(Tsutomu and Kayoko Okamoto) Development of Gait B

Development of Gait
by Electromyography
Application to Gait Analysis and Evaluation
Tsutomu Okamoto, Ph.D.
Kayoko Okamoto, Ph.D.
Walking Development Group

Osaka, Japan
Copyright 2007 by Okamoto & Okamoto
Published by

~qTOO3;PJf~pJT
llJ
WALKING
G-S04 Tenno 2-6, Ibaraki-shi,

Osaka 567-0S76, JAPAN
All rights reserved. No part of this publication may be
reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopy, recording,
or any information storage and retrieval system, without
permission in writing from the publisher.
ISBN978-4-902473-05-6
Printed in Japan
Preface
The gait of a human being continues to change over the course of a
lifetime. The first stage is that of neonatal reflex stepping, which is
thought to be the origin of bipedal upright walking in human beings.
This then develops into young infant stepping at the age of one to two
months, followed by inactive stepping and then by voluntary infant
supported walking at the age of six to twelve months. Infants then
acquire independent walking at around the age of one and begin to
acquire mature adult walking at around the age of three.
We have analyzed the detailed changes in the development of human
gait employing electromyography (EMG) which has enabled us to carry
out motion analysis impossible with conventional methods. At present
very little longitudinally analyzed post natal gait development data is
available anywhere in the world because of the difficulty of carrying
out the necessary experiments. It is even more difficult to record
electromyographically the neonatal reflex stepping of newborn babies
or the moment when babies begin upright, independent walking. Even
today the papers that I wrote on this subject in the 1970s and 1980s
continue to be cited.
We have continued up to the present to carry out additional crosssectional and longitudinal experiments concerning gait development
from the newborn baby stage to that of infant independent walking and
have in the process accumulated much electromyographical data. The
results of our analysis of normal gait development suggest that it can
not only contribute to the explanation and clarification of human bipedal
upright walking, but also be applied to various areas of research such as
the diagnosis of and therapy for various walking disorders and the
evaluation of the level of gait function restoration and improvement.
We have gathered together in this book the results of our study and
analysis of gait carried out over the last 40 years, in the hope that this
rare elctromyographical data concerning gait development will
contribute to the further development of this field.
Part I contains our analysis, based on movement and muscle activity,
of the development and changes in gait from birth until the age of eight,
that is from the stage of neonatal reflex stepping, thought to be the
origin of bipedal upright walking in human beings, through that of the
iii
acquisition and mastering of infant independent walking to that of the

acquisition of mature adult walking.
Part II introduces our application of this to the analysis and evaluation
of gait. We have created "An Index of Gait Instability" based on the
results of our analysis of the gait development of infant independent
walking, which we apply to research into the nature of human stepping
and the evaluation of the level of restoration of walking functions in the
elderly.
We hope that this book will prove useful to those engages in gait
studies, not only as a basic reference material analyzing the development of gait, but also as a basis for research, analysis and application
in various fields that will help to generate new ideas about human gait.
Tsutomu Okamato
iv
Contents
Preface
Contents
Part I
iii
v
Development of Gait -Birth to Age Eight-
1 . Newborn Stepping in Neonates and Young Infants

Early neonatal period (1 - 2 weeks)
Late neonatal period (3 - 4 weeks)
Onset of infant period (1- 2 months)
Initial infant period (3 - 4 months)
Discussion
2 . Independent Walking in Infants

1st day of learning to walk
2 weeks after learning to walk
At around 1 month after learning to walk
From 2 to 3 months after learning to walk
Subsequent development
Standing posture on the 1st day of walking
Discussion
8
12
16
18
20
25
28
30
32
34
36
38
40
3. From Newborn Stepping to Mature Walking

- Developmental Changes in One Individual-
45
Neonatal stepping
Young infant stepping
Infant supported walking
Infant walking
Immature child walking: unsettled muscle activity
Mature walking: toward a mature pattern
Developmental period of gait
Discussion
48
50
52
54
56
58
59
61
Part II
Application to Gait Analysis and Evaluation

-An Index of Gait Instability-
67
4. An Index of Gait Instability
-Based on the Development of Independent Walking-
69
EMG findings during the development of gait

EMG activity in unstable walking
Criteria for Instability
An Index of Gait Instability
73
79
84
86
5. Application of an Index of Gait Instability (1)

-Supported Walking in Normal Neonates and Infants-
89
Until the 1st month of age

92
From 1 to 4 months of age
94
96
Developmental changes in EMG patterns
98
Application of an index of gait instability to supported walking
in babies
99
Discussion
101
6. Application of an Index of Gait Instability (2)
-Recovery of Walking in an Elderly Man after Stroke- 107

1 month after the stroke
7 months after the stroke
1 year 7 months after the stroke
EMG evaluation of walking stability
Discussion
References
Appendix
Acknowledgements
About the Authors
vi
110
112
114
116
117
121
125
131
133
Development of Gait by Electromyography
EMG experiment of infant walking
The electromyographic (EMG) recordings were done with a pen-writing mUltipurpose

electroencephalograph, using surface electrodes 5 mm in diameter. The skin at each electrode
locus was scratched lightly with a needle, reducing the resistance between pairs of electrodes
to less than 5000
(Okamoto et aI., 1987).
Neonatal stepping at 3 weeks afte birth
The purpose of this study was to examine the developmental changes in the functional mechanisms of leg
muscles in newborn stepping over the first 4 months in ten
normal neonates. Neonatal stepping in the first month
showed excessive co-activation, that is, co-contraction patterns of mutual antagonists appeared especially during
stance phase. The discharge patterns of co-contraction in
neonatal stepping began to change to reciprocal patterns in
young infant stepping (after the first month), but excessive
muscular activities associated with a slightly squatted
posture and forward lean still remained . Strong muscle
activities of leg extensors due to a parachute reaction of the
legs before floor contact, not seen in the neonatal period,
began to appear in the young infant period from 1 month
of age to 3 months. We suggest that these gradual changes
of leg muscular activity in newborn stepping are evoked
by development of balance, postural control, and strength,
thereby modulating the neonatal stepping reflex.
When a newborn infant is held under the arms in an upright

position, well-coordinated walking movements (stepping reflex) appear
to be elicited by tactile stimuli on the soles of the feet as they are
placed on the floor (Fig. 1-1).
McGraw (1940) and Zelazo et al. (1972) have discussed the
significance of early stepping movements for development of adult gait.
Newborn stepping has been an object of study for a long time. Only a
few attempts so far, however, have been made to study characteristics
of newborn stepping by electromyography (EMG).
Forssberg (1985) noted that the lateral gastrocnemius showed
strong activity just before the foot reached the floor (Fig. 1-2). Because
this was like a digitigrade pattern, he concluded that man is born
with a quadrupedal locomotor program. Thelen (1982, 1987), however,
did not find any strong activity in the gastrocnemius before floor
contact (Fig. 1-2) . To further study this problem of the EMG pattern
in the gastrocnemius before foot contact in stepping, it would be
instructive to record EMG data during stepping not only in the neonatal
period (up to 1 month of age), but also in the young infant period
(after 1 month of age).
We have thus closely examined the characteristics of newborn
stepping in ten babies during both neonate and young infant periods in
terms of the functional mechanisms of leg muscles.
Four male and six female neonates were observed from 1 to 4 weeks
after birth. Criteria used for selecting the subjects were that they be
full-term with birth weight between 2500 g and 4200 g. They were
screened by pediatricians to rule out abnormalities and illnesses.
Motor development of each subject was within normal limits.
Fig. 1-1.
Newborn stepping at 2 days after birth.
4 Development of Gait
EMGs of all subjects were recorded from the neonatal period (up to
the 1st month of age) to the young infant period (from 1 to 4 months
after birth) at intervals of 1 to 4 weeks.
To induce newborn stepping, the examiner held the neonate under
the arms with the soles of the feet touching a horizontal flat surface.
Well-coordinated walking movements were observed from around 1
week after birth to around 3 months. We could not induce stepping
simply at will, but tended to be successful when the infants were lively,
crying, hungry, or slightly excited (Figs. 1-1 and 1-2). For analysis we
selected well-coordinated walking movements consisting of three or
more steps.
The EMGs were recorded from six muscles in the right leg (Fig.1-2):
tibialis anterior (fA), lateral gastrocnemius (LG) , vastus medialis (VM),
rectus femoris (RF), long head of biceps femoris (BF) , and gluteus
maximus (GM), and from two to six muscles in the left leg, usually the
TA, LG, RF, and BE
RF: rectus femoris

GM : gluteus maximus
(Knee extensor, Hip flexor)
(Hip extensor)
Mutual antagonist: SF
SF : biceps femoris
(Knee flexor, Hip extensor)
Mutual antagonist: RF
LG : lateral gastrocnemius
VM: vastus medialis

(Knee extensor)
TA : tibialis anterior
(Ankle plantar flexor)
(Ankle dorsiflexor)
Mutual antagonist: TA
Mutual antagonist : LG
Fig. 1-2. Muscles chosen for recording EMG.
Newborn Stepping in Neonates and Young Infants 5
Surface electrodes 5 mm in diameter were used. To attenuate

artifacts in the surface electrode recordings, skin impedance was
lowered by scratching loci of the electrodes lightly with a needle before
the electrodes were applied (Okamoto et al., 1987). The EMG
recordings were done with an 18-channel pen-writing electroencephalograph (60 mm/sec) with the gain set at 12 mm/0.5mV. An
analog pulse signal from the video recording camera (60 frames/sec)
was recorded simultaneously with the EMGs. The walking cycle was
divided into swing phase (SW) and stance phase (ST) by the video
recordings.
Movement and EMG recordings obtained during newborn stepping
showed some variations both within and among subjects. Variations in
stepping form and EMG patterns appeared to depend to some extent
on how the infant was supported. We thus selected as representative
data those movements and EMG patterns of stepping that were seen
relatively frequently in the neonatal or young infant period being
observed. For purpose of analysis, longitudinal observations were
divided into early neonatal period (from 1 to 2 weeks after birth), late
neonatal period (from 3 to 4 weeks), onset of infant period (from 1 to 2
months), and initial infant period (from 3 to 4 months).
Mature adult walking pattern

We need to examine normal stable adult walking to compare with
gait in terms of developmental processes. Figure 1-3 shows a typical
EMG of adult walking (the subject is a female 29 years of age). From
the basogram, stance and swing phases can be demarcated.
The discharge patterns of the TA and LG, which participate in
movement of the ankle joint, showed an almost reciprocal relationship.
The TA (an ankle dorsiflexor) discharged through most of swing
phase and at the beginning of stance phase, whereas the LG (an ankle
plantarflexor), which participates in push off motion, discharged in a
strong burst in the latter part of stance phase. The hip and knee
muscles, VM, RF, BF, and GM, acted for shock absorption during the
transition from swing phase to stance.
Typical EMG pattern of normal adult walking
0.5
k~j~AA:j~A
326
mv]
333
339
346
356
365
372
378
388
Tibialis anterior (TA)
Lateral gastrocnemius (LG) -+---M~"""~I M'H~---1<Y,\,~-...."I(NJIN't--t-
Vastus medialis (VM)

Rectus femoris (RF)
Biceps femoris (BF)
Gluteus maxim us (GM)

Ankle
Plantar flexion
Dorsiflexion
Knee
Extension
Hip
Extension
Flexion
Flexion
Swing
Stance Phase
Pho.e (SW)
(ST)
Swing
Stance
Basogram
Foot contact (FC)
VTR signal
HC FF
HO
TO
1 sec
.J._"'-"'''''''''''''''Nt'''",,,,'''',_.J,--*-JIJtr-,,_,,f.,.,,,JtUlt.,,--,~
300
350
400
l~( ~~ ~
HC
(Heel Contact)
FF
HO
TO
(Foot Flat)
(Heel Off)
(Toe Off)
Fig. 1-3. Typical adult EMG pattern in leg muscles during walking.
Swing phase (SW: short phase), Stance phase (ST: long phase), Basogram: Foot contact (He,
FF, HO, TO).
Newborn Stepping in Neonates and Young Infants
Early neonatal period

(1 - 2 weeks)
ST
sw
ST
~ 'I'\'fJtW~~t;~i~
(R)
TA
",""
LG
,
VM
RF
BF
GM
,--
(l)
TA
LG
VM ----~~~--------~------+_----------~~~~~------+_
BF
-------------------+------+-------------------~------~.
STANCE (ST)
SWING (SW)
--'I
___
1_
s e_c __
0.5 mv
1 week (Y.T.)
NEONATAL STEPPING
Fig. 1-4. EMGs of stepping at 1 week after birth (Y.T.).

SW: swing phase, ST: stance phase, (R): right leg, (L): left leg, TA: tibialis anterior, LG: lateral
gastrocnemius, VM: vastus medialis, RF: rectus femoris , SF: biceps femoris, GM: gluteus maximus.
Stepping in this period was characterized by quick hip flexion in

which the thigh became approximately horizontal in the first part of
swing phase. The foot was raised forward and dorsiflexed strongly, as
shown in Figures 1-4, 1-5, and 1-6. In the middle part of swing phase,
the leg was often held in the flexed position. Then the foot began to
reach the floor slowly, the knee extending passively along with the hip.
The foot usually contacted the floor with the heel or sole first (Fig.
1-7), but in a few instances toe contact was seen. A fairly squatted
posture was often observed during stance phase.
1 week after birth (Fig. 1-4): In the beginning of stance phase, no

notable activity was seen in the leg muscles examined. In single stance
phase, continuous discharges were frequently observed in the TA,
VM, RF, and GM, activities not usually seen in adult gait. The
discharge patterns of the VM and GM were highly consistent, but
activities at the ankle (the TA and LG) and of two-joint muscles
crossing the knee and hip (the RF and BF) showed slight variations.
That is, at the ankle, a reversed reciprocal (TA+, LG-) pattern was
observed in many cases, but co-contraction (TA+, LG+) and reciprocal
(TA-, LG+) patterns were seen in some of the subjects. Across the
knee and hip, a reversed reciprocal (RF+, BF-) pattern was observed
in many cases, but co-contraction (RF+, BF+) and reciprocal (RF-, BF+)
patterns were observed in some of subjects. In the first part of swing
phase, continuous discharge in the TA was seen in most cases.
Sometimes slight activity of the RF was observed in the same phase.
In the latter part of swing phase, when the leg was extending, activities
were hardly seen in the LG, VM, RF, BF, and GM.
Fig. 1-5. Foot contact of newborn stepping at 2 days after birth.
ST
(R)
SW
V, , , , / " " 1,.1/ " I
TA
'i~
LG--________
ST
" " , ,
W. , / , . , , W; " " I I " I WI I I ;
.. , I
......r+~
r-__~--~--------------------+_--+_-----
VM
RF
BF
GM
( L)
TA-""""\
LG
RF
BF
I
SW
ST
1 sec
2 weeks (A. I.)

NEONATAL STEPPING
Fig. 1-6. EMGs of stepping at 2 weeks after birth (A.I.).
0.5 mv
2 weeks after birth (Fig. 1-6): In the beginning of stance phase,

continuous discharges in many leg muscles were observed more
often than at 1 week after birth, especially in single stance phase.
Continuous discharge patterns of the VM and GM were again
consistent, but EMG patterns of the TA and LG and of the RF and BF
showed some variations. That is, reciprocal, reversed reciprocal, and
co-contraction patterns were seen in those muscles. In the first part of
swing phase, continuous discharge patterns of the TA were similar to
those of 1 week after birth. In contrast to the first week, continuous
discharge in the RF was observed during swing phase in many cases.
In the latter part of swing phase, activity in the BF frequently began to
appear before foot contact.
Foot contact (Fe)

Heel contact with slow leg extension
Foot flat with slow leg extension

EMGs of leg extensors before floor contact
Early neonatal period (1-2 weeks after birth)
Fig. 1-7. Foot contact of stepping in neonatal period (1-2 weeks after birth).
(-): no activity.
11
Late neonatal period

(3 - 4 weeks)
rt~flflft
sw
ST
'-H'"""1'-1V.--r--r--t-.'-I"~'-I--''-h'-l--i-W....,~t-H-1'''''''''''i,....J'""'h''''~I'''''''~~h-h''
(R)
TA_ ..Wi
.,'
LG
VM
RF ,
BF-!\
,II~
...,.
...
Jill/ILl
'n'
''''r-
GM
(LiA"'M~~~~~I'J'
LGW~~~ ,
RF.~"" "
...
~~~
.\
.';':;'1.;'.;, 1r""'"'Wt""'~
'.Oi.~i~~'jI~""'\""~
....
w..~t~
BF~~
SW
ST
1 sec
3 weeks (H. YJ
NEONATAL STEPPING
Fig. 1..8. EMGs of stepping at 3 weeks after birth (H.Y.).
0.5 mv
--
SW
hS-rT-r-r--rr-h h----r-r-r--r-,,-...,-,
(R)~
TA
II"
1ft.
""
LG Ib",
"'I'
RF
SF
II.
In
'-\iNo'~
"""rIP
r-"
'W
VM ~~7i
1
~.,~'f
".'
.'\'"',..
~,
.J.
\1"'"
/-h-
,I
I\,
GM
(L)
,I.
",1."1".1,,,1,,,11,
"r'lf
'''''I'
TA /JftOoMiI\--~--"aWJ"'M--..-j>,~~
.., ........~,''I'--
LG~~~----r-----~--~~~~~-
VM~~#~~I~\\\-\fi\'oJ--.'tol\--+III~~1'~~~tr
qj.-'~~~4~~~~ftJt!\,jr\""-
IIIWIlfIM~.r------+---1'-~,~fr""\"\- GMMM~vN(i~/""""'---+--~~~~'4\t'+.'/i'UIf,~,*,-ST
SW
1 sec
I 0.5 mv
4 weeks (T. YJ
NEONATAL STEPPING
Fig, 1-9, EMGs of stepping at 4 weeks after birth (T.Y,),
13
As in the early neonatal period, leg flexion was very active in the
first part of swing phase in this period (Figs. 1-8 and 1-9). The thigh
was outwardly rotated as it was raised diagonally in a forward and
lateral direction and the foot dorsiflexed strongly. Then the foot began
to approach the floor slowly, the knee extending passively along with
the hip. The foot usually contacted the floor with the lateral border
first (Figs. 1-10, 1-11, and 1-12), but sometimes the heel, sole, or
forefoot made initial contact. The fairly deep squatting posture of the
early neonatal period began to become less pronounced during stance
phase.
3 and 4 weeks after birth (Figs. 1-8 and 1-9): Throughout stance
phase, continuous discharges of leg muscles were observed in many
cases. EMG patterns of the VM and GM were consistent as in the
early neonatal period. The reversed reciprocal ankle pattern during
stance, seen in neonatal former period, was hardly evident, whereas
the reciprocal and co-contraction patterns became more frequent.
Discharge patterns of the two-joint knee and hip muscles showed
reversed reciprocal, reciprocal, and co-contraction patterns as in the
early neonatal period. In the first part of swing phase, continuous
activity was seen in the TA as in the early neonatal period, but weak
bursts of the RF and BF were seen often at the beginning of swing
phase. In the latter part of swing phase, activities began to be seen in
the LG, VM, RF, BF, and GM in some of the neonates.
VM
LG (-), partly (+) f----+--r~
H , partly (+)
Vastus medialis
(Knee extensor)
Lateral gastrocnemius
Lateral border with slow leg extension

Late neonatal period (3-4 weeks after birth)
Fig. 1-10. Foot contact of stepping in neonatal period (3-4 weeks after birth).
(-): no activity, (+): noticeable activity.
Fig. 1-11. Neonatal stepping at 26 days after birth.
Fig. 1-12. Foot contact of neonatal stepping at 22 days after birth.
15
Onset of infant period

(1 - 2 months)
SW
ST
"-"-.-v " , Wr,,....,,,....,,r-r-.,.-,,.-,.,. . "W
U ' , ,, ,
( R)
TA
BF
GM~--...
(Ll
TA'IIA(/~""",~-""'t"''''r'TM~~.~~i''J,rIII..~~.....~IJ1fIt4>~Ji'.
ST
SW
1 sec
0.5 mv
1.5 months (T. YJ

YOUNG INFANT STEPPING
Fig. 1-13. EMGs of stepping at 1.5 months after birth (T.Y., same subject as in Fig. 1-9).
After 1 month, as shown in Figure 1-13, leg flexion was performed

strongly in the first part of swing phase as in the neonatal period, but
the degree of hip flexion tended to decrease slightly. We found mostly
plantarflexion of the foot before floor contact rather than dorsiflexion,
which had been more prevalent in the neonatal period. The foot
usually contacted the floor with the lateral border of the forefoot first
(Fig. 1-14). Knee extension began to be performed more actively than
in the neonatal period. A half-squatting posture during stance phase
tended to increase.
1.5 months after birth (Fig. 1-13): During stance phase, continuous
discharges of the VM and GM were seen as in the neonatal period.
The ankle muscles likewise exhibited reciprocal and co-contraction
patterns as in the late neonatal period. The two-joint knee and hip
muscles showed reversed reciprocal, reciprocal, and co-contraction
patterns, similar to the neonatal period. In the first part of swing
phase, continuous activity of the TA was observed in many instances,
as in the neonatal period. In the beginning of swing phase, weak
activities of the RF and BF were seen often, but not always. In the
latter part of swing phase, activities of the LG, VM, RF, BF, and GM
appeared often before foot contact.
LG (-), (+) 1---+--+-.1

Foot contact (Fe)

Lateral border of forefoot with fast leg extension

Onset of young infant period (1-2 months after birth)
Fig. 1-14. Foot contact of stepping in young infant period (1-2 months after birth).
(-), (+): instances of no activity and of noticeable activity intermingled.
17
Initial infant period

(3 - 4 months)
h,J
SW
ST
W t'
' ihoIh'th
'
f ...... ; i .. Wrlr'Ir'Ih'.h
I' "
i' "
..
Wit' , .
t'
of
\-T
( R)
TA---4~~~------~~~~--~--------~~~~
~
TA~:;~~
(0
LG~
ST
"'~';~
SW
t sec
3 months
0.5 mv
CA. I.)
Fig. 1-15. EMGs of stepping at 3 months after birth (A.I.. same subject as in Fig. 1-6).
In this period (Fig. 1-15), the lower limb flexed strongly in the first
part of swing phase as in the neonatal period, but the total degree of
hip flexion tended to decrease slightly. The foot usually approached
the floor with a more rapid and vigorous extension of the lower limb,
with the toes initially contacting the floor (Fig. 1-16). Knee extension
and ankle plantarflexion were visibly active in many cases. A halfsquatting posture during stance phase became more frequent.
3 months after birth (Fig. 1-15): During stance phase, continuous
discharges in the VM and GM were observed until onset of the infant
period, as mentioned above, but continuous discharges in the
anteriorly situated TA and RF tended to decrease or disappear, leading
to reciprocal patterns (TA- LG+ and RF- BF+) in most cases. In the
first part of swing phase, strong bursts in the TA, RF, and BF were
frequently observed. In the latter part of swing phase, strong activities
of the LG, VM, BF, and GM appeared often. Although strong activity in
the LG and VM were observed shortly before foot contact, activities in
the BF and GM were not generally seen. In the course of this period,
co-contraction patterns of the ankle (TA+, LG+) and the knee and hip
(RF+, BF+), seen fairly often in stance phase in the late neonatal period
and onset of the infant period, gave way to reciprocal patterns (TALG+ and RF- BF+) . Strong activities of the LG and VM in the latter part
of swing phase, hardly observed during the neonatal period, became
remarkably more frequent.
VM (+), partly (-)

LG (+), partly (-) f-+-+-~
Vastus medialis
(Knee extensor)
Forefoot with fast leg extension

Initial young infant period (3-4 months after birth)
Fig. 1-16. Foot contact of stepping in young infant period (3-4 months after birth).
(+): noticeable activity, (-): no activity.
19
Discussion
Although Thelen et al. (1982) reported that when held upright,
newborn infants show well-coordinated walking movement that
normally cannot be elicited after about 2 months of age, we could
induce infant stepping until around 3 months of life in a number of
cases. Forssberg (1985) and Thelen et al. (1987) pointed out from
movement patterns and EMGs, that the locomotor pattern of the
newborn differs markedly from that of an adult.
From our results, newborn stepping was characterized by active leg
flexion with the thigh becoming horizontal, a somewhat squatted
posture, and variable forms of foot contact with the surface (Figs. 1-17
and 1-18). Leg muscle activities in newborn stepping are usually
irregular and involve more co-activation than in adult walking,
especially in stance phase. For example, in single stance continuous
discharge patterns were seen in the knee and hip extensors (VM and
GM) in neonatal and infant stepping, associated with a progressively
decreasing but ever present squatted posture. These activities in the
leg extensors appear to be attributable to the squatted posture itself
and would thus not be seen in adult gait.
On the other hand, we did observe some similarities in leg muscle
activity between newborn stepping and adult gait. As swing phase was
beginning, for example, bursts were usually observed in the TA during
newborn stepping. Muscle activation seen in flexors of the lower limb
at the onset of the stepping cycle becomes incorporated into supported
walking seen prior to independent walking, thence into early
independent walking, and so on to adult gait. These results suggest
that mature walking may evolve from the newborn movement pattern.
We could see a developmental trend across the neonatal and young
infant periods in stance phase and at the end of swing phase. In
stance phase, contractile activity between mutual antagonists varied
among co-contraction (TA+ LG+ and RF+ BF+), reciprocal (TALG+ and RF- BF+) , and reversed reciprocal (TA+ LG- and RF+ BF-)
patterns. The reciprocal pattern tended to appear more often if the
baby happened to be inclined forward and the reversed reciprocal
pattern when the baby was inclined backward. Co-contraction might
be viewed as an intermediate situation between these two tendencies.
Developmental changes in the pattern of newborn stepping

1week (Y. T.)
Neonatal period
(Early. 1- 2 weeks)
~~~~j
2 weeks (A.I.)
Neonatal period
(Early. 1-2 weeks)
jjlttf1
3 weeks (H.Y.)
Neonatal period
(Late. 3-4 weeks)
fttfttttt
1.5
Infant period
(Onset. 1-2 months)
months (H.Y.)
frftfflf
3
Infant period
(Initial. 3-4 months)
iiw~~
3.5
Infant period
(Initial. 3-4 months)
months (A.I.)
months (H.Y.)
rl~~~(f
Fig. 1-17. Developmental changes in the pattern of newborn stepping.

