Sie sind auf Seite 1von 41

See

discussions, stats, and author profiles for this publication at:


https://www.researchgate.net/publication/223348184

Joint stiffness: Myth or reality?

Article in Human Movement Science · December 1993


DOI: 10.1016/0167-9457(93)90010-M

CITATIONS READS

218 1,117

2 authors, including:

Vladimir M Zatsiorsky
Pennsylvania State University
350 PUBLICATIONS 10,865 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related
projects:

Poster 12 Therapeutic Effects of Gait Training and Gait-Related


Training with Functional Electrical Stimulation for Chronic Stroke
Patients: Prospective Observational Study View project

All content following this page was uploaded by Vladimir M Zatsiorsky on 14 October 2017.

The user has requested enhancement of the downloaded file.


Human Movement Science 12 (1993) 653-692 653
North-Holland

Joint stiffness: Myth or reality?


Mark L. Latash *3a, Vladimir M. Zatsiorsky b
‘I Department of Physiology, Rush-Presbyterian St.Luke S Medical Center,
1753 W. Congress Parkway, Chicago, IL 60612, USA
h Biomechanics Luborutory, Penn State lJnit,er.sity, Unil,ersity Park, PA 16802, USA
and State Institute of Physical Culture, Moscow, Russia

Abstract

Latash, M.L. and V.M. Zatsiorsky, 1993. Joint stiffness: Myth or reality. Human Movement
Science 12. 653-692.

The notion of joint stiffness as commonly studied in biomechanics and motor control is
compared with the physical definition of stiffness. The importance of elastic deformation and
storage of elastic energy is stressed. Different terms are suggested in order to differentiate
between experimentally observed relations between joint angle and torque that are likely to have
different nature. A review of studies measuring stiffness of joint subcomponents and intact joints
is presented. We suggest to either abandon the term ‘joint stiffness’ as misleading or to state up
front stiffness of which of the joint components or subsystems is analyzed in each particular
study. We also suggest that each study of ‘joint stiffness’ should clearly state to what extent the
results are defined by the system’s properties and to what extent they are reflections of the
particular experimental procedure.

‘I am appalled. They have taken any sort of length


step, measured the consequent tension step or Llice
wrsa, and called the result a stiffness. It is rubbish, of
course, even for one-element situations, and mostly
they are not one-element situations.’

D.R. Wilkie (1979, p. 634)

1. Introduction

The main purpose of this paper is to discuss the notion of ‘joint


stiffness’ and the problems associated with its use and abuse in the
biomechanics and motor control literature.

* Corresponding author

0167.9457/93/$06.00 0 1993 - Elsevier Science Publishers B.V. All rights reserved


Spring-like behavior of muscles and joints has been known since the
middle of the last century (Weber, 1847). The importance of muscle
spring properties was emphasized by such classics of biomechanics
and motor control as Bernstein (1947, 19671, Hill (1950, 19531, and
Sherrington (1906). In particular, spring-like properties of muscles and
joints are believed to play an important role in maintaining human
vertical posture (Feldman, 1966; Gurfinkel et al., 19651, in storing and
recoiling elastic energy over a stretch-shortening muscle cycle
(Alexander and Vernon, 1975; Alexander and Bennet-Clark, 1977;
Alexander et al., 1982; Alexander, 1988; Huijing, 1992; Komi, 19921,
particularly during locomotion (Gleim et al., 19901, and in control of
muscular activity (Feldman, 1966, 1986; Feldman et al., 1990; Bizzi,
1980; Bizzi et al., 1982, 1992; Latash and Gottlieb, 1991a; Hogan,
1990; Latash, 1992).
In order to describe and study these properties, researchers in the
fields of biomechanics and motor control frequently use the well-
established physical notion of stiffness. However, the applicability of
this term for describing such complex objects as muscles and joints is
not obvious. Its usage in many of the studies is likely to make a
physicist nervous and the emergence of such expressions as ‘negative
stiffness’ in serious scientific publications may even cause a nervous
breakdown.
The notion of stiffness has been introduced in physics to character-
ize properties of certain types of deformable bodies under an influ-
ence of external forces. In the absence of external forces, these bodies
are supposed to maintain constant shape. Muscles are not such
bodies, and joints can hardly be considered bodies at all. They are
rather links between the bodies or conglomerates of bodies. In partic-
ular, Hasan and Enoka (1985) reported unstable angle ranges in the
human elbow when changes in the force arms led to a seemingly
negative stiffness in the joint while both flexors and extensors behaved
like classical springs with positive stiffness. In order to apply the
notion of stiffness to muscles and joints, one needs to redefine this
notion, clarify how it can be measured in an ideal mental experiment,
and explicitly state the differences between this ideal experiment and
real procedures that are frequently used for measuring muscle and
joint stiffness.
This paper consists of three major sections. In the first section, the
notion of stiffness, as it is used in physics, is considered. The second
ML. Laiash, VWf. Zatsiorsky / Joint stiffness 655

section deals with the stiffness of the joint constituent tissues (tendons
and muscles). Since a comprehensive review on the properties of
muscles and tendons has been recently published by Zajac (19891, we
shall just touch these issues and elaborate only on those that have
important implications for control of voluntary movements. In the
third section, joint stiffness is analyzed.

1.1. The notion of stiffness in physics

In physics, the notion of stiffness is introduced for objects that


deform under the influence of an external force, generate force to
oppose the external force, and can store elastic energy. For an ideal,
unidimensional spring, according to Hooke’s law, this force is propor-
tional to spring deformation and is directed along the same coordi-
nate:

F,= -kx,

where x is coordinate of the tip of the object. If one imposes a force


vector directed along the spring, waits until the spring comes to a new
equilibrium state, and then measures changes in force and length,
stiffness, k can be defined as:

k = -AF/Ax, (2)

where AF is change in force and Ax is change in length. Note that


this particular mental experiment involves two important factors: (a)
the spring is unidimensional and (b) the measurements are performed
at equilibria. Eq. (2) becomes inapplicable if the measurements are
performed while the object is still moving.
If the spring does not have inertia but has an inertial component
attached to its end, it can be described as:

m d2x/dt2 = -hx, (3)

where m is inertia. If the system also involves a viscous element,


acting in parallel to the spring, that develops force proportional to
656 M.L. Latash, V.M. Zatsiorsky /Joint stiffness

velocity and directed against the velocity vector, the equation will be:

m d2x/dt2 + b dx/dt + hx = 0, or

d*x/dt* + 1 dx/dt
7
+ WAX= 0, (4)

where b is coefficient of viscosity, 7 = m/b, and o, = ik/m . If such a


system is oscillating in the absence of an external force, the frequency
of oscillations will be defined by w1 = o. - (b2/4m2)(f = w1/27r),
while the decrement in the amplitude of oscillations will be defined by
7. Eq. (4) suggests a way of measuring stiffness by observing the
behavior of the system, i.e. by analyzing the system’s free oscillations.
Note, however, that this method is also based on a number of
assumptions that are generally false for many biological objects. First,
the spring should be unidimensional and massless. Second, inertia
should be concentrated at one end of the object. Third, values of the
coefficients k, b, and m should be time-independent which means
that they should also be length- and velocity-independent.

1.2. Stiffness and quasi-stiffness

Measurement of stiffness of even the simplest of biological objects


is a tricky business because of the major problems emphasized earlier,
in particular time-dependence of the coefficients k, b, and m. A
number of more subtle confounding problems will be discussed later.
Therefore, Eq. (2) cannot be directly applied. Let us write a more
general equation for a hypothetical unidimensional system including
inertia, a damping element, and a spring:

F(t)=m(t) d2x/dt2+b(t) dx/dt+k(t)x(t), (5)

where m(t) is inertia, b(t) is viscosity, k(t > is stiffness, x is length, and
t is time. A number of researchers have studied the dependencies of
F upon x, and termed the derivative dF/dx stiffness. Let us differen-
tiate both sides of Eq. (5) by t and then divide both sides by dx/dt.
M.L. Latash, V.M. Zatsiorsky / Joint stiffness 657

Let us also use V = dx/d t for velocity, A = d2x/d t 2 for acceleration,


and J = d3x/d t 3 for jerk:

dF/dx = dm/dt * A/V+m(t)* J/V+ db/dt + b(t)* A/V

+ dk/dt * x/V+ k(t). (6)

This equation is quite different from Eq. (2). Its right side is not just
k(t) but a complicated expression. Therefore, using the term ‘stiff-
ness’ for the derivative dF/dx is misleading even if m, b, and k are
not time dependent. This derivative is dependent not only upon time
changes in m, b, and k, but also upon movement kinematics (x, V, A,
and J). That is, it rej7ects both feutures of the system and the method of
testing. In an ideal mechanical system, high inertia and/or viscosity
may lead to high values of d F/dx independently of k(t), and even
when there is no elastic element at all, if the measurements are
performed not at equilibria.
Since dF/dx is frequently used in the literature for describing the
properties of biological objects, let us denote it with a special term
‘quasi-stiffness’, q( t >. By definition:

q(t) = dF/dx. (7)

Quasi-stiffness is an ability of the system to resist externally imposed


displacements disregarding the time course of the displacement. This
ability is not necessarily related to the ability of the system to deform
or to store elastic energy. Note that there is a term in mechanics,
‘mechanical impedance’ that reflects properties of a system deter-
mined by its inertial, viscous, and elastic elements. However,
impedance is defined as the ratio of force to particle velocity (En-
cyclopaedia of Science and Technology, 1977) and, therefore, is differ-
ent from quasi-stiffness.

1.3. Two simple physical examples

(1) If a time-dependent force is acting on a small heavy ball


(m = const.), and there is no friction or any other forces (Fig. lA),
movement of the ball can be described by:

F(t) = m *d2x/dt2. (8)


6.58 M.L. Latash, VM. Zatsiorsky / Joint stiffness

m
Fig. 1. Two simple physical examples when a derivative dF/dx can be calculated but the noiton
of stiffness is inapplicable. More detail in text.

Formally, one may calculate a function

q(t) = dF/dx = m * J/V. (9)

which satisfies the definition for quasi-stiffness but apparently has


nothing to do with stiffness. Note that comparing functions q(t) from
different studies should not be expected to make much sense since
these functions reflect the experimental procedure as much as the
object.
(2) Let us consider movements of a free pendulum (a heavy ball on
an ideal weightless thin metal rod) in the field of gravity (Fig. 1 B).
There will be changes in the vector of force F(t) with which the rod
acts on the ball and in the angle a(t) between the rod and direction of
the gravity. Generally, nothing prevents one from taking a derivative
dF/da. However, apparently it cannot be considered stiffness of the
pendulum since no elastic energy is being stored and no deformation
is taking place.

1.4. The importance qf elasticity

Elastic deformation rather than displacement is a crucial character-


istic of mechanical systems that allows one to use the notion of
stiffness. Such deformation is related to storage of elastic energy, so if
a displacement does not lead to a change in elastic energy, the notion
of stiffness should not be invoked. For example, when an active
muscle is being stretched by an external force, its resistance can be
due to both elastic and non-elastic forces. When the resistance is due
to elastic forces, (a) potential energy of deformation increases and (b)
the muscle does not spend metabolic energy to resist the imposed
M.L. Latash, VM. Zutsiorsky / Joint stiffness 659

displacement. When non-elastic forces provide resistance to an im-


posed muscle stretch, (a> the muscle spends metabolic energy for
resistance (which is referred to as negative muscle work) and (b) this
energy is not stored but dissipated.

