ness and viscous damping according to ®fth
power of body height con®rms the proportional growth
of human body segments in children aged from 6 to 18
years. The dynamics of the passive lower leg in chil-
dren's population is described by the following equa-
tion:
a00 ‡ 2 a0 ‡ 165 a ‡ 18:1 Hÿ1 sin a ˆ text ‡ tint :
When the motion is restricted to the horizontal plane
and/or when the torque generated by the muscles is
much larger than the gravitational torque (tint > 18.1
Hÿ1 ) (fast movements), then similar equations describe
the motion of the body segment, despite the body size.
Thus, the dynamics is an invariant property of the
growing human motor system. This is an important
feature of the growing human motor system. Similar
kinetic activity with respect to the shape, amplitude and
duration can be expected as a result of the proportional
propelling of a segment. The established patterns of the
movements can be used without any adaptation to the
geometrically increasing but dynamically similar bio-
mechanical segment system. It means that there is no
need to change the activation patterns in children 6±18
yrs. Therefore, in human gait, similar kinematic gait
patterns are accompanied by similar patterns of muscle
activation [37].
In our study, small range (less than 0.26 rad) passive
sti€ness and viscous damping do not change in spastic
658 M.K. Lebiedowska, J.R. Fisk / Clinical Biomechanics 14 (1999) 653±660

Table 3
Results of small range knee passive sti€ness and viscous damping and moment of inertia of lower leg measurements (group of children with
hemiplegia of di€erent origin) a

Sti€ness K Damping B Inertia I Kvar

(N/rad m4 ) (N s/rad m4 ) (kg/m3 )

A N A N A N A N

P1 3.56 4.23 0.039 0.056 0.037 0.039 0.98 0.99

P2 4.63 7.32 0.079 0.099 0.04 0.058 0.97 0.99
P3 2.3 5.24 0.037 0.066 0.024 0.049 0.99 0.98
P4 5.9 ÿ 0.068 ÿ 0.042 ÿ 0.96 ÿ
P5 3.25 1.88 0.056 0.035 0.037 0.024 0.96 0.99
P6 9.46 8.86 0.116 0.127 0.07 0.056 0.98 0.99
P7 5.26 6.02 0.059 0.07 0.031 0.031 0.98 0.99
P8 6.06 4.39 0.057 0.104 0.05 0.061 0.95 0.99
P9 3.33 5.05 0.042 0.039 0.022 0.030 0.97 0.98
P10 6.49 9.4 0.052 0.064 0.026 0.033 0.97 0.99
P11 5.86 3.78 0.029 0.027 0.039 0.025 0.99 0.99
P12 6.96 7.57 0.011 0.010 0.045 0.042 0.98 0.98
P13 7.7 7.9 0.022 0.02 0.036 0.039 0.98 0.99
P14 9.25 7.5 0.075 0.09 0.035 0.027 0.99 0.99
P15 5.02 5.59 0.067 0.086 0.036 0.032 0.82 0.99
P16 3.47 4.30 0.032 0.028 0.019 0.014 0.97 0.99
P17 5.6 6.9 0.009 0.013 0.032 0.024 0.98 0.98
P18 5.6 5.4 0.031 0.042 0.026 0.031 0.98 0.99
P19 3.33 4.97 0.04 0.042 0.023 0.018 0.97 0.98
P20 4.25 ÿ 0.052 ÿ 0.027 ÿ 0.99 ÿ
P21 3.11 3.36 0.043 0.042 0.031 0.027 0.97 0.99
P22 4.8 4.6 0.052 0.058 0.026 0.022 0.99 0.99
P23 4.56 4.32 0.056 0.052 0.033 0.029 0.99 0.99
n 23 21 23 21 23 21 23 21
mean 5.2 5.65 0.048 0.056 0.02265 0.0344 0.971 0.986
SD 1.89 1.87 0.023 0.031 0.011 0.012 0.027 0.007
a
Note: All parameters normalized by ®fth power of body height. A: a€ected side of the body, N: nona€ected side of the body.