B: same subject as E. C: same subject as D and F.
The reversed reciprocal pattern was seen relatively often in the

early neonatal period (first 2 weeks), but the other two patterns became
more frequent in the late neonatal period (3rd and 4th weeks) and
as the infant period began (2nd month). In the initial infant period
(3rd and 4th months) the reciprocal pattern became more dominant
than the other patterns, although all three patterns could still be
observed. Interestingly, this trend anticipates the changes in pattern
between mutual antagonists seen as a baby first begins to walk
independently and becomes more stable in the ensuing months.
At the end of swing phase in the neonatal period, the LG and VM
exhibited no activity until the foot actually touched the floor (Fig. 1-18).
The foot reached the floor in a relatively passive action of the lower
limb, contacting the floor variously with the heel, entire sole, or lateral
border. Thelen et al. (1982, 1987) did not find any strong activity in the
gastrocnemius before floor contact in the neonatal period. In the
second month, at the onset of the infant period, the LG and VM began
to become active before actual contact of the foot with the floor, with
the lateral part of the forefoot generally touching the floor first. Thelen
et al. (1987) and Forssberg (1985) reported that the gastrocnemius
showed strong activity just before the foot reached the floor in the
young infant period. The activities of the LG and VM subsequently
became more pronounced shortly before and during floor contact in
the 3rd and 4th months, to the point that one might associate such
activity with the parachute reaction. Milani-Comparetti et al. (1967)
observed from movement analysis that the parachute reaction of the
lower limbs begins to appear at about 4 months after birth. Our
observations, if they are of the same phenomenon, suggest that the
beginnings of the parachute reaction can be found by EMG much
earlier than by visual observation of behavior.
These changes in muscle activity during the stance and swing
phases of newborn stepping represent what might be considered as
the first developmental changes in human bipedal locomotion. Further
research would be necessary to elucidate the extents to which these
changes can be attributed to maturation of balance, postural control,
and strength, as well as to emergence and disappearance of the
neonatal stepping reflex itself.
Neonatal and
young infant period
Foot contact
with leg extension
EMGs of VM and LG
before floor contact
Early neonatal
VM(-)
LG(-)
1-2 weeks
after birth
Heel contact or foot flat
with slow leg extension
Late neonatal
VMH, partly (+)
LG(-), partly (+)
3-4 weeks
after birth
Lateral border
with slow leg extension
Young infant (onset)

VMH, (+)
LGH. (+)
1-2 months
after birth
Lateral border of forefoot
with fast leg extension
Young infant (initial)

VM( +), partly (-)
LG(+), partly (-)
3-4 months
after birth
Forefoot
with fast leg extension
Fig. 1-18. Developmental changes of foot contact in newborn stepping.

VM : vastus medialis, LG: lateral gastrocnemius, (-): no activity, (+): noticeable activity,
(-), (+): instances of no activity and of noticeable activity intermingled.
Conclusion
In ten neonates first seen at 1 to 4 weeks after birth, EMGs of
stepping were recorded at 1 to 4 week intervals until around 4 months
of age.
During stance phase in neonatal stepping, many leg muscles showed
excessive continuous discharges compared with the adult walking
pattern. Continuous activity was seen in the vastus medialis and
gluteus maximus to maintain a partially squatted posture. Mutual
antagonists in the lower limbs variously showed reciprocal and cocontraction patterns during the neonatal period, but the EMG patterns
began to shift toward predominantly reciprocal patterns in the young
infant period, associated with leaning forward.
In the first part of swing phase, activity in the tibialis anterior was
observed in most cases. During neonatal stepping, in the latter part of
swing phase, muscular activity was not seen in the lateral gastrocnemius or vastus medialis, but during young infant stepping
EMG activity in these two muscles became marked before the foot
reached the floor, suggesting that muscular activities participating in
active ankle plantarflexion and knee extension began to act as a
precursor to the parachute response of the lower limb.
In summary, these muscular activities of the lower limb characterize
the EMG features of newborn stepping. Changes in EMG patterns
during newborn stepping, detectable well before corresponding
changes can be visually observed in movement analysis, may be the
first signs of development in human locomotion.
In order to elucidate electromyographic (EM G) characteristics of infant walking at the onset of independent gait, we
longitudinally recorded EMGs from muscles of both legs
during the learning process of walking in an infant, from 10
months after birth until about 3 years of age. We found EMG
characteristics of infant gait up to around 1 month after
learning to walk that are not usually seen in adult gait. In
stance phase from foot contact until push off, the role of the
vastus medialis for maintaining stability became clear as a
slightly squatted position was used to lower the center of
gravity. Orderly reciprocal or co-contraction patterns of activity
in the rectus femoris and biceps femoris or in the tibialis
anterior and gastrocnemius were found to be related to
returning the body's center of mass toward its initial position.
In the latter half of swing phase, the vastus medialis and
gastrocnemius showed strong activities with the knee extending and ankle plantarllexing for active leg extension to prevent
falling. These characteristically excessive muscle activities
in infant walking are considered to express weak muscle
strength and an immature balancing system. As months
and years pass, the muscles become stronger and balance
matures, obviating the need for so much myoelectric activity.
Normal human infants begin to walk independently when they are

about 1 year of age. Thelen et al. (1989) noted that independent
walking emerges when a threshold has been reached for muscle
strength and ability to balance, but the baby who has just become able
to walk independently exhibits a pattern notably different from adult
gait. Although a great deal of investigation has been done on
development of gait, there are few EMG studies in the area. Crosssectional kinesiological EMG studies on the development of
independent gait in babies have been performed by Sutherland et al.
(1980), Forssberg (1985), and Thelen et al. (1987), but we have not
seen much longitudinal EMG study on the acquisition of gait outside
of that by Okamoto et al. (1972, 1983, 1985, 2001, 2003). By means of
both longitudinal and cross-sectional EMG and cinematographic
findings, we have reported that specific changes can be observed at
certain times in the course of that development. That is, during the
early stage of independent walking, a baby squats slightly while
leaning forward and takes steps with strong active extension of the
legs, exhibiting considerable instability. After this early stage of
independent walking, the baby exhibits increased stability with the
body tilted only slightly forward (childhood walking pattern), and by 3
years of age the body is upright as in adult walking (adult walking
pattern). What seem to be most lacking, however, are EMG studies
during the very early stage of independent walking in the infant. The
purpose here is to explore a little further the onset of independent
walking in the infant and to determine EMG characteristics of infant
walking by longitudinal observations.
The subject was one baby who first began to walk independently at
306 days after birth. We made longitudinal observations on this child
from the time she first began to walk independently at 10 months after
birth until a stable adult-like walking pattern was achieved at around 3
years of age.
Figure 2-1 shows a representative form of infant walking at the onset
of independent walking, when the infant succeeded to walk 5 to 10
steps without support. Slight knee flexion was often observed in the
supporting leg, the foot base in the double support period was very
wide, and the body's center of gravity was lowered during stance. The
arms were spread apart and elevated.
The gait in this baby first learning to walk was characterized by

quick hip and knee flexion in which the thigh became almost
horizontal in the first part of swing phase. The foot was raised forward
and slightly outward, then the foot began to reach the floor quickly,
the knee extending actively along with the hip. The foot usually
contacted the floor with the foot flat and forefoot first, but in a few
instances the heel made initial contact. A squatting posture with the
body inclined forward was often observed during stance phase. We
noticed several other characteristics that differ from adult gait, such as
a wide base at the feet and a "high guard" position of abducted arms
(Figs. 2-3, 2-5, 2-7, and 2-9).
Figures 2-2, 2-4, 2-6, 2-8, and 2-10 show longitudinal developmental
changes of EMG activity in the learning process of walking. Compared
with corresponding muscular activities of the adult walking pattern,
excessive muscular activities and variations appeared during the
learning process of infant walking from the 1st day of learning to walk
until 2 or 3 months after learning to walk. In the description that
follows, we focus attention on EMG activity patterns seen in the infant
that deviate from normal adult walking and examine developmental
changes in muscle activity related to infant independent walking.
Fig. 2-1. Gait pattern at the onset of independent walking.
Independent Waiking in infants 27
fffffr
(R)
TA
LG
VM
RF
BF
GM
(U
TA
LG
VM
RF
BF
GM
sw
ST
( R)
KNEE~
(L)
KNEE
EXT. ~
FLEX.
-----------
V
t sec
0.5 mv
Fig. 2-2. EMGs on the 1st day of independent walking (at 10 months of age).
ST: stance phase, SW: swing phase, (R): right leg, (L): left leg, TA: tibialis anterior, LG:
lateral gastrocnemius, VM: vastus medialis, RF: rectus femoris, BF: biceps femoris , GM:
gluteus maximus, KNEE EXT: extension, KNEE FLEX: flexion.
Figure 2-2 shows a representative excerpt of the EMG patterns of

the infant's independent walking on the day when she succeeded in
walking 5 to 10 steps for the first time, at 10 months after birth.
In stance phase, at the ankle, a pattern of two or three alternating
bursts between the TA and LG was most prevalent, but co-contraction
of both muscles was also seen frequently. At the knee, the VM was
continuously active from foot contact until push off. At the hip and
knee, three types of discharge pattern were seen in the biarticular RF
and BF muscles. One was a reciprocal (RF-, BF+) pattern in which
discharge of the RF tended to decrease or disappear while that of the
BF increased. A second was a reversed reciprocal (RF+, BF-) pattern
in which discharge of the BF tended to decrease or disappear while
that of the RF increased. The third was a co-contraction (RF+, BF+)
pattern of the two muscles. When the infant became able to walk
continuously, we generally found a reciprocal or co-contraction pattern,
although we occasionally observed a reversed reciprocal pattern. At
the hip, the GM was continuously active.
In swing phase, the LG and VM often showed strong activity in the
latter half of the phase.
Fig. 2-3. Foot prints on the 1st day of independently walking (at 1 year 1 month).
sw
ST
10.5 months
( R)
TA
LG
VM
RF
SF
GM
(L)
TA
LG
VM
RF
SF
GM
ST
( R)
KNEE
( L)
KNEE
sw
EXT
.
FLEX.
1 sec
Fig. 2-4. EMGs at 2 weeks after learning to walk (at 10.5 months).
0.5 mv
Figure 2-4 shows representative EMG patterns of infant walking at

about 2 weeks after learning to walk (at 10.5 months after birth), when
the infant was able to take more than 20 steps.
In stance phase, at the ankle, the earlier pattern of two or three
alternating bursts between the TA and LG changed to one or two
alternating bursts, but co-contraction of the two muscles was also seen
frequently. At the knee and hip (the VM, RF, BF, and GM), EMG
patterns in this period did not differ from those on the 1st day of
learning to walk (Fig. 2-2).
In swing phase, the LG and VM frequently showed strong activities
in the latter half of that phase, as on the 1st day of learning to walk.
Fig. 2-5. Unstable infant independent walking.
At around 1 month after learning to walk
11 months
( R)
TA~~~~~~~~~~~~mM~~~~~~-+~WM~
GM+-~~~~~~~~~~--~--~~~-r~~~~+-~~~
(L)
TA~~~~NH~~~~~~~~~~~~~~~~~
( R)
KNEE
EXT. ..
( Ll FLEX.
KNEE ---~
/---1 sec
1 month after learning to walk
Fig. 2-6. EMGs at 1 month after learning to walk (at 11 months).
0.5 mv

around 1 month after learning to walk (at 11 months after birth). At
this point, the infant began to walk by herself for long periods.
In stance phase, at the ankle, the previous pattern of one or two
alternating bursts between the TA and LG disappeared, but cocontraction of both muscles was also seen frequently. The reciprocal
(fA-, LG+) pattern tended to increase, and co-contraction (TA+, LG+)
of both muscles tended to be seen at about the same frequency as at 2
weeks after learning to walk (Fig. 2-4), but the reverse reciprocal (TA+,
LG-) pattern began to decrease or disappear. At the knee, VM activity
tended to decrease or disappear. At the hip and knee, although the
reciprocal (RF-, BF+) pattern increased, the reverse reciprocal (RF+,
BF-) and co-contraction (RF+, BF+) patterns tended to occur much less
frequently than at the onset of independent walking. At the hip, activity
of the GM in this period did not change from the pattern at the onset
of independent walking (Figs. 2-2 and 2-4).
In swing phase, discharges of the VM began to decrease in intensity
or even disappear in the latter half of that phase, in contrast to the
situation at the onset of independent walking (Figs. 2-2 and 2-4).
Discharges of the LG, on the other hand, still remained strong in the
latter half of swing phase.
21st day after learning to walk
43rd day after learning to walk
Fig. 2-7. Foot prints of initial infant walking on the 21 st and 43rd days after learning to walk (at
1 year 1 month).
From 2 to 3 months after learning to walk
( R)
!il"~
~11
1'1'
11\'
ld'llLtk
LG ,~.
VM
RF
GM
.I.
l' \
LG 1.1
I.
VM
111
3'I.lj~h'.
hUll,
1'1'
'If"
.L IdlL~.!
"I fU""llr
~II"
!'I
I.
.1,lhlLJ
"11.1.
"1'''
I'
JL
.~.
"
.~ J~u
"d. ,1J.Jlil, I
"'.Jj.
'Ir, I""",!,'
'11r'
'JI.J...iIIo,
,1I n"r'
~~h:'
llllll
I,~I.
~" .
I''''' ['or
UUi,
:r'll"
I~"
,t...
II II~jJll
'r'l'II"I"
~rJl
1.1 'ILl,
~L~
I 'lI'm
"
1,\ """'1..".,
"111 1
ST
~T.
'"I:'~U
.oiJ
'r~
~,UJ III.
'I'
~..
dj~J.
l"'lr
~
r
~LlLl.
11"'\lf
11/ It
""r'l"
",i"
RF
,1
..J.'ll...
kL.L 1\
'w
c., .. Ji!b
'\ "r\'
11"
r .".
Ilr!\1~ql
iJI.l
1'1"1
TI' I"
,I.
l....i1,dl
11'
"1'
II
"'11"
~,~I .Il,lil",
(L)
~I jJo..
I'
I.o!.
.~ 1..11,
'II " II'" '11'
".Ii""L
SF
I~ IJI
IU., !1~
Ill' !'r
'r
'f'
111''1I\~J.ilal'
l~h.J,
1\111
'1'''''
-"
TA 11/
Il.Iil.
frr".r'"'
11'
GM
1!111 I~.
TA
SF
12 months
SW
ST
~""""'\J
'IIITI'H
,r.I~I~1
ItJ,
.L
rm"
'1"
...
.1
I
It.
Ilfll'
'./.~j ,JIJld.
II 'I'T''''
..L [,IIJljA,
~(l
,~~r'
,lIl.
'I
.lJ,.
,1"""
~r'"
SW
(R)
KNEE
, sec
2 months after learning to walk
Fig, 28, EMGs at 2 months after learning to walk (at 12 months),
lJ,,,
,I.
~~, '~f l'll i~u ~ Jl:

1 ,.
~~
'1'
j~lI..lI..1c
'It
'''I

about 2 months after learning to walk (at 12 months after birth). The
infant had acquired comparatively stable walking.
In stance phase, at the ankle, the co-contraction (TA+, LG+) pattern
began to decrease in frequency, whereas the reciprocal (TA-, LG+) and
reversed reciprocal (TA+, LG-) patterns were the same as at 1 month
after learning to walk (Fig. 2-6). At the knee and hip, there were no
obvious changes in EMG patterns of the VM, RF, BF, and GM.
In swing phase, although strong discharges of the VM decreased or
even disappeared in the latter half of that phase, discharges of the LG
still remained the same in the latter half of swing phase as at 1 month
after learning to walk (Fig. 2-6).
TO
sw
Fe
1 year
ST
TO
SW
1 sec
FC
1 year 3 months
ST
0.5 mv
1 week after learning to walk
1 sec
0.5 mv
3 months after learning to walk
Fig. 2-9. EMGs in mutual antagonists (TA versus LG) of infant independent walking.
TO: toe off, Fe: foot contact, SW: swing phase, ST: stance phase,
Left: at 1 week after
learning to walk (at 1 year), Right: at 3 months after learning to walk (at 1 year 3 months).
Muscle activity progressed from excessive co-contraction of mutual antagonists to reciprocal
patterns.
Subsequent development
TO
He
TO
HC
. 11.,
TA
.1 ""
!,...~
I I~'II'J\'
~/~
LG
'JliLJ~jd ~
..Il ,U~
"11' '11"
~ ~.
VM
.I
RF
.l/k.
BF
'\1'''
. 1,
rr
"1'"
GM
SW
ST
1 sec
SW
10.5mv
-------'
1 year 9 month s
IMMATURE CHILD
WALKING PATTERN
ST
__,_,e_c_->I mv
0.5
3 years 2 months
MATURE ADULT
WALKING PATTERN
Fig. 210. EMGs of the learning process of walking.