1.5. The terminology

Let us suggest the following terms for the derivative dF/dx de-
pending upon the physical nature of the system and method of
measurement:
(1) Stiffness (k). The measurements are performed at equilibria. Re-
sistance to the external force is provided by elastic forces, and
potential energy is being stored.
(2) Apparent stiffness (K). The measurements are performed at equi-
libria. The physical nature of the resistive forces is being disre-
garded.
(3) Quasi-stiffness (q). The measurements are performed not at equi-
libria.
Further, we are going to consider systems with lumped parameters
including inertia, viscosity, stiffness, and external forces. That is, in
each case, a simplified mechanical model of the actual system can be
drawn that includes an explicit spring or, more frequently, several
springs. Stiffness will be considered a characteristic of these springs
rather than of the whole system. As we are going to show, the
presence of spring-like components does not assure that the system as
a whole can be attributed a characteristic termed stiffness. If the
models are realistic or too simplistic is another question which is
beyond the scope of the present paper.
It may seem that the main theme of this paper deals mostly with the
problems of terminology rather than with those of motor control or
biomechanics. However, the border is not that strict. First, inappropri-
ate usage of a term borrowed from physics commonly leads to misun-
derstanding if the reader skips the small print in the Methods section.
In extreme cases, it leads to expressions like ‘negative stiffness’
(Dyhre-Poulsen et al., 1991) that do not make physical sense. Second,
a number of contemporary models of motor control, in particular
different versions of the Equilibrium-Point (EP-1 hypothesis, are based
on central regulation of properties of the joints that include joint
stiffness (Feldman, 1980, 1986; Bizzi, 1980; Bizzi et al., 1982, 1992;
660 ML. Lurash, V.M Zutsiortvky / Joint stif@exs

Latash and Gottlieb, 1991a; Latash, 1992). Therefore, understanding


what is ‘joint stiffness’ becomes a necessary step in critical assessment
of these models. Third, recent studies of postural control and locomo-
tion have been revealing the importance of ‘passive’ elasticity (see the
next two sections) placing greater burden on stiffness of tendinous
structures in these types of motor activity.

2. Stiffness of tendons and muscles

Measurement of mechanical properties (stiffness, viscosity,


impedance, etc.) of passive biological tissues (as well as passive joints),
albeit not easy, does not meet with conceptual difficulties. The regular
approaches for describing the corresponding parameters for inanimate
materials have been widely used for this purpose.

2. I. Tendon stiffness

The mechanical properties of tendons (isolated as well as whole


tendons including intramuscular portion) have been described in
several extensive reviews (Yamada, 1970; Crisp, 1972; Alexander,
1981, 1988; Woo, 1986; Proske and Morgan, 1987; Zajac, 19891. The
main function of a tendon is to transmit tensile force. Before the
nineteen-sixties, the role of tendons, as structural components, has
been generally ignored. Implicitly, they have been assumed to be
inextensible and frictionless. Elliot (1965) was one of the first to note
the importance of tendons in storage and recoil of the potential
energy of deformation as well as in shock attenuation,
The macroscopically observed tendon deformation is a function of
two factors:
(a) the geometrical arrangement of tendon fibers which is described
as a network, a waiving pattern, a rope-like twisting or a woven
mesh; and
(b) elastic properties of the fibers.
A tendon force-deformation curve is concave and has two charac-
teristic regions. During the early stages of stretching, tendons are
rather compliant. They are deformed due to straightening out wavi-
ness in the collagen fibrils. The initial region of the tendon force-de-
formation curve (‘toe’) is influenced by geometrical reorientation of
M.L. Lutush, V.M. Zutsiorsky / Joint stij~tms Ml

the tendon fibers rather than their elongation. When a greater tensile
force is applied, the elastic fibers are forcibly stretched, the tendon
stiffness increases and is kept almost constant. The estimates of the
tangential modulus of the stress-strain curve are in the range of
800-1500 MPa for different human and wallaby muscles (Ker et al.,
1986).
In most of the more recent studies, tendon is considered to be
elastic or viscoelastic with a linear elasticity (Ducati et al., 1982;
Oguztoreli and Stein, 1982, 1983). Morgan (1977) has demonstrated
that apparent tendon stiffness is independent of muscle length and
tension induced by electrical stimulation of the motor nerve. Further-
more, Proske and Morgan (1984) have later shown that tendon stiff-
ness remains the same when only parts of the muscle are activated.
On the other hand, changes in tendon stiffness with an increase in
the applied tensile force have been reported by many authors. Rack
and Westbury (1984) stimulated efferent fibers in such a way that
there were virtually no changes in muscle fiber length and all the
movement was supposed to happen in the tendinous component.
Tendon stiffness changed significantly (from 2 to 25 N/mm) with an
increase of muscle force. An increase in tendon stiffness with force
was also shown by Diamant et al. (1972), Ker (19811, Bennett et al.
(19861, and Shadwick (1990).
Note the following mechanical characteristics of tendons:
(a) Energy dissipation in tendons, determined by the area inside the
hysteresis loop curves, is small (Alexander, 1988). It means that
almost all the deformation energy stored in the tendons during
lengthening returns to the system afterwards. Tendons are good
accumulators of the mechanical energy.
(b) The normal physiological range of load lies within the ‘elastic toe’
of the force-deformation curve and is less than 4% of strain which
is considered a limit of tendon reversibility (Harris et al., 1966;
Crisp, 1972).
(cl Above the ‘elastic toe’ range, the tendon stiffness can be regarded
as constant.

2.2. Stiffness of passive muscles

When a passive muscle is at resting length or less, it is rather


compliant. Actin and myosin filaments can move past each other with
662 M.L. Latush, KM. Zutsiorsky / Joint stif@ess

little resistance, and the connective tissue surrounding contractile


elements is in a slack state with no tension. At resting length, stiffness
of human elbow flexors was reported to be about 600 N/m (Ralston
et al., 1947). The stiffness of a relaxed muscle is about 100 times less
than the tendon stiffness (Arndt, 1976). Consequently, during relaxed
passive movements, the tendons are practically not deformed. The
passive muscle tension has been thought to reflect properties of the
connective interfibrillar tissue (Borg and Caulfield, 1980; Alnaqeeb et
al., 1984) although it has also been suggested to be due to myofibrillar
elasticity (Magid and Law, 1985).
When length of a passive muscle exceeds the resting length, the
resistance is provided by the connective tissue (parallel elastic compo-
nents, PEC, according to the well-known Hill model). By definition,
PEC are responsible for muscle stiffness when contractile components
do not generate force.

2.3. Stiffness of actille muscles

Measurement of muscle stiffness when muscles are active is consid-


erably harder than measuring stiffness of passive elements. First, the
length of an active muscle is not directly prescribed by the level of its
activation. It is also a function of external resistance. Second, muscle
reactions are time dependent. At least three time characteristics are
important: (a> time of mechanical disturbance with respect to an
initial stimulus triggering muscle activity; (b) duration (velocity) of
mechanical disturbance; and cc> time after the mechanical perturba-
tion. Third, a muscle is composed of many elements that have differ-
ent mechanical characteristics.

2.3.1. Stiffness of muscle fibers


In studies of muscle fibers, small stretches or releases are applied
to one end of the fiber. A wealth of important scientific knowledge is
collected with this method (for review see Woledge et al., 1985;
Brenner, 1990). Note, however, that in the framework of this direction
of research, the very term ‘stiffness’ is used differently from its
common usage in physics. According to Woledge et al. (1985) ‘The
word stiffness means the instantaneous (!> dependence of tension (P>
on length (1). . . The stiffness of a muscle fiber can be measured by
very rapidly (!) changing its length while recording the tension’ (ex-
M.L. Lutush, KM. Zatsiorsky / Joint stif$ess 663

clamation marks are added by the authors; compare with Eq. (6) and
the epigraph by Wilkie, 19791. This definition corresponds to quasi-
stiffness as it has been defined earlier. Consequently, caution should
be exercised when data from such muscle fiber studies are used to
describe the mechanical behavior of more complex objects, including
joints.
When a muscle fiber is stimulated, its stiffness (or quasi-stiffness) is
proportional to the overlap between the thick and thin filaments and
changes with time together with fiber tension. In addition, fiber
tension drops to zero when the fiber is permitted to shorten quickly
only 6 nm per half-sarcomere (this distance is short compared to the
size of the crossbridge). Therefore, it has been concluded that in
active muscles, actin and myosin filaments are restricted in their
relative motion and, also, that fiber stiffness results from deformation
within the attached crossbridges. The compliance of the filaments
themselves is very low and, in many cases, may be disregarded (cf.
Jung et al., 1988, 1989).
Stiffness of activated extrafusal muscle fibers increases markedly
(Hoffer and Andreassen, 1981; Proske and Morgan, 1984) to the
extent that it may exceed the stiffness of the serial tendinous compo-
nent (Morgan, 1977; Morgan et al., 1978).
Muscle fiber stiffness is frequently assumed to reside in cross-
bridges and to lead to purely elastic storage of energy. It is assumed to
increase with muscle force (Huxley, 1974; Morgan et al., 1978; Julian
and Sollins, 1975; Walmsley and Proske, 1981; Proske and Morgan,
1984; Rack and Westbury, 1984; Sinkjaer et al., 1988). Based on a
crossbridge model, Morgan (1977) predicted a linear relation between
short-range muscle stiffness and muscle tension. He also experimen-
tally observed such a relation by applying small fast stretches to an
isometrically contracted cat soleus muscle. Short-range stiffness was
found to be dependent upon force but not upon operating length. A
similar conclusion has been reached by Hoffer and Andreassen (1981).

2.3.2. Stiffness of a whole acticated muscle ’


According to the well-known classical Hill model (1938, 1949, 1953,
1960, 19701, muscle stiffness is determined by its parallel (PEC) as
well as series (SEC) elastic components. PEC is much more compliant

’ In these experiments, the level of muscle activation is maintained or assumed constant


than SEC (Jewel1 and Wilkie, 1958) and, in studies of active muscles,
SEC is typically the main object of interest.
Studies in this area (unlike others!) are usually performed taking
into account the physical nature of the studied phenomena, in particu-
lar invoking the notion of stiffness (elasticity) only when storage and
recoil of potential energy of deformation takes place. When contribu-
tion of parallel elasticity is ignored, muscle deformation is, at a first
approximation, regarded as a combination of telescopic sliding of the
thick and thin filaments past each other and, in addition, elastic
length changes. The term ‘stiffness’ is used to characterize only the
elastic component of deformation (where deformation energy is stored)
but not the telescopic motion albeit this latter component changes the
muscle force-length curve and its AF/Ax values.
In this framework, muscle stiffness is synonymous to SEC stiffness.
SEC stiffness can be determined when:
(a> values of F and x are measured at least at two points of the
force-length curve, and
(b) length of contractile components (CCs> is not changed during the
entire measurement period.
This can be achieved when SEC is permitted to shorten or is
forcibly lengthened. At least two well-known methods are based on
applying releases to one end of a muscle: ‘quick release’ (Wilkie,
1956a; Jewel1 and Wilkie, 1958; Bahler, 1967) and ‘controlled release’
(Hill, 1950, 1953; Huxley and Simmons, 1970). In the first case, a load
applied to a muscle is decreased in a step-like manner, and in the
second, a stimulated muscle is permitted to shorten over a preset
small distance. Fast changes in muscle length (in the quick release
method) or in muscle tension (during the controlled release proce-
dure) are attributed to changes of SEC only. The CC, by assumption,
does not change its length and tension during and immediately after
rapid releases.
In a method suggested by Wilkie (1950), SEC stiffness is deter-
mined from the course of force development. During the force devel-
opment, SEC component is subjected to stretching and CC compo-
nent is shortening. If the CC force-velocity curve is known, SEC
stiffness can be calculated. The CC force-velocity curve is assumed to
be: (a) the same as the muscle force-velocity curve and (b) identical in
isometric and isotonic conditions.
M. L. Latash, V.M. Zatsiorsky / Joint st(ffiw.s.s 665