Table 4
Results of small range knee passive sti€ness and viscous damping and moment of inertia of lower leg measurements (group of children with tet-
raplegia, diplegia or paraplegia of di€erent origin) a

Sti€ness K Damping B Inertia I Kvar

(N/rad m4 ) (Ns/rad m4 ) (kg/m3 )

R L R L R L R L

P1 10.3 4.1 0.1 0.048 0.032 0.028 0.8 0.96

P2 3.29 2.7 0.023 0.042 0.028 0.026 0.99 0.99
P3 6.93 3.2 0.07 0.036 0.049 0.022 0.99 0.99
P4 9.5 8.4 0.096 0.1 0.047 0.05 0.97 0.99
P5 7.47 6.9 0.1 0.069 0.026 0.03 0.97 0.97
P6 3.37 1.32 0.042 0.017 0.031 0.016 0.97 0.96
P7 3.11 3.36 0.043 0.042 0.031 0.027 0.97 0.99
P8 4.8 4.6 0.052 0.058 0.026 0.022 0.99 0.99
P9 4.56 4.32 0.056 0.052 0.033 0.029 0.99 0.99
n 18 18 18 18
mean 5.12 0.0581 0.0307 0.97
SD 0.058 0.026 0.009 0.04
a
Note: All parameters normalized by ®fth power of body height. R: right side of the body, L: left side of the body.

children treated as a group; although in some individ- lV was observed, con®rming the well known fact that in
uals, di€erences were observed. The larger ¯uctuation of spasticity full relaxation is dicult. These results agree
the passive sti€ness at level of EMG activation below 50 with Douglas et al. [33] and Stein et al. [5] obtained for a
 M.K. Lebiedowska, J.R. Fisk / Clinical Biomechanics 14 (1999) 653±660 659
X X
knee joint. They did not ®nd an increase in the sti€ness Ia00 ‡ Ba0 ‡ Ka ‡ G ˆ Text ‡ Tint …1†
in spastic patients when the muscles were electrically
silent. Douglas et al. [33] found a decrease in the sti€ness or
X X
of paraplegics with upper motorneuron lesions com- a00 ‡ ca0 ‡ x2 a ‡ b ˆ text ‡ tint …2†
pared to the control subjects. This ®nding correlated to
a decrease in the thigh circumference and atrophy of the where:
muscles. Contrary to those results, Lehman et al. [30] G ˆ MRg
and Hufschmidt and Mauritz [40] found that the sti€-
ness in the knee joint in spastic patients was larger than b ˆ GIÿ1 …3†
in controls (40±80 N m/rad). To exclude the in¯uence of
muscle activation, Lehman et al. [30] con®rmed their c ˆ BIÿ1 …4†
results with nerve blocks. Similar results were obtained
for ankle joint muscles by Gottllieb et al. [2]. Bromberg x2 ˆ KIÿ1 …5†
and Grimby [18] reported a velocity dependent increase
of resistance torque, but they did not measure muscle text ˆ Text Iÿ1 …6†
activity It should be noted that patients with muscle
contractures were involved the Hufschmidt and Mauritz tint ˆ Tint Iÿ1 …7†
[40] study. It is not clear if muscle contactures were where B, K is the tensors of viscous damping and sti€-
present in Lehman et al. [30] patients; however, the 5 ness of the muscles acting on the joint: passive and ac-
degree range of motion criterion seems to give place for tive (including voluntary and re¯ex muscle activity), M
such an assumption. Contractures were neither present the matrix of the segment masses, R the matrix of the
in our studies nor in the Stein et al. [5] study. distances of center of mass from joint axis, I the matrix
The inconsistency of the passive viscoelastic param- of the segmentÕs moment of inertia around the joint axis,
eters reported for healthy subjects and in patients may Text;int the tensor of the external and external moment of
be addressed in future studies. The following issues forces (respectively), a; a0 ; a00 the matrix of the general-
should be taken into account: the measurement method ized coordinate of motion and their time derivatives.
applied, the detailed clinical status of the patients (2) The biomechanical model applied in the small
(contractures, atrophy of muscles) and the ability of oscillation techniques:
subjects to relax.
…mr ‡ me †x00 ‡ …br ‡ be †x0 ‡ …kr ‡ ke †x ‡ …m ‡ me †gx
ˆ d…t† …8†
5. Conclusions
where mr ; kr ; br is the mass of the lower leg (with foot),
During growth of the human body segment, the sti€ness and viscous damping of knee joint (reduced to
segment passive dynamics remain invariant. Thus, it is the point of attachment), me ; ke ; be the external loading
not necessary to adapt the motor control parameters to mass, sti€ness and viscous damping of the apparatus,
the changes in biomechanical parameters of the growing x; x0 ; x00 the transitional position, velocity and accelera-
limb segment. The small range passive sti€ness and tion at the point of attachment (perpendicular to the
viscous damping in patients a€ected by spastic paresis long segment's axis), m the segment's mass, me the ex-
are not changed. Thus the small range viscoelastic ternal mass, ke the sti€ness of arrangement (ke ˆ 148 N/
properties cannot be responsible for abnormal control in m), d(t) the Dirac pulse function.
human spastic paresis and cannot be used as a de- The standard solution of Eq. (7) leads to the ex-
scriptor of spasticity. pression
x2 ˆ ‰kr ‡ ke ‡ g…m ‡ me †Š=…mr ‡ me † …9†