TO: toe off, He: heel contact, SW: swing phase, ST: stance phase, Left: at 1 year 9 months
(immature child walking pattern), Right: at 3 years 2 months (mature adult walking pattern).
Figure 2-10 (left panel) shows representative EMG patterns of

immature childhood walking at 1 year 9 months of age. The infant
acquired comparatively stable walking with the body inclined forward
(Fig. 2-11) .
In stance phase, at the ankle, a reciprocal (fA-, LG+) pattern was
observed most often. Reciprocal (RF-, BF+) patterns were also seen at
the hip and knee, and continuous activities of antigravity muscles (LG,
BF, and GM) were found. In swing phase, discharges of the LG seen at
around 2 or 3 months after learning to walk decreased or disappeared
in the latter half of swing phase and more greatly resembled the usual
adult walking pattern.
Discharge patterns of the leg muscles did not appreciably change
from 3 months after learning to walk until approaching the third year
of age.
Figure 2-10 (right panel) shows representative EMG patterns at 3
years 2 months of age, resembling mature adult walking. At this point,
the infant appeared to have acquired the adult walking pattern using a
strong push-off of the foot with the body erect (Fig. 2-11).
In stance phase, at the ankle, reciprocal (TA-, LG+) patterns
previously found in the first half of stance phase decreased or
disappeared and strong bursts were observed instead in the latter part
of stance phase, as in adult walking. At the knee and hip, reciprocal (RF-,
BF+) patterns decreased or disappeared. Strong continuous discharges
of the LG, BF, and GM, that had been seen until about the end of 2
years of age, began to decrease or disappear. EMG activity patterns
that decreased or disappeared at around 3 years of age were closely
approximating adult forms.
IMMATURE
INFANT WALKING
PATTERN
..
IMMATURE
CHILD WALKING
PATTERN
up to 3 months
after learning to walk
3 months - 2 years
after learning to walk
1 year - 1.3 years
1.3 years - 3 years
..
MATURE
ADULT WALKING
PATTERN
after 2 years
of learning to walk
3 years -
Fig. 2-11. Development of gait pattern from infant walking to mature walking.
GM~~~n~~~--~~~~~~~~MM~~~~~~~~~
( L)
TA
::~~~
GM..
... 11,
~~.
",., ..
~'" I",."""~,,,~ .Jr~!..."l~,
"'-1111-'....._ _ __
( R)
KNEE-------."
( L)
EXT .
FLEX.
K N E E - - - - -......
( R)
FC
( L)
FC
FF
HC
TC
1 sec
0 .5 mv
Fig. 2-12. EMGs of standing posture with a slight squat on the 1st day of independently
walking (at 10 months).
(R): right leg, (L): left leg, FF: foot flat with the body erect, HC: heel contact with the body
inclined backward, TC: toe contact with the body inclined forward.
Figure 2-12 shows EMGs of standing with a slight squat on the 1st
day of independent walking. These discharge patterns were similar to
those during stance phase on the same day (Fig. 2-2).
During maintenance of standing posture at the ankle, alternative
bursts between the TA and LG generally showed a reciprocal (TA-,
LG+) pattern at toe contact (fC) with the body inclined forward, and a
reversed reciprocal (fA+, LG-) pattern at heel contact (HC) with the
body inclined backward. Occasionally a co-contraction (TA+, LG+)
pattern was seen at toe contact (TC) with the body inclined forward.
At the knee, the VM showed continuous strong activity during
maintenance of slight knee flexion. At the hip and knee, the three
discharge patterns (reciprocal, reversed reciprocal, and co-contraction)
between biarticular muscles (RF and BF) could be seen. At the hip,
the GM generally showed continuous activity during the maintenance
of standing.
Fig. 2-13. Standing posture just before independent walking at 1 year of age.
Discussion
When a baby is just beginning to walk, characteristic EMG patterns
can be seen that are excessive when compared to the corresponding
patterns in adults. We consider here certain EMG patterns that
gradually changed from the time of first learning to walk, principally
those in stance phase and in the latter part of swing phase.
In stance phase, we have found that excessive muscular activity and
patterns peculiar to gait in an infant who has just begun to
independently walk, strongly resemble lower limb activity during
maintenance of an upright standing posture in the same period of
development (Figs. 2-2, 2-4, and 2-12), suggesting that a common
mechanism operates both in standing and in the initiation of gait. From
a mechanical point of view, at this very early stage, both activities
require a low center of gravity and a wide base of support to assure
maximum stability. Generally these tasks can be accomplished, even
though strength and balance are yet undeveloped, by spreading the
legs apart to widen the base of support and by maintaining the knees
in slight flexion to lower the center of gravity. During knee flexion in
stance phase, continuous discharges of the VM are generally seen
until around 1 month after learning to walk (Figs. 2-2 and 2-4). In
stationary standing, the VM is continuously active as the baby stands
fairly squatted on the 1st day of independently walking (Fig. 2-12). The
VM activity seen at the onset of independent gait thus appears to
contribute to holding a posture with slight knee flexion, permitting the
body's center of gravity to be lowered so that balance is easier to
maintain. Mer the first month of walking, such continuous discharges
of the VM tend to decrease or disappear (Figs. 2-6, 2-8, and 2-10). This
agrees with observations by Okamoto et al. (1985, 2001, 2003) that the
load at the knees decreases as strength and balance develop.
Another important factor to consider is keeping the vertical
projection of the body's center of gravity well within the bounds of the
base of support. In our study, the baby who had just begun to walk
independently exhibited control over inclination of the trunk during
walking or standing, thus keeping the center of gravity within the base
of support, by orderly patterns of activity in the leg muscles
(Figs. 2-2, 2-4, and 2-12). As mentioned above, three types of discharge
patterns were seen in the biarticular RF and BF muscles. First, the
reciprocal (RF-, BF+) pattern is considered to be necessary for gait
with an anteriorly inclined trunk. Before strength and balance have

matured to the point that push off can be effectively used with the
trunk upright, as in adult gait, this pattern tends to increase after 1
month of learning to walk. This pattern is similar to a child's walking
pattern (Fig. 2-10, left panel). Second, the reversed reciprocal (RF+,
BF-) pattern is considered to help control displacement of the body's
center of mass by participating in maintenance of posterior inclination
ofthe trunk. Third, a co-contraction (RF+, BF+) pattern is considered
to keep balance control with the body erect. The reversed reciprocal
and co-contraction patterns are normally seen during the very unstable
period of the first month after beginning to walk (Figs. 2-2 and 2-4),
but not thereafter. These patterns are not seen in the child or adult
walking pattern. These EMG patterns thus suggest that excessive
muscular activity is a characteristic feature of balance control when the
baby takes steps for the very first time. While these two muscles (RF
and BF) act at the hip and knee, Nashner et al. (1985) have pointed
out that ankle strategy is the most efficient for returning the body's
center of mass to its initial position. Indeed, in our study the TA and
LG exhibited alternating reciprocal patterns of activity, thus affording
anteroposterior control over the center of gravity to help maintain
upright stability. That is, activity of the TA is considered to participate
in maintenance of posterior inclination of the trunk, while activity of
the LG is considered to be necessary for gait with an anteriorly
inclined trunk. We also found variations in the alternating reciprocal
patterns of the ankle muscles at about 2 weeks after learning to walk
(Fig. 2-14). From the viewpoint of the developmental process, it clear
that two or three alternating bursts of these muscles (TA and LG),
seen in the very unstable period at the onset of independent walking
and stationary standing, disappear at around 1 month after learning to
walk (Figs. 2-2 and 2-6). The fact that this alternating burst pattern
becomes attenuated with experience of walking further suggests that it
is a characteristic EMG feature of balance control when the baby takes
steps for the very first time. The TA and LG have previously been
reported to co-contract in many instances at the onset of independent
walking, and McGraw (1940) pointed out that co-contraction of these
mutual antagonists is indicative of maintaining balance by strongly
stabilizing the ankle. However, basograms recording using foot contact
switches during stationary standing (Fig. 2-12), when the trunk was
markedly inclined forward, suggest that the TA in synchrony with the
LG acts for inversion to actively prevent falling. In addition to the
pattern of two or three alternating bursts of the TA and LG, as mentioned above, co-contraction of ankle muscles can be considered the
expression of an immature balancing system.
It would be very difficult for an infant to maintain a prolonged single
stance phase at the onset of independent walking. In the adult walking
pattern, strong myoelectric discharges during single leg support are
hardly seen from foot contact until push off. In contrast, excessive
discharges at the onset of independent walking in infant are often
observed during single leg support. During single leg support, as
shown in Table 2-1, up to around 1 month of learning to walk, the
anteriorly located muscles of the lower limb (TA, VM, and RF) are just
as active as the posteriorly located muscles (LG, BF, and GM). But
after a full month of walking, activity of the anterior muscles tend to
disappear. On the other hand, reciprocal EMG (TA-, LG+ and RF-,
BF+) patterns seen in childhood gait become more prevalent. Reversed
reciprocal EMG (TA+, LG- and RF+, BF-) patterns disappear and are
not seen in child and adult gait patterns. This suggests that excessive
activity of the anterior muscles indicate marked instability, whereas
excessively activity of the posterior muscles should be associated with
a lesser degree of instability.
In swing phase, up to the first month of walking, the VM (a knee
extensor) is generally active from the middle of swing phase until the
subsequent foot contact (Figs. 2-2 and 2-4). The LG (an ankle
plantarflexor) is likewise active in this part of swing phase during
about the first three months of independent gait (Figs. 2-2, 2-4, 2-6, and
2-8). Compared to the situation of standing on both feet, these patterns
occur when only the contralateral leg is providing a very small base of
support, and the airborne foot is being actively plantarflexed while the
knee is being actively extended, suggestive of operation of the
protective parachute reflex to prevent falling.
It thus becomes clear that when a baby first begins to walk, muscle
activity plays a relatively great role in providing stability to maintain
posture and to keep the body's center of gravity low and within the
base of support. From the early stages of walking, the muscles become
stronger and balance matures as months and years pass, obviating the
need for so much myoelectric activity. Thus some patterns of EMG
activity can be identified that are present in infant walking but
are subsequently no longer present in child or adult gait. As the baby
matures, these excesses gradually become refined until, at about three
years of age, they very much resemble muscle activities of adults.
Table2-1. Developmental changes of EMG pattern during single leg support

Joint
Ankle
EMG pattern
1st day
2 wks
1 mon
2-3 mons
Reciprocal (TA-, LG+)
(+)
(+)
(++)
(++)
Reversed Reciprocal (TA+, LG-)
(++)
(+)
(-)
(-)
Co-contraction (TA+, LG+)
(+)
(+)
(+)
()
Continuous (VM+)
(++)
(++)
()
()
Reciprocal (RF-, BF+ )
(+)
(+)
(++)
(++)
Reversed Reciprocal (RF+, BF-)
()
()
(-)
(-)
Co-contraction (RF+, BF+)
(+)
(+)
()
()
Continuous (GM + )
(++)
(++)
(++)
(++)
Knee
Knee & Hip
Hip
Frequency of occurrence, (++): very much, (+ ): much, (): a little, (- ): little.
sw
ST
sw
ST
sw
ST
1 sec
0.5 mv
Fig. 2-1 4. Variations in EMG pattern of ankle joint muscles at 2 weeks after learning to walk (at
10.5 months of age).
ST: stance phase, SW: swing phase, TA: tibial is anterior, LG: lateral gastrocnemius, Left (ST-l):
two or three alternating bursts between the TA and LG, Center (ST-2): one or two alternating
bursts between the TA and LG , Right (ST-3): one continuous discharge pattern.
Conclusion
To determine EMG characteristics of infant walking, we longitudinally recorded EMGs using surface electrodes from twelve
muscles of both legs in an infant from 306 days after birth.
Up to around 1 month after learning to walk, in stance phase the
VM showed activity associated with holding a slightly flexed knee
joint. Alternating reciprocal patterns between the RF and BF muscles
came into playas the body inclined backward and forward, whereas a
co-contraction pattern of both muscles appeared when the body was
erect. Alternating reciprocal patterns between the TA and LG helped
to maintain balance and to prevent falling backward or forward. Cocontraction patterns of these two muscles were seen to stabilize the
ankle joint to maintain body balance, preventing strong forward falling.
In the latter half of swing phase, the VM and LG showed strong
activities with the knee extending and the ankle plantarflexing to
prevent falling.
These characteristically excessive discharge patterns of infant gait
were not seen in subsequent childhood gait or in adult gait, and they
began to decrease or disappear after about 1 month of learning to
walk. It is in this sense that these leg muscle activities are considered
EMG characteristics of infant walking at the onset of independent
walking.
Electromyographic (EMG) recordings of the lower limbs

were made from a girl from 3 weeks after birth until 8 years of
age to determine EMG changes in the development of human
bipedal locomotion. Recordings were taken from the tibialis
anterior (TA), lateral gastrocnemius (LG), vastus medialis
(VM), rectus femoris (RF), biceps femoris (BF), and gluteus
maximus (GM) muscles. In each of three developmental
stages of gait, primitive walking, supported walking, and
independent walking, muscle activity progressed from
excessive co-contraction of mutual antagonists to reciprocal
patterns. For the stance limb, the predominant reciprocal
pattern to emerge was continuous activity of the posteriorly
located LG and BF as opposed to the anteriorly located TA and
RF In independent walking this preponderance of maintained
activity by the LG and BF in stance phase gradually waned
over the first 2 years of walking to focused bursts of activity.
The developmental changes observed in this girl appear to
have been attributable to changes in posture reflecting
increased strength and to improvements in control of balance
reflecting neuromaturation.
During the first three years of life, human bipedal locomotion

develops gradually toward mature walking throughout a series of
phases: newborn stepping, infant supported walking, infant
independent walking, and child walking (Fig. 3-1). In the 20th century,
some studies have provided detailed technical descriptions (kinematics,
kinetics, temporal events, and electromyography) of the developmental
process of infant locomotion, although to study gait in babies using
adult techniques is very difficult.
McGraw (1940) analyzed seven selected phases in the development
of erect locomotion from newborn stepping to mature erect walking,
using film analysis, and pointed out the relations between several
reflexes and the development of motor behavior. Touwen (1976)
clarified the interactions between reflexes and the development of
motor behavior, emphasizing the longitudinal study of motor development. Using EMG can provide information about the maturation of
gait that is both significant and otherwise unavailable in conventional
motion analysis.
Although the study of human locomotion in infants using EMG is
difficult, some cross-sectional and longitudinal EMG studies on the
development of gait have been done. Forssberg (1985), Thelen et al.
(1987), and Okamoto et al. (1972, 1985,2001,2003), have studied the
developmental process from newborn stepping until infant supported
walking prior to independent walking, and Sutherland et al. (1980) and
Okamoto et al. (1972, 1985, 2001, 2003) have researched the learning
process from early infant independent walking to mature walking.
These studies have generally described developmental changes of
various leg muscular activities in both supported and unsupported
walking. We are unaware, however, of any studies that have described
EMG developmental changes from newborn stepping all the way to
mature walking longitudinally in the same individual.
The purpose of this study was to study longitudinal developmental
changes of human locomotion in terms of leg muscle activity. EMGs of
the same subject were recorded over a period of 8 years, from 3 weeks
after birth to 8 years of age, so that the entire span of gait development
could be examined in one individual.
We made longitudinal observations on a female infant from 3 weeks

after birth until 8 years of age. At the beginning, to induce newborn
stepping, the examiner held the infant under the arms with the soles
of the feet touching a horizontal flat surface. Well-coordinated walking
movements were observed fairly consistently from shortly after birth
to around 3 or 4 months. Although newborn stepping could not simply
be arbitrarily elicited at the will of the examiner, we were able to
induce selected well-coordinated walking movements of three or more
steps during this period.
From 3 weeks after birth to 3 years of age, EMGs were recorded 38
times, at intervals ranging from 2 weeks to 2 months. After that, from
3 to 8 years of age, EMGs were recorded 10 times, about every 6
months.
Based on the longitudinal EMG findings of the present investigation,
as well as those from previous studies (Okamoto et al.,1972, 1985,
2001), we divided the early development of gait into the following four
phases: neonatal stepping, onset of young infant stepping, initial young
infant stepping, and infant supported walking. Subsequent maturation
of gait was also divided into four phases: onset of infant walking, initial
infant walking, immature child walking, and mature walking. The data
in Figures 3-2, 3-4, 3-5, 3-6, 3-7, and 3-8 show representative EMG
patterns and forms from our longitudinal observations of the same
subject (Figs. 3-1 and 3-3).
t1tft~f frffff
Itlrftfi
llUli ~ i I f~e It~l~~

Jilit/fA j 1111l\~lk
Fig. 3-1. Developmental changes of gait in one individual (birth to age eight).
Top : neonatal and infant stepping, Middle: infant supported and independent walking,
Bottom: child walking (same subject).
From Newborn Stepping to Mature Walking 47
Neonatal stepping
(up to 4 weeks after birth)
TO
FC
11 .
TA
..".
\." ..
LG
VM
RF
BF
j,
..1
,J.
,.I..
~~
." ....,"
GM
SWING (SW)
STANCE (ST)
1 sec
I 0.5 mv
3 weeks
Fig. 3-2. EMGs of neonatal stepping (at 3 weeks after birth).

TO: toe off, Fe: foot contact, SW: swing phase (short phase), ST: stance phase (long
phase), TA: tibialis anterior, LG: lateral gastrocnemius, VM: vastus medialis, RF : rectus
femoris, BF: biceps femoris, GM: gluteus maximus.
Fig. 3-2 shows EMG patterns of leg muscles at 3 weeks after birth.
The stepping in this period was characterized by quick hip and knee
flexion in which the thigh became horizontal in the middle part of
swing phase. The foot dorsiflexed strongly as it was brought forward.
The foot then approached the floor more slowly, the knee extending
relatively passively as the hip extended. The foot usually contacted the
floor with the lateral border first, but sometimes the heel, sole, or
forefoot made initial contact instead. The supporting leg was relatively
flexed during stance phase.
The TA, RF, and BF exhibited notable myoelectric activity as the
ipsilateral foot was leaving the floor to begin swing phase. The TA
continued to be active throughout much of swing phase, whereas the
RF showed no more than sporadic weak activity during that period,
and the BF was relatively silent until stance phase was being
approached. The LG, VM, and GM did not show any remarkable
activity during swing phase. During stance phase, the LG, BF, and GM
showed relatively continuous activity as antigravity muscles. The VM
and RF tended to be active when knee flexion was not very pronounced, that is, during the double-stance phases. In single-stance
phase, activities of mutually antagonistic muscles (TA versus LG and
RF versus BF) showed reciprocal (TA- and LG+, RF- and BF+) , cocontraction (fA+ and LG+, RF+ and BF+), and reversed reciprocal (fA+
and LG-, RF+ and BF-) patterns, but activities of certain mutually
antagonistic muscles were variable and inconsistent (Fig. 3-12).
Fig. 3-3. Neonatal stepping at 3 weeks after birth.

(from 1 to 5 months of age)
ffffrJf Ir~~I~
TO
TA
Fe
TO
Fe
lilii!
TO
Fe
U'"'
.1,
LG
VM II'
,j;..
RF
BF
.,1,
F
II
GM
.."
sw
1 sec
I 0.5 mv
1.5 months
SW
ST
SW
ST
1 sec
3.5 months
I 0.5 mv
ST
t sec
I 0.5 mv
5 months
Fig. 3-4. EMGs of young infant stepping (Left: at 1.5 months after birth, Center: at 3.5
months after birth, Right: at 5 months after birth).
1 ) Onset of young infant stepping (from 1 to 2 months of age)
Fig. 3-4 (left) shows EMGs at 1.5 months after birth. Step frequency
was more regular than during neonatal stepping. Hip flexion was
pronounced in the first part of swing phase as in neonatal stepping,
but leg extension began to be more vigorous in the latter part of swing
phase. The lateral border of the forefoot initially contacted the floor for
the most part.
In swing phase, continuous activity of the TA tended to terminate
sooner in the latter part of swing phase than during the neonatal
period, but bursts of the RF and BF during the phase showed
the same tendencies as before. The LG and VM, which had
been quiet during swing in neonatal stepping, began to exhibit
activity shortly before foot contact, and sometimes activity in the
GM was also observed before foot contact. During stance phase,
continuous bursts or discharges of the VM were seen often, as in
neonatal stepping. The posteriorly located LG, BF, and GM showed

continuous discharges that were comparatively variable in intensity. In
single stance, the reciprocal pattern was observed as in neonatal
stepping. Incidence of the co-contraction pattern decreased from that
at the time of neonatal stepping, and the reversed reciprocal pattern
was hardly seen (Fig. 3-12).
2) Initial young infant stepping (from 3 to 5 months of age)
Fig. 3-4 (center) shows EMGs at 3.5 months after birth. Stepping in
this period was performed actively as foot contact was audible. Hip
flexion was again conspicuous in the first part of swing phase as it was
in the neonatal period and in the onset of young infant stepping. Leg
extension in the latter part of swing phase was performed in a more
active manner than in the onset of young infant stepping, often quite
visibly so. Initial contact of the floor was, for the most part, by the
forefoot. The supporting leg, especially the knee, tended to be
relatively extended in this period.
In swing phase, the TA, RF, and BF became less active at toe off
than they did in the onset period of young infant stepping. This was
especially true for the TA, but that muscle would always exhibit at
least a weak burst in the first part of swing phase. In the latter part of
swing phase, discharge patterns of the LG and VM tended to increase
frequently. During stance phase, continuous bursts or discharges of
the VM were seen often, and the LG, BF, and GM showed continuous
discharges of variable intensity, as they did in the onset of young infant
stepping. Mutually antagonistic muscles in single stance exhibited the
reciprocal pattern more readily than previously. The co-contraction
pattern was hardly seen, and the reversed reciprocal pattern was not
seen at all in this period.
Fig. 3-4 (right) shows EMGs at 5 months after birth. Stepping was
actually difficult to elicit at that age, but we did manage to induce it
and the resulting EMG pattern was basically similar to the initial
period of young infant stepping at around 3.5 months after birth.