These methods have been broadly used to determine elastic charac-


teristics of active human muscles (Fenn and Garvey, 1934; Pertuzon,
1968, 1972; Goubel and Pertuzon, 1973; Goubel, 1974; Hof and Van
den Berg, 1981). The SEC force-deformation curve is found to be
non-linear; SEC stiffness increases at high forces (Hill, 1950; McPher-
son, 1953; Reichel et al., 1956). The SEC force-deformation curve can
be described with an exponential equation (Sandow, 1954).
Several methods were elaborated to distinguish between the contri-
bution of muscle fibers and tendinous structures into SEC stiffness. In
the alpha method (Morgan, 19771, the muscle-tendon complex is
modeled by two serially connected springs. The first spring, with a
constant stiffness, represents the tendon, and the second one, with a
stiffness proportional to muscle tension, represents muscle fiber with-
out tendinous structures. During quick length changes, the stiffness of
tendons as well as fiber stiffness can be estimated. The method was
recommended for force levels above 20% of maximal isometric ten-
sion, at which the tendon stiffness is out of the ‘toe’ region and can be
regarded as almost constant. The spindle-null method (Rack and
Westbury, 1984) is based on applying small sinusoidal stretches with
simultaneous sinusoidal modulation of muscle activity. Discharges
from muscle spindles are registered. When muscle spindles are silent
(the null-point), fiber length is assumed not to change, and all the
length changes are attributable to the tendon.
All the methods for measuring series elastic properties of human
muscle are indirect. We agree with Hof (1990) that these methods rely
on a number of unproved assumptions and are not very accurate. The
situation is even worse when vibration methods based on Eq. (4) are
applied. In principle, Eq. (4) is valid only for linear systems with
lumped parameters, in which elastic (spring) and viscous (dashpot)
elements are acting in parallel (Fig. 2A).
Hill’s three- and two-element models are based on the idea that
contractile (telescopic) and elastic components are connected serially
rather than in parallel. If the changes in CC force and/or length take
place during measurement, the behavior of the system shown in Fig.
2B cannot be described by Eq. (4). When force/length changes of the
CC during the measurement period are neglected, the ‘stiffness’
values can formally be determined, however conceptually they are not
identical to muscle stiffness measured, for instance, by the quick
release method. The stiffness values measured with the vibration
A

Fig. 2. Examples of mass-spring mechanical models. A: A mass-spring system with a viscous


darning element. The damping element and the spring are acting in parallel. This system can be
described with Eq. (4). B: A three-element muscle model (Maxwell model). A viscous element is
also assumed but not shown. SEC elasticity (a spring) and CC are connected in series. This
model has two degrees of freedom which reside in CC and SEC. Eq. (4) cannot be applied to this
model if movement (deformation) in CC is permitted.

method are rather different for various modifications of this method,


for instance, for the free (damped) oscillation and the forced oscilla-
tion methods (Aruin et al., 1979; Zatsiorsky et al., 1981; Zatsiorsky
and Aruin, 1984)
Vibration methods for measuring muscle stiffness are based on
assumptions that:

(a> level of muscle activation is constant during the entire period of


measurements;
(b) parallel stiffness does not contribute; and
(c) CC does not change its length and tension during the measure-
ments (or these changes are known and incorporated into the
equations).

The last assumption looks dubious. In addition, contrary to the


three-element Hill model, vibration methods assume that SEC are
damped and the damping is viscous (proportional to velocity). Con-
trary to that, for such methods as, for instance ‘quick release’, the
absence of damping is essential. Consequently, stiffness measurements
based on two different models presented in Fig. 2A and 2B may
provide different results.
Zajac (1989) has shown that elastic energy storage in muscles with
long tendons resides in tendon and not in muscle. Even for extremely
stiff muscle-tendon complexes, a sizable proportion of the total energy
is being stored in the tendons. Therefore, the next step is to consider
the muscle-tendon complex.

2.4. Stiffness of the muscle-tendon complex

Most of the investigators use the classical model of muscle devel-


oped by Hill (1938) and Wilkie (1956a). Hill (1938, 1949) recognized
the importance of tendon elasticity in the development of muscle
force, however, experimental studies of these effects have been initi-
ated relatively recently (Morgan et al., 1978; Rack et al., 1983; Rack
and Westbury, 1984; Ker et al., 1986; Proske and Morgan, 1987). In
particular, the force-length and force-velocity properties of the muscle
are being frequently considered analogous to those properties of
muscle fibers and of sarcomeres (for review see Zajac, 1989).
Energy stored in the external and internal parts of the tendon is
much bigger than that stored in the crossbridges (Alexander and
Bennet-Clark, 1977; Rack et al., 1983; Rack and Westbury, 1984).
Thus, for many muscles, tendon compliance dominates and the series
elastic element can be ignored (note that in some models, the series
elastic element is assumed to include tendon compliance). Dynamics
of muscle contraction is, therefore, largely defined by tendon and
muscle aponeurosis elasticity and dynamics of muscle activation may
no longer be rate-limiting (Zajac, 1989).
Discrepancies between changes in muscle length and spindle length
during walking in cats, including shortening of spindles during length-
ening of the muscle have been reported by Hoffer et al. (1989). The
authors attributed these discrepancies to both muscle architecture and
compliance of the long tendinous element in series with the spindles.
That is, length changes on muscle spindles are not necessarily simply
related to changes in parental muscle length as assessed by changes in
joint angles. For human thumb movements, it has been shown that
only a portion of an imposed movement reaches the muscle fibers
(Rack and Ross, 1984) so that length of the muscle fibers does not
determine joint angle in any positive sense. A similar result has been
reported for the ankle joint (Rack et al., 1983). Compliant tendons
may play an important role in control of gripping and handling
movements. The assumption that changes in muscle fiber length
mimic changes in length of the muscle-tendon complex may be invalid
for many of the muscles during many natural movements (Hof et al.,
1983; Bobbert et al., 1986) despite the fact that most movements are
performed at submaximal forces. Force-length muscle relations should
be considered very carefully because even when length of the musclc-
tendon complex is held constant, changes in muscle force are likely to
lead to changes in length of muscle fibers (cf. Hill, 1938; Ritchie and
Wilkie, 1958; Zajac, 1989).
Because of the compliant tendon in series, during ‘isometric’ con-
tractions, there is a shortening of muscle fibers of medial gastrocne-
mius muscle in the cat by as much as 18%-28% (Griffiths, 1991). The
tendons are also a major contributor to ‘creep’, i.e. a slow rise of
tension during ‘isometric’ contractions at muscle lengths much longer
than the optimal muscle length. Muscle fibers can actually shorten as
a result of imposed slow stretches while the stretch is fully absorbed
by the tendon. The same result was observed in walking cats during
the stance phase of the walking cycle. These observations suggest that
tendons act as mechanical buffers protecting muscle fibers from
damage during eccentric contractions. The role of tendinous compo-
nent is likely to be even more important in biarticular muscles where
muscle fibers typically occupy only a fraction of the total length
(Griffiths, 1984; Walmsley and Proske, 1981).
Storage of elastic energy in the muscle-tendon complex is assumed
to be an important mechanism during locomotion and some other
tasks (Hof et al., 1983; Bobbert et al., 1986; Ker et al., 1986 MacMa-
hon, 1984; Bennett et al., 1986; Alexander et al., 1982). Kangaroos are
consuming less oxygen than quadruped mammals of the same weight
running at the same speed. This is thought to be achieved by storage
of elastic energy in tendons that are longer and more compliant than
those of cats (Alexander and Vernon, 1975). Ankle extensor muscle
fibers are generally much shorter than the tendons, with the ratio of
1: 5 for the cat (Walmsley and Proske, 1981) and 1: 7 for the wallaby
medial gastrocnemius (Griffiths, 1984). In the camel plantaris, the
muscle fibers are vestigial so that the entire length is being taken by
the tendon (Alexander et al., 1982). This can be only if muscle fibers,
that are in series with the tendon, are very stiff. When a wallaby
muscle was developing close to its maximum isometric tension, up to
eight times as much movement occurred in the tendon as in the
M.L. Latash, KM. Zatsiorsky / Joint stiffness 669

muscle fibers (Morgan et al., 1978). With a decrease in tendon length,


a larger proportion of movement occurred in the muscle fibers, and
there was a steep rise in work absorption by the muscle leading to an
energy loss. Alexander and Bennet-Clark (1977) have shown that
wallaby’s Achilles tendon can store up to ten times as much energy as
the muscle.
During locomotion, muscle fibers can actually shorten at the early
stance phase after the hind feet are placed in the ground while the
ankle extensor muscles stretch (Griffiths, 1984, 1989). This is due to
high compliance of the tendons at low tensions. So, muscle stretch
during locomotion can be accompanied by shortening of muscle fibers
and spindles.
All these examples illustrate that, even without central reflex con-
nections, the muscle-tendon complex cannot be considered a single
spring and, therefore, cannot be assigned a value of stiffness. Depend-
ing on the type of the muscle, level of muscle activation, operating
length of the muscle-tendon complex, and parameters of external
perturbation, resistance to the perturbation will receive different
contributions from the tendon serial compliance and compliance of
muscle fibers. However, we are going to commit this common crime
(cf. Hasan, 1992) and assume that there is such a thing as ‘peripheral
spring’ whose stiffness characterizes the reaction of the muscle-tendon
complex to externally imposed length changes in the absence of reflex
action.

3. Joint stiffness

Most of the voluntary movements are controlled by a number of


muscles and represent rotations in joints. Therefore, analysis of
single-joint movements becomes an important step towards analysis of
control of natural movements. Unfortunately for the experimenters,
most of the commonly studied joints of human limbs (e.g., shoulder,
wrist, and ankle) have more than one degree-of-freedom and are
controlled by more than two muscles some of which are bi-articular,
i.e. their contractions directly induce changes in joint torque and/or
movements in two adjacent joints. As a result, single-joint movement
becomes a fiction or an experimental artefact. We however think that
this fiction is an important intermediate step in analysis of voluntary
670 M.L. L&ash, KM. Zatsiorsky / Joint stiffness

movements representing a paradigm for testing hypotheses which may,


in future, appear helpful for analysis of more complex and more
natural movements.

3.1. Stiffness of passive joints

Passive joint stiffness can be determined when all muscles crossing


the joint are relaxed. In experiments, subjects are asked to relax the
muscles and ‘not to intervene voluntarily’. The absence of muscle
activity is verified by electromyogram (EMG). When a relaxed body
link is passively moved by an external force, no muscle electric activity
is seen in a broad range of joint angles (Hoefer and Putnam, 1939;
Buchtal and Clemmesen, 1946; Buchtal, 19571.
At rest, the summed distance between the points of insertion of
antagonistic muscles is somewhat longer than the summed resting
length of these muscles (Gurfinkel et al., 1965). Therefore, muscles
acting around a joint are usually slightly stretched and generate a
certain tension. A balance of passive forces from antagonistic muscles
determines the so-called neutral position in a joint. To determine the
joint neutral position, the relaxed limb is placed horizontally on a
special manipulandum (McKinley and Berkwitz, 1928, 1933; Broman,
1949) or subjects, while being submerged into water, are asked to
relax, and joint angles are measured from underwater photographs
(Lehman, 1940).
Neutral position is about 90” for the elbow joint (McKinley and
Berkwitz, 1928, 1933; Broman, 19491 and about 133” and 134” for the
knee and hip joints (Lehman, 1940). If one of the muscle groups
crossing the joint is impaired, the joint neutral position shifts towards
the healthy side (Berkwitz, 1932). When both muscle groups are
impaired (Parkinson’s disease, hemiplegia), the neutral joint position
is not changed.
Because many of the joints are also crossed by biarticular muscles,
the neutral positions in adjacent joints are interrelated. For instance,
a knee joint neutral position changes linearly when the hip joint
position is changed from 180” to 60” (Shabashova, 1947). The follow-
ing empirical formula is valid: knee neutral angle = 60.4” + 0.57p
(where p is hip angle).
Both direct and indirect methods of measuring passive joint stiff-
ness have been used in the literature. Direct measurements, in which
M.L. Latash, V.M. Zatsiorsky / Joint stiffness 671

external load is applied to a body link and the joint angle is measured,
have been used since 1896 (Mosso, 1896; Mosso and Benedicenti,
1896; Rieger, 1906; Reijs, 1921; Filimonoff, 1925; Herzog, 1927;
Spiegel, 1929; Werestchagin, 1931). Vibration measurements were
pioneered by Schalterbrand (1929, 1935, 1937, 1940) and Maikov
(1930) and, then, were used by others (Joyce et al., 1974; Gottlieb and
Agarwal, 1978; MacKay, 1984; MacKay et al., 1986). The underlying
assumption of the method is that the system under consideration can
be described as a linear system with lumped parameters and Eq. (4) is
valid. Passive joint stiffness values, measured in a narrow range of
joint motion, close to the neutral joint position, are usually very low.
They increase sharply near the limits of the joint range of motion.
However, the stretched muscles are typically activated in this range via
the stretch-reflex mechanism, and the joint cannot be considered
passive anymore. Joint flexibility is also influenced by factors different
from stiffness and should be studied as a special phenomenon
(Zatsiorsky, 1977; Hutton, 1992).
Stiffness of relaxed human joints measured from the frequency
response is under 4 Nm/rad for the elbow (Zahalak and Heyman,
1979; Kwan et al., 1979; MacKay et al., 19861, about 1.2 Nm/rad for
the wrist (Lakie et al., 19791, and about 19 Nm/rad for the ankle
(Gottlieb and Agarwal, 1978; Kearney and Hunter, 1982). Note,
however, that in many of these studies, the influence of viscosity upon
natural frequency of the joint was disregarded. Lower values of
passive elbow stiffness of about 1 Nm/rad were reported by Feldman
(1979) based on experiments with smooth unloadings followed by the
joint torque and angle measurements at new equilibria.
Generally, passive joint stiffness:

(a) is higher at a small range of limb displacement (short-range


stiffness) (Lakie et al., 1981; Walsh and Wright, 1982; MacKay et
al., 1986; Kearney and Hunter, 1982).
(b) depends on the initial joint position, in particular increases with
elbow flexion (Lestienne and Pertuzon, 1974; MacKay et al., 1986)
as well as ankle dorsiflexion (Gottlieb and Agarwal, 1978);
Cc) decreases over a period of 1 s due to viscous relaxation (Kwan et
al., 1979; MacKay et al., 1986; although, this is rather a feature of
quasi-stiffness, according to our terminology introduced in Section
1.5);
672 h4.L. Lafash, V.M. Zatsiorsky / Joint stiffkss

(d) depends on the direction of joint motion (hysteresis is evidently


seen on the torque-angle curves (Fig. 4 in Hof and Van den Berg,
1981; Figs. 3-5 in Yoon and Mansour, 1982).

3.2. Stiffness of an isolated joint

Let us consider a pin joint with only one degree of freedom. Eq. (3)
will undergo minor transformations:

d2a( t) Wt)
T(t) = m(t)7 + “(t)dt + W)[4t) - %Wl~
where T is torque around the joint, (Y is joint angle, and a<, is resting
angle of the joint. This equation is analyzed with different degrees of
simplifications in most of the mass-spring models of single-joint motor
behavior. Experimental analysis of systems described by Eq. (6) looks
relatively straightforward: An experimenter tries to control time
changes of one of the two measurable variables, torque T(t) or angle
a(t), measures time changes of the other variable, and, with a suffi-
cient number of measurements, is able to calculate time changes of
the coefficients m(t), b(t), and k(t) with some degree of accuracy.
There are numerous ways of performing such experiments including
application of singular small torque perturbations at certain times and
measuring length changes (Ma and Zahalak, 1985; Gottlieb et al.,
1986), using sinusoidal perturbations over a range of frequencies
(Joyce et al., 1974; A garwal and Gottlieb, 1977; Rack et al., 1978;
Cannon and Zahalak, 19821, or using randomized perturbations with
subsequent correlation of torque and angle changes (Hunter and
Kearney, 1982; Lacquaniti et al., 1982; Weiss et al., 1988; Bennett et
al., 1989, 1992).
Attempts at direct application of Eq. (6) for modeling and analysis
of voluntary joint movements lead to considerable complications if
one tries to incorporate at least minimal information about muscles
and their central connections. The first of these problems is that Eq.
(6) models a damped loaded spring with an instantaneous reaction to
external load. Thus, this approach ignores reflex time delays that are
likely to play an important role in movements (especially, fast ones).
Another problem is due to at least two different mechanisms giving
rise to viscous behavior of intact muscles (Feldman et al., 1990). First,
M.L. Latush, KM. Zatsiorsky /Joint stiffness 673

there is a well-known classical relation between muscle force and


speed of its length changes (Hill’s equation, Hill, 1938, 1953):

where T, and b, are constants.


Besides that, dynamic sensitivity of muscle spindle afferents makes
the tonic stretch reflex sensitive not only to muscle length but also to
speed of its changes. One may accept or reject the h-version of the
equilibrium-point hypothesis (Feldman, 1966, 1986) which considers
the tonic stretch reflex mechanism the basis of the single-joint motor
control. However, velocity-dependent autogenic reflex effects of mus-
cle spindles have never been seriously questioned. Formally, they also
imply presence of a viscous component in muscle reaction to load (or
length) changes. Let us consider these problems in detail.

3.3. Stiffness of a joint with muscle reflexes

What will happen if length of an intact activated muscle is rapidly


changed by some external factor ? One can single out at least three
components in the muscle reaction:

(1) A muscle will oppose the length change by an increase in its force
due to purely mechanical factors, as if it were a rubber band.
(2) There may be a distinct short-latency monosynaptic reaction lead-
ing to a phasic muscle contraction also opposing the external force
but short in duration (Loop 1 in Fig. 3).
(3) There will be a tonic increase in muscle force due to a hypotheti-
cal length-sensitive mechanism commonly associated with the no-
tion of tonic stretch reflex (Loop 2 in Fig. 3).
The lower part of Fig. 3 illustrates all three components as separate
springs with different m, b, and k. Now, imagine that a rapid
‘step-like’ perturbation is applied at a time t, leading to a virtually
instantaneous increase in external torque (load). Muscle length will
start to change. During the first several tens of milliseconds, this
process will depend only on the first factor, i.e. muscle reaction as a
rubber band. After a time delay corresponding to the latency of
monosynaptic reaction for the muscle (t,), there will be a burst of
614 M.L. Latash, KM. Zatsiorsky / Joint stiffness

Force
*

Fig. 3. A schematic drawing of three components that can contribute to spring-like behavior of
an intact muscle (and, consequently, to an intact joint). Inertial and viscous elements are
assumed but not shown. There are time delays in both monosynaptic reflex (Loop 1) and
polysynaptic tonic stretch reflex (Loop 2) arcs.

activity leading to a short phasic contraction also resisting the pertur-


bation. Then, the time will come for the longer-latency tonic reflexes
which will lead to development of tonic muscle contraction also
resisting the perturbation. Let us assume that the latency of the tonic
stretch reflex is t,.
So, if a perturbation is applied at t,, the kinematics will be defined
by different factors at different times from the moment of perturba-
tion. A measurement at a time t (t, < t < t,) would only supply
information about the muscle’s ‘rubber-band’ reaction and is equiva-
lent to testing a deafferented muscle (cf. ‘intrinsic stiffness’ measured
by Sinkjaer et al., 1988). A measurement at a time t’ (t, <t’ < t2)
would also supply information on the monosynaptic reactions (which
is, by the way, not easy to extract!), i.e. on the muscle with intact
monosynaptic reflex connections but without the hypothetical longer-
latency tonic reflex connections. And eventually, at a time t” (t” > t,>,
measurement would supply information relevant to all three compo-
nents. Apparently, if an experimenter applies a perturbation at t, and
measures muscle length immediately after the perturbation, muscle
M.L. Latash, VM. Zatsiorsky / Joint stiffness 615

length changes will reflect properties of only the first of the three
mechanisms providing for spring muscle properties.
It seems absolutely necessary to modify Eq. (8) in order to take into
account the three mechanisms contributing to muscle reaction to
external perturbations. The major problem is not in the presence of
three parallel springs illustrated in Fig. 3, but in their different time
delays. It is also very important to realize what is the cause and what
is the consequence in muscle reactions. If there is an external change
in torque around a joint, it will cause a muscle length change due to
the first, purely peripheral mechanism:

d*a(t) d4)
T(t) - m,,(t)7 -b,,(t)dt = k”(t)bw - a”41 3 (9)

where (Y, corresponds to ‘zero’ angle for the first, peripheral spring.
We will use (Y, and (Ye for ‘zero’ angles for the two other springs. Note
that the viscous component b,(t)(da(t)/d t> in Eq. (9) may be consid-
ered equivalent to Hill’s viscosity (see above).
Changes in muscle length will cause monosynaptic reactions, and
torque changes become a consequence! Taking into account the delay
t,, one gets:

d*a(t + t,) da(t + t,)


T(t+t,)-m,(t+t,) dt2 -b@+b) dt

= h(t)[4) - 41. (10)


Changes in muscle length will also provide for the tonic stretch
reflex leading to a tonic torque change. The following equation
describes behavior of this third spring:

d2a(t + t2) da(t + t2)


T(t + t2) -m2(t + t2) &2 - b2(f + t2) dt

= kZWC4f) - %@)I. (11)


Now we are ready to substitute Eq. (8) with another one taking into
676 M.L. Lutash, KM. Zatsiorsky /Joint st(ffness

account all three components of muscle reaction to perturbation. It


seems fair to suppose that m,(t) = m,(t) = m,(t) = m(t). Then,

d2a( t) dff(t) d2a( t + t,)


T(t) -m(t)7 -‘&)F + T(t + 6) -dt +b) dt2

da(t + t,) d%( t + tJ


-h@+b) dt +T(t+t,)-m(t+t,) dt2

da( t + t2)
-b2(t+t2) dt = k,(t)[4t) - %Wl + W)

x [4t) - 4)l + ~2ww - 4f)L

where external, inertial, and viscous torque components of the three


springs are superimposed.
This equation illustrates the complexity of such a system and how
hard it is to analyze its behavior without introducing simplifications.
Obviously, one cannot perform such analysis by directly correlating
values of a(t) and T(t). Eq. (12) is non-linear with time-dependent
coefficients and therefore cannot be directly analyzed with traditional
methods for analyzing stiffness, e.g. by driving the system sinusoidally
and determining its natural frequency (cf. Joyce et al., 1974; Agarwal
and Gottheb, 1977; Rack et al., 1978; Cannon and Zahalak, 1982).