Acknowledgements Comparing Eq. (8) with the de®nition of free oscillation

The authors thank A. Polisiakiewicz M.Sc. for help-
ing in data collection and Dr. M. Lebiedowski for his
clinical evaluations.
Appendix A
(1) Human limb dynamics are described by the fol-
lowing equations.
period x ˆ 2p=T where T is the periods of free oscilla-
tions, the function
mr ‡ me ˆ T 2 =4p2 ‰kr ‡ ke ‡ g…m ‡ me †Š …10†
can be determined (Fig. 2A).
(3) Determination of viscous damping of muscles
acting on the joint.
The transitional viscous damping b of muscles acting
on a joint was calculated from the decrement of dam-
ping (Fig. 2B) according to the formula
660 M.K. Lebiedowska, J.R. Fisk / Clinical Biomechanics 14 (1999) 653±660
b ˆ 2…mr ‡ me †… ln Ai ÿ ln Aiÿ1 †=nDt; …11†
where Dt ˆ Ti ÿ Tiÿ1 , n is the number of oscillations used
in calculations, Ti the amplitude of i oscillation used in
calculations.
References
[1] Hill AV. The dynamic constant of human muscles. Proc R Soc (B)
1940;128:263±74.
[2] Gottlieb GL, Agarwal GC, Penn R. Sinusoidal oscillations of the
ankle as a means of evaluating of spastic patients. J Neurol
Neurosurg Psychiatry 1978;41:32±9.
[3] Zahalak GI, Heyman SJ. A quantitative evaluation of the
frequency-response characteristics of active human skeletal muscle
in vivo. Trans ASME 1979;101:28±37.
[4] Lacquaniti F, Soechting JF. Simulation studies on the control of
posture and movement in a multi-jointed limb. Biol Cybern
1986;54:367±78.
[5] Stein RB, Zehr P, Lebiedowska MK, Popovic D, Scheiner A,
Chizeck HJ. Estimating mechanical parameters of leg segments in
individuals with and without physical disabilities. IEEE Trans
Rehabil Eng (3) 1996;4:201±11.
[6] Robinson CJ. Elasticity and viscosity variations about the knee as
a function of joint angle and perturbation intensity. Personal
communications 1996.
[7] Dempster WT. Space requirements of the seated operator.
WADC Technical Report No. 87892 Wright Patterson Air Force
Ohio 1955:55±159.
[8] Hatze H. A mathematical model for the computational determi-
nation of parameter values of anthropometric segments. J
Biomech 1980;13:833±43.
[9] Lebiedowska M, Polisiakiewicz A. Changes in the lower leg
moment of inertia due to child's growth. J Biomech 1997;30:723±8.
[10] Wilkie DR. The relation between force and velocity in human
muscle. J Physiol 1950;110:249±80.
[11] Webster DW. The dynamic quantitation of spasticity with
automated integrals of passive motion resistance. Clin Pharmacol
Ther (6) 1964;5:900±8.
[12] Such CH, Unsworth A, Wright V, Dowson D. Quantitative study
of sti€ness in the knee joint. Ann Rheum Dis 1975;34:286±91.
[13] Mansour JM, Audu AM. The passive elastic moment at the knee
joint and its in¯uence on human gait. J Biomech 1986;19:369±73.
[14] Hunter IW, Kearney RE. Dynamics of ankle sti€ness: variation
with mean ankle torque. J Biomech 1982;15:747±52.
[15] Wright V, Johns RJ. Physical factors concerned with the sti€ness
of normal and diseased joints. Bull John Hopkins Hosp
1960;106:216±31.
[16] Knutsson E, Martensson A. Dynamic motor capacity and its
relation to prime motor dysfunction, spastic re¯exes and antag-
onistic coactivation. Scand J Rehabil Med 1980;12:93±106.
[17] Chesworth BM, Vandervoort AA. Age and passive ankle sti€ness
in healthy women. Phys Ther 1989;69:217±8.
[18] Bromberg C, Grimby G. Measurement of torque during passive
and active ankle movements in patients with muscle hypertonia ±
a methodological study. Scand J Rehabil Med Suppl 1983;(9