(from 6 to 12 months of age)
TO
TO
Fe
ft~jtrtt ~~liil
Fe
TO
Fe
"",
'I
TA+-~--------~~
.lIi,/"Il.
LG+-~~~~~~~
, d,~
'1"'1"'1"
'II'
,',!",'
VM +-~--------~~
RF+-~--------~~
.1....,
"11 "1
BF+-~~~~--~~
I~
GM+-~--------~~
sw
sw
ST
1 sec
0.5 mv
6 months
sw
ST
1 sec
9 months
.J
'\1'
11
I 0.5 mv
ST
1 sec
O.S my
11.5 months
Fig. 3-5. EMGs of infant supported walking (Left: at 6 months after birth, Center: at 9 months
after birth, Right: at 11,5 months after birth),
Fig. 3-5 shows EMGs at 6, 9, and 11.5 months after birth. After
around 6 months after birth, the infant began to crawl after toys and
tended to show comparatively stable stepping when supported upright.
At around 11 months, 1 month before walking independently, the
infant became able to stand by herself and to walk with one-handed
support. The relatively pronounced flexion of the hip seen in the first
part of swing phase of the previous period was slightly reduced. Active
leg extension tended to disappear in the latter part of swing phase.
The heel usually contacted the floor first. The supporting leg was
extended.
In the first part of swing phase, activity of the RF was sporadic and
bursts of the BF disappeared. In the latter part of swing phase, the LG
and VM showed minimal activity at 6 and 9 months, but shortly before
independent walking (at 11.5 months after birth), strong myoelectric
discharges of the LG and VM were sometimes seen. During stance
phase, the continuous activity of the VM seen in earlier stages of

development no longer appeared. The LG and BF, antigravity muscles,
were markedly active in stance phase throughout this period of
supported walking (Fig. 3-6). Mutually antagonistic muscles during
this period seldom exhibited co-contraction patterns, but rather worked
in reciprocal patterns as a rule.
621
385 days
0.5 mv
TA
LG
VM
RF
8F
GM
KNEE
Ext.
Flex.
FC~
600
700
1st day of independent walking
Fig. 3-6. EMGs on the 1st day of independent walking (at 1 year of age).
SUPPORT: supported walking, INDEPENDENT: independent walking.
Infant walking
(from 1 week to 2 months after learning to walk)
j~R'~~ i~A~~'
TO
Fe
jl9Ajil
TO
Fe
"I.
TA
,II. '",1, IIIh~LL. [1'1'.1,
LG
1'1
VM
RF
""
~Ii.
BF~~\~~~~~-+~'
GM~~~~~~~--~~
sw
ill.
"~:~',r
sw
ST
1 sec
0.5 mv
1 year
HI
111I~I'r
UI.._
TI~"
.11,
'I,'
sw
ST
1 sec
0.5 mv
1.1 years
ST
_-,-,1s"",
ee'----.J
0.5 mv
1.3 years
Fig. 3-7. EMGs of learning process of infant independent walking (Left: 1 week after learning
to walk at 1 year, Center: at 1 year 1 month, Right: at 1 year 3 months).
1) Onset of infant walking (up to 4 weeks after learning to walk)
Fig. 3-7 (left) shows EMGs of 1 week after learning to walk at

around 1 year of age. The infant was able to walk more than 10 steps
without external support. The walking in this period was characterized
by quick hip flexion in which the thigh was raised forward diagonally
in the first part of swing phase. Then the foot reached the floor
quickly, the knee extending actively along with the hip. The foot
usually contacted the floor with the forefoot or mid-sole first, but
occasionally the heel made initial contact. Slight knee flexion was often
observed in the supporting leg, the foot base in the double support
period was very wide, and the body's center of gravity was lowered
during stance. The arms were spread apart and elevated (Fig. 3-6).
At the beginning of swing phase, the TA and RF were strongly

active, often accompanied by moderate activity of the BF that had
continued from the previous stance phase. Later on in swing phase,
the LG, VM, BF and GM all generally became active and, just before
foot contact the TA might exhibit a burst of activity. Stance phase was
characterized by large amounts of activity in the muscles investigated,
variously intermittent or continuous for a given muscle, so that mutual
antagonists displayed not only reciprocal patterns in single stance, but
also co-contraction and reversed reciprocal patterns (Fig. 3-12).
2) Initial infant walking (from 1 to 2 months after learning to walk)
Fig. 3-7 (center) shows EMGs of 1 month after learning to walk, at 1

year 1 month of age. The infant was becoming accustomed to walking
and began to walk by herself for extended periods of time. The femur
was not lifted so high in the middle part of swing phase as at the onset
of independent walking. Leg extension in the latter part of swing phase
was active. The forefoot initially contacted the floor for the most part,
but occasionally the entire sole made initial contact. There was a
diminution in base width and a gradual decrease in the slight flexion at
the knee of the supporting leg, and the body's center of gravity was
slightly higher during stance. The upper extremities, which had been
extended and abducted earlier, now began to approach the body.
At toe off, just as swing phase was beginning, most muscles
exhibited the same patterns of activity as at the onset of independent
walking 1 month before. During stance phase, activities of the muscles
become less sporadic. Continuous discharge of the VM began to
appear less and less (Fig. 3-9). Strong maintained activities of the TA
and RF seen at onset of infant walking became markedly moderated,
whereas in the LG and BF such activities remained unabated though
more consistent in nature during stance phase (Fig. 3-9). In singlestance, the discharge patterns of mutual antagonists described in
Figure 3-12 showed the reciprocal pattern more often than in the initial
period of infant independent walking. The co-contraction pattern
decreased comparatively, but was sometimes still seen, while the
reversed reciprocal pattern was hardly seen at all in this period.
Immature child walking

Unsettled muscle activity
(from 3 months to 2 years after beginning to walk)
AJAlA II ~ll l ~~U~

TO
TA
He
TO
He
TO
He
,.1
IH.
Ill'
IJI~
LG
1,/
' '''"
AL.
'lr'
'WI
W-
VM
RF
BF
,
'\ I"""
,,\,/,,,
I'll
.....1
GM
SW
sw
ST
1 sec
0.5 mv
2 years
ST
1 sec
3 years
SW
0.5 mv
ST
1 sec
0.5 mv
7 years
Fig. 3-8. EMGs of learning process of child walking (Left: at 2 years, Center: at 3 years,
Right: at 7 years).
Fig. 3-7 (right) shows EMGs of the child 3 months after learning to
walk, at 1 year 3 months of age. The infant had begun to shift to a
comparatively stable walking pattern of her own. The thigh was no
longer strongly lifted up in the first part of swing phase. Subsequent
knee extension, previously prominent in the latter part of swing phase,
now began to become more passive. The heel and toe began to touch
almost simultaneously. The heights of the hip joints were higher than
in infant walking, and the child began to exhibit more force to propel
her body forward. The foot base was narrowed to the width of the
shoulders. The upper extremities were still held away from the body,
although they were only slightly elevated now.
At the beginning of swing phase, the TA and RF worked in the same
way at 3 months of independent walking as at 1 month. The VM no
longer exhibited activity in the middle of swing phase after 3 months
of walking, and activity in the LG was essentially absent throughout

swing phase until the very end. During stance phase, activities in the
leg muscles changed little between 1 and 3 months after beginning to
walk, except for the GM, which exhibited less activity especially in the
latter part of this phase.
Figure 3-8 (left) shows EMGs at 2 years of age. The child had
acquired a comparatively stable walking pattern at that time and was
gaining control of movements related to running and fast walking. The
thigh was no longer lifted up prominently in the first part of swing
phase, nor did the lower leg extend rapidly in the latter part of swing
phase. The heel and toe contacted the floor almost simultaneously.
Slight knee flexion in the supporting leg was prolonged and the child
propelled herself forward with a pumping action of the thighs, the
trunk leaning forward slightly and the foot base narrowed. The
forearms were slightly elevated during gait at ordinary speed.
The TA was notably active as swing phase was about to begin, but
neither the RF nor the BF exhibited activity as the foot left the ground.
Just before the foot returned to the ground, the TA showed very little
activity. During stance phase, continuous discharge patterns of the
antigravity LG, BF, and GM were similar to those seen at 1 to 3
months after learning to walk.
VM+
LG+
Terminal SW
Terminal SW
VM+
Ak ~rt
0-

lyear
H M+
Knee extension Plantarflexion
Onset of infant walking
~~:~ ~ :~~s_.~~r~~~n~n~ ~o_~.~)_____

Initial infant walking
LG+, BF+
~F+
TA+
BFl
LO+
Squat
Backward
Forward
++)
(++ )
(++)
(++ )
(+ )
(_)
(+)
(_)
(_)
(++)
(-)
(-)
(-)
(-)
(++)
U -------n-------D.-------.[}-------. ----
(1-2 months after learning to walk)
(3 months - 2 years after learning to walk)
TA+, RF+
Throughout ST Throughout ST Throughout ST
Fig. 3-9. Developmental changes of EMG patterns in leg muscles from onset of infant
walking to immature child walking.
Frequency of occurrence, (++): very much, (+): much, (-): little.
Mature walking
Toward a mature pattern
(after 2 years of learning to walk)
Fig. 3-8 (center and right) shows EMGs at 3 and 7 years. The child
had begun to acquire stable walking resembling that of an adult in this
period. The thigh showed minimal flexion in the first part of swing
phase. The foot usually contacted the floor with the heel first and the
toes lifted like an adult. Walking with the body inclined forward was
seen until around the end of 2 years of age, when the body began to
become more erect. The child exhibited strong pushing-off motions of
the foot, and the upper extremities were no longer held in any degree
of elevation.
In swing phase, the TA was consistently active as the foot was
leaving the ground, sometimes accompanied by slight activity of the
RF. The TA also began to show marked activity just before the foot
touched the floor in many instances (Fig. 3-10) . During stance phase,
continuous activity of the LG previously found in the first half of stance
phase decreased or disappeared and strong bursts were observed
instead in the latter part of stance phase. Strong continuous discharges
of the BF and GM previously seen in stance phase began to decrease
or disappear, thus assuming activity patterns similar to those of adults.
TA+
Just before ST
X~TA+
Dorsiflexion
1.3 years
(3 months - 2 years after lea rning to walk)
3 years
Mature walking
(after 2 years of learning t o walk)
(-)
(++)
LG+, BF+
Throughout ST
al
LG+Forward
(++)
D-(-)
Fig, 3-10, Developmental changes of EMG patterns in leg muscles from immature child
walking to mature walking.
Frequency of occurrence, (++): very much, (-): little.
After birth
Development of gait
Birth
Neonatal reflex stepping
Neonatal stepping
1 month
Onset of young infant stepping
(1-2 months after birth)
2 months
(Inactive stepping)
6 months
Initial young infant stepping


1 year
1.1 years
Infant walking

(up to 4 weeks after learning to walk)
1.3 years
It

3 years
Mature walking
Mature walking
Fig. 3-11 . Developmental period of gait during newborn stepping, infant supported walking,
and independent walking.
Neonatal
st epping
Young infant
stepping
Infant supported
walking
Infant
walking
Immature child Mature

walking
walking
:: ~ ~:-
Reciprocal pattern
:: ~~:
Co-contraction pattern
:::~:::
Reversed reciprocal pattern
Fig. 3-12. Developmental changes of EMG patterns in mutual antagonists (TA versus LG and
RF versus BF) during ipsilateral single stance.
TA: tibialis anterior, LG: lateral gastrocnemius, RF: rectus femoris, BF: biceps femoris,
+: noticeable activity, - : no activity.
Reciprocal pattern: posterior muscle is active while anterior muscle is inactive, associated
with forward inclination of the trunk. At the hip, BF is active and RF is inactive. At the
ankle, LG is active and TA is inactive.
Reversed reciprocal pattern: anterior muscle is active while posterior muscle is inactive,
associated with backward inclination of the trunk. At the hip, RF is active and BF is
inactive. At the ankle, TA is active and LG is inactive.
Co-contraction pattern: muscles on both sides of a joint are simultaneously active.
Discussion
During the first 3 years of life (Fig. 3-11), movements related to
walking appear to begin with gross patterns of muscle activation,
frequently including co-activation of mutual antagonists. Not only in
supported walking and then in subsequent independent walking, but
even in neonatal primitive walking one can see over time progression
from excessive gross activation to more efficient and economical
production of muscle activities in the lower limbs (Figs. 3-12 and 3-13).
Such findings are evident in both stance and swing phases of the
walking patterns. Interestingly, even some specific changes noted in
newborn stepping over the first couple of months recur as the baby
later masters voluntary supported walking and then independent gait.
As the foot leaves the floor, activities of the TA, and to some extent
the RF, remain relatively consistent across primitive, supported, and
independent modes of walking. The latter two modes might thus be
characterized as containing a "primitive" component at toe-off. Activity
of the BF, on the other hand, at this same point in the gait cycle, varies
from mode to mode as well as within a given mode. In both neonatal
primitive walking and independent walking the BF is active along with
the TA and RF when the child first performs these modes of
locomotion, but the BF subsequently works in a reciprocal pattern
with the other two muscles as the child gains experience.
Just before the foot returns to the floor, the LG and VM are active at
around 3 or 4 months after birth and again during the first few months
of independent walking. The appearance of this activity coincides with
gradual emergence of the parachute reaction, which Milani-Comparetti
(1967) describes as appearing at about 4 months after birth, so we
need to consider the possibility that activity of the LG and VM as the
foot approaches the floor may be closely related to the parachute
reaction at 3 or 4 months. As supported walking becomes more
voluntary in subsequent months, this activity of the LG and VM is no
longer manifest, nor is it seen 1 to 3 months after first learning to walk
independently. The absence of such muscle activity appears to reflect
development of balance and postural control. These changes in activity
of the LG and VM around floor contact might be interpreted as
development from simple reflexes and subcortical motor responses
through cortical inhibition of these reflexes to a growing influence of
voluntary or cortical motor control.
The manner in which the foot contacts the floor undergoes a similar
progression in both supported walking and independent gait. At first
the forefoot initially contacts the floor, but as development progresses,
the sole of the foot makes initial contact with the floor and subsequent
to that the heel makes initial contact. A burst of activity from the TA
just before touchdown becomes more distinct as this sequence
proceeds, so it might be interpreted as an indicator of stability in gait.
In this connection, other behaviors change during this progression
that likewise reflect incremental achievement of stability in gait. For
example, width of foot placement gradually decreases as the walking
pattern becomes more stable, and a "high guard" position of abducted
arms becomes "medium guard" and eventually an adult-like "low
guard" in the process.
After the foot contacts the floor, the muscles in that lower limb can
be subjected to greater loads than is possible in swing phase. This is
particularly apparent in the VM. In the primitive mode of gait, the VM
exhibits considerable activity as the infant pushes the foot against the
floor in the extension phase of the primitive pattern. From 6 months,
however, the baby is in a supported mode of gait wherein the VM no
longer impulsively pushes against the floor and the baby simply relies
on the supporting person to bear weight during gait. The VM is not
very active even at 11 months, when the baby is close to graduating
from the supported mode to independent walking, apparently because
she has learned to passively bear weight through the knee on the
stance side when the center of gravity of the superincumbent body
segments has been brought anterior to the knee joint. At the beginning
of the independent mode of gait, however, the situation dramatically
changes as the baby suddenly finds herself solely responsible for both
maintaining balance and bearing weight. When the foot contacts the
floor and the lower limb on that side accepts body weight, the knee
remains slightly flexed, presumably to keep the center of gravity of the
superincumbent segments low, and thus make the task of balancing
easier. Only after gait in the independent mode has progressed to the
point when the baby can skillfully bring her weight over and just
anterior to the knee on the stance side, can the VM display a brief
focused burst during weight acceptance and otherwise be silent or
minimally active in stance phase.
As the baby moves forward from double support to single support
in stance phase, she encounters a more sophisticated task of
dynamically maintaining balance. The interplay between the LG and
TA, mutual antagonists at the ankle, as well as between the RF and BF,
mutual antagonists at both the hip and the knee, becomes very
important during single support. When a baby first begins to walk
without external support, co-contraction between each pair of muscles
provides gross stability to make this difficult task feasible for the
uninitiated. Because this co-contraction pattern first arises from trial
and error, the other possibilities of reciprocal and reversed reciprocal
patterns also appear (Fig. 3-12). Interestingly, all three of these
patterns appear in primitive walking as well, and we have observed
that the reciprocal pattern at this very early stage tends to appear
when the trunk is inclined forward and the reversed reciprocal pattern
when the trunk is leaning back, suggesting that muscles naturally
respond to mechanical loading. Since walking can be characterized as
a succession of incomplete forward falls, the reciprocal pattern
eventually becomes more dominant than the co-contraction pattern as
the baby attains dynamic stability in walking forward. By 3 years of
age, the refinements of activity between these mutual antagonists are
fully in place and the child propels herself forward with precisely
measured doses of muscular activity and full dynamic control of
balance.
One finding to emerge from longitudinal observations was that
developmental changes and refinements of excessive muscular activity
during newborn stepping and supported walking appear again during
the learning process of independent walking. As strength and balance
improve in a normal infant, unnecessary muscle activation disappears
leading to a series of developmental stages of bipedal locomotion in
both supported and unsupported walking (Fig. 3-13). We suggest that
the refinement of excessive co-activation, which can serve as a barometer
to indicate increasing level of skill in human locomotion, comes from
changing posture by improvement of strength and balance control
reflecting neuromaturation.
After birth
Birth
RF+
Initial SW
Dorsiflexion
I".
Hip flexion
Neonatal reflex stepping

(++)
(++)
Onset of young infant stepping

(++)
(++)
(++)
(++)
(++)
(+),(-)

(up to 4 weeks after learning to walk)
(++)
(++)

(++)
(++)

(++)
(+),(-)
(++)
(+),(-)
Development of gait
TA+
Initial SW
~~.
Neonatal stepping
1 month
2 months
................................ _.....
Initial young infant stepping


(Inactive stepping)
6 months


1 year
1.1 years
Infant walking
1.3 years
......................... _.......
3 years
Mature walking
Mature walking
Fig. 3-13. Developmental changes of EMG patterns in leg muscles during newborn stepping,
infant supported walking, and independent walking.
BF+
VM+
LG+
Initial SW
Terminal SW
Terminal SW
"'~
If L.,
Knee flexion Knee extension Plantarflexion
TA+
VM+
TA+, RF+
LG+, BF+
Just before ST Throughout ST Throughout ST Throughout ST
~,~
Dorsiflexion
-};"
~F+
TA+
BFl
LG+
Squat
Backward
Forward
(++)
(-)
(-)
(+),(-)
(++)
(+)
(+)
(++)
(+)
(+)
(-)
(++)
(-)
(+)
(+)
(++)
(++)
(-)
(++)
(-)
(++)
(-)
(-)
(-)
(-)
(- )
(-)
(++)
(+),(-)
(++)
(++)
(+),(-)
(++)
(++)
(+)
(+),(-)
(-)
(+)
(+),(-)
(-)
(-)
(++)
(-)
(-)
(-)
(-)
(-)
(-)
(++)
(-)
(-)
(-)
(++)
(- )
(-)
(-)
TA: tibialis anterior, RF: rectus femoris, BF: biceps femoris, VM: vastus medialis, LG: lateral
gastrocnemius,
SW: swing phase, ST: stance phase. Frequency of occurrence, (++): very
much, (+): much, (-): little, (+),(-): instances of noticeable activity and of no activity
intermingle.
Application to EMG biofeedback training
Infant independent walking at 1 year of age
To develop an index of gait instability from electromyographic (EMG) information, we made observations on infants
from the time they first began to walk independently at about
1 year of age until around 3 years of age. From our findings
we obtained the following criteria.
(1) Very unstable gait: As seen in a child within the first
month of learning to walk, the vastus medialis is active in the
latter half of swing phase, the tibialis anterior and rectus
femoris are active during stance phase, and activity of the
vastus medialis is continuous. These EMG characteristics
are not usually seen in subsequent childhood gait or in adult
gait, and they serve as markers of very unstable gait.
(2) Unstable gait: Activity of the gastrocnemius in the latter
half of swing phase is generally noted only within the first 3
months after the child learns to walk, and that activity is
interpreted as a sign of unstable gait.
(3) Slightly unstable gait: Activity of the gastrocnemius in
the first half of stance phase and the continuous activities of
the biceps femoris and gluteus maximus from initial contact
with the floor until push off are found in children until 3 years
of age. These activities are considered EMG markers of
slightly unstable gait.
Generally a baby becomes able to perform bipedal upright walking

at about 1 year of age. Compared to quadrupedal crawling, walking
involves maintenance of a mechanically unstable upright position and
keeping one's balance while transporting the body's center of gravity.
This requires a highly developed antigravity mechanism and operative
balance reactions.
Thelen et al. (1989) noted that independent walking emerges when a
threshold has been reached for muscle strength and ability to balance,
but the baby who has just become able to walk independently exhibits
a pattern notably different from adult gait. McGraw (1940) and
Okamoto et al. (1985, 2001, 2003), for example, have both noted that,
although initial contact of the heel on the floor can be found in babies
first learning to walk, as might be found in adult gait, they also often
contact the floor first with the forefoot, which is not characteristic of
adult gait at all. Novice walkers have been noted to have several
characteristics that differ from adult gait, such as increased cadence,
decreased step length, excessive rotation or flexion in stance phase, a
pattern of circumduction in place of the hip and knee flexion that
arises (in the adult) at the very beginning of stance phase, a lack of
accompanying arm movement, or strong co-contraction or other form
of muscle hyperactivity.
McGraw (1940), Burnett et al. (1971), and Okamoto et al. (1985,
2001, 2003) have analyzed the development of independent walking
and noted a regular progression from an initially wide base at the feet
and a "high guard" position of abducted arms as evidence of instability
in walking, giving way to a slightly narrower base and "medium guard"
posture as the walking becomes more stable, finally to an adult-like
"low guard" posture associated with stable gait.
The mechanism of the motor development of babies has usually
been studied in relation to postural development, often with, for
example, the motor development evaluation form of Milani-Comparetti
et al. (1967), which relates primitive reflexes to postural and motor
development. To study gait in babies using electromyography (EMG) ,
goniometers, and force plates as is done with adults is very difficult,
although some investigators have used film analysis to study the
development of gait. Longitudinal study of motor development,
emphasized by Touwen (1976) as very important, has been rare.
McGraw (1940) studied the relations between several reflexes and the
development of motor behavior, and Touwen (1971, 1976) has clarified
the interactions between reflexes and the development of motor
70 Application to Gait Analysis and Evaluation
behavior. Cross-sectional kinesiological EMG studies on the

development of infant walking have been performed by Sutherland et
al. (1980), Forssberg (1985), and Thelen et al. (1987), but we have not
seen much in the way of longitudinal EMG study on the acquisition of
gait outside of that by Okamoto et al. (1972, 1983, 1985, 2001, 2003).
TA
LG
RF
SF
.EMG patterns began to decrease or disappear at about 1-3 months after learning to walk
EMG patterns began to decrease or disappear at about 3 years of age
EMG patterns began to appear at about 3 years of age
IMMATIURE CHILD WALKING PATTERN
.+11]+0
11]+0
MATURE ADULT WALKING PATTERN
IMMATURE INFANT WALKING PATTERN
Fig. 4-1. Schematic diagram of EMG activity as indication of unstable walking.