3.4. Stiffness of an intact joint under central control

Until now, we have not considered which variables are specified by


the central control system during control of voluntary movements. Let
us now become more specific and, according to our own preferences,
turn to the equilibrium-point (EP-) hypothesis, according to which
motor control is based on central regulation of parameters of the
human tonic stretch reflex. Eq. (11) represents action of this loop and
therefore, the components in Eq. (32) with the argument (t + t2> in no
circumstances can be ignored since such a simplification would mean
ignoring the basic mechanism of control. From this view, Eq. (8)
M.L. Latash, V.M. Zabrsky / Jomr sriffnm 677

seems absolutely inadequate for studying the processes of motor


control, although it may be useful for analyzing issues dealing with
mechanical properties of the peripheral apparatus under different
states of muscle activation.
Regulation of parameters of the tonic stretch reflex arc can lead to
two basic effects upon the joint compliant characteristic (JCC), a
translation parallel to the angle axis and a change in the slope. Two
variables have been introduced corresponding to these two effects, Y
and c (Feldman, 1980; Feldman and Latash, 1982). Changes in the
reciprocal variable Y lead to the JCC translations along the angle axis
while changes in the coactivation variable c induce changes in the
JCC slope. This can be formally expressed as:

where f is a monotonically decreasing function. Central control can


be described with time functions r(t) and c(t) that lead to translations
and/or rotations of the JCC. The movement result from disparity
between current joint torque and angle values and equilibrium values
corresponding to a new position of the JCC.
Getting back to Eq. (121, note that central control by specifying r(t)
and c(t) is analogous to defining time changes of a,(t) and k,(t). It
seems to be proper time to confess that Eq. (12) is also wrong. Note
that the latency of the TSR is a sum of three components: transmis-
sion in the afferent path, central delay, and transmission in the
efferent path. It is virtually impossible to assess the central delay time.
For the present purposes, let us assume that this time is zero. Then,
fat.i = feff = flat/2 where t,Z,t is the total latency of the TSR.
Central control signals are supposedly supplied to a central proces-
sor. If at a time t,, instantaneous position of the JCC is described by
central variables c, and Y,, the generated efferent signal will reflect
position of the JCC and an afferent volley that was in fact generated
t,,[/2 ago. When the efferent signal will come to the muscles, it will
convey an outdated information reflecting position of the JCC that
was centrally generated t,;,/2 ago and afferent information that was
generated by the peripheral receptors tlat ago.
Eq. (12) takes into account the hypothetical delay in the TSR arc t,
(analogous to flat). It, h owever, fails to reflect the fact that there is
also a delay between generation of an afferent volley and its interac-
678 M.L. Latash, KM. Zatsiorsky /Joint stiffness

tion with the central command which occurs when the afferent signal
comes to the hypothetical central processing unit. This factor can
easily be introduced by changing the argument of cu,(t> and k,(t)
functions:

d2Cr(t) Wt) d2a( t + t,)


T(t) -M(t)7 - b”(t)dt + T(t + b) -m(t + h) dt2

da(t + tJ d2a(t + t2)


-htt + b> dt +T(t+t2)-m(t+t2) dt2

da( t + t2)
-b2(t+t2) dt = W)b(t) - a,(t)1 + W)

The methods for analysis of Eq. (14) have certainly been based on a
number of assumptions and simplifications. It is important, however,
to make sure that these simplifications are explicitly stated, their
possible effects on the quality of the results are assessed, and they do
not eliminate the basic features of the hypothetical motor control
process.

3.5. Experimental analysis of joint stiffness

Let us remind two major factors that make traditional methods


used for analysis of stiffness and viscosity in physics inapplicable to
analysis of intact muscles and joints:

(1) Muscles have central reflex connections that provide for delayed
spring-like reactions.
(2) Stiffness and viscosity in each of the hypothetical loops are likely
to be length-, velocity-, and time-dependent.

For example, assessments of joint stiffness and viscosity based on


resonant frequency (Viviani et al., 1976; Cannon and Zahalak, 1982;
M. L. Larash, V.M. Zafsiorsky / Joint stiffness 679

Hunter and Kearney, 1982; MacKay et al., 1986) or on changes in the


kinematics following a perturbation in a time interval considerably
longer than the reflex delays (MacKay et al., 1986; Weiss et al., 1988;
Blanpied and Smidt, 1992) ignore both factors. All these studies
measured quasi-stiffness, i.e. a function q(t) = dF/dt (see Eq. (5)
above) that reflects both properties of the joints and muscles and the
experimental procedure.
Blanpied and Smidt (1992) studied ‘intrinsic stiffness’ of human
plantarflexors. They measured slope of the force-angle dependence
during the first 62 ms after an externally imposed relatively rapid joint
dorsiflexion. Note that during the first 62 ms, a large proportion of the
total force was matched by the inertial forces accelerating the foot
which is obvious from Fig. 2 of Blanpied and Smidt (19921, and also
maybe by viscous forces. Therefore, the ratio dF/dcu cannot be
considered stiffness, and this is one more example of measuring
quasi-stiffness, 4( t >.
Viviani et al. (1976) attempted to assess average stiffness and
viscosity of the elbow flexor and extensor muscle groups. They studied
phase relations between EMG and inertial torque at different fre-
quencies of sinusoidal joint voluntary movement. EMG was consid-
ered a measure of central command equivalent to total muscle force.
Although EMG does correlate well with muscle force in static condi-
tions (Bigland and Lippold, 1954; Crago et al., 1976; Buchanan et al.,
19861, this relation breaks down during movements (Cord0 and Rymer,
1982; Yoneda et al., 1983). Also, equating EMG with central com-
mand ignores possible reflex contribution to the observed patterns of
muscle activation (cf. Abdusamatov and Feldman, 1986; Feldman et
al., 1990; Latash and Gottheb, 1991b). Joint stiffness was assessed
based on the resonant frequency of the movements. In the study by
Viviani et al., average stiffness was considered constant for move-
ments at different frequencies. The authors state that ‘ . . . the data
presented here were obtained under open-loop conditions.. . ’ (p. 55)
without pointing out how the reflex contribution could be eliminated
in voluntary movements of healthy human subjects.
Characteristic values of peripheral (or intrinsic) joint stiffness corre-
sponding to spring 1 in Fig. 2 were measured in a number of studies
when the analysis window was chosen so that not to include possible
reflex-mediated muscle reactions to external perturbations (Ma and
Zahalak, 1985; Bennett et al., 1989; Sinkjaer et al., 1988; MacKay et
al., 1986). Reflex contribution to stiffness has been shown by Sinkjaer
et al. (1988) to start approximately 50 ms after the perturbation.
Analysis during the first 50 ms has shown a nearly linear increase in
ankle joint stiffness with muscle force from approximately 10 Nm/rad
to over 120 Nm/rad. Higher values of stiffness were seen for smaller
perturbations that induced joint movements of up to 1” (see also
MacKay et al., 1986).
Note also that muscles behave like low-pass filters (for review see
Zajac, 1989). As a result, reflex delay times t, and t, (Fig. 3) may be
different for different time patterns of force perturbation and/or
limb displacement. In particular, experimental procedures with fast
external perturbations are likely to provide data reflecting both prop-
erties of the system and parameters of the applied perturbations.
Performing measurements only at equilibria or using slow changes in
external force allow one to attenuate or even avoid this problem.
Experimental analysis of reflex-mediated joint stiffness (k2 in Eq.
(14)) and possible influence of central control signals started with
so-called ‘static’ experiments (Asatryan and Feldman, 1965; Feldman,
1966). In these experiments, the subjects were instructed to maintain a
position in a joint against an external load. The subject received an
instruction ‘not to intervene voluntarily’, and it was assumed that he
did not change the voluntary motor command when changes in the
external load were introduced. These changes led to joint movements.
Joint angle and torque were measured after the limb reached a new
equilibrium state and plotted as points on a torque-angle plane.
Interpolation of the points was considered a JCC. Characteristic
values of stiffness in such experiments were about 10 Nm/rad.
Similar experiments were later performed by other researchers (Davis
and Kelso, 1982; Vincken et al., 1983; Gottlieb and Agarwal, 1988;
Latash and Gottlieb, 1990; De Serre and Milner, 1991). If the subjects
are asked to co-contract their flexor and extensor muscles in the initial
position and ‘not-to-intervene-voluntarily’, a similar loading-unload-
ing procedure leads to more steep JCC corresponding to higher
reflex-mediated joint stiffness (Davis and Kelso, 1982). A similar
increase in joint stiffness was observed by De Serre and Milner (1991)
when the subjects acted against an unstable load which probably
forced them to use higher levels of muscle coactivation. In all these
experiments, the experimenters assumed and relied upon the subject’s
ability to fix a voluntary motor command.
A4.L. Lutush, PX Zutsicmky / Joint stiffitess h81

Note that JCC reconstructed with such a method represents summed


action of two springs, the peripheral one and the TSR one (Fig. 31.
Action of the monosynaptic reflexes is phasic, i.e. present only during
the process of changes in muscle length. Since the measurements were
performed at equilibria, the contribution of this spring can be ignored.
Recently, a method has been introduced for reconstructing the
patterns of joint stiffness changes during voluntary movements (Latash
and Gottlieb, 1991a). The method is based on comparing the changes
in joint torque and kinematics during relatively slow changes in the
external conditions of the movement execution, i.e. loadings and
unloadings. It relies upon the ability of the subjects to reproduce the
same motor command (‘to do the same’) when the external conditions
change. Time shifts of JCCs enable reconstruction of time changes in
joint stiffness and equilibrium position. In particular, during slow
elbow flexion movements, joint stiffness did not demonstrate visible
changes. During fast single-joint movements, joint stiffness had a
transient peak in the middle of the movement (an increase from
approximately 5 Nm/ rad to over 50 Nm/ rad) and a slightly increased
steady level after the movement termination. These observations are
in a good correspondence with optimization modeling by Lan and
Crago (1992).

4. Stiffness of multi-joint systems


The difficulties and ambiguities in trying to use the concept of
stiffness for individual joints suggest that making another step up to
the more complex multi-joint systems is likely to make this concept
even less applicable (although this seems hardly possible). All the
problems originating from non-unidimensionality, presence of multi-
ple serial and parallel springs with different time delays, and time-de-
pendence of the coefficients of stiffness and viscosity continue to play
their confounding role. New problems are likely to emerge, in particu-
lar, due to the presence of bi-articular muscles. The complexity of the
multi-joint systems suggests that, generally, they cannot be described
with one parameter related to deformation and storage of elastic
energy, i.e. stiffness. Theoretically, in certain special experimental
conditions, a multi-joint system can behave like a spring that can be
adequately described with a single stiffness parameter. However, we
have failed to find examples of such precisely controlled studies.
682 M.L. Latash, KM. Zatsiorsky / Joint st#ness

MacMahon and Greene (1979; Greene and MacMahon, 1979)


measured ground reaction forces (F) during running and knee bend-
ing and compared them with changes in ‘leg length’ (L) defined as the
distance from the hip to the ankle joint. ‘Leg stiffness’ was calculated
as a ratio dF/dL (cf. Eq. (51) thus allowing the authors to solve a
number of theoretical and applied problems including optimization of
stiffness of the running surface. Since maximal force at take-off
corresponds to maximal acceleration of the body center of mass, its
peak was found close to the time of maximal flexion of the knee joint.
It was minimal at the time when the leg was nearly completely
extended. However, if one measures maximal force production in leg
extension, the relation would be the opposite, i.e. the shorter the leg
the lower the force (for a review see Kulig et al., 1984). Peak isometric
forces are exerted when the leg is almost completely extended. In
order to account for these two findings, a notion of negutiue stiffness
was introduced. Partially, this situation occurred because the inertial
component was not taken into account in the aforementioned studies,
and q(t) rather than k(t) was calculated. 2
Another example of ignoring inertial and viscous components and
calculating ‘ankle joint stiffness’ during different phases of hopping is
presented by a study by Dyhre-Poulsen et al. (1991). Their method
also assessed not joint stiffness but its quasi-stiffness, q(t). Not
surprisingly, negative values of muscle stiffness in the ankle joint have
been reported just after touch down when landing, and positive value
were observed during the take-off. These findings can be compared to
negative values of muscle ‘stiffness’ measured during muscle stretches
at high velocities (Joyce et al., 1969; Rack, 1981).

