Gait and Posture 6 (1997) 163 – 170

Review Paper
The development of mature gait

D. Sutherland *
Motion Analysis Laboratory, Childrens Hospital, San Diego, CA 92123 -4282, USA

Received 5 July 1997; received in revised form 4 August 1997; accepted 5 August 1997

Abstract

We must understand the natural history of immature walking in order to define and interpret pathological gait in young
children. Step length, cadence and walking velocity show evidence of both central nervous system maturation and growth until
approximately 4 years of age. Until 4 years-of-age the step factor (step length divided by leg length) increases, but thereafter the
step factor remains constant and adult-like. After 4 years-of-age the changes in velocity, cadence, step length in normal children
are appropriately attributed to changes in limb length. Dynamic joint angle measurements indicate stabilization of gait by 312 to
4 years-of-age. However, there is greater inter-subject and intra-subject variability of gait in children of all ages than of adults.
Force plate measurements of subjects 2 to 7 years-of-age show an increasing definition of the mid-stance trough in the vertical
force curve. The second peak of the vertical force curve is deficient, by adult standards, particularly in the younger age groups.
Moment and power curves of hip, knee and ankle in subjects 4 years and older show patterns not unlike those of adult subjects.
There are differences in magnitude, suggesting that the youngest children use their hip flexor and extensor muscles more than their
ankle plantar flexors for power generation. A study from another center shows no age-related differences in hip, knee and ankle
powers in children 5–16 years of age. All of the studies reported are very helpful in understanding gait maturation but researchers
are needed to address remaining questions. © 1997 Elsevier Science B.V.

Keywords: Pathological gait; Step length; Cadence; Walking velocity

1. Introduction are beginning to come clearly into focus. The question

of what factors control the maturation of gait will be a
We all can agree that the walking patterns of very research focus for physiologists, neuro-scientists, devel-
small children differ from those of adults, but why is opmental pediatricians, physical therapists, and gait
this important and why present childrens’ gait data? laboratory teams for many years ahead. It has only
The answer is that one must understand the normal or been in the past 2 or 3 decades that the tools for
natural history of a studied phenomenon before at- dynamic investigation of gait have become both avail-
tempting to describe and study the pathological or able and sufficiently well developed to permit serious
abnormal. It follows that there are three important studies of gait maturation in children.
questions to ask: (1) How does the gait of children From an observational standpoint, the general se-
differ from that of adults? (2) When do children achieve quence of the early development of human mobility
an adult-like gait pattern? (3) What are the factors following birth is well known. The newborn infant is
controlling the maturation of gait? Much has been helpless and fully dependent. The infant is able to sit
written on this subject, but only questions (1) and (2) upright at about 6 months after birth, begins to crawl
around 9 months, and walks without support at around
* Fax: + 1 619 6147494. 1 year of age [1]. There is evidence that the control

0966-6362/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved.

PII S 0 9 6 6 - 6 3 6 2 ( 9 7 ) 0 0 0 2 9 - 5
164 D. Sutherland / Gait and Posture 6 (1997) 163–170

system is in place, though poorly co-ordinated, well

before upright independent walking and it was discov-
ered that pre-walking, kicking movements closely re-
semble the lower extremity movements during the early
phase of independent walking [2].
To appreciate the changes that occur during the
process of gait maturation, we must examine data and
it will be useful to use the measurement categories
commonly used in gait analysis; time – distance parame-
ters, kinematics, electromyography, kinetics, and en-
ergetics.