TA: tibialis anterior, LG: lateral gastrocnemius, VM: vastus medialis, RF: rectus femoris, BF:
biceps femoris, GM: gluteus maximus.
An Index of Gait Instability 71
Knowing the developmental changes in the EMG features of the

process of acquiring gait in normal babies should be helpful in better
understanding the diagnosis and treatment of abnormal gait in
developmentally delayed children. EMG enables one to view the ability
to walk and to maintain balance and control during gait in ways that
would not be noticed otherwise. Even for the elderly individual, the
stages of motor development can be considered in relation to
therapeutic exercise for maintaining gait function.
To examine the role of muscle activity in the developmental process
of normal gait in babies, we used EMG with surface electrodes to
longitudinally study babies from the very unstable stage of first
learning to walk at 1 year of age to the attainment of adult-like stability
in gait at about 3 years of age. In addition to reconfirming the forward
lean with shuffling in the early stages of gait and the more erect gait
with decisive push off in stable adult-like gait, we were able to derive
from these longitudinal EMG observations an index for the instability
of gait based on developmental processes (Figs. 4-1 and 4-2) .
..
......
INFANT
PATTERN
++.
:......_-..A
IMMATURE
INFANT WALKING
IMMATURE
CHILD WALKING
MATURE
ADULT WALKING
Fig. 4-2. Refinement of excessive muscular activity during the development of gait.
Excessive muscle activities in infant walking are considered to express weak muscle
strength and an immature balancing system. As months and years pass, the muscles become
stronger and balance matures, obviating the need for so much myoelectric activity.
EMG findings during the development of gait

The subjects were three babies who first began to walk independently at about 1 year of age (at 306 days, 375 days, and 385 days after
birth). We made longitudinal observations on these three children
from the time they first began to walk independently at about 1 year of
age until a stable adult-like walking pattern was achieved at around 3
years of age. EMGs were recorded every 1 or 2 weeks in the initial
period of independent walking and after that every 1 or 2 months. For
purposes of comparison, these observations were supplemented with
cross-sectional observations taken from fifteen babies in the first year
of age (five infants at the onset of independent walking, five at 1 month
after learning to walk, and five at 3 months after learning to walk), five
babies in the second year, five more in the third year of age, and five
adults. To more precisely search for EMG characteristics of gait
stability, we also observed maintenance of standing posture in infants
in the initial period of learning to walk.
1. Learning process of walking

Fig. 4-3 shows longitudinal changes in EMG activity patterns in the
learning process of walking for subject A, who began to walk on the
385th day after birth for the first time. Fig. 4-4 shows EMGs in the
learning process of walking for subject B, who began to walk on the
306th day. In all subjects tested, excessive muscular activities appeared
from the initial period of learning to walk until around 3 years of age,
as compared with the corresponding muscular activities of adult
walking. In the description that follows, we focus attention on the
peculiar EMG activity patterns seen in the infant that deviate from
normal adult walking.
We looked for EMG patterns that gradually changed and were
refined in the transition between first learning to walk and achieving
an adult-like pattern, finding such refinements principally in stance
phase and in the latter part of swing phase. (See chapters 2 and 3,
Figs. 4-1 and 4-2) (Okamoto et al. 1972, 1983, 1985,2001,2003).
TO
TO
12.5 months
Fe
Fe
1 year 1 month
TA (Tibialis anterior)
LG (Gastrocnemius)
VM (Vastus medialis)
RF (Rectus femoris)
BF (Biceps femoris)
GM (Gluteus maximus) +--4I+</-t-"'-M.w.-'-I'{o/'I'IHI'+~#I.t~
sw
ST
ST
1 sec
I 0.5 mv
TO
He
it
He
"
T A '- I'--NIH'I'
LG
,.......-\IrIoI~'f
~1IlM-v,~II'II-f---..,-llMi\YiI
2 years 1 month
11
_----'1'-',:::,0'---------110.5 mv
1 month after learning to walk (A-2)
1st day of learning to walk (A-l)
TO
SW
He
TO
2 years 11 months
I--'
RF ........O-+\\I>.-.....-Ho!--.,.."..\O__
BF
ifi---I---IIU/If,-I-/IfM'-
GM
WIir"---t~<fr'iI'IIr-t--'fMVrSW
ST
1 sec
j,t
'['I
I'
'"
SW
10.5 mv
3 months after learning to walk (A-3)
[I"
ST
SW
_ _1:.;':::'::..
0 _ --,10.5 mv
2 years 1 month (A-4)
ST
_ --,Ic:,::::.o, ---,p.S mv
2 years 11 months (A-5)
Fig. 4-3. EMGs of learning process of infant walking (subject A).

Top: Left (A-1); 1st day of learning to walk (at 12,5 months), Right (A-2); 1 month after
learning to walk (at 1 year 1 month), Bottom: Left (A-3); 3 months after learning to walk (at 1
year 3 months), Center (A-4); at 2 years 1 month, Right (A-5); at 2 years 11 months.
11 months
~'"'I..J - -
T A (Tibialis anterior)
LG (Gastrocnemius)
VM (Vastus medialis)
RF (Rectus femoris)
B F (Biceps femoris)
GM (Gluteus maxim us) -+--'~""'i\III~~~-t---"""+-1h'-"""--l
SW
ST
ST
_---',-"
,,"'c_
SW
2 weeks after learning to walk (8-1)
1 month after learning to walk (B-2)
TO He
3 years 5 months
Jil.
TA 1-lI\I\It~~.~~
LG
10.5 mv
1 sec
-'10.5 mv
1""1
-,.~"
1!t\l't-l'/li{W-+~/Wo--r-+
1".,11.. II'r'
r,1I'
'I'~j'
,.",1. "",.
VM -+~;"'--+---f--+
RF
SW
ST
_--,-I="~C_~I 0.5 mv
3 months after learning to walk (8-3)
SW
ST
SW
ST
1 sec
_-----"-.::,=.:'c'-----'10.5 mv
1 year 9 months (8-4)
/0.5 mv
3 years 5 months (8-5)
Fig, 4-4. EMGs of learning process of infant walking (subject B).

Top: Left (B-1); 2 weeks after learning to walk (at 10.5 months), Right (B-2); 1 month after
learning to walk (at 11 months). Bottom: Left (B-3); 3 months after learning to walk (at 1 year
1 month), Center (B-4); at 1 year 9 months, Right (B-5); at 3 years 5 months.
2. Standing posture at the onset of independent walking
SF --''''*~''''''''''''ri-+1...../ -.....................'-~-tlf',""'-Ifc.. ~\I\,~,~~,,,,,,,,~,,,,\",,I''1'1<~'''''''''---~-+-
~~\Wlu.'tI~lil'
GM
KNEEEXT.~_--_ _ _ _ _ _ _ _ _ - - -FLEX.
-------
.-----,
FC--------------"
(Foot Contact)
Foot Flat
Toe Contact
t sec
Foot Flat
I 0.5 mv
Heel Contact
1 sec
I 0.5 mv
STANDING POSTURE
Fig, 4-5, EMGs of standing posture at the onset of independent walking,

Left: standing posture with pronounced squat on the 1st day of independently walking (subject
B, at 10 months), Right: standing posture with slight squat at 2 weeks after learning to walk
(subject B, at 10,5 months), Foot Flat: foot flat with the body erect, Toe Contact: toe contact
with the body inclined forward, Heel Contact: heel contact with the body inclined backward,
1) On the 1st day of independent walking
Fig. 4-5 (left panel) shows EMGs of standing posture on the 306th
day after birth, just as the baby was starting to walk for the first time
(subject B, same as in Fig. 4-4).
During maintenance of standing posture, at the ankle, alternative
bursts between the TA and LG generally showed a reciprocal pattern.
At the knee, the VM showed continuous strong activity during the
maintenance of standing posture. At the hip and knee, three discharge
patterns (reciprocal, reversed reciprocal, and co-contraction) between
biarticular muscles (the RF and BF) could be seen, similar
to those reported for stance phase at the onset of independent walking
(Figs. 4-3 A-1 and 4-4 B-1).
2) At about 2 weeks after learning to walk
Fig. 4-5 (right panel) shows EMGs of standing posture at about 2

weeks after learning to walk, on the 318th day after birth (subject B,
same as in Fig. 4-4).
During standing, at the ankle, alternative reciprocal activity between
the TA and LG began to decrease or disappear. More specifically,
strong continuous discharge patterns of the TA tended to decrease or
disappear, whereas the LG showed continuous activity. At the knee,
activity in the VM tended to decrease or disappear. At the hip and
knee, whereas the reciprocal pattern increased, the reversed reciprocal
and co-contraction patterns tended to decrease or even disappear
compared to what was seen at the onset of independent walking.
Discharges of the RF tended to decrease or disappear while those of
the BF generally increased. The GM showed continuous discharges,
similar to the pattern on the 1st day of independent walking. However,
when the infant maintained balance with the body inclined backward,
losing her body balance control momentarily, strong bursts of the TA
and RF emerged (Fig. 4-5, right panel, just before heel contact).
KNEE EXTENDED
KNEE JOINT
BACKWARD
ERECT
KNEE FLEXED
FORWARD
BACKWARD
ERECT
FORWARD
STANDING
POSTURE
'--
Tibialis anterior
(+)
(- )
Gastrocnemius
(-)
(-)
(+)
(+),H
(-)
H.Cl
(+)
(-)
Vastus med ialis
(+)
(+),H
(+)
(-l.C)
(- )
(+)
(+l.C)
(+)
Rectus femoris
(+)
(-)
(-)
(+)
(+),()
(-)
Biceps femoris
(-)
(-l.C)
(+)
(-)
(-)
(+)
Gluteus maximus
(+)
(+)
(+)
(+)
(+)
(+)
Fig. 4-6. EMG patterns in leg muscles of standing posture at the onset of infant
independent walking.
(+): noticeable activity, (-): no activity, (): slight activity, (+),(-): instances of noticeable
activity and of no activity intermingled.
3. Supported walking
TO Fe
TC FC
~ 1'j1-1"
""\...J"""'"""r-'""r'""
r-~ ~\.
TA
W,
~1
LG
....... ~I"
Ji.!..
10.5 months
~
. ~
tuh. 111",
1,1
1m '"
VM
RF
~1.
,I
SW
ST
ST
SW
SW
ST
1 sec!
INDEPENDENT GAIT
111/
l'
0.5 mv
SUPPORTED GAIT
SUPPORTED GAIT
(FORWARD SWAY)
(ERECT POSTURE)
Fig. 4-7. EMGs of independent and supported walking at the onset of learning to walk (subject
B, at 10.5 months).
Left: independent walking at 2 weeks after learning to walk, Center: supported walking with
the body inclined forward (immature child walking pattern), Right: supported walking with the
body erect (mature adult walking pattern).
Fig. 4-7 (center panel) shows EMGs of supported walking with the
body inclined forward. The right panel of this figure shows supported
walking with the body erect at about 2 weeks after learning to walk in
a 10 month old baby (subject B, same as in Fig. 4-4) .
When the infant walked with external support, strong discharges of
the LG and VM disappeared in the latter half of swing phase, and
continuous discharges of the TA, VM, and RF decreased or
disappeared during stance phase. With forward sway of the trunk, the
excessive activities in the LG, BF, and GM were seen during stance
phase, and these discharge patterns closely resembled the independent
walking pattern of a child usually seen from 3 months to 3 years of
age. With the body erect, discharges of the LG in the first part of
stance phase as well as activities of the BF and GM in stance phase
tended to decrease or disappear, closely resembling the stable adult
walking pattern (Fig. 4-7, right panel).

Based on EMG studies on the development of independent gait in
babies, Okamoto et al. (1972, 1983, 1985, 2001, 2003) (Fig. 4-1), Kazai
et al. (1976) and Sutherland et al. (1980) have reported that specific
changes can be observed at certain times in the course of that
development. In the present study we have confirmed, by means of
both longitudinal and cross-sectional EMG and cinematographic
findings, that during the first month of independent walking, a baby
squats slightly while leaning forward and takes steps with strong
active extension of the legs, exhibiting considerable instability. After
about 3 months of independent walking, a baby exhibits increased
stability with the body tilted slightly forward, and by 3 years of age the
body is upright as in adult walking. The activities of muscles during
these different stages show a similar transition from instability to
stability. At first, when the baby is just beginning to walk, characteristic
discharge patterns can be seen that are excessive when compared to
the corresponding patterns in adults. As the baby matures, these
excesses gradually become refined until, at about 3 years of age, they
very much resemble muscle activities of adults.
As mentioned at the beginning of this chapter, we have been able to
derive an index of instability in walking based on our EMG findings
and previous studies (Okamoto et al. 1972, 1983, 1985, 2001, 2003) on
the developmental process of the acquisition of gait. We did this by
identifying EMG patterns found in the very unstable walking of babies
first learning to walk but not seen in the stable walking of adults. We
looked for EMG patterns that gradually changed and were refined in
the transition between first learning to walk and achieving an adult-like
pattern, finding such refinements principally in stance phase and in the
latter part of swing phase.
1. Stance Phase
ST-TA, ST-RF: Bursts or continuous activity of the tibialis anterior
during stance phase (ST-TA) and activity of the rectus femoris in

synchrony with the tibialis anterior (ST-RF) are normally seen during
the very unstable period of the first month after beginning to walk (Figs.
4-3 A-1 and 4-4 B-1), but not thereafter (Figs. 4-3 A-2 and 4-4 B-2).
When enough external support is given to the infant during stance

phase at the onset of independent walking, strong discharges of the
TA and RF tend to decrease or disappear (Fig.4-7, same subject as in
Fig. 4-4 B-1). On the other hand, we noticed that when one infant
maintained balance with the body inclined backward, losing her body
balance control momentarily during stationary standing at the initial
period of independent walking, strong bursts of the TA and RF
emerged (Fig.4-5, right: just before heel contact). These findings
suggest that the ST-TA and ST-RF patterns help control displacement
of the body's center of mass by participating in maintenance of
posterior inclination of the trunk.
ST-VM: Continuous discharges of the VM during stance phase are
seen until around 1 month after learning to walk in many instances, as
well as those of the TA and RF (Figs. 4-3 A-1 and 4-4 B-1), as
mentioned above. After the first month of walking, continuous
discharges of the VM tend to decrease or disappear (Figs. 4-3 A-2 and
4-4 B-2). That such activity is related to instability is further suggested
by disappearance of the ST-VM pattern when enough external support
is given to the baby to make the slightly squatted position unnecessary
(Fig. 4-7, same subject as in Fig. 4-4 B-1), although the ST-VM pattern
is seen in the initial period of independent walking when support is
withdrawn (Fig. 4-4 B-1). Considering stationary standing, the VM is
continuously active as the baby stands fairly squatted on the 1st day of
independent walking (Fig. 4-5, left), whereas 2 weeks later the
squatting is much shallower and activity of the VM is minimal or
absent (Fig. 4-5, right). This agrees with observations by Okamoto et
al. (1985, 2001, 2003) that the load at the knees decreases as strength
and balance develop. The ST-VM activity seen at the onset of
independent gait thus appears to contribute to holding a posture with
slight knee flexion, permitting the body's center of gravity to be
lowered so that balance is easier to maintain.
ST-LG, ST-BF, ST-GM: Activity of the gastrocnemius in the first
half of stance phase (ST-LG) and continuous discharges of the biceps
femoris and gluteus maximus during stance phase (ST-BF, ST-GM) are
found in the slightly unstable gait typically seen from 1 month after
having learned to walk until about the middle of 2 years of age (Fig.
4-3 A-2, A-3, A-4, 4-4 B-2, B-3, and B-4). After that these activities are no
longer seen, as the pattern of walking closely resembles that of an
adult (Figs. 4-3 A-5 and 4-4 B-5). A look at muscle activities when
external support is given to a child who has just begun to walk
independently (Fig. 4-7) reveals that when the support is provided with
the trunk inclined forward, the ST-LG, ST-BF, and ST-GM patterns
appear as they would in an independently walking child with at least 3
months of walking experience but not yet 3 years of age (Fig. 4-7,
center). If support is provided so that the trunk is upright, on the
other hand, the activities found with the trunk inclined forward are not
present and the results look more like mature adult gait (Fig. 4-7,
right). These findings suggest that the ST-LG, ST-BF, and ST-GM
patterns help control displacement of the body's center of mass by
participating in maintenance of anterior inclination of the trunk. Before
strength and balance have matured to the point that push off can be
effectively used with the trunk upright, as in adult gait, these three
muscle activation patterns are considered to be necessary for gait with
an anteriorly inclined trunk (Figs. 4-5, 4-6, 4-7, and 4-8).
We have found that excessive muscular activity in the stance phase
of gait in a child who has just begun to independently walk strongly
resembles lower limb activity during maintenance of an upright
standing posture for the same period of development (Figs. 4-3 A-l, 4-4
B-1, 4-5 left, and 4-6), suggesting that a common mechanism operates
both in standing and in the initiation of gait. From a mechanical point
of view, at this very early stage, both activities require a low center of
gravity and a wide base of support to assure maximum stability.
Generally these tasks can be accomplished, even though strength and
balance are yet undeveloped by spreading the legs apart to widen the
base of support and by maintaining the knees in flexion to lower the
center of gravity. In our study, the role of the uniarticular vastus
medialis for maintaining stability became clear as a slightly squatted
position was used to lower the center of gravity.
Another important factor to consider is keeping the vertical
projection of the body's center of gravity well within the bounds of the
base of support. In our study, babies who had just begun to walk
independently exhibited two-way control over inclination of the trunk
during walking or standing, thus keeping the center of gravity within
the base of support, by orderly patterns of activity in the biarticular
rectus femoris and biceps femoris (Figs. 4-3 A-l and 4-4 B-1). While
these two muscles act at the hip and knee, Nashner et al. (1985) have
pointed out that ankle strategy is the most efficient for returning the
body's center of mass to its initial position. Indeed, in our study the
gastrocnemius and tibialis anterior exhibited reciprocal patterns of
activity, thus affording anteroposterior control over the center of
gravity to help maintain upright stability. The fact that this EMG
pattern becomes attenuated in standing after 2 weeks' experience of
walking (Fig. 4-5) further suggests that it is a characteristic feature of
balance control when the baby takes steps for the very first time. In
some instances these mutual antagonists co-contract, indicative of
maintaining balance at the ankle by strongly stabilizing the ankle, as
pointed out by McGraw (1940) (Fig. 4-4 B-1).
These findings illustrate how intricately the muscles across the
ankle, knee, and hip joints contribute to maintenance of balance during
stance phase in the initial period of independent walking (See ST in
Fig. 4-10). During this most unstable period, the anteriorly located
muscles of the lower limb (TA, VM, RF) are just as active as the
posteriorly located muscles (LG, BF, GM), but after a full month of
walking, balance has matured to the point that activity of the posterior
muscles tends to become more dominant. This suggests that excessive
activity of the anterior muscles should indicate marked instability,
whereas excessively strong activity of the posterior muscles should be
associated with a lesser degree of instability. When gait is performed
without much activity even from the posterior muscles, a high degree
of intrinsic stability can be inferred to be present.
TO
TA
He
TO
,.
He
.il.
.,111.
LG
"1"
-"I
VM
RF
BF
~I
~~ ,
GM
sw
ST
I
!
""1
u/u>
'"""T"
sw
ST
ERECT POSTURE
sw
ST
_ _' ,_ec_ --,'
ADULT WALK
FORWARD
0.5
FORWARD
WITH KNEE FLEXED
Fig. 4-8, EMGs of adult walking under differing conditions,
mv
2. Swing Phase
SW-LG, SW-VM: Up to the first month of walking, the vastus
medialis is active from the middle of swing phase until subsequent foot
contact (SW-VM, Figs. 4-3 A-1, A-2, and 4-4 B-1) in many instances.
The gastrocnemius is likewise active in this part of swing phase for
about the first 3 months of independent gait (SW-LG, Figs. 4-3 and
4-4) . As seen in studies by Okamoto et al (1985) and by Kazai et al (1976)
as well as in the present investigation, even in the most unstable
period of the onset of independent walking (Fig. 4-4 B-1) the provision
of external support turns off the SW-LG and SW-VM patterns (Fig.
4-7), implying that these two patterns of muscle activity are definitely
associated with intrinsic instability in gait. Compared to the situation of
standing on both feet, these patterns occur when only the contralateral
leg is providing a very small base of support, and the airborne foot is
being actively plantarilexed (SW-LG) while the knee is being actively
extended (SW-VM), suggestive of the operation of the protective
parachute reflex to prevent falling (Fig. 4-9).
j{/i~jJffM'Ai
1
STANCE
10
11
SWING
LG --;-'.
KNEEFLEX.
FC~L--li
VTR -
HC FF
HO
STUMBLE
TO
FF HO
lit-
10
+11
FOOT CONTACT
Fig. 4-9. EMGs of leg extensors after a stumble at 3 years of age.

