Gait & Posture 25 (2007) 380–384

www.elsevier.com/locate/gaitpost

Medial gastrocnemius muscle behavior during human

running and walking
M. Ishikawa *, J. Pakaslahti, P.V. Komi
Neuromuscular Research Center, Department of Biology of Physical Activity, P.O. Box 35(LL220),
FIN-40014, University of Jyväskylä, Finland
Received 20 December 2005; received in revised form 28 April 2006; accepted 2 May 2006

Abstract

Utilization of elastic energy in the tendinous tissues (TT) of the human skeletal muscle may be task dependent. The present study was
designed to investigate this problem by comparing the fascicle-TT interaction of the medial gastrocnemius muscle (MG) during ground
contact of running and walking. Seven subjects ran and walked with a natural cadence. Ankle and knee joint angular data were recorded by
electrogoniometers for estimating the entire MG muscle-tendon unit (MTU) length, together with the ground reaction forces. The MG fascicle
length was measured by using the high-speed ultrasound image scanning during movements. The results showed that in running, after the
rapid early fascicle stretching (0–10% of the contact period), the fascicles shortened throughout the ground contact while TT was stretched
prior to shortening. In walking, the fascicles shortened initially (0–15% of the contact period) due to sudden plantar-flexion. Thereafter, the
fascicles and TT lengthened slowly until the end of single support (15–70% of the contact period.). The fascicles then shorted during the push-
off phase (70–100% of the contact period). These results demonstrate that the MG fascicles behaved differently between running and walking
and did not follow the length change pattern of the MTU during the ground contact period. The estimated working range of active muscle
fibers in force–length relationship could shift more to an ascending limb (shorter length) phase in running than in walking. These results
suggest that MG fascicles can work within the optimal working range of the sarcomeres in the force–length relation but are responsible for the
effective utilization of the TT elasticity during human running.
# 2006 Elsevier B.V. All rights reserved.

Keywords: Gait; Locomotion; Ultrasonography; Elastic energy

1. Introduction [3–6]. The question arises whether it is possible to store and

Elasticity of the skeletal muscle plays an important role
by improving the power output and efficiency in human
locomotion, such as hopping and running, where the muscle
is actively stretched (the braking phase) prior to the
concentric contraction (the push-off phase). This supports
the elastic utilization concept [1,2]. This behavior is called a
stretch-shortening cycle (SSC) action [3] on the entire
muscle-tendon unit (MTU) level. However, muscle fiber
displacement, pennation angle of muscle fibers and/or
tendon strain can substantially affect the entire MTU length,
and has different implications for muscle-tendon function
* Corresponding author. Tel.: +358 14 260 2086; fax: +358 14 260 2071.
E-mail address: masaki.ishikawa@sport.jyu.fi (M. Ishikawa).
subsequently utilize the tendinous elastic energy in the same
way during the different SSC exercises, for example
between hopping, walking and running. Much of what we
know about human locomotion and the interaction between
fascicles and tendinous tissues (TT), which include the
aponeuroses and the free length of tendon, is derived from
studies examining a single locomotory task only.
Studies on animal locomotion have revealed that
skeletal muscle function may be different for different
conditions and tasks [7,8]. It has been proposed that the
interaction between fascicles and TT can be changed via
alterations in the pattern of muscle activations (timing,
duration and intensity) when speed, gait task or environ-
ment change [7]. In human movements, it is unknown how
the muscle fascicles and TT interact in different locomotion

0966-6362/$ – see front matter # 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.gaitpost.2006.05.002
 M. Ishikawa et al. / Gait & Posture 25 (2007) 380–384 381