2. Time-distance parameters

Walking velocity is the product of step frequency and

step length, or stride frequency and stride length. If
right and left step lengths are approximately equal, as is Fig. 1. Cadence (steps/min) vs. age (n = 420). The lines connect the
25th, 50th, and 75th percentiles. The dots outside the lines indicate
the case with normal children, the choice of step length minimum and maximum. The numbers above the bottom horizontal
and step frequency is appropriate for the calculation of line of the graph show the numbers for each age group represented.
velocity calculations. However, if the right and left Reproduced with permission [12].
steps are unequal, even in normal children, as reported
by Wheelwright et al. [3], stride frequency and stride old children by Sutherland et al. [12]. In the same
length must be used. Alternatively, the velocity must be report, cadence was reported as dropping sharply over
measured over a distance. In our laboratory, we rou- the next 2.5–3 years (Fig. 1). Stride length increased
tinely measure stride frequency and stride length, as rapidly until 4 years of age, then continued to increase
many, if not most, of the patients for pathological at a slower rate (Fig. 2). Velocity increased rapidly, in
studies have unequal step lengths. spite of the decreasing cadence, until 3.5–4 years of
What are the factors controlling step frequency, step age. (Fig. 3). Todd et al. [11] compiled published data
length, and walking velocity in children? What are the from multiple authors on time-distance parameters
relative roles of musculo-skeletal growth and central through full skeletal growth. It is obvious from their
nervous maturation? If maturation of the control sys- report that most of the time-distance parameters are
tem is important, at what age does it cease to be a highly related to musculo-skeletal growth. In other
contributing factor? There are now some answers to
words, we cannot say that step frequency, walking
these questions. To consider the influence of growth
velocity, and step length are fully mature until skeletal
and limb length on time – distance parameters, it is
growth has been completed.
useful to apply the most basic knowledge of pendulum
motion. We know that the lower limbs act as com-
pound pendulums during swing phase, with muscular
contractions controlling the initiation of swing and the
excursion of the shank [4,5]. It is common knowledge
that a pendulum motion is dependent on the length of
the pendulum and the center of mass. Therefore, it
follows that the length of the limb and the location of
the center of mass are the physical properties that,
combined with the action of muscles crossing the joints,
are responsible for the angular rotations and time of
swing.

2.1. Influence of body height

Body height has been demonstrated to be systemati-

cally related to stride length in adults [6]. There is
further evidence that limb length influences step fre-
quency, step length, and walking velocity [3,7 – 12]. A Fig. 2. Stride length (cm) vs. age (n =420). Reproduced with permis-
mean cadence of 176 steps/min was reported for 1 year sion [12].
 D. Sutherland / Gait and Posture 6 (1997) 163–170 165

Fig. 5. Normalized stride length s/l (crosses, left-hand scale) and

normalized cadence n̄ =n(g/l) − 1/2 (squares, right-hand scale) as a
function of age. Children’s data from Sutherland et al. [12]. Adult
Fig. 3. Walking velocity (cm/s) vs. age (n= 420). data (right-most point) from Murray et al. [30], age 30 – 35 years, leg
length assumed at 1.00 m. Reproduced with permission [16].
2.2. Dimensionless gait measurements
incorrect. By dividing step length by leg length, the
How closely does step length match height or leg dimensionless step length shows change until around 4
length? This question introduces the subject of dimen- years of age. Scrutton [10] reported that the step factor
sionless gait measurements, designed to normalize for (step length divided by leg length) increases from ages 1
differences in stature [8,11,13 – 15]. Data from the tread- through 4 years. Hof and Zilstra [16], in a letter to the
mill studies of normal children between 4 and 10 years editor of the Journal of Biomechanics using data from
of age gives convincing evidence that, when scaling for our laboratory, showed changes in dimensionless step
leg length is used, non-dimensionless numbers of stride length and cadence in children up to 4 years of age
length and cadence are essentially adult values. This is (Fig. 5). O’Malley [14], also taking data from our
not unexpected. However, children under 4 years of age publication, used more complex statistical techniques,
are not included in Zilstra’s report. ‘detrending and simultaneous normalization’ to make a
Our laboratory reported a scatter plot of leg length straight line for stride length, contrasting with the
and step length showing a direct linear relationship changes up to 4 years with geometric scaling (Fig. 6). If
between step length and leg length in children, ages 1 the goal is simply to compare dimensionless stride
through 7 years of age (Fig. 4) [12]. This would appear length or step length to the normal for children 1–7
to rule out anything but growth as the determining years of age, O’Malley’s method can be selected, but
factor for step length. However, this assumption is

Fig. 6. Stride length normalized by geometric scaling, compared with

Fig. 4. Step length (cm) vs. leg length (cm) (n=414). Reproduced normalization by detrending and simultaneous normalization. O’Mal-
with permission [12]. ley, reproduced with permission [29].
166 D. Sutherland / Gait and Posture 6 (1997) 163–170

angles will follow. A full description can be found in

the research monograph by Sutherland et al. [12]. The
dynamic joint angles show clear evidence of maturation
in the ages up to 3.5–4 years. The maturational
changes are in both the curve patterns and the predic-
tion regions for curve variability.