5. Concluding remarks

(1) The notion of stiffness has been introduced in physics to


characterize properties of certain types of deformable bodies under an
influence of external forces. In the absence of external forces, these

’ Recently, MacMahon (1990) suggested to differentiate two stiffnesses during the support
phase in locomotion, one due to knee bending, and the other due to leg rotation. Leg rotation
can be compared to an inverted pendulum rotation. Therefore, there will be changes of the ratio
of vertical component of ground reaction force to vertical displacement of the body center of
gravity (cf. Fig. l).However, the notion of stiffness can hardly be applied to such phenomena.
M.Ld. Latash, VM. Zatsiorslcy /Joint stiffness 683

bodies are supposed to maintain constant shape. Muscles are not such
bodies, and joints are complex system consisting of several bodies.
(2) The following terms for the derivative dF/dx are suggested:
(a) Stiffness (k). The measurements are performed at equilibria.
Resistance to the external force is provided by elastic forces, and
potential energy is being stored.
(b) Apparent stiffness (K). The measurements are performed at
equilibria. The physical nature of the resistive forces is being disre-
garded.
(c) Quasi-stiffness (4). The measurements are performed not at
equilibria.
Apparent stiffness (K) and quasi-stiffness (4) are easier to measure.
However, these indices do not reflect participation of elastic forces
and accumulation of elastic energy. In particular, 9 can easily be
negative while genuine stiffness is always positive. We suggest that
studies involving experimental procedures that lead to calculation of K
or 4 should clearly state the difference between these measures and
the physical notion of stiffness.
(3) Stiffness measurement of passive biological tissues (tendons,
muscles, and passive joints), albeit not easy, does not meet with
conceptual difficulties.
(4) In studies of active muscle fibers, a wealth of important scien-
tific knowledge has been collected with methods of small stretches
and releases. However, these methods measure quasi-stiffness (or, at
best, apparent stiffness) of muscle fibers. Consequently, caution should
be exercised when data from these studies are used to describe the
mechanical behavior of more complex objects, joints included.
(5) Muscle stiffness is determined, according to the well-known Hill
model, by its parallel (PEC) as well as series (SEC) elastic compo-
nents. Vibration methods for measuring muscle stiffness in humans
are based on several assumptions which are in contradiction with the
Hill’s model. Consequently, ‘stiffness’ measurements based on two
different models provide different results.
(6) The muscle-tendon complex cannot be considered a single
spring and, therefore, cannot be assigned a value of stiffness. In
general, elastic energy storage in muscle-tendon complexes resides
primarily in tendons rather than in muscles. However, depending on
the type of the muscle, level of muscle activation, operating length of
the muscle-tendon complex, and parameters of external perturbation,
resistance to the perturbation will receive different contributions from
the tendon serial compliance and compliance of muscle fibers.
(7) Passive joint stiffness is determined when all muscles crossing
the joint are relaxed. Passive joint stiffness: (a) is higher at a small
range of limb displacement (short-range stiffness); (b) depends on the
initial joint position; cc> decreases over a period of 1 s due to viscous
relaxation (although, this comment refers to frequently measured
quasi-stiffness); (d) depends on the direction of joint motion.
(8) An intact joint is a complex, non-linear system whose angle
changes under a changing external force cannot be described with just
one parameter. One can probably take into account the inertial
component with a reasonable degree of accuracy. The presence of two
components in the velocity-dependent force generation by the muscle
(see above), one of which has a reflex time delay, makes analysis of
viscous forces quite complicated. The same is true for the elastic
components of muscle reaction. Therefore, we feel that experimental
analysis is very likely to lead to assessments of joint stiffness and
viscosity dependent upon both properties of the studied system and
particular method used for making these assessments. Therefore, we
think that intact human joint cannot be assigned a parameter termed
‘stiffness’.
(9) This does not mean, however, that stiffness (and viscosity) of
joint subsystems and subcomponents cannot be studied. Considerable
success in the studies of passive structures suggests that at least some
of the joint subcomponents can be assigned values of stiffness (and
viscosity) that comply with the strict definitions introduced in physics.
We also feel optimistic towards the introduced method of study of
stiffness of the tonic stretch reflex subcomponent that contributes to
the overall joint behavior and is assumed to play a major role in the
process of single-joint motor control. We believe that all the other
subcomponents should be studied separately from each other by
designing special experimental manipulations that are able to separate
them according, for example, to their characteristic time constants.
(10) We suggest to either abandon the term ‘joint stiffness’ as
misleading or to state upfront stiffness of which of the joint compo-
nents or subsystems is analyzed in each particular study. We also
suggest that each study of ‘joint stiffness’ should clearly state to what
extent the results are defined by the system’s properties and to what
extent they are reflections of the particular experimental procedure.
M.L. Latash, KM. Zatsiorsky / Joint stiffness 685

Because of the ambiguities related to the notion of joint stiffness,


stiffness of multi-joint intact systems (e.g., limbs) seems even more
impossible to define and measure.

Acknowledgments

The authors express their gratitude to Dr. Ziaul Hasan for his
profound and helpful comments. Preparation of this manuscript was
in part supported by an NIH grant HD 30128.

References

Abdusamatov, R.M. and A.G. Feldman, 1986. Description of the electromyograms with the aid
of a mathematical model for single joint movements. Biophysics 31, 549-552.
Agatwal, G.C. and G.L. Gottlieb, 1977. Compliance of the human ankle joint. Journal of
Biomechal Engineering 99, 166-170.
Alexander, R.McN., 1981. ‘Mechanics of skeleton and tendons’. In: B.B. Brooks (ed.), Handbook
of physiology. Section a. The nervous system. Vol. 2 Motor control (pp. 17-42). Bethesda,
MD: American Physiological Society.
Alexander, R.McN., 1988. Elastic mechanisms in animal movement. Cambridge: Cambridge
University Press.
Alexander, R.McN. and H.C. BennettClark, 1977. Storage of elastic strain energy in muscle and
other tissue. Nature 265, 114-l 17.
Alexander, R.McN. and A. Vernon, 1975. Mechanics of hopping by kangaroos (Macropodidae).
Journal of Zoology (London) 177, 265-303.
Alexander, R.McN., G.M.O. Maloiy, R.F. Ker, AS. Jayes and C.N. Warui, 1982. The role of
tendon elasticity in the locomotion of the camel (Camelus dromedarius). Journal of Zoology
(London) 198, 2933313.
Alnaqeeb, M.A., N.S. Al Zaid and G. Goldspink, 1984. Connective tissue changes and physical
properties of developing and ageing skeletal muscle. Journal of Anatomy 139, 6777689.
Arndt, K.H., 1976. Achillesehnenruptur and Sport. Leipzig: Johann Ambrosius Barth.
Aruin. A.S., V.M. Zatsiorsky, L.M. Raitsin, N.I. Volkov and E.A. Shirkovets, 1977. Influence of
elastic forces of the muscles upon the efficiency of muscle work. Human Physiology 3,
420-426.
Aruin, AS., V.M. Zatsiorsky, G.J. Panovko and L.M. Raitsin, 1979. Equivalent biomechanical
characteristics of the ankle joint muscles. Human Physiology 4, 862-868.
Asatryan, D.G. and A.G. Feldman, 1965. Functional tuning of the nervous system with control of
movements or maintenance of a steady posture. I. Mechanographic analysis of the work of
the limb on execution of a postural task. Biophysics 10, 925-935.
Bahler, A.S., 1967. Series elastic component of mammalian skeletal muscle. American Journal of
Physiology 213, 1560-1564.
Bennett, M.B., R.F. Ker, N.J. Dimery and R.McN. Alexander, 1986. Mechanical properties of
various mammalian tendons. Journal of Zoology (London) 209, 5377548.
686 M.L. Latash, VM. Zatsiorsky / Joint stiffness

Bennett, D.J., Y. Xu, J.M. Hollerbach and I.W. Hunter, 1989. Identifying the mechanical
impedance of the elbow joint during posture and movement. Abstracts of the Society for
Neuroscience 15, 396.
Bennett, D.J., J.M. Hollerbach, Y. Xu and I.W. Hunter. 1992. Time-varying stiffness of human
elbow joint during cyclic voluntary movement. Experimental Brain Research 88, 433-442.
Berkwitz, NJ., 1932. Quantitative studies on the human muscle tonus. II. An analysis of
eighty-two normal and pathological cases. Archives of Neurology and Psychiatry 28, 603-614.
Bernstein, N.A., 1947. On the construction of movements. Moscow: Medgiz (in Russian).
Bernstein, N.A., 1967. The co-ordination and regulation of movements. Oxford: Pergamon Press.
Bigland, B. and 0. Lippold, 1954. The relation between force, velocity and integrated electrical
activity in human muscles. Journal of Physiology 123, 214-224.
Bizzi, E., 1980. ‘Central and peripheral mechanisms in motor control’. In: G.E. Stelmach and J.
Requin (eds.), Tutorials in motor behavior (pp. 131-143). Amsterdam: North-Holland.
Bizzi, E., N. Accornero, W. Chapple and N. Hogan, 1982. Arm trajectory formation in monkeys.
Experimental Brain Research 46, 139-143.
Bizzi, E., N. Hogan, F.A. Mussa-Ivaldi and S. Giszter. 1992. Does the nervous system use
equilibrium-point control to guide single and multiple joint movements? Behavioral and
Brain Sciences 15, 603-613.
Blanpied, P. and G.L. Smidt, 1992. Human plantarflexor stiffness to multiple single-stretch trials.
Journal of Biomechanics 25, 29-39.
Bobbert, M.F.. P.A. Huijing and G.J. van Ingen Schenau, 1986. A model of the human triceps
surae muscle-tendon complex applied to jumping. Journal of Biomechanics 19, 887-898.
Borg, T.K. and J.B. Caulfield, 1980. Morphology of connective tissue in skeletal muscle. Tissue
and Cell 12, 197-207.
Brenner, B., 1990. ‘Muscle mechanics and biochemical kinetics’. In: J.M. Squire (ed.), Molecular
mechanics in muscular contraction (pp. 77-150). Boca Raton, FL: CRC Press.
Broman, T., 1949. Electra-mechanographic registrations of passive movements in normal and
pathological subjects. Acta Psychiatrica et Neurologica Suppl. 53, 1.
Buchanan, T.S., D.P.J. Almdale. J.L. Lewis and W.Z. Rymer. 1986. Characteristics of synergetic
relations during isometric contractions of human elbow muscles. Journal of Neurophysiology
56, 1225-1241.
Buchtal, F., 1957. Introduction to electromyography. Kobenhavn: Stand. University.
Buchtal, F. and S. Clemmesen, 1946. Action potentials in pathological postural reflex activity
(spacticity and rigidity). Acta Psychiatrica et Neurologica 21, 151-162.
Cannon, S.C. and G.I. Zahalak, 1982. The mechanical behavior of active human skeletal muscle
in small oscillations. Journal of Biomechanics 15, 111-121.
Cordo, P.J. and W.Z. Rymer, 1982. Motor unit activation patterns in lengthening and isometric
contractions of hindlimb extensor muscles in the decerebrate cat. Journal of Neurophysiology
47, 782-796.
Crago, P.E., J.C. Houk and Z. Hasan, 1976. Regulatory actions of human stretch reflex. Journal
of Neurophysiology 39, 925-935.
Crisp, J.D.C., 1972. ‘Properties of tendon and skin’. In: Y.C. Fung. N. Perrone and M. Anliker
(eds.). Biomechanics. Its foundations and objectives (pp. 141&180). Englewood Cliffs, NJ:
Prentice-Hall.
Davis, W.R. and J.A.S. Kelso, 1982. Analysis of ‘invariant characteristics’ in the motor control of
Down’s syndrome and normal subjects. Journal of Motor Behavior 14, 194-212.
De Serre, S.J. and T.E. Milner. 1991. Wrist muscle activation patterns and stiffness associated
with stable and unstable mechanical loads. Experimental Brain Research 86. 451-458.
Diamant, J., A. Keller, E. Baer, M. Litt and R.G.C. Arridge, 1972. Collagen: Ultrastructurre and
its relation to mechanical properties as a function of ageing. Proceedings of the Royal Society
B 180, 239-315.
M.L. Latash, V.M. Zatsiorsky / Joint stiffness 687