Activities of the LG (an ankle plantarflexor) and VM (a knee extensor) in the latter part of
swing phase (before foot contact) tended to increase after a stumble (VTR @).
Criteria for Instability
ST-TA
VERY
UNSTABLE
UNSTABLE
TRANSITION TO
ADULT WALKING
PATTERN
II
EMG patterns began to decrease or disappear at about 1month after learning to walk
EMG patterns began to decrease or disappear at about 3 month s after learning to walk
EMG patterns began to decrease or disappear at about 3 years of age
Fig. 4-10. Schematic diagram of EMG activity as indication of unstable walking.

ST: stance phase, SW: swing phase, TA: tibialis anterior, LG: lateral gastrocnemius, VM:
vastus medialis, RF: rectus femoris, BF: biceps femoris, GM: gluteus maximus.
It thus becomes clear that when a baby first begins to walk, muscle
activity plays a relatively great role in providing stability to maintain

posture and to keep the body's center of gravity low and within the
base of support. As months and years pass, the muscles become
stronger and balance matures, obviating the need for so much
myoelectric activity. Thus some patterns of EMG activity can be
identified in the early stages of walking that are no longer present in
adult gait, and they can be validly associated with intrinsic instability.
More specifically, these "excessive" patterns of myoelectric activity can
be classified into the following categories of association with instability
in gait (Fig. 4-10, Table 4-1).
(1) Very unstable gait: As seen in a child within the 1st month of
learning to walk the vastus medialis is active in the latter half of

swing phase, the tibialis anterior and rectus femoris are active
during stance phase, and activity of the vastus medialis is
continuous. These EMG characteristics are not usually seen in
subsequent childhood gait or in adult gait, and they serve as
markers of very unstable gait.
(2) Unstable gait: Activity of the gastrocnemius in the latter half of
swing phase is generally noted only within the first 3 months after
the child learns to walk, and that activity is interpreted as a sign of
unstable gait.
(3) Slightly unstable gait: Activity of the gastrocnemius in the first half
of stance phase and continuous activities of the biceps femoris and

gluteus maximus from initial contact with the floor until push off are
found in children until 3 years of age . These activities are
considered EMG markers of slightly unstable gait.
As a practical application, we have been able to apply this EMG
index of gait instability to EMG patterns noted during recovery of
walking in an elderly man after cerebral infarction, demonstrating the
validity and usefulness of this index of gait instability. This index of
instability is thus proposed as a basic tool for analyzing the
developmental process of gait as well as for electro myographically
assessing clinical progress in acquisition of the ability to walk.
An Index of Gait Instability

Table 4-1. EMG activity in unstable walking
Region
Code
ST-TA
Ankle
SW-LG
ST-LG
SW-VM
Knee
ST-VM
ST-RF
Knee & Hip
ST-BF
Hip
ST-GM
Interpretation
Activity of the TA
in stance phase
Activity of the LG
in the latter part of swing phase
Activity of the LG
in the first half of stance phase
Activity of the VM
Activity of the VM
in stance phase
Activity of the RF
in stance phase
Activity of the BF
in stance phase
Activity of the GM
in stance phase
Activity decreases
or disappears at
1 month after
learning to walk
3 months after
learning to walk
3 years of age
1 month after
learning to walk
1 month after
learning to walk
1 month after
learning to walk
3 years of age
3 years of age
Indication
Very
unstable
Unstable
Slightly
unstable
Very
unstable
Very
unstable
Very
unstable
Slightly
unstable
Slightly
unstable
ST : stance phase, SW: swing phase, TA: tibialis anterior, LG: lateral gastrocnemius,
VM : vastus medialis, RF : biseps femoris, BF : biseps femoris, GM : gluteus maximus.
Conclusion
To obtain an index of gait instability from EMG information, we
made longitudinal observations on three children from the time they
first began to walk independently at about 1 year of age until a stable
adult-like walking pattern was achieved at around 3 years of age. For
purposes of comparison, these observations were supplemented with
cross-sectional observations taken from fifteen babies in the first year
of age, five babies in the second year, and five more in the third year
of age. From all of these observations we were able to construct an
index of gait instability (Table 4-1) .
As seen in a child within the first month of learning to walk, the
vastus medialis is active in the latter half of swing phase, the tibialis
anterior and rectus femoris are active during stance phase, and activity
of the vastus medialis is continuous. These EMG characteristics are
not usually seen in subsequent childhood gait or in adult gait, and they
serve as markers of very unstable gait.
Activity of the gastrocnemius in the latter half of swing phase is

generally noted only within the first 3 months after the child learns to
walk, and that activity is interpreted as a sign of unstable gait.
Activity of the gastrocnemius in the first half of stance phase and
continuous activities of the biceps femoris and gluteus maximus from
initial contact with the floor until push off are found in children until 3
years of age. These activities are considered EMG markers of slightly
unstable gait.
We propose this index of instability as a way to analyze gait in terms
of developmental processes and also as a way to electromyographically
assess clinical progress in gait training.
Fig. 4-11 . Very unstable infant walking at 1 year of age.
When does a baby feel instability from?
The purpose of this study was to see whether an electromyographic (EMG) index of gait instability is applicable to
the developmental process of supported walking in normal
neonates and infants. In six neonates ranging in age from 14
to 26 days after birth, EMGs of stepping were recorded at
approximately from 1 to 4 week intervals until around 4 months.
Additionally, longitudinal EMGs of one subject were recorded
at 1 or 2 week intervals until just before independent walking.
EMG patterns of the lateral gastrocnemius (an ankle plantar
flexor) and vastus medialis (a knee extensor) in the latter
part of swing phase indicating unstable walking, not seen in
the neonatal period up to the first postnatal month, tended to
increase in young infants at around 3 postnatal months. These
results suggest the addition of voluntary infant stepping to
reflex neonate stepping from around 3 months. From 6 to 12
months, these marked activities tended to decrease, gradually
coming to resemble adult stable walking through development
of strength, balance, and postural control. In conclusion,
muscular activities of the lateral gastrocnemius and vastus
medialis in the latter part of swing phase indicate unstable
walking, findings which are applicable to developmental
changes during newborn stepping and infant supported
walking.
Human gait without support is associated with muscle strength,

balance, and postural control. Although some people have to resort to
supported walking by virtue of immaturity, aging, postoperative status,
or disease, supported walking generally resembles the normal adult
walking pattern except for the substitution of missing postural and
balance elements with support. However, to decide whether gait is
stable or not is very difficult from motion analysis alone, for example,
by comparison with a normal pattern. Although supported walking
may show no perceptible change during rehabilitation, the underlying
EMG pattern may be changing during the recovery of gait. To evaluate
gait stability during supported walking, it is therefore very important
to examine muscle function.
We have determined signs of unstable walking from EMG activity
patterns based on the developmental process of normal infant walking
and normal adult walking (Chapter 4), and have used them to analyze
gait in terms of development processes and also as a way to
electro myographically assess clinical process in gait training. They
have reported that the vastus medialis is active in the latter half of
swing phase, the tibialis anterior and rectus femoris are active during
stance phase, and that activity of the vastus medialis is continuous, as
seen in a child within the first month of learning to walk. These EMG
characteristics are not usually seen in subsequent childhood gait or in
adult gait, and they serve as markers of very unstable gait. Activity of
the gastrocnemius in the latter half of swing phase is generally noted
only within the first 3 months after the child learns to walk, and that
activity is interpreted as a sign of unstable gait. Activity of the
gastrocnemius in the first half of stance phase and continuous activities
of the biceps femoris and gluteus maximus from initial contact with the
floor until push off are found in children until 3 years of age. These
activities are considered EMG markers of slightly unstable gait.
We were interested in applying the index of gait instability to the
developmental process of supported walking in normal neonates and
infants. Newborn stepping has been an object of study for a long time.
McGraw (1940) and Zelazo et al. (1972) have discussed the significance of early stepping movements for development of adult gait.
Only a few attempts so far, however, have been made to study
developmental changes of stepping by EMG, particularly in relation to
gait instability. In this study, we applied the idea of the index of gait
instability to developmental changes during newborn stepping and
infant supported walking.
Four male and two female infants were observed at 14, 18, 19, 22,
23, and 26 days after birth. We initially observed developmental
process of supported walking in neonates and infants using an index of
gait instability. In interpreting the EMG pattern of supported walking
in all subjects tested, we focused on the discharge pattern in the latter
part of swing phase and in stance phase. From this, we could see
developmental changes in the EMG patterns of swing phase and
relatively wide variations in those of stance phase. Most of the figures
in this chapter are from longitudinal representative EMG patterns in
subject A (Fig. 5-1; 22 days after birth).
Fig. 5-1. Stepping at 22 days after birth.
Application of an Index of Gait Instability 91
Until the 1st month of age
ST
(Al
rtrtftft ft
sw
I~
TA
ST
.....
~~
"0'
'.
LG '
VM
I'!
RF
SF .l,
,ul"I,li
,~~
''''''1'
GM
(U
TA"
.L,wI,,~.
"",rr"1tr~
LG
'.;.~fI/I~\1i!I, ~"'14,....
RF
SF
""""-
..
#-'
SWING
\~~o\iiI~i;.<""'I"
~/~---
STANCE
1 sec
0.5 m~
22 days
NEONATAL STEPPING
Fig, 5-2, EMGs of newborn stepping at 22 days after birth (same subject as in Fig, 5-1),
ST: stance phase, SW: swing phase, (R): right leg, (L): left leg, TA: tibialis anterior, LG:
lateral gastrocnemius, VM: vastus medialis, RF: rectus femoris, SF: biceps femoris, GM:
gluteus maximus,
Figure 5-2 shows representative EMG patterns of stepping at 22

days after birth. This is one of the more rhythmical walking-like
patterns that we obtained in this neonate. The discharge patterns of
leg muscles during newborn stepping are generally more irregular
than in adult walking.
The walking in this period was characterized by quick hip and knee
flexion in which the thigh became almost horizontal in the first part of
swing phase. The foot was raised forward and dorsiflexed strongly.
Then the foot began to reach the floor slowly, the knee extending
passively along with the hip. The foot usually contacted the floor with
the heel first, but in a few instances the forefoot made initial contact (Fig.
5-3) . A squatting posture was often observed during stance phase. At
around 1 month, such squatting began to become less pronounced.

As the ankle and knee were extending, activities in the LG (an ankle
plantarflexor) and VM (a knee extensor) were hardly seen in the
second half of swing phase. In stance phase continuous activities were
seen in the ankle, knee, and hip muscles, especially in single stance.
The discharge patterns of many leg muscles (TA, LG, VM, RF, BF, and
GM) indicated unstable walking showing reciprocal and co-contraction
patterns, with wider variations and more excessive activity of discharge
patterns than in adult walking. The TA and LG showed reciprocal
patterns on the right and co-contraction on the left. Activity of the TA
decreased or disappeared while that of the LG increased in many
instances. Conversely, in some of the subjects activity of the LG
decreased or disappeared while that of the TA increased. In the knee
and hip muscles, the RF and BF showed both reciprocal and cocontraction patterns. In some subjects, the reciprocal pattern was one
in which activity of the RF decreased or disappeared while that of the
BF increased, whereas in other subjects the converse reciprocal
pattern prevailed. Among the knee muscles, continuous activity of the
VM was often observed, but sometimes no activity was seen at all.
~~
20 days after birth
Fig. 5-3. Foot prints of neonatal stepping at 20 days after birth.
ST
(R)
~ffl:fff
ST
SW
TA 'fi\>:iI
f'it>l
~r\{"
LG '!ri,"
VM~'Mt~..
~,~,""
1
1 -
,.'#' .,!"",~
"""
01<
..
RF
BF~
,- , -
GM
(L)i~lUl
'1"1
TA
LG 't-'/"
J,
RF '~
BF--
r.
SWING
--"'@t~
-:.: f1
I"
........
STANCE
1 sec
O.5mv
44 days
Fig. 5-4. EMGs of newborn stepping at 44 days after birth (same subject as in Figs, 5-1 and
5-2).
Figures 5-4 and 5-5 show representative EMG patterns of stepping at

44 and 105 days after birth in the same infant. We found in six infants
tested that we could induce a stepping pattern after 1 month of age,
but not so easily as before 1 month.
After 1 month, leg flexion was performed strongly in the first part of
swing phase as in the neonatal period, but the degree of hip flexion
tended to decrease slightly. We found mostly plantarflexion before
floor contact rather than dorsiflexion which was more prevalent in the
newborn period. From 3 to 4 months, the toe initially contacted the
floor for the most part. Knee extension was performed more actively
than in the neonatal period. After around 1 month, a half-squatting
posture during stance phase tended to increase.
5W
5T
(Rl
TA---4--~---4~--~~--~--~-r--~
LG __~~~~~~~~~~--~~~-
VM~~-r~~'~~~~~~~__+-~~_
RF--~--~~-~~--_4~_.--~_r----
,.."," ""--+--t-...~--+-+,~-
GM
';~ : )" 'f ::~'~I=~:n~ll~~:n

SWING
STANCE
1 sec
I O.S mv
105 days
Fig, 5-5. EMGs of newborn stepping at 105 days after birth (same subject as in Figs. 5-1,
5-2, and 5-4).
From 1 to 3 months, the LG and VM began to show activity before

the foot actually touched the floor (Fig. 5-4). From 3 to 4 months, such
activity become more pronounced (Fig. 5-5). On the other hand, we
found wide variations in the discharge patterns of many leg muscles (TA,
LG, VM, RF, BF, and GM) during stance phase (Figs. 5-4 and 5-5)
similar to the neonatal period (Fig. 5-2). The stepping did not cease at
around 4 or 5 months after birth, as shown in the longitudinal
observations of subject A.
(R)
ST
SW
LG
~lilii
-.
_....,
",'
"I'
TA
J.L . I.,
VM
'1'~Il'f1
RF
SF
.I~"
..lJ1l,
''1~1
-'I
'''1''
GM
( L)
TA
II
1J,
k'l
If" r~'!""r
LG
~, L!a...
1,11..11.,[1
" II"
.J,
RF
.L" "'1
SF
SWING
"'.
STANCE
1 sec
O.5 mv
351 days
INFANT SUPPORTED WALKING
Fig. 5-6, EMGs of supported walking at 351 days after birth, 34 days before independent
walking (same subject as in Figs, 5-1 , 5-2, 5-4, and 5-5),
During infant supported walking (6-12 months of age), step

frequency was more regular than during the newborn period. The
relatively pronounced flexion in the hip and knee seen in the previous
period was slightly reduced (Fig. 5-7).
Figure 5-6 shows representative EMG patterns of supported walking
at 351 days after birth, but before independent walking was achieved.
From 11 to 12 months after birth, the infant became able to stand by
herself and to walk with one-handed supported walking. The femur
was not actively lifted up in the first part of swing phase, and active
ankle plantarflexion and knee extension tended to disappear in the
latter part of swing phase. The heel usually contacted the floor first.
This subject began to walk without support at 385 days after birth.
Strong myoelectric activity of the LG and VM seen in earlier periods

(Figs. 5-4 and 5-5) tended to disappear in the latter part of swing phase
in this period (Fig. 5-6). The marked variations in the discharge
patterns of both these muscles seen previously were no longer seen
during infant supported walking, and more closely resembled adult
walking. Co-contraction patterns of ankle, knee, and hip muscles seen
in newborn stepping also tended to decrease or disappear during
stance phase, but reciprocal patterns of mutual antagonists generally
remained. Nevertheless we did find some instances of excessive
muscular activity in many leg muscles during stance phase, similar to
what was seen in earlier periods (Figs. 5-2, 5-4, and 5-5).
Fig. 5-7. Stable infant supported walking before independent walking.
Developmental changes in EMG patterns
ffftJttftl
r~' I
TO
STANCE (ST)
SWING (SW)
1 sec
14 days
0.5 mv
NEONATAL STEPPING
~fllff~i
~:="I =t,~:It::
TO
SW
FC
ST
83 days
1 sec
0.5 mv
Fig. 5-8. EMGs of newborn stepping at 14 and 83 days after birth (subject 8).
Figure 5-8 shows the EMG patterns of the ankle and knee extensors
before foot contact during stepping at 14 and 83 days after birth (subject
B), illustrating the differences between neonate and young infant
periods.
In the neonatal period (up to 1 month), we could not see strong
discharges of the LG and VM in the latter part of swing phase (Fig.
5-8, top). However, in the young infant period (1-4 months), we could
see strong discharges of both these muscles in the same phase in
many instances, as mentioned above (Fig. 5-8, bottom) . These
developmental changes in muscle activity from the neonatal period to
the young infant period were similar to those of subject A (Figs. 5-2,
5-4, and 5-5) .
Application of an index of gait instability

to supported walking in babies
Table 5-1. EMG index of gait instability
Joint
Sign of instability
Interpretation
Ankle
SW-LG
Activity in the LG during SW (+)
Unstable
Knee
SW-VM
Activity in the VM during SW (+)
Unstable
SW: latter part of swing phase, LG : lateral gastrocnemius, VM: vastus medialis,
(+): noticeable activity.
Table 5-1 shows the EMG index of gait instability in the latter part of
swing phase (Chapter 4, Fig. 5-9). Table 5-2 shows the results of
evaluating developmental changes in LG and VM muscular activities
during newborn stepping of each subject using the criteria for gait
instability (Table 5-1). Table 5-3 shows developmental changes of
muscular activities in the LG and VM before foot contact, from
newborn stepping to supported walking just prior to independent
walking in subject A. From the results obtained in the swing phase, it
was discovered that we were able to apply the idea of the index of gait
instability to developmental changes during newborn stepping and also
infant supported walking.
Ankle plantar flexion
Knee extension
Activity of the LG
in the latter part of SW
Activity of the VM
in the latter part of SW
Fig. 5-9. EMG index of gait instability in the latter part of swing phase.
SW: swing phase, LG: lateral gastrocnemius, VM : vastus medialis, +: noticeable activity.
Table 5-2. Activities of the lateral gastrocnemius (LG) and vastus medialis (VM) in the latter
part of swing phase of stepping induced during very early development
Months after birth

Subject
Muscle
0-1
1-3
3-4
VM
H
H
(+), H
(+),H
(+), partly H
(+), partly H
LG
LG
(-)
VM
H, partly (+)
(+), H
(+), (-)
(+), partly H
LG
(-)
(-)
H, partly (+)
(-), partly (+)
(+), (-)
(+), (-)
H
(-)
(+), H
(+), (-)
(+), partly H
(+), partly (-)
H
(-)
(+), H
(+), (-)
(+), partly H
(+), partly (-)
VM
LG
VM
LG
VM
LG : lateral gastrocnemius, VM : vast us medialis, (+) : noticeable activity, (-) : no activity,

(+), (-): instances of noticeable activity and of no activity intermingled.
Table 5-3. Activities of the lateral gastrocnemius (LG) and vastus medialis (VM) in the latter
part of swing phase of stepping during the first year of development in subject A
Months after birth

Muscle
0-1
1-3
3-4
LG
VM
(-)
(+), H
(+), (-)
(+), partly H
(+), partly (-)
6-12
H, partly
H , partly
(+)
(+)
(+): noticeable activity, (-): no activity, (+), (-): instances of noticeable activity and of no
activity intermingled.
Discussion
McGraw (1940) reported that infant stepping can be elicited shortly
after birth and during the first months, and that thereafter it usually
disappears. Thelen et al. (1987) and Forssberg (1985) pointed out
from movement patterns and EMGs that the locomotor pattern of the
newborn differs markedly from that of an adult. Usually the leg muscle
activities of newborn stepping are irregular and include a high degree
of co-activation compared with the adult walking pattern.
We will first interpret the meaning of wide variations in stance
phase, especially in the EMG patterns of mutual antagonists of the leg
muscles during stepping (Figs. 5-2, 5-4, and 5-5). Diminution of activity
in the TA and RF accompanied by greater activity in the LG and BF
may have been due to forward leaning of the body. When, conversely,
activity in the LG and BF decreased or disappeared while that in the
TA and RF increased, this may have resulted from leaning backward.
Co-contraction of these muscles was also observed in many instances,
probably related to maintaining a standing posture with the body erect
or to stabilizing of the ankle, knee, and hip joints. These variations of
leg muscle activities may be caused by the changing posture during
supported newborn stepping.
During stepping, sustained discharges of the VM are probably
attributable to bearing body weight with the knee flexed. When VM
activity was low or absent during stance phase, the manner of bearing
weight on the knee joint may have resulted in a smaller load. We thus
suggest that the degree of activity in the VM may be regarded as an
indication of magnitude of load on the knee joint.
Even if the EMG patterns of leg muscles in stance phase appear to
indicate unstable walking during stepping, we hesitate to consider
them as reliably consistent signs of instability because variations in
discharge patterns of the leg muscles are probably closely related to
magnitude of joint load, which can be influenced by many factors other
than instability. We thus find it undesirable to apply the EMG index of
gait instability to discharge patterns in stance phase, during which
considerable variations are seen in newborn stepping.
Secondly, we will focus on developmental changes in the EMG
patterns of supported walking during swing phase. Up to the first
month of age, muscular activities of the LG and VM were hardly seen
in the latter part of swing phase (Figs. 5-2, 5-8 top, Tables 5-2, 5-3), as
also reported by Thelen et al. (1987) and Okamoto et al. (2001, 2003) .
The leg extends passively and the foot contacts the floor usually with
the heel first, as mentioned in chapter 1. These findings clearly show
that muscular activities for knee extension and ankle plantarflexion are
not observed in this period. As mentioned above (Table 5-1), we
determined that activity of the leg extensors before floor contact
indicates gait instability, but it is risky to judge the presence of gait
stability from an absence of activity in these two muscles in the
neonatal period, because stepping in the neonatal period is a reflex
movement performed under the control of lower (spinal) levels of the
central nervous system (eNS), and equilibrium reflexes are yet
immature.
Neonatal newborn
stepping
( until 1 month after birth)
~j~~ll
Slow leg extension
Young infant
stepping
( 1-3 months after birth)
fflll
Fast leg extension
Fig. 5-10. Developmental changes in newborn stepping.