3.1. Sagittal plane mo6ements

The easiest abnormalities to observe are the sagittal

movements at the ankle and knee.

3.1.1. Ankle
Initial heel strike, absent in 1 year old subjects, is
only slightly developed in 1.5 year olds. The ankle
plantar/dorsiflexion curve is notable at age 1 year, by
absence of plantar flexion movement after initial foot
Fig. 7. Duration of right, single-limb stance as percentage of gait vs. contact. There is also less dorsiflexion in swing at age 1
age (n=420). Reproduced with permission [12].
year, which can be observed as a relative drop-foot [12].

this method masks the maturation process. The geo- 3.1.2. Knee
metric scaling process, advocated by Hof and Zilstra, is The primary change in the knee flexion/extension
simpler, but it should only be applied for children who curve by age is gradual development of an initial knee
are 4 years and older and thus beyond the age of flexion wave. For those unfamiliar with the term initial
normal maturation. Both growth and a maturation knee flexion wave, it is used to describe the flexion of
process influence time-distance parameters between 1 the knee during loading response and subsequent exten-
and 4 years of age. sion during mid stance. This initial knee flexion wave is
brought about by an external knee flexion moment
2.3. Percent of single stance controlled by an eccentric contraction of the quadriceps
muscle. Its purpose is to act as a shock absorber,
The percent of single stance is another useful mea- producing a more gradual elevation of the body’s cen-
surement in the evaluation of walking maturity. The ter of mass and reducing the energy requirement for
ability to balance on one lower extremity is the hall- walking [6]. The wave is present early, but not well
mark of human bipedal gait. By definition, the time of developed until 4 years of age [12]. The second, and
single limb stance and the time of opposite swing are much stronger, flexion wave begins well before toe-off
identical. A normal single stance time is an indication and prepares the limb for advancement. There is very
of stability and an insurance of adequate time for little difference in swing phase knee flexion across the
swing. The percent of time in single limb support is ages 1 to 7 years of age [12]. Peak knee flexion occurs
restricted for the young child. A gradual increase in the early in swing phase. The mean peak knee flexion of
percent of single limb support occurs until a level near 415 subjects is 72°. A slight reduction in peak knee
that of a normal adult is reached, by 3.5 – 4 years of age flexion (65°) can be observed in the 1 year old subjects
(Fig. 7) [12]. [12].

3.1.3. Hip
3. Kinematics The changes in dynamic hip flexion/extension with
age are also minimal. The dynamic range in ages 1, 1.5,
Kinematics is defined as the study of movements and 2 years are about 9° less than ages 2.5–7 years of
without respect to the forces producing them. Kinetics age [12].
is the study of forces producing movements. It should
be obvious that accurate measurement of movements is 3.2. Coronal plane mo6ements
required to provide a foundation on which to build the
super-structure of kinetic analysis. Joint angles are Pelvic obliquity is quite abnormal until 3 years of age
widely used to show the rotational changes between [12]. Until 3 years of age, there is less elevation in
limb segments. We can use the changes in dynamic joint stance of the ipsilateral anterior superior iliac spine.
angles as one method of assessing gait maturity. A brief The children up to that age also show less adduction of
description of the most important changes in joint the hip during stance phase. The wide base of support
 D. Sutherland / Gait and Posture 6 (1997) 163–170 167