Ducati, A., F. Parmiggiani and M. Scieppati, 1982. Simulation of post-tetanic potentiation and
fatigue in muscle using a visco-elastic model. Biological Cybernetics 44, 129-135.
Dyhre-Paulsen, P., E.B. Simonsen and M. Voigt, 1991. Dynamic control of muscle stiffness and
H reflex modulation during hopping and jumping in man. Journal of Physiology 437,
287-304.
Elliot, D.H., 1965. Structure and function of mammalian tendon. Biological Reviews 40,
3922421.
Encyclopaedia of Science and Technology, 1977. Impedance, Mechanical, 7 (p. 44). New York:
McGraw-Hill.
Feldman, A.G., 1966. Functional tuning of the nervous system with control of movement or
maintenance of a steady posture. II. Controllable parameters of the muscle. Biophysics 11,
565-578.
Feldman, A.G., 1979. Central and reflex mechanisms of motor control. Moscow: Nauka (in
Russian).
Feldman, A.G., 1980. Superposition of motor programs. I. Rhythmic forearm movements in
man. Neuroscience 5, 81-90.
Feldman, A.G., 1986. Once more on the equilibrium-point hypothesis (I model) for motor
control. Journal of Motor Behavior 18, 17-54.
Feldman, A.G., S.V. Adamovitch, D.J. Ostty and J.R. Flanagan, 1990. ‘The origin of electromyo-
grams - Explanations based on the equilibrium point hypothesis’. In: J.M. Winters and
S.L.-Y. Woo teds.), Multiple muscle systems. Biomecuanics and movement organization (pp.
195-213). Berlin: Springer-Verlag.
Feldman, A.G. and M.L. Latash, 1982. Afferent and efferent components of joint position sense:
Interpretation of kinaesthetic illusions. Biological Cybernetics 42, 2055214.
Fenn, W.O. and P.H. Carvey, 1934. The measurement of the elasticity and viscosity of skeletal
muscle in normal and pathological cases; a study of so-called ‘muscle tonus’. Journal of
Clinical Investigations 19, 383-397.
Filimonoff, J.N., 1925. Klinische Beitrage zum Tonusproblem. Zeitschrift fur die gesamte
Neurologie und Psychiatric 96, 368-384.
Gleim, G., N. Stachefeld and J. Nicholas, 1990. The influence of flexibility on the economy of
walking and jogging. Journal of Orthopaedic Research 8, 814-823.
Gottlieb, G.L. and G.C. Agarwal, 1978. Dependence of human ankle compliance on joint angle.
Journal of Biomechanics 11, 177-181.
Gottlieb, G.L. and G.C. Agarwal, 1988. Compliance of single joints: Elastic and plastic charac-
teristics. Journal of Neurophysiology 59, 937-951.
Gottlieb, G.L., G.C. Agarwal and A.S. Tanavde, 1986. A method of estimating human joint
compliance which is insensitive to reflex, triggered or voluntary reactions. Abstracts of the
Society for Neuroscience 12, 468.
Goubel, F., 1974. Les proprietes mechaniques du muscle au tours du mouvement sous-maximal
(p. 217). These Doctorat d’Etat, Fat. Sci. Lille, France.
Goubel, F. and E. Pertuzon, 1973. Evaluation de I’elasticite du muscle in situ par une methode
de quick-release. Archives Internationales de Physiologie et Biochimie 81, 697-701.
Greene, P.R. and T.A. MacMahon, 1979. Reflex stiffness of man’s antigravity muscles during
kneebends while carrying extra weights. Journal of Biomechanics 12, 881-891.
Griffiths, R.I., 1984. Mechanical properties of an ankle extensor muscle in a freely hopping
wallaby. Ph.D. Thesis. Monash Univ. Victoria, Australia.
Griffiths, R.I., 1987. Ultrasound transit time gives direct measurement of muscle fibre length in
vivo. Journal of Neuroscience Methods 21, 159-165.
Griffiths, R.I., 1991. Shortening of muscle fibres during stretch of the active cat medial
gastrocnemius muscle: The role of tendon compliance. Journal of Physiology 436, 219-236.
688 M.L. Latash, V.M. Zatsiorsky / Joint stiffness

Gurfinkel, VS., Y.M. Kots and M.L. Schick, 1965. Control of body posture. Nauka, Moscow (in
Russian).
Hasan, Z., 1992. Is stiffness the mainspring of posture and movement? Behavioral and Brain
Sciences 15, 7566758.
Hasan, Z. and R.M. Enoka, 1985. Isometric torque-angle relationship and movement-related
activity of human elbow flexors: Implications for the equilibrium-point hypothesis. Experi-
mental Brain Research 59, 441-450.
Heerkens, Y.F., R.D. Woittiez, P.A. Huijing, A. Huson and G.J. van Ingen Schenau, 1981.
Inter-individual differences in passive resistance of the human knee. Human Movement
Science 4, 167-188.
Herzog, F., 1927. Uber die myotatische Innervation antagonistischer Muskeln. Deutsche
Zeitschrift fin die Nervenheilkunde 96, 22-37.
Hill. A.V., 1938. The heat of shortening and the dynamic constants of muscle. Proceedings of the
Royal Society of London B 126, 1366195.
Hill, A.V., 1949. The heat of activation and the heat of shortening in a muscle twitch.
Proceedings of the Royal Society of London B 136, 195-211.
Hill, A.V., 1950. The series elastic components of muscle. Proceedings of the Royal Society of
London B 137, 2733280.
Hill, A.V., 1953. The mechanics of active muscle. Proceedings of the Royal Society of London B
141, 104-117.
Hill, A.V., 1960. Production and absorbtion of work by muscle. Science 131, 897-903.
Hill, A.V.. 1970. First and last experiments in muscle mechanics. Cambridge: Cambridge
University Press.
Hoefer, P.F.A. and J.J. Putnam, 1939. Action potentials of muscles in normal subjects. Archives
of Neurology and Psychiatry 42, 201-210.
Hof, A.L., 1990. ‘Effects of muscle elasticity in walking and running’. In: J.M. Winters and
S.L.-Y. Woo teds.), Multiple muscle systems: Biomechanics and movement organization (pp.
591-607). Berlin: Springer-Verlag.
Hof. A.L. and J.W. van den Berg, 1981. EMG to force processing. II. Estimation of parameters
of the Hill muscle model for the human triceps surae by means of a calf ergometer. Journal
of Biomechanics 14. 759-770.
Hof, A.L.. B.A. Geelan and J. van den Berg, 1983. Calf muscle movement. work and efficiency in
level walking; roles of series elasticity. Journal of Biomechanics 16, 5233537.
Hoffer, J.A. and S. Andreassen, 1981. Regulation of soleus muscle stiffness in premammillary
cats: Intrinsic and reflex components. Journal of Neurophysiology 45, 267-285.
Hoffer, J.A., A.A. Caputi, I.E. Pose and RI. Griffiths, 1989. Roles of muscle activity and load on
the relationship between muscle spindle length and whole muscle length in the freely walking
cat. Progress in Brain Research 80, 75085.
Hogan, N., 1990. ‘Mechanical impedance of single- and multi-articular systems’. In: J.M. Winters
and S.L.-Y. Woo teds.), Multiple muscle systems: Biomechanics and movement organization
(pp. 149-164). Berlin: Springer-Verlag.
Huijing, P.A., 1992. ‘Elastic potential of muscle’. In: P.V. Komi ted.), Strength and power in
sport (pp. 151-168). Oxford: Blackwell.
Hunter, I.W. and R.E. Kearney, 1982. Dynamics of human ankle stiffness: Variation with mean
ankle torque. Journal of Biomechanics 15, 747-752.
Hutton, R.S., 1992. ‘Neuromuscular basis of stretching exercises’. In: P.V. Komi (ed.1, Strength
and power in sport. The encyclopaedia of sports medicine (pp. 29-38). Oxford: Blackwell.
Huxley, A.F., 1974. Review lecture: muscular contraction. Journal of Physiology 243, l-43.
Huxley, A.F. and R.M. Simmons, 1970. Rapid ‘give’ and the tension ‘shoulder’ in the relaxation
of frog muscle fibres. Journal of Physiology 210, 32233.
M.L. L&ash, V.M. Zatsiorsky / Joint stiffness 689

Jewell, B.R. and D.R. Wilkie, 1958. An analysis of the mechanical components of frog’s striated
muscle. Journal of Physiology 143, 515-540.
Joyce, G.C., P.M.H. Rack and D.R. Westbury, 1969. The mechanical properties of cat soleus
muscle during controlled lengthening and shortening movements Journal of Physiology 204,
461-474.
Joyce, G.C., P.M.H. Rack and H.F. Ross, 1974. The forces generated in the human elbow joint
in response to imposed sinusoidal movements of the forearm. Journal of Physiology 240,
351-374.
Julian, F.J. and M.R. Sollins, 1975. Variation of muscle stiffness with force at increasing speeds
of shortening. Journal of General Physiology 66, 287-302.
Jung, D.W.G., T. Blangi, H. de Graaf and B.W. Treijtel, 1988. Elastic properties of relaxed,
activated, and rigor muscle fibers measured with microsecond resolution. Biophysical Journal
54, 897-908.
Jung, D.W.G., T. Blange, H. de Graaf, and B.W. Treijtel, 1989. Weakly attached cross-bridges in
relaxed frog muscle fibers. Biophysical Journal 55, 6055619.
Kearney, R.E. and I.W. Hunter, 1982. Dynamics of human ankle stiffness: Variation with
displacement amplitude. Journal of Biomechanics 15, 753-756.
Ker, R.F., 1981. Dynamic tensile properties of the plantaris tendon of sheep (Ovis aries). Journal
of Experimental Biology 93, 283-302.
Ker, R.F., N.J. Dimery and R.McN. Alexander, 1986. The role of tendon elasticity in hopping in
a wallaby (Macropus rufogriseus). Journal of Zoology (London) 208, 417-428.
Komi, P.V., 1992. ‘Stretch-shortening cycle’. In: P.V. Komi (ed.), Strength and power in sport
(pp. 169-179). Oxford: Blackwell.
Kulig, K., J. Andrews and J.G. Hay, 1984. Human strength curves. Exercise and Sport Science
Reviews 12, 417-466.
Kwan, H.C., J.T. Murphy and M.W. Repeck, 1979. Control of stiffness by the medium latency
electromyographic response to limb perturbation. Canadian Journal of Physiology and
Pharmacology 57, 277-285.
Lacquaniti, F., F. Licata and J.F. Soechting, 1982. The mechanical behavior of the human
forearm in response to transient perturbations. Biological Cybernetics 44, 67-77.
Lakie, M., E.G. Walsh and G. Wright, 1979. Wrist compliance. Journal of Physiology 295,
9x-99P.
Lakie, M., E.G. Walsh and G. Wright, 1981. Measurements of inertia of the hand, and the
stiffness of the forearm muscles using resonant frequency methods, with added inertia or
position feedback. Journal of Physiology 310, 3-4P.
Lan, N. and P.E. Crago, 1992. Equilibrium-point hypothesis, minimum effort control strategy
and the triphasic muscle activation pattern. Behavioral and Brain Science 15, 769-771.
Latash, M.L., 1992. Virtual trajectories, joint stiffness, and changes in natural frequency during
single-joint oscillatory movements. Neuroscience 49, 209-220.
Latash, M.L. and G.L. Gottlieb, 1990. Compliant characteristics of single joints: Preservation of
equifinality with phasic reactions. Biological Cybernetics 62, 331-336.
Latash, M.L. and G.L. Gottlieb, 1991a. Reconstruction of joint compliant characteristics during
fast and slow movements. Neuroscience 43, 697-712.
Latash, M.L. and G.L. Gottlieb, 1991b. An equilibrium-point model of dynamic regulation for
fast single-joint movements: 1. Emergence of strategy-dependent EMG patterns. Journal of
Motor Behavior 23, 163-177.
Lehman, B., 1940. Zur Physiologie des Liegens. Arheitsphysiologie 11, 253-264.
Lestienne, F. and E. Pertuzon, 1974. Determination, in situ, de la visco-elasticit& du muscle
humain inactive. European Journal of Applied Physiology 32, 159-170.
Ma, S.-P. and G.I. Zahalak, 1985. The mechanical response of the active human triceps brachii
muscle to very rapid stretch and shortening. Journal of Biomechanics 18, 585-598.
690 M.L. Latash, KM. Zutsiorsky / Joint stiffness