Strong muscle activities of leg extensors (LG and VM) due to a parachute reaction of the
legs before floor contact, not seen in the neonatal period, began to appear in the young
infant period from 1 month of age to 3 months.
From 1 to 3 months, the leg extensors begin to show some activity

before the foot reaches the floor (Figs. 5-2, 5-8 bottom, Tables 5-2, 5-3)
and the forefoot begins to contact the floor first more often. When the
infant begins to actively perform knee extension and ankle
plantarflexion, the strong muscle activities of leg extensors begin to
participate before floor contact in this period (Fig. 5-10). The marked
discharge patterns of these two muscles are similar to those of
unstable supported walking and very unstable independent walking for
the first time. From these findings, we believe that the activities of the
leg extensors before floor contact in this period indicate lack of
stability, as identified by Okamoto et al. (2001, 2003). We thus believe
that infants in this period begin to feel instability (Fig. 5-11).
From around 3 to 4 months old, marked activity of the leg extensors
are observed (fables 5-2 and 5-3). Active ankle plantarflexion and knee
extension before foot contact become the mode of performance in this
period. Milani-Comparetti (1967) reported that the parachute reaction
of the legs begins to appear at about 4 months after birth. The
parachute reaction is an equilibrium reflex performed under control of
higher (cortical) levels of the CNS. Strong muscle activities of the leg
extensors before the floor contact may arise as parachute reactions
(self-protection) with the maturation of the CNS. Although McGraw
(1940) pointed out that it is difficult sometimes to tell whether the
active leg extension before floor contact is deliberate or of reflex
quality, we would suggest that it is the beginning of voluntary infant
stepping (supported walking) added upon reflex neonate stepping at
this period.
From 6 to 12 months, absence of activity in these muscles before
floor contact becomes the rule, and marked activity tends to disappear
in the remaining instances (Table 5-3). Lack of activity in these
muscles indicates that leg extension before foot contact is not
performed actively in this period, in contrast with the active leg
extension of around 3 or 4 months. The developmental change of
infant stepping in this period is due mainly to maturation of the
equilibrium system and development of strength. We interpret the
absence of activity in the leg extensors in the latter part of swing phase
in this period to resemble adult stable walking and thus to have
resulted from development of strength, balance, and postural control.
Conclusion
We made longitudinal observations on six normal neonates to see
whether an EMG index of gait instability derived from the developmental process of normal infant walking is applicable to EMG patterns
of supported walking in neonates and infants.
Muscular activities of the LG (an ankle plantarflexor) and VM (a
knee extensor) in the latter part of swing phase indicate unstable
walking, findings which are applicable to developmental changes
during newborn stepping and infant supported walking.
In stepping during the first month, muscular activities were not seen
in the LG or VM. It would be misleading to consider such gait to be
"stable" simply because these muscles are inactive in the neonatal
period, especially since stepping in the neonatal period is under the
control of lower levels of the central nervous system.
At around the third postnatal month, the LG and VM showed strong
activity just before the foot reached the floor, suggesting that muscular
activities participating in active ankle plantarflexion and knee extension
act as a parachute reaction (Figs. 5-10 and 5-11). This may be the
beginning of superimposition of voluntary infant stepping on top of
reflex neonate stepping.
From 6 to 12 months, when the infant was becoming able to
maintain standing without support, marked activities of the LG and
VM before floor contact tended to disappear as in the adult pattern.
We thus presume that absence of activity in the LG and VM at that
time suggests stable walking, reflecting development of strength,
balance and postural control during that period.
In summary, our observations of developmental changes in newborn
stepping and infant supported walking, combined with an EMG index
of gait instability based on subsequent stages of development, lead us
to believe that activities of the LG and VM in the latter part of swing
phase can be interpreted in terms of a scheme of early development of
stability in walking.
Young infant
period
Neonatal
period
1.------+
4----------------------------------+.
VM: vastus medialis

(Knee extensor)
Frequency
of
occurrence
LG: lateral gastrocnemius

(EMG)
/-----'/
'/
'/
'/
"'1....
'/
'/
'/
~--.~
Birth
1 month
.....
TA: tibialis anterior
.............~~~.~I~.~~:~!~;.~~:~.
2 months 3 months 4 months after birth
Fig. 5-11. Developmental changes of EMGs in leg muscles before floor contact of newborn
stepping.
EMG patterns of the LG (an ankle plantarflexor) and VM (a knee extensor) in the latter part
of swing phase indicating unstable walking, not seen in the neonatal period up to the first
postnatal month, tended to increase in young infants at around the third postnatal month.
These results suggest the addition of voluntary infant stepping to reflex neonate stepping
from around 3 months after birth.
We believe that infants at around 3 months after birth begin to feel instability.
To study the recovery of walking in an 85 year old man who

had right hemiplegia after suffering a cerebral infarction,
electromyograms (EMGs) were recorded from his leg
muscles. We used signs for instability that we derived from
EMG patterns seen in the developmental process of normal
infants. The myoelectric activity at 1 month after the stroke
showed many patterns indicative of unstable walking, closely
resembling activity patterns seen in very unstable independent
gait of a 1 year old baby in the first month of learning to walk.
7 months later, these patterns indicating unstable gait tended
to have decreased or disappeared, although some marked
activity betraying instability still remained. However, when the
patient walked with the support of a hand cart and was able to
hold his trunk upright, these excessive muscular activities
decreased or disappeared, closely resembling the stable adult
walking pattern. We recommend further study of the evaluation of recovery of walking after stroke by comparing the
patient's EMG patterns to those not only of normal adult human
gait, but also of the development of human walking in early
childhood.
To recover gait in hemiplegic patients, it is desirable to start

rehabilitation as soon as possible after the stroke. Gait analysis after
stroke has been studied by film, temporal patterns, footprints, gas
metabolism, electromyography (EM G) , and other methods of
mechanical, anatomical, and physiological analysis. However, few
studies have thus far been done on EMG evaluation of recovery of
walking after stroke.
We have studied EMG activity in normal infants from the time they
first walk independently until they display adult-like walking patterns.
We have become able to determine signs of unstable walking from
EMG activity patterns based on the developmental process of normal
infants (Chapter 4, Table 6-1). We were interested in examining gait in
elderly persons, especially those undergoing a process of rehabilitation
in walking, and we wanted to consider the appropriateness of using
our signs for instability derived from the EMG patterns seen in the
developmental process. The purpose of this study was to consider
electro myographically recovery of walking in an elderly man undergoing rehabilitation after suffering a cerebral infarction.
Fig. 6-1. Form of walking in an elderly man and an infant.

Gait patterns in the recovery of walking after the stroke closely resembled gait patterns seen
in the very unstable independent gait at 1 year of age.
The patient was an 85 year old man who had right hemiplegia after
suffering a cerebral infarction. Slight spasticity was seen at 1 month
after onset of the stroke, but coordination in walking was almost
normal. He usually displayed heel contact and a heel-to-toe pattern in
gait. His walking posture was characterized by a markedly forward
lean of the trunk and a slight squat (Fig. 6-1).
After the stroke he had slight dementia, so it was hazardous to let
him live by himself. He needed assistance or supervision in his daily
life.
In the first month of recovery, the patient was given training to
enhance activity of the tibialis anterior. He was instructed to do heel
walking with support. Training also included trunk function and
balance. At 2-3 weeks after the stroke, he could perform supported
walking using a parallel bar, and by 1 month he had recovered
independent walking. We recorded EMGs of the independent walking
at 1 month after the stroke.
To more closely examine the EMG characteristics of the walking
stability in the patient, he was instructed to walk with a hand cart for
support, keeping his trunk upright.
Table 6-1. EMG activity in unstable walking
Region
Code
ST-TA
Ankle
SW-LG
ST-LG
SW-VM
Knee
ST-VM
ST-RF
Knee & Hip
ST-SF
Hip
ST-GM
Interpretation
Activity of the T A
in stance phase
Activity of the LG
Activity of the LG
in the first half of stance phase
Activity of the VM
Activity of the VM
in stance phase
Activity of the RF
in stance phase
Activity of the SF
in stance phase
Activity of the GM
in stance phase
Activity decreases
or disappears at
1 month after
learning to walk
3 months after
learning to walk
3 years of age
1 month after
learning to walk
1 month after
learning to walk
1 month after
learning to walk
3 years of age
3 years of age
Indication
Very
unstable
Unstable
Slightly
unstable
Very
unstable
Very
unstable
Very
unstable
Slightly
unstable
Slightly
unstable
ST : stance phase, SW : swing phase, TA : tibialis anterior, LG : lateral gastrocnemius,

VM: vastus medialis, RF: rectus femoris , SF: biseps femoris, GM: gluteus maximus.
1 month after the stroke
SW
fffr{
ST
(Rl
TA
r-J11M~'I'flIIl\l1
GM '---I-'tI--Mi'M.m.
(Ll
TA~~~~~--~~~~~~~-r--~WHW~~~--~~
GM
ST
SW
1 sec
0.5 mv
85.0 years
Fig. 6-2. EMGs of walking at 1 month after stroke.

(R): Right leg; affected limb, (L): Left leg; normal limb, ST: stance phase, SW: swing phase,
TA : tibialis anterior, LG : lateral gastrocnemius, VM: vastus medialis, RF : rectus femoris, SF :
biceps femoris , GM : gluteus maximus.
The subject had difficulty walking, writing, and eating by himself,

but recovered independent walking by 1 month after the stroke. Fig.
6-2 shows representative EMGs of walking at 1 month after the stroke.
The walking velocity was about 50 - 55 m/min. His walking form was
characterized by a marked forward lean of the trunk, a slight squat,
and a short step length. He generally contacted the floor with the heel
first, barely clearing the floor with the toes, but his foot occasionally
hit the ground with the toes first or with the entire sole (Fig. 6-3).
In swing phase, the LG of either leg showed activity in the latter half
of that phase, but in some instances this activity was absent on either
side. At the knee, the VM on the right affected side also showed
strong activity in the latter half of swing phase. On the sound left side,
however, the VM did not exhibit such activity, but rather it showed
discharges just prior to heel contact as in normal adult walking. In
stance phase, at either ankle, the TA was continuously active from heel
contact until push off. The LG was frequently active, on either side, in
the first half of stance phase, but instances of no discharge were also
seen on either side. At the knee and hip, the VM, RF, BF, and GM
bilaterally showed strong co-contractile patterns in stance phase.
Fig. 6-3. Form of walking at 1 month after the stroke (left) and at 1 month after
learning to walk (right).
7 months after the stroke
fff/f
( R)
TA
LG
VM
RF
SF
GM
( L)
TA
LG
VM
RF
SF
GM
ST
sw
1 sec
85.6 years
Fig. 6-4. EMGs of walking at 7 months after stroke.

(R): Right leg; affected limb, (L): Left leg; normal limb.
0.5 mv
Fig. 6-4 shows representative EMGs of walking at 7 months after the

stroke. The gait pattern was similar to that at 1 month after the stroke,
but heel contact became very consistent, with rare instances of initial
contact by the entire sole or toes.
In swing phase, the LG of the affected right leg continued to exhibit
excessive activity in the latter half of the phase, but the corresponding
myoelectric discharge pattern of the left LG had begun to disappear.
The strong burst of activity in the VM of right leg observed at 1 month
after the stroke was no longer seen in the latter half of swing phase,
but instead a burst was observed just prior to heel contact as in normal
adult walking. In stance phase, the strong continuous discharge
patterns of the TA seen in both legs 1 month after the stroke tended to
decrease or even disappear. These discharge patterns were similar to
the normal adult pattern, but marked activity could still occasionally be
seen in the TA during stance phase on either side. The LG on both
sides showed strong activity at the beginning of stance phase. The
VM, BF, and GM on either side still exhibited marked co-contractions
throughout stance phase, but the excessive continuous discharge
patterns of the RF on either side seen at 1 month after the stroke
tended to decrease.
Fig. 6-5. Diagram illustrating the differences between an adult and an elderly man during
walking.
LG: lateral gastrocnemius, VM: vastus medialis, BF: biceps femoris, GM: gluteus maxim us, (+):
noticeable activity.
1 year 7 months after the stroke
ST
SW
(R)
(L)
TA~~~~1~~~~~~~n~~~~~~;H~~~~~-
SW
ST
1 sec
86.6 years
Fig. 6-6. EMGs of walking at 1 year 7 months after stroke.

(R): Right leg; affected limb, (L): Left leg; normal limb.
0.5 mv
Fig. 6-6 shows representative EMGs of walking at 1 year 7 months

after the stroke. The gait pattern resembled that at 7 months after the
stroke.
In swing phase, the LG of the affected right leg still showed
discharges in the latter half of swing phase as in the recordings of the
previous two occasions, but the magnitude of that activity was
decreased or sometimes even absent. On the other hand, activity of
the LG on the sound left side was not usually seen, thus resembling
the pattern seen in normal adult gait. In stance phase, excessive
discharge of the LG seen at the beginning of the phase on either side
looked like the pattern at 1 month and at 7 months after the stroke.
The VM, BF, and GM on either side, as before, showed excessive
continuous co-contractile patterns in stance phase.
fl ~
86.6 years
TA
LG
VM
RF
SF
GM
KNEE
EXTENSION
FLEXION
SmlNG
1 sec
SQATTING
0.5 mv
STANDING
Fig. 6-7. EMGs of the right leg during standing at 1 year 7 months after stroke (same subject
as in Fig. 6-6).
EMG patterns of squatting posture before standing closely resemble the excessive
continuous patterns of activity in the LG, VM , BF, and GM associated with the forward sway
of the trunk with the knees flexed in stance phase of elderly walking.
EMG evaluation of walking stability
Table 6-2. EMG evaluation of walking stability
Right leg: affected limb

EMG pattern of
unstable walking
1 month
7 months
after stroke
after stroke
ST-TA
(++)
SW-LG
(+), partly (-)
(+), partly (-)
(+), partly (-)
(+)
(+)
H, partly
(+)
1 year 7 months
after stroke
H , partly
ST-LG
(++), partly (-)
SW-VM
(+)
ST-VM
(++)
(++)
(++)
(+)
ST-RF
(+)
(+)
(+)
ST-BF
(++)
(++)
(++)
ST-GM
(++)
(++)
(++)
1 month
after stroke
7 months
after stroke
1 year 7 months
unstable walking
ST-TA
(++)
SW-LG
(+), partly (-)
Left leg : normal limb

EMG pattern of
ST-LG
(++), partly
H, partly
H
(+)
after stroke
H, partly
(-)
(-)
(++)
(++)
SW-VM
(-)
(-)
ST-VM
(++)
(++)
(++)
ST-RF
(++)
(+)
(+)
ST-BF
(++)
(++)
(++)
ST-GM
(++)
(++)
(+)
(+)
(++): marked activity, (+): noticable activity, (-): no activity.
See Table 6-1 for descriptions of the EMG patterns.
Table 6-2 shows in brief the results of evaluating the patient's gait
based on our criteria for walking stability. The myoelectric activity at 1
month after the stroke showed many patterns indicative of unstable
walking. 7 months later, these patterns indicating unstable gait tended
to decrease or disappear, but the overall picture was still one of a more
unstable walking pattern than normal adult gait.
Discussion
One purpose of gait evaluation is to describe how a patient's
performance differs from "normal" gait. Generally, normal adults show
regular reciprocal discharge patterns of agonists and antagonists
during the gait cycle. If an individual cannot maintain dynamic stability
because of a central nervous disorder or aging, some EMG activities
not usually seen in the adult pattern should appear. In our case study,
excessive muscular activities appeared early in the recovery period
after a stroke (Fig. 6-2).
In the latter half of swing phase, the VM and LG showed activities not
ordinarily seen in the adult. These discharge patterns were similar to
those seen at the onset of very unstable independent walking in early
child development. That is, activity of the VM is observed in the latter
half of swing phase until about 1 month after first learning to walk,
whereas activity from the LG in the latter half of swing phase continues
until about 2 or 3 months after learning to walk. As Okamoto et al. (1985,
2001, 2003) have pointed out, these patterns may be considered to come
from the knee extending and the ankle plantartlexing to prevent falling.
These excessive muscular activities decrease or disappear when the
child is given external support, so these EMG patterns in the latter half
of swing phase suggest unstable walking.
In a patient who has suffered a stroke, excessive plantartlexion on
the affected side is one typical problem. In our patient, activity of the
LG was observed in the latter half of swing phase bilaterally 1 month
after the stroke. At 7 months after the stroke, the excessive activity
was still found in the right LG, but no longer in the left LG. Stability
had thus returned to the extent that indirect effects of the stroke on
the sound side had disappeared and the patient was experiencing
disability more from the direct effects on the affected side alone.
In stance phase, the TA showed strong activity not usually seen
except at the beginning and end of the phase in the adult. This
discharge pattern was also similar to that seen in the first month of
independent walking by a baby. This activity might be interpreted as
an effort to maintain balance with the toes gripping the surface. The
LG also exhibited strong activity not seen at the beginning of the
phase in normal adult gait, similar to what occurs in small children
until around 3 years of age. Since the co-contraction of the TA and LG
in stance phase decreases or disappears in stable supported walking
of a baby and the patient (Fig. 6-8), those patterns of excessive

myoelectric activity are indicative of an unstable walking pattern. To
efficiently control body balance in various postures, the TA and LG
should operate in reciprocal patterns. Co-contraction of these two
muscles suggests effort to stabilize the ankle joint. In our patient
strong continuous discharges of the TA were observed bilaterally in
stance phase 1 month after the stroke. At 7 months after the stroke,
such excessive patterns of the TA during stance phase, suggesting
very unstable walking, rarely appeared on either side, indicating that
walking at this stage was much closer to normal.
The elderly subject in this study had a forward posture with slightly
flexed knees, a tendency pointed out by Crithley (1956).
In stance phase, activity of the VM not seen in normal adult gait was
strong and continuous, resembling the pattern seen in a baby during
the very first month of independent walking. This muscle activity
might be interpreted as helping to maintain balance with slightly flexed
knees. Similarly, strong continuous activity of the LG, BF, and GM in
our patient not seen in the normal adult were like the activities of
these muscles seen in small children until around 3 years of age. Such
activities of these antigravity muscles might be attributable to the
patient's forward posture (Figs. 6-5 and 6-7). When the patient used
support while walking and attained an upright posture, the excessive
activities of these muscles tended to decrease or disappear, looking
instead like the stable normal adult pattern (Fig. 6-8). These excessive
discharge patterns of the VM, LG, BF, and GM in stance phase thus
suggest an unstable walking pattern.
Just as Finley et al. (1969) suggested that the elderly persons in
their study were trying to preserve stability, the increased EMG of our
patient not seen in normal adult walking may reflect an attempt to
control progression and stability or the age-related changes in gait
posture. "Age-related" changes in walking behavior may be caused
either by the normal aging process or by disease, but it is difficult to
determine which changes in gait occur as part of the normal aging
process and which occur as the sequelae of pathological processes. In
either case, it is necessary to evaluate all age-related changes in gait
and to compare them with normal human gait, including the
development of walking.
Our patient showed excessive muscular activities compared with