of the younger children is linked with these changes. O. unpuu et al. [18] collected and analyzed kinematic
Narrowing of the base of support accompanies the and kinetic data of 31 normal children at the Newing-
maturational changes in these movements. ton Children’s Hospital Gait Analysis Laboratory. The
age range of their subjects was 5 to 16 years. They
3.3. Trans6erse plane mo6ements concluded that the pediatric data were similar to that of
normal adults.
3.3.1. Pel6ic rotation
The dynamic range of pelvic rotation varies only 3.6. Variability of kinematic data
slightly across the ages from 1 to 7 years. There is a
trend toward a lower dynamic range over the age of 3 Children have greater inter-subject variability than
years [12]. adults [12,19]. Extensive data showing prediction re-
gions for dynamic joint motion for children 1–7 years
of age, prepared with a boot strapping technique, have
3.3.2. Hip rotation
been published [12]. These data show decreasing vari-
The curve pattern for hip rotation is remarkably
ability with age. A related subject is that of repeatabil-
similar for ages 1 through 7 years, but there is a clear
ity. In a study of normal children between 5 and 16
external rotation bias in the magnitude of external
years of age, Gorton et al. [20] found that, although the
rotation at ages 1 and 1.5 years. This increase in
pattern of walking was unchanged, the day to day
external rotation correlates with the changes seen in the
repeatability was lower for the younger children.
passive range of motion of the hip [12].

3.3.3. Foot rotation 4. Kinetics

Foot rotation (line of progression) is characterized by
a number of qualitative changes. Only at age 1 year is Force measurements are very difficult to obtain on
there no internal rotation movement during pre-swing. children under 2 years of age. One of the causes for this
There is irregularity during swing phase up to the age difficulty is that children under 2 years of age usually
of 2.5 years. There is greater stance phase external contact the same force plate with the successive step. In
rotation up to age 2 years and the prediction regions general, the vertical, medial-lateral, and fore-aft curves
are quite broad, emphasizing the greater variability in of children 2 years and older have a pattern similar to
foot rotation in the youngest children [12]. those of adults, but there are some differences. The
vertical-force curves show an increasing definition of
3.4. The pel6ic-span/ankle spread ratio the mid-stance trough and the second peak does not
exceed 100% body weight in the age range of 2–7 years.
Immature children walk with a wide base of support.
As they mature, they are better able to control the body
mass over the supporting limb, because of greater co-
ordination and improved balance. The pelvic-span/an-
kle spread ratio gives a relative measure of the dynamic
base of support in double stance phase. The ratio
increases, indicating a narrowing base of support, until
3.5 years and thereafter remains constant (Fig. 8) [12].

3.5. Completion of maturational changes in joint angles

It is now feasible to establish the time when all of the

changes in dynamic lower extremity joint angles are
complete and thus establish a time window for comple-
tion of central nervous system maturation. Biafore [17],
using a neural network technique, determined the age
at which the development of gait stabilizes. Using mo-
tion data from 415 studies of normal individuals, ages
1 – 7 years [12], he included 12 different lower extremity
Fig. 8. Pelvic-span/ankle spread ratio in each of the ten age groups
kinematic variables. The neural network was trained
studied. The square indicates the median value, the vertical bar
using a set of 25 simulated individuals of age 7 years. encompasses the middle 50% of subjects, and the upper and lower
The results provided evidence that gait stabilizes be- marks respectively indicate the greatest and least values recorded.
tween the ages of 3.5 and 4.0 years of age [17]. Reproduced with permission [12].
168 D. Sutherland / Gait and Posture 6 (1997) 163–170