MacKay, W.A., 1984. Resonance properties of the human elbow. Canadian Journal of Physiol-
ogy and Pharmacology 62, 802-808.
MacKay, W.A., D.J. Crammond, H.C. Kwan and J.T. Murphy, 1986. Measurements of human
forearm viscoelasticity. Journal of Biomechanics 19, 231-238.
McKinley, J.C. and N.I. Berkwitz, 1928. Quantitative studies on human muscle tonus. Archives
of Neurology and Psychiatry 19, 1036-1045.
McKinley, J.C. and N.I. Berkwitz. 1933. Electric action potentials in muscles during recording of
mechanical tonus tracing. Archives of Neurology and Psychiatry 29, 272-280.
MacMahon, T.A., 1984. Muscles. reflexes, and locomotion. Princeton, NJ: Princeton University
Press.
MacMahon, T.A., 1990. ‘Spring-like properties of muscles and reflexes in running’. In: J.M.
Winters and S.L.-Y. Woo (eds.), Multiple muscle systems. Biomechanics and movement
organization (pp. 380-390). Berlin: Springer-Verlag.
MacMahon, T.A. and P.R. Greene, 1979. The influence of track compliance in running. Journal
of Biomechanics 12, 893-904.
McPherson, L., 19.53. A method of determining the force-velocity relation of muscle from two
isometric contractions. Journal of Physiology 122, 172-177.
Magid, A. and D.J. Law, 1985. Myofibrils bear most of the resting tension in frog skeletal
muscle. Science 230, 1280-1282.
Maikov, S.S., 1930. Study of muscular tonus. Reviews in Psychiatry, Neurology and Reflexology
3-4, 44-55 (in Russian).
Morgan, D.L., 1977. Separation of active and passive components of short-range stiffness of
muscle. America1 Journal of Physiology 232, C45-C49.
Morgan, D.L., U. Proske and D. Warren. 1978. Measurements of muscle stiffness and the
mechanism of elastic storage of energy in hopping kangaroos. Journal of Physiology 282.
253-261.
Mosso, A., 1896. Description d’un myotonometre pour Ctudier la tonicite des muscles chez
I’homme. Italian Biological Archive 25, 385.
Mosso, A. and A. Benedicenti. 1896. La tonicite des muscles Ctudiee chez I’homme. Italian
Biological Archive 25, 385.
Oguztoreli, M.N. and R.B. Stein, 1982. Analysis of a model for antagonistic muscles. Biological
Cybernetics 45, 177-184.
Oguztoreli, M.N. and R.B. Stein, 1983. Optimal control of antagonistic muscles. Biological
Cybernetics 48, 91-99.
Pertuzon, E., 1968. Un dispositif pour les experiences de quick release. Le Travail Humaine 31,
303-308.
Pertuzon, E., 1972. La contraction musculaire dans le mouvement volontaire maximal (p. 208).
These Doctorat d’Etat, Fat. Sci., Lille, France.
Piziale. R.L. and J.C. Rastegar, 1977. Measurement of the non-linear, coupled stiffness charac-
teristics of the human knee. Journal of Biomechanics 10, 45-51.
Proske, U. and D.L. Morgan, 1984. Stiffness of cat soleus muscle and tendon during activation of
part of muscle. Journal of Neurophysiology 52, 459-468.
Proske, U. and D.L. Morgan, 1987. Tendon stiffness: Methods of measurement and significance
for the control of movement. A review. Journal of Biomechanics 20, 75-82.
Rack, P.M.H., 1981. ‘Limitations of somatosensory feedback in control of posture and move-
ment’. In: V.B. Brooks (ed.), Handbook of physiology, section 1: The nervous system, vol. 2:
Motor control (pp. 229-256). New York: American Physiological Society.
Rack, P.M.H. and H.F. Ross, 1984. The tendon of flexor pollicis longus: Its effects on the
muscular control of force and position at the human thumb. Journal of Physiology 351,
999110.
M.L. Latash, KM. Zatsiorsky / Joint stiffness 691

Rack, P.M.H. and D.R. Westbury, 1984. Elastic properties of the cat soleus tendon and their
functional importance. Journal of Physiology 347, 479-495.
Rack, P.M.H., H.F. Ross and T.I.H. Brown, 1978. ‘Reflex responses during sinusoidal movement
of human limbs’. In: J.E. Desmedt ted.), Cerebral Motor control in man: Long loop
mechanisms (pp. 216-118). Base]: Karger.
Rack. P.M.H., H.F. Ross, A.F. Thilmann and D.K.W. Walters, 1983. Reflex responses at the
human ankle: The importance of tendon compliance. Journal of Physiology 347, 479-495.
Ralston, H.J., V.T. Inman, L.A. Strait and M.D. Shaffrath, 1947. Mechanics of human isolated
voluntary muscle. American Journal of Physiology 151, 612-620.
Ralston, H.J. and B. Libet, 1953. The question of tonus in skeletal muscle. American Journal of
Physical Medicine 2, 85-91.
Reichel, H., A. Bleichert and F. Zimmer, 1956. Die elastischen Eigenschaften des Skelets and
Herzmuskels. Zeitschrift fir Biologie 108, 188-195.
Reijs, I.H.O., 1921. Uber die Veranderung der Kraft wahrend der Bewegung. Pfliiger’s Archiv
fir die gesamte Physiologie 91, 2344249.
Rieger, C., 1906. Untersuchungen iiber Muskelzustande. Jena
Ritchie, J.M. and D.R. Wilkie, 1958. The dynamics of muscular contraction. Journal of
Physiology 143, 104-124.
Sandow, A., 1954. On the active state of striated muscle. Science 127, 760-769.
Schalterbrand, G., 1929. Messung des Dehnungswiderstandes am menschlichen Muskel bei
Gesunden, Spastikern und Parkinsonismusfallen. Deutsche Zeitschrift fir die Nerven-
heilkunde 109, 231-250.
Schalterbrand, G., 1935. Verhalten der myotatischen Reflexe bei verschiedenen Tonusstorungen.
Deutsche Zeitschrift fir die Nervenheilkunde 136, l-18.
Schalterbrand, Cl., 1937. Myographische Untersuchungen in der Klinik. 1. Geschichte der
Myographie und Beschreibung eines neuen Myographen. Deutsche Zeitschrift fur die Ner-
venheilkunde 142, l-24.
Schalterbrand, G., 1940. Die Muskelspannungsmessung (Myographie) und ihre Bedeutung far
die klinische Diagnostik. Zentralblatt fir die innere Medizin 61, 1-25.
Shabashova, A.S., 1947. Balance of tonus of antagonistic muscle groups. PH.D. thesis, Moscow
(Cited after Gurfinkel, Kots and Shick, 1965).
Shadwick, R.E., 1990. Elastic energy storage in tendons: Mechanical differences related to
function and age. Journal of Applied Physiology 68, 1033-1040.
Sherrington, ChS., 1906. Integrated activity of the nervous system. New Haven, CT: Yalen
University Press.
Sinkjaer, T., E. Toft, S. Andreassen and B.C. Hornemann. 1988. Muscle stiffness in human ankle
dorsiflexors: Intrinsic and reflex components. Journal of Neurophysiology 60, 11 IO- 1121.
Spiegel, E.A., 1929. Zur Methodik der Tonusuntersuchung am Menschen. Zeitschrift fir die
gesamte Neurologie und Psychiatric 122, 475-484.
Such, C.H., A. Unsworth, V. Wright and D. Dowson, 1975. Quantitative study of stiffness in the
knee joint. Annals of Rheumatoid Diseases 34, 286-291.
Tardieu, C., P. Colbeau-Justin, M.D. Bret, A. Lespargot and G. Tardieu, 1981. Effects on
torque-angle curve of differences between therecorded tibia-calcaneal angle and the true
anatomical angle. European Journal of Applied Physiology 46, 41-45.
Vincken. M.H., C.C.A.M. Gielen and J.J. Denier van der Gon, 1983. Intrinsic and afferent
components in apparent muscle stiffness in man. Neuroscience 9, 529-534.
Viviani, P., J.F. Soechting and C.A. Terzuolo, 1976. Influence of mechanical properties on the
relation between EMG activity and torque. Journal of Physiology (Paris) 72, 45-5X.
Walmsley, B. and U. Proske, 1981. Comparison of stiffness of soleus and media1 gastrocnemius
muscles in cats, Journal of Neurophysiology 46, 250-259.
692 M.L. Latash, KM. Zatsiorsky /Joint stiffness

Walsh, E.G. and G.W. Wright, 1982. A hanging-hand tremorograph. Journal of Physiology 329.
I-2P.
Weber, E., 1847. Wagners Handworterbuch der Physiologie 3, 81.
Weiss, P.L., I.W. Hunter and R.E. Kearney, 1988. Human ankle joint stiffness over the full range
of muscle activation levels. Journal of Biomechanics 21, 539-544.
Werestchagin, N., 1931. Untersuchung iiber Musketonus und Ermiidung (Neue Methode zur
Bestimmung des Dehnungswiderstandes des Muskelsl. Pfliiger’s Archiv fiir die gesamte
Physiologie 227, 1X8-201.
Wilkie, D.R., 1950. The relation between the force and velocity in human muscle. Journal of
Physiology 110, 249-280.
Wilkie, D.R., lY56a. The mechanical properties of muscle. British Medical Bulletin 12, 1777182.
Wilkie, D.R., 1956b. Measurement of the series elastic component at various times during a
single muscle twitch. Journal of Physiology 134, 527-530.
Wilkie, D.R., 1979. ‘General discussion’. In: H. Sugi and G.H. Pollach ted.), Cross-bridge
mechanism in muscle contraction (p. 6341. Baltimore: University Park Press.
Woledge, R.C., N.A. Curtin and E. Homsher, 198.5. Energetic aspects of muscle contraction.
Monographs of the Physiological Society No. 41. London: Academic Press.
Woo, S.L.-Y., 1986. ‘Biomechanics of tendons and ligaments’. In: G.W. Schmid-Schonbein,
S.L.-Y. Woo SL-Y and B.W. Zweifach Frontiers in biomechanics (pp. 180-195). Berlin:
Springer-Verlag.
Yamada, H., 1970. Strength of biological materials (Ed. by F.G. Evans). Baltimore: Williams and
Wilkins.
Yoneda, T., K. Oishi and A. Ishida, 1983. Variation of amount of muscle discharges during
ballistic isometric voluntary contraction in man. Brain Research 275, 305-309.
Yoon, Y.S. and J.M. Mansour, 1982. The passive elastic moment at the hip. Journal of
Biomechanics 15, 905-910.
Zahalac, G.I. and S.I. Heyman, 1979. A quantitative evaluation of the frequency-response
characteristics of active human skeletal muscle in viva. Journal of Biomechanical Engineering
101, 28-37.
Zajac, F.E., 1989. Muscle and tendon: Properties, models, scaling, and application to biomechan-
its and motor control. CRC Critical Reviews in Biomedical Engineering 17, 359-411.
Zatsiorsky (Zaciorskij), V.M., 1977. Die kijrperlichen Eigenschaften des Sportlers (3rd ed.1.
Berlin: Bartels u. Wernitz.
Zatsiorsky, V.M., A.S. Aruin, L.M. Raisin and G.J. Panovko, 1981. ‘The determination of the
equivalent biomechanical characteristics of the ankle joint muscles by vibration tests’. In: G.
Bianchi, K.V. Frolov and A. Oledzki teds.), Man under vibration. Suffering and protection.
International CISM-IFToMM-Symposium. Udine, Italy, April 336, 1979. Proceedings, PWN
(pp. 166-175). Amsterdam: Elsevier.
Zatsiorsky, V.M. and A.S. Aruin, 1984. Biomechanical characteristics of the human ankle joint
muscles. European Journal of Applied Physiology 52, 400-406.

View publication stats

Das könnte Ihnen auch gefallen