normal adult walking at 1 month after a stroke. At 7 months after the
stroke, the muscle activities indicative of unstable walking tended to
decrease or disappear, although signs of excessive muscular activity
still remained. When stable supported (hand cart) walking with the
trunk upright was performed, the EMG patterns were closely similar
to those of a stable adult walking pattern (Fig. 6-8). From the EMG
data in this study, we saw that we could elicit a more normal adult
walking pattern by substituting for missing postural and balance
elements. These results suggest the validity and usefulness of
evaluating the recovery of walking after a stroke by comparing EMG
data with patterns seen in normal human gait, including the
development of walking in early childhood.
~
HC
TO
TA
LG
VM
RF
BF
GM
sw
ST
1 sec
CANE
I 0.5 mv
sw
ST
1 sec
HAND CART
sw
ST
1 sec
0.5 mv
0.5 mv
HAND CART
(ERECT POSTURE)
Fig. 6-8. EMGs of the right leg during supported walking at 1 year 7 months after stroke (same
subject as in Fig. 6-6).
TO: toe off, He: heel contact, SW: swing phase, ST: stance phase, TA: tibialis anterior, LG:
lateral gastrocnemius, VM: vastus medialis, RF: rectus femoris,
SF: biceps femoris, GM:
gluteus maximus.
Conclusion
To study the recovery of functional mechanisms of the leg muscles
of an elderly man following a cerebral infarction, we studied the
activities of his leg muscles during gait at 1 month, 7 months, and 1
year 7 months after his stroke.
The muscle activities at 1 month after the stroke closely resembled
activity patterns seen in the very unstable independent gait of a 1 year
old baby in the first month of learning to walk.
At 7 months after the stroke, these abnormal patterns of activity in
the tibialis anterior (TA) and lateral head of the gastrocnemius (LG)
tended to decrease or disappear, although activity of the LG in the
affected right leg during the latter half of swing phase still remained.
Muscle activities associated with flexed knees in the vastus medialis
(VM) and with a forward inclined trunk in the LG, biceps femoris (BF),
and gluteus maximus (GM) were still marked in stance phase
bilaterally. These EMG patterns were similar to slightly unstable
walking seen in small children until around 3 years of age.
When the patient walked with the support of a hand cart and was
able to hold his trunk upright, the strong burst of activity in the LG in
the latter half of right swing phase disappeared, and the excessive
continuous patterns of activity in the LG, VM, BF, and GM associated
with forward sway of the trunk with the knees flexed in stance phase
decreased or disappeared, closely resembling the stable adult walking
pattern.
We recommend further study of the evaluation of recovery of
walking after stroke by comparing patients' EMG patterns to those not
only of normal adult human gait, but also those of the development of
human walking in early childhood.
References
Basmajian, J. v., & Deluca, C. J. (1985). Human locomotion. In
Muscles Alive (pp.367-388). Baltimore: Williams & Wilkins.
Burnett, C. N., & Johnson, E. W. (1971). Development of gait in
childhood. Part II. Develop. Med. Child. Neurol., 13(2),207-215.
Crithley, M. (1956). Neurologic changes in the aged. ]. Chron. Dis., 3,
459-477.
Finley, F. R; Cody, K A; & Finizie, R V. (1969). Locomotion patterns
in elderly women. Arch. Phys. Med., 50, 140-146.
Forssberg, H. (1985). Ontogeny of human locomotor control. 1. Infant
stepping, supported locomotion and transition to independent
locomotion. Exp. Brain. Res., 57, 480-493.
Kazai, N.; Okamoto, T.; & Kumamoto, M. (1976) . Electromyographic
study of supported walking in infants in the initial period of
learning to walk. In P. V. Komi (Ed.), Biomechanics V-A (pp.
311-318) . Baltimore: University Park Press.
McGraw, M. B. (1940). Neuromuscular development of the human
infant as exemplified in the achievement of erect locomotion. ].
Pediat., 17,747-771.
Milani-Comparetti, A, & Gidoni, E. A (1967). Routine developmental
examination in normal and retarded children. Develop. Med.
Child. Neurol., 9, 631-638.
Nashner, L. M., & McCollum, G. (1985). The organization of human
postural movements: A formal basis and experimental synthesis.
Behavior Brain Sci., 8, 135-172.
Okamoto, T.; Okamoto, K; & Andrew, P. D. (2003). Electromyographic
developmental changes in one individual from newborn stepping
to mature walking. Gait and Posture, 17, 18-27.
Okamoto, T.; Okamoto, K; & Andrew, P. D. (2001). Electromyographic
study of newborn stepping in neonates and young infants.
Electromyogr. Clin. Neurophysiol., 41, 289-296.
Okamoto, T., & Okamoto, K (2001). Electromyographic characteristics
at the onset of independent walking in infancy. Electromyogr. Clin.
Neurophysiol., 41, 33-41.
121
Okamoto. T.; Tsutsumi, H.; Goto, Y.; & Andrew, P. D. (1987). A simple
procedure to attenuate artifacts in surface electrode recordings by
painlessly lowering skin impedance. Electromyogr. Clin.
Neurophysiol., 27, 173-176.
Okamoto, T., & Goto, Y. (1985). Human infant pre-independent and
independent walking. In S. Kondo (Ed.), Primate Morphophysiology, Locomotor Analyses and Human Bipedalism (pp.
25-45) . Tokyo: University of Tokyo Press.
Okamoto, T.; Goto, Y.; & Kumamoto, M. (1983). Electromyographic
study of the bifunctional leg muscles during the learning
process in infant walking. In H. Matsui & K. Kobayashi (Eds.),
Biomechanics VJll-A (pp 419-422). Illinois: Human Kinetics
Publishers.
Okamoto, T., & Kumamoto, M. (1972). Electromyographic study of the
learning process of walking in infants. Electromyography, 12,
149-158.
Sutherland, D. H.; Olshen, R.; Cooper, L.; & Woo, S. L. (1980). The
development of mature gait. ]. Bone. Joint. Surg., 62-A, 3, 336-353.
Thelen, E.; Ulrich, B. D.; & Jensen, ]. L. (1989). The developmental
origins of locomotion. In M. H. Woollacott & A Shumway-Cook
(Eds.), Development of Posture and Gait Across the Life Span (pp.
25-47). South Carolina: University of South Carolina Press.
Thelen, E., & Cooke D. W. (1987). Relationship between newborn
stepping and later walking: A new interpretation. Develop. Med.
Child. Neurol., 29, 380-393.
Thelen, E., & Fisher, D. M. (1982). Newborn stepping: An explanation
for a "disappearing" reflex. Dev. Psychol., 18, 760-775.
Touwen, B. C. L. (1976). Neurological Development in Infancy. London:
Spastics International Heinemann.
Touwen, B. C. L. (1971). A study on the development of some motor
phenomena in infancy. Develop. Med. Child. Neurol., 13, 435-446.
Winter, D. A (1991). The Biomechanics and Motor Control of Human
Gait. Ontario: University of Waterloo Press.
Zelazo, P. R.; Zelazo, N. A; & Kolb, S. (1972). "Walking" in the
newborn. Science, 176,314-315.
122
The books which were quoted
Primale Morphopbysiology,
LocomOior Anal)' cs
and Human Bipedali m
1) Basmajian, ]. V. (1974) . Muscles Alive. Baltimore : Williams &

Wilkins.
2) Kondo, S. (Ed.) (1985). Primate Morphophysiology, Locomotor
Analyses and Human Bipedalism. Tokyo: University of Tokyo Press.
3) Lois, B. (1994) . Motor Skills Acquisition in the First Year. Texas:
Therapy Skill Builders.
4) Leonard, C. T. (1998). The Neuroscience of Human Movement.
St. Louis: Mosby-Year Book.
5) Woollacott, M. H., & Shumway-Cook, A. (Eds.) (1989) . Development
of Posture and Gait Across the Life Span. South Carolina: University
of South Carolina Press.
123
Mature adult walking pattern
The development and degradation of gait
TO Fe
TO
1.9 years
He
""' I'"'"'"
r.J
+-H
~r'
TO
He
3 years
r---r--'"~r-
~..~ .
.I"
.,0,
""..
" /"
UI"
r1l"
.L
I~
,,,-
'~'r
VM -Hl<\lJo1rn!1/1J.+JI~~-...f.II
RF~~__-m~__~~
..~
SF '*'*~'*""ftIJl
OM -t--Mi!I>---h-+'!flII'
sw
ST
SW
~O.5mv
INFANT
WALKING
.....
n ST
'T
I
~ O.5 mv
TA
He
29 y ears
TO
ST
~ O.5mv
IMMATURE
CHILD WALKING.....
kU~k A~jtA
TO
SW
He
60 years
TO
MATURE
WALKING
ffff{
He
85 y ellr s
/'""
JiI,
LO
nl~
VM
RF
SF
'II
1.11
OM
sw
ST
I sec
I 0.5 mv
ADULT
WALKING
sw
ST
I sec
I 0.5 mv
SW
ST
~ O.5 mv
ELDERLY
WALKING
TO: toe off, FC: foot contact, HC: heel contact, SW: swing phase, ST: stance phase, TA:
tibialis anterior, LG: lateral gastrocnemius, VM : vastus medialis, RF: rectus femoris , SF: biceps
femoris, GM : gluteus maximus.
126 A ppendix
IMMATURE
INFANT WALKING
PATTERN
IMMATURE
CHILD WALKING
PATTERN
MATURE
ADULT WALKING
PATTERN
[[}srn~rn\
[1J
ADULT WALKING
r:i,~~~NI+1 E1RI~;1:r~N
UNSTABLE
SUGHTLY UNSTABLE
I)
STABLE

Region
Ankle
Code
Interpretation
Indication
ST-TA
Activity of the TA in stance phase
Very unstable
SW-LG
Activity of the LG in the latter part of swing phase
Unstable
ST-LG
Activity of the LG in the first half of stance phase
Slightly unstable
SW-VM
Activity of the VM in the latter part of swing phase
Very unstable
ST-VM
Activity of the VM in stance phase
Very unstable
ST-RF
Activity of the RF in stance phase
Very unstable
ST-BF
Activity of the BF in stance phase
Slightly unstable
ST-GM
Activity of the GM in stance phase
Slightly unstable
Knee
Knee & Hip
Hip
ST: stance phase, SW: swing phase, TA : tibialis anterior, LG : lateral gastrocnemius,
VM: vastus medialis, RF: rectus femoris , BF : biseps femoris, GM : gluteus maximus.
Appendix 127
EMG experiment of infant walking

EMG signals were recorded via silver electrodes coated with silver
chloride. The electrodes were 5 mm in diameter. Muscles chosen for
recording, based on our previous studies of gait (Okamoto et aI., 1972,
1985, 2001, 2003), were the long head of the biceps femoris (BF) ,
gluteus maximus (GM), lateral head ofthe gastrocnemius (LG), rectus
femoris (RF), tibialis anterior (TA), and vastus medialis (VM). Two
electrodes were placed on each muscle 5 mm apart from edge to edge
of the electrode collars in the direction of the muscle fibers, midway
between the proximal and distal ends of the superficially palpable part
of the muscle belly. A reference electrode was placed over the patella.
The skin at each electrode locus was scratched lightly with a needle,
reducing the resistance between pairs of electrodes to less than 5000Q
(Okamoto et aI., 1987). The signals were amplified one million times by
biological amplifiers and recorded on a pen-writing electroencephalograph with the paper speed set at 60 mm/s and an amplitude of 12
mm representing a voltage change of 0.5 mV. The upper limit of the
frequency response of the pens was 120 Hz at -3 dB.
The child's gait was recorded with a video camera at 60 frames/s, and a
voltage pulse synchronizing each frame was recorded with the EM Gs. From
the video recordings we determined when and, if feasible, how the foot
contacted the floor and made qualitative observations about joint motions.
With this instrumentation we were able to divide gait into stance and
swing phases and judge presence and absence of muscle activity. We did
not generally need to assess relative intensity of activity for a given
muscle, and when that was done it was only for such obvious
distinctions as "large" versus "small". At the beginning of each session,
clean baselines were established for each channel of EMG. Because of
the low skin impedance at each electrode locus, no artefact appeared
due to movement of electrode cables (Okamoto et aI., 1987). An artefact
was producible only by striking the electrode itself with a flick of the
finger, a procedure used to confirm proper channel assignments for the
muscles recorded. Small voltage changes beyond the baseline were
generally interpreted as cross talk from nearby muscles. Activity from
the target muscle was crisp and large in contrast, and accorded with
reasonable expectations in terms of conceivable motor activity. A stricter
criterion would have been difficult to establish, especially when the
child was in the earlier stages of development.
128 Appendix
Certain technical limitations in our study are noteworthy. We used a

pen-writer to record EMG, so the higher frequency components of the
myoelectric signal were lost. Since our observations were qualitative
rather than quantitative and since the significant content of EMG was
contained below 100 Hz, we believe that our assertions based on data
thus recorded are fundamentally valid. Another potential problem in
this study was the possibility of cross talk. Had this been an obvious
problem it would have been apparent in correlated patterns among the
numerous recordings examined. Muscles were deemed to be active
only if crisp and clear waveforms were recorded rather than lowamplitude voltages with rounded peaks. Since the observations of the
present study are based on a few babies, our findings should be
confirmed in longitudinal studies by other investigators, hopefully with
other muscles recorded as well and with more precise methods of
kinematic recording.
My first EMG experiment in infant walking (1967)

The subject is my daughter (co-author), and my wife is providing strong cooperation.
Appendix 129
An infant's first steps alone are an achievement in the struggle for the development
of balance, postural control , and strength during the first of year of life.
Acknowledgements
In 1967, with the help and cooperation of my wife, I decided to make
my two daughters the subjects of study. I learned that no cross-sectional
and longitudinal electro myographic studies of infant walking were being
carried out anywhere in the world and I decided to use my daughters,
two-year old Kayoko and six-month old Emi, as they then were, as the
subjects of my study, and I succeeded in creating electromyographic
records of the development process in which infants acquired independent walking and developed further skills.
My studies attracted almost no attention in Japan. However, Dr. John V.
Basmajian, Professor at the School of Medicine, Emory University,
Atlanta, U.S.A., the world authority on gait, recognized the value of my
studies when I presented a report at the International Congress of
Electromyography held in Brussels in 1971, and he invited me to join
the gait research project at the Center for Rehabilitation Medicine
at Emory University. Since then I have made electromyographical
recordings of the walking of more than a thousand people, ranging from
the newborn to the elderly, and accumulated much valuable data on
gait. This time too, thanks to the help of my wife and daughters, I was
able to analyze this data and to gather the results into a book that would
introduce them for the first time to the world.
This study of gait carried out by one Japanese family may seem a
very small step in the eyes of the world, but I firmly believe that our
study will contribute the development of gait studies in the next
generation.
I would like to express my gratitude to all the children and their
parents, together with many supporting staff, who took part in and
played a vital role in the difficult experiments involving the
electromyographical recording of infant walking. I would also like to
express my sincere gratitude to my wife and daughters who supported
my studies right from the first experiment up to the publication of this
book, and who will continue, I am sure, to give me their support.
131
Last but not least, I would like to express my gratitude and affection
to my darling grandchildren, who appear in this book and who will be
the bearers of the next generation. I would like you to follow your own
chosen paths with full confidence into the future, just as I plan to
continue in "my way" until the end of my own life.
Tsutomu Okamoto
"Let's leave something wonderful for the children!"
132
About the Authors

Director of Walking Development Group
Professor Emeritus of Kansai Medical University
Dr. Tsutomu Okamoto has researched electromyographical gait studies

for forty years. He was a member of the gait research project at the
Center for Rehabilitation Medicine at Emory University, Atlanta,
U. S. A. His papers have been published in many professional journals
and his electromyographical data on infant walking is cited widely
internationally. He won the All Japan Canoe Championships (1961),
served as one of the Tokyo Olympics Special Coaches and the President
of the All Japan Intercollegiate Canoe Fderation (1990-2002). He has
received many awards, including the Minister of Education, Culture,
Sports, Science and Technology Award, Prince Chichibu Memorial
Award and the Osaka Governor Award.
133
Kayoko Okamoto, Ph.D.

Researcher of Walking Development Group
Part-time Lecturer at Kyoto University
Dr. Kayoko Okamoto became the first subject of her father's electromyographical experiments when she was one year old. She has continued
gait research and development together with her father, Dr. Tsutomu
Okamoto, and won the Incentive Prize from the Japanese Society of
Biomechanics. She has researched and developed a unique walking
method and exercise program for preventing falls and avoiding becoming
bed-ridden based on her own grandfather's experience, who regained
independent walking power with rehabilitation training from a bedridden
condition caused by cerebral infarction. Her lectures on anti-aging
walking are extremely popular among middle-aged and elderly persons.
She teaches the theory and practice of "exercise for a heaithy life" at
university.
134
EMG
WALKING
The Walking Development Group aims through its lectures, educational

activities, and publications to spread knowledge of international
research into walking in an easily understandable form and hopes
through its activities to increase the number of people in the world who
are overflowing with youth and health.
Our message is "Let's leave something wonderful for the children!"
We aim to create a society where the children who will carry the world
forward are able actively to pursue their aims in a spirit of vitality.
Director of Walking Development Group

~lTIm~liJf~pJf
G-804 Tenno 2-6, Ibaraki-shi, Osaka 567-0876, JAPAN

http://www13.ocn.ne.jpFhokou/
135
Development of Gait
Same subject as cover photograph (at 3 years of age)
The children acquire mature adult walking pattern, which indicates

strong push off motions of the foot, with the body erect, at about 3
years of age. This foundation will provide them with a basis to
continue walking across the life span.
Okamoto & Okamoto

(Tsutomu and Kayoko Okamoto) Development of Gait B

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Tsutomu and Kayoko Okamoto) Development of Gait B

Uploaded by

Copyright:

Available Formats

Development of Gait

Walking Development Group

Copyright 2007 by Okamoto & Okamoto

Walking Development Group

G-S04 Tenno 2-6, Ibaraki-shi,

acquisition and mastering of infant independent walking to that of the

Development of Gait -Birth to Age Eight-

1 . Newborn Stepping in Neonates and Young Infants

2 . Independent Walking in Infants

3. From Newborn Stepping to Mature Walking

Application to Gait Analysis and Evaluation

4. An Index of Gait Instability

-Based on the Development of Independent Walking-

EMG findings during the development of gait

5. Application of an Index of Gait Instability (1)

Until the 1st month of age

6. Application of an Index of Gait Instability (2)

-Recovery of Walking in an Elderly Man after Stroke- 107

Development of Gait by Electromyography

EMG experiment of infant walking

The electromyographic (EMG) recordings were done with a pen-writing mUltipurpose

(Okamoto et aI., 1987).

Neonatal stepping at 3 weeks afte birth

When a newborn infant is held under the arms in an upright

Newborn stepping at 2 days after birth.

RF: rectus femoris

(Knee extensor, Hip flexor)

VM: vastus medialis

(Ankle plantar flexor)

Fig. 1-2. Muscles chosen for recording EMG.

Newborn Stepping in Neonates and Young Infants 5

Surface electrodes 5 mm in diameter were used. To attenuate

Mature adult walking pattern

Typical EMG pattern of normal adult walking

Tibialis anterior (TA)

Lateral gastrocnemius (LG) -+---M~"""~I M'H~---1<Y,\,~-...."I(NJIN't--t-

Vastus medialis (VM)

Biceps femoris (BF)

Gluteus maxim us (GM)

Newborn Stepping in Neonates and Young Infants

Early neonatal period

Fig. 1-4. EMGs of stepping at 1 week after birth (Y.T.).

Stepping in this period was characterized by quick hip flexion in

1 week after birth (Fig. 1-4): In the beginning of stance phase, no

Fig. 1-5. Foot contact of newborn stepping at 2 days after birth.

Newborn Stepping in Neonates and Young Infants 9

V, , , , / " " 1,.1/ " I

W. , / , . , , W; " " I I " I WI I I ;

2 weeks (A. I.)

Fig. 1-6. EMGs of stepping at 2 weeks after birth (A.I.).

2 weeks after birth (Fig. 1-6): In the beginning of stance phase,

Foot contact (Fe)

Foot flat with slow leg extension

Newborn Stepping in Neonates and Young Infants

Late neonatal period

Fig. 1..8. EMGs of stepping at 3 weeks after birth (H.Y.).

Fig, 1-9, EMGs of stepping at 4 weeks after birth (T.Y,),

Newborn Stepping in Neonates and Young Infants

Lateral border with slow leg extension

EMGs of leg extensors before floor contact

Fig. 1-11. Neonatal stepping at 26 days after birth.

Fig. 1-12. Foot contact of neonatal stepping at 22 days after birth.

Newborn Stepping in Neonates and Young Infants