The second peak of the vertical-force curves is well

below 100% in the youngest children [12], suggesting
that there is a maturation factor relating to the concen-
tric action of the ankle plantar flexor muscles during
terminal stance [21]. Oeffinger et al. [22] reported in-
creasing amplitude in the ankle sagittal plane external
dorsiflexion moments in ages 4 years and above. In
addition, their study showed a decrease in the absorp-
tion portion of the ankle power curve (A1) and an
increase in power generation (A2) with increasing age.
An additional finding was a decreasing power genera-
tion in hip motion, both in the initial hip extensor (H1),
and pre-swing and early swing hip flexor (H3) power
generation, as defined by Winter [23]. This suggests that
the youngest walkers use their hip flexor and extensor
muscles more than their ankle plantar flexors for power
Fig. 9. Mean times of onset and cessation of EMG activity in the
generation. These maturational changes reported by
tibialis anterior in each of the ten age-groups. The on and off times
Oeffinger et al. appear to contradict the study of are expressed as percentages of stance and swing phases. All measure-
O. unpuu et al. [24] who reported that significant change ments were made using surface electrodes. In each case, the number
in age is not reflected in the normal moment and power of subjects in each age-group is noted following the point of cessation
curves of 31 normal subjects from 5 to 16 years of age. of stance-phase activity. Stance time has been changed to represent
100% and swing time has been changed to represent 100%, in order to
Both studies confirm that the general shape of the
normalize for the changing toe-off time. Reproduced with permission
curves is constant, but Oeffinger et al. report differences [12].
in the magnitude of moments and powers with increas-
ing age. incomplete. Ideally, the database should include all the
muscles and should be collected with fine-wire elec-
trodes, as surface recordings include some cross-talk
5. Electromyography between muscles [25]. This is impractical for children.
As a matter of policy, we exclude the use of fine-wire
There are practical problems in using EMG for de- electrodes for children under 4 years of age. Further-
termination of gait maturation. For ethical reasons, more, for children over 4 years and for adults, we
fine-wire electromyography has not been used on very reserve the fine-wire electrodes for deep muscles, very
young children. The available database is limited to small muscles, and for those muscles that are consid-
surface recordings of a limited number of muscle ered for transfer. Thus, for the foreseeable future, we
groups. The database from our laboratory includes the must rely on surface recorded electromyographs of
following muscles or muscle groups; gluteus maximus, children. If fine-wire recorded telemetry is obtained, we
gluteus medius, medial hamstrings, lateral hamstrings, must consult an adult database [25]. Thus, we acknowl-
vastus medialis, gastroc-soleus, and tibialis anterior edge that our grasp of the total muscle activity in
[12]. Maturational changes are present in the phasic children, as obtained by electroymographs, is limited.
activity of the vastus medialis, tibialis anterior, and the The limitations of electromyography should not dis-
gastroc-soleus between the ages of 1 and 2 years. The courage us from its use, even in studies of very young
vastus medialis shows premature onset in swing phase, children. The surface recordings of muscle activity,
as compared with the older age groups in the study. when compared with a normal database, provide
The stance phase activity of the tibialis anterior is unique information that can be used in the interpreta-
prolonged beyond the normal period of loading re- tion of moment and power measurements.
sponse and swing phase onset is delayed, as compared
with the children over 2 years of age (Fig. 9). The
gastroc-soleus fires prematurely in swing phase in a
high percentage of the subjects 2 years and younger and 6. Energetics
in a much smaller percentage of the subjects more than
2 years of age. Subtler changes are evident in the Oxygen consumption is higher for children, 6–12
gluteus maximus, gluteus medius, medial hamstrings, years, than for adolescents, 13–19 years. The rate for
and lateral hamstrings [12]. adults is even lower than that of the adolescents (Fig.
By comparison with the normal database for time– 10) [26]. Given the fact that the oxygen consumption
distance parameters, kinematics, and kinetics, the elec- rate, plotted against walking velocity, diminishes pro-
tromyography database for young children is gressively during the growth years and is still lower in
 D. Sutherland / Gait and Posture 6 (1997) 163–170 169