Suppl):108±117.
[19] Brown TIH, Rack PMH, Ross HF. Forces generated at the
thumb interphalangeal joint during imposed sinusoidal move-
ments. J Physiol 1982;332:69±85.
[20] Joyce GC, Rack PMH, Ross HF. The forces generated at the
human elbow joint in response to imposed sinusoidal movements
of the forearm. J Physiol 1974;240:351±74.
[21] Johns RJ, Wright V. Relative importance of various tissues in
joint sti€ness. J Appl Physiol 1962;17:824±8.
[22] Walsh EG, Wright GW. Postural thixotropy at the human hip. Q
J Exp Physiol 1988;73:369±77.
[23] Carter RR, Crago PE, Keith MW. Sti€ness regulation by re¯ex
action in the normal human hand. J Neurophysiol 1990;66:105±
17.
[24] Gielen CCA, Houk JC. Nonlinear viscosity of human wrist. J
Neurophysiol 1984;52:554±69.
[25] Jahnke MT, Proske U, Struppler A. Measurements of muscle
sti€ness, the electromyogram and activity in single muscle spindles
of human ¯exor muscles following conditioning by passive stretch
or contraction. Brain Res (1) 1989;493:103±12.
[26] Akazawa K, Milner TE, Stein RB. Modulation of Re¯ex EMG
and Sti€ness in Response to Stretch of Human Finger Muscle. J
Neurophysiol (1) 1983;49:16±27.
[27] Allum JHJ, Mauritz KH. Compensation for Intrinsic Muscle
Sti€ness by Short-Latency Re¯exes in human Triceps surae
Muscles. J Neurophysiol (5) 1984;52:797±818.
[28] Sinkjaer T, Toft E, Andreassen S, Hornemann BC. Muscle
sti€ness in human ankle dorsi¯exors: intrisinic and re¯ex compo-
nents. J Neurophysiol 1988;60:1110±21.
[29] Evans CM, Fellows SJ, Rack P, Ross HF, Walters DKW.
Response of the normal human ankle joint to imposed sinusoidal
movement. J Physiol 1983;344:483±502.
[30] Lehman JF, Price R, de BJ, Lateur S, Hinderer C. Spasticity:
quantitative measurements as a basis for assessing e€ectiveness
of therapeutic intervation. Arch Phys Med Rehabil 1989;70:
6±15.
[31] Lakie M, Walsh EG, Wright GW. Measurement of inertia of the
hand and the sti€ness of the forearm using resonant frequency
methods with added inertia or position feedback. J Physiol
1981;310:3P±4P.
[32] MacKay WA, Crammond DJ, Kwan HC, Murphy JT. Mea-
surement of human forearm viscoelasticity. J Biomech 1986;19:
231±8.
[33] Douglas AJ, Walsh EG, Wright GW. Edmond tone around the
human knee in paraplegia. Q J Exp Physiol 1989;74:897±905.
[34] Rack PMH, Ross HF, Thilmann AF. The ankle stretch re¯exes in
normal and spastic subjects. Brain 1984;109:637±54.
[35] Zhang LQ, Nuber G, Butler GN, Bowen M, Rymer WZ. In vivo
human knee joint dynamic properties as functions of muscle
contraction and joint position. J Biomech 998;31:71±6.
[36] Bullock D, Grossberg S. The vite model: a neural command
circuit for generating arm and articular trajectories. In: Kelso
JAS, Mandell AJ, Schlesinger MF, editors. Dynamic Patterns in
Complex Systems. Singapore: Worlds Scienti®c, 1988;315±326.
[37] Winter D. Biomechanics and Motor Control of Human Move-
ment. 2nd ed. New York: Wiley, 1990;56.
[38] Lebiedowska M, Syczewska M, Gra€ K, Kalinowska M. Appli-
cation of biomechanical growth models in the quantitative
evaluation of the child motor system. Disabil Rehabil (3)
1996;18:137±42.
[39] Hufschmidt A, Schwaller I. Short-range elasticity and resting
tension of relaxed human lower leg muscles. J Physiol
1987;391:451±65.
[40] Hufschmidt A, Mauritz KH. Chronic transformation of muscle in
spasticity: a peripheral contribution to increased tone. J Neurol
Neurosurg Psychiatry 1985;48:676±85.