the adult population sampled, it is difficult to try to 7. Discussion

separate the influence of growth and other factors in
this progressive decline. The measurement of oxygen The data presented firmly establishes the fact that, in
consumption does not assist in determining when cen- addition to growth, there is a maturation process bring-
tral nervous system maturation has occurred. Nonethe- ing about gait stabilization between 3.5 and 4 years.
less, if we are comparing the oxygen consumption rate Both growth and central nervous maturation control
for patients with pathological gait to normal subjects, it the gait changes up until 3.5–4 years. However, growth
is clearly important to consult normal age related val- alone explains the majority of changes through the
ues or take body size into consideration. Bowen et al. remainder of the growing years. The changes that occur
[27] collected oxygen consumption and oxygen cost after the maturation period are primarily found in the
time–distance parameters, which are correlated well
data for 96 normal children, ages 5 – 15 years. The
with limb length or body height. Of these two measure-
subjects were allowed to walk at their self-paced walk-
ments, body height has many supporters as a scaling
ing speed. They reported that height, weight, and body
factor, as it is an easier measurement to make. Still,
surface area all were correlated with oxygen cost, but
there is a theoretical advantage to using leg length.
inverse body surface area and oxygen cost proved to Arguably, limb length is more directly related to lower
have the best linear correlation. From this they calcu- limb function than height. Geometric scaling should be
lated an oxygen cost index, which they found useful in reserved for children 4 years and older, in order to
multiple tests from one child, regardless of the time avoid the changes in time–distance parameters which
between analyses. The publications that have been men- are influenced by central nervous system maturation.
tioned all give evidence that age must be included Recognition of the changes which occur normally as a
and/or scaling factors introduced for energy measure- part of growth and development is essential to prevent
ments of children, if normalization is to be carried out. mislabeling such deviations from adult gait as evidence
Rose and associates [28] investigated the relationship of of gait pathology. Though a great deal of information
heart rate to oxygen consumption, both for children is now available about the gait changes, we are still
with cerebral palsy and for normal children. They unable to fully understand or separate the contributions
found a good correlation between oxygen consumption of the various components of central nervous system
and heart rate through a wide range of walking speeds development which drive the maturation process.
and concluded that heart rate could be used as an index
of energy expenditure. Because of its ease of application
the heart rate, as an estimate of energy expenditure, is 8. Future Studies
widely used. However, caution is required in interpreta-
tion as factors other than oxygen consumption can What are some of the future directions for research?
effect the heart rate. (1) The kinematic studies of normal children be-
tween the ages of 1 and 4 years should be repeated
using current hardware and software.
(2) Studies should be undertaken to fill the gap in the
available database of moments and powers in the ages
of 2 through 4 years.
(3) Electromyographic studies need to be expanded.
Additional muscles should be included in the database
and quantitative representation of the EMG plots as
percent of maximal manual or functional testing would
aid in the interpretation [25].
(4) Studies are needed to gain more information
about the actual operation of the motor control system
in human subjects. This is the greatest challenge of all.
Success will require close collaboration with neuro-sci-
entists and physiologists. Questions should be asked
such as ‘Is muscle activation in certain of the lower
extremity muscles brought on by a stretch response?’ If
the answer is yes, how does this process change during
normal maturation? What roles do myelination, prolif-
Fig. 10. Oxygen consumption plotted against walking velocity of
eration of synapses, integration of sensory input, and
children ages 6 – 12 years, adolescents 13–19 years, and adults. Re- learning, play in the maturation process?
produced with permission [26]. A fertile field of investigation is open.
170 D. Sutherland / Gait and Posture 6 (1997) 163–170

Acknowledgements [13] Hof AL. Scaling gait data to body size. Gait Posture
1996;4:222 – 3.
[14] O’Malley MJ. Normalization of temporal-distance parameters in
The comments and suggestions of David Winter were pediatric gait. J Biomech 1996;29(5):619– 25.
very helpful in the completion of this paper. My co-au- [15] Zijlstra W, Prokip T, Berger W. Adaptability of leg movements
thors in the publication most frequently referenced in during normal treadmill walking and split-belt walking in chil-
this paper, Richard Olshen, Edmund Biden and Mari- dren. Gait Posture 1996;4 212-221.
[16] Hof AL, Zijlstra W. Comment on ‘normalization of temporal-
lynn Wyatt, provided energy, enthusiasm and multidis- distance parameters in pediatric gait’. J Biomech 1997;30(3):299.
ciplinary skills in our early work. Sherill Marciano and [17] Biafore S, Cottrell G, Focht L, Kaufman K, Wyatt M, Suther-
Jill Jordano gave invaluable assistance with typing and land D. Neural network analysis of gait dynamics. In: Trans
editing. 37th Annu Meet Orthop Res Soc, Anaheim, CA, 1991.
[18] O. unpuu S, Gage JR, Davis RB. Three-dimensional lower ex-
tremity joint kinetics in normal pediatric gait. J Pediatr Orthop
1991;11:341 – 9.
[19] Lasko-McCarthey P, Beuter A, Biden E. Kinematic variability
References and relationships characterizing the development of walking.
Dev Psychobiol 1990;23(8):809– 37.
[1] Sheridan MDS. The Developmental Progress of Infants and [20] Gorton GE, Stevens CM, Masso PD, Vannah WM. Repeatabil-
Young Children. London: H.M. Stationery Office, 1960. ity of the walking patterns of normal children. Gait Posture
1997;5(2):155.
[2] Thelen E, Cooke DW. Relationship between newborn stepping
[21] Sutherland DH, Cooper L, Daniel D. The role of the ankle
and later walking: a new interpretation. Dev Med Child Neurol
plantar flexors in normal walking. J Bone Jt Surg (Am)
1987;29(3):380– 93.
1980;62(3):354– 63.
[3] Wheelwright EF, Minns RA, Law HT, Elton RA. Temporal and
[22] Oeffinger DJ, Augsburger S, Cupp T. Pediatric kinetics: age
spatial parameters of gait in children. I: Normal control data.
related changes in able-bodied populations. Gait Posture
Dev Med Child Neurol 1993;35(2):102–13.
1997;5(2):155 – 6.
[4] Mena D, Mansour JM, Simon SR. Analysis and synthesis of
[23] Winter DA. Biomechanics and Motor Control of Human Move-
human swing leg motion during gait and its clinical applications.
ment. New York: Wiley, 1990.
J Biomech 1981;14(12):823–32. [24] O. unpuu S, Davis RB, DeLuca PA. Joint kinetics: methods,
[5] Piazza SJ, Delp SL. The influence of muscles on knee flexion interpretation and treatment decision-making in children with
during the swing phase of gait. J Biomech 1996;29(6):723– 33. cerebral palsy and myelomeningocele. Gait Posture
[6] Inman V, Ralston HJ, Todd F. Human Walking. Baltimore, 1996;4(1):62 – 78.
MD: Williams and Wilkins, 1981. [25] Perry J. Dynamic electromyography. In: Gait Analysis: Normal
[7] Beck RJ, Andriacchi TP, Kuo KN, Fermier RW, Galante JO. and Pathological Function. SLACK Inc., Thorofare, New Jersey
Changes in the gait patterns of growing children. J Bone Jt Surg 1992:381 – 411.
(Am) 1981;63(9):1452–7. [26] Waters RL, Hislop HJ, Thomas L, Campbell J. Energy cost of
[8] Grieve DW, Gear RJ. The relationships between length of stride, walking in normal children and teenagers. Dev Med Child
step frequency, time of swing and speed of walking for children Neurol 1983;25:184 – 8.
and adults. Ergonomics 1966;5(9):379–99. [27] Bowen T, Cooley S, Castagno P, Lennon N, Miller F, Richards
[9] Norlin R, Odenrich P, Sandlund B. Development of gait in the J. A technique for normalization of oxygen cost during self
normal child. J Pediatr Orthop 1981;1(3):261–6. paced walking. Gait Posture 1997;5(2):141.
[10] Scrutton DR. Footprint sequences of normal children under five [28] Rose J, Gamble JG, Lee J, Lee R, Haskell WL. The energy
years old. Dev Med Child Neurol 1969;11:44–53. expenditure index: A method to quantitate and compare walking
[11] Todd FN, Lamoreux LW, Skinner SR, Johanson ME, St. Helen energy expenditure for children and adolescents. J Pediatr Or-
R, Moran SA, Ashley RK. Variations in the gait of normal thop 1991;11:571 – 8.
children. J Bone Jt Surg (Am) 1989;71A(2):196–204. [29] O’Malley M. Authors response. J Biomech 1997;30(3):301–2.
[12] Sutherland DH, Olshen RA, Biden EN, Wyatt MP. The Devel- [30] Murray MP, Drought B, Ross C, Kory MD. Walking patterns
opment of Mature Walking. London: MacKeith Press, 1988. of normal men. J Bone Jt Surg 1964;46A:335 – 60.

. .