tasks. Consequently, the issue of how skeletal muscles
function might change with different tasks has remained
relatively unexplored. In human drop jump exercises, the
fascicles of the medial gastrocnemius muscle (MG)
shortened throughout the ground contact [9]. On the other
hand, recent human walking studies showed that MG
fascicles contract isometrically [10] or even lengthen [11]
during the single support phase (15–70% of the contact
period) when MTU and TT are stretching. These different
and perhaps contradicting MG fascicle behaviors during
SSC exercises may suggest existence of a movement
specific fascicle behavior. Consequently, the fascicle-TT
interaction during human locomotion needs further
clarification. To begin to address issues relating to
variability in human skeletal muscle function, the purpose
of the present study was to examine how the fascicle-TT
interaction together with muscle activation varies in medial
gastrocnemius muscle (MG) due to the different tasks of
walking and running.
2. Methods
Seven healthy male volunteers (age 26.6  2.7 years,
height 183.4  6.0 cm, body mass 80.9  15.2 kg;
means  S.D.) participated after being informed all the
risks associated with this study, which was approved by the
Ethic Committee of the University of Jyväskylä. The
subjects walked with a natural cadence (1.48  0.12 m s1)
and ran (2.74  0.21 m s1) twice on a 10 m long force
platform system (Raute Inc, Finland) (Fig. 1a). The ground
reaction forces on the right and left sides were recorded
separately to define the ground contact phases. Real-time
ultrasound devices (SSD-5500, Aloka, Japan) were used to
measure the fascicle length of MG during running and
walking [9–11]. A 7.5 MHz linear-array ultrasound probe of
6 cm long was put over the MG muscle and fixed securely.
The longitudinal ultrasonographic images in MG were
scanned (96 images s1) during the movements (Fig. 1b).
The MG fascicles were identified in each image along their
length from the superficial and deep aponeuroses. One end
of the fascicle, which was extended off the acquired
ultrasound image, was sometimes estimated. Surface bipolar
Ag–AgCL electromyographic (EMG) electrodes (5 mm
diameter with interelectrode distance 20 mm) were used to
record the EMG activities in MG (Glonner electronic,
Munich, Germany; input impedance >25 MV, common
mode rejection ratio >90 dB). Ankle and knee joint
rotations were recorded by electrogoniometers attached
over the right lateral melleoli and popliteal crease,
respectively. These signals were stored simultaneously by
a 12-bit A/D converter with a sampling frequency of 1 kHz.
2.1. Analyses
To combine the analysis of the various parameters, the
MTU length of MG data was re-sampled at 100 Hz and the
fascicle length data at 96 Hz was interpolated at 100 Hz so
that the time scales matched. The regression equations of
Hawkins & Hull [12] were used to estimate MTU length in
MG from measurements of ankle and knee joint flexion
angles. The TT length was calculated by subtracting the
fascicle length multiplied by the cosine of the fascicle angle
from the MTU length [9–11]:
LTT ¼ LMTU  Lfa  cos a
where LTT is the TT length, LMTU is the muscle-tendon unit
length, Lfa is the fascicle length and a is the angle created by
the fascicle line and its insertion into the aponeurosis lines
(Fig. 1b). The length changes were calculated from the time
of initial foot contact to the peak length and from the peak
length to the time of toe-off, respectively.
After the EMG signals were rectified, 10 step cycles of
the signal data (EMG, ground reaction forces and joint angle
data) were averaged for each subject. In the ultrasonographic
data the averaging was done for three consecutive step

Fig. 1. (a) Set-up for the walking and running experiments: the ultrasound images of fascicles and EMG of the medial gastrocnemius muscle (MG) were
recorded simultaneously on a unique 10 m long force plate system, composed of two rows of individual force plates. The ultrasound apparatus was pushed at the
speed of the subject outside the force plate area. The ankle and knee joint angles were also measured simultaneously by using goniometers. (b) The calculation of
the fascicle length (Lfa) and the angle from medial gastrocnemius (MG) muscle (the width and depth scanning areas are 58.5 and 60.0 mm, respectively).
382 M. Ishikawa et al. / Gait & Posture 25 (2007) 380–384

Fig. 2. Records of the measured and calculated parameters during the contact phase of running and walking. Vertical (A; Fz) and horizontal (B; Fy) ground
reaction forces, EMG activities of medial gastrocnemius (C; MG), the length of muscle-tendon unit (D; MTU), fascicle (E) and tendinous tissues (F; TT). The
contact period was normalized as 100%. Bars show the values of the 1 S.E. The vertical lines denote the contact, the transition point from the braking to push-
off phases and the take-off.

cycles. Coefficient of variation for fascicle length and angle tively) than in walking (48.9  2.5 and 44.1  3.8 cm,
of three step data were in the range of 1–6% and 1–7%, respectively). Table 1 shows the length changes from the
respectively. The EMG signals were integrated and then initial ground contact to the peak lengthening point. Both
averaged (aEMG) in the following three phases; preactiva- MTU and TT demonstrate greater stretching amplitudes in
tion, braking and subsequent push-off phases. The pre- running than in walking (Table 1, p < 0.05). However, the
activation was defined as the 100 ms preceding ground MG fascicle behavior was different between running and
contact and the braking and push-off phases were divided by walking (Fig. 2E). In running, after the initial ground contact,
the zero level of the horizontal ground force. the fascicles were lengthened by 0.4  0.3 cm immediately
after ground contact (0–10% of the contact period). They then
2.2. Treatment of data shortened throughout the contact period. Lfa in running was
greater at the initial ground contact (5.9  1.8 cm) than at the
Mean (S.D.) values were calculated for seven subjects toe-off (4.3  1.3 cm) ( p < 0.05). In contrast, the fascicles in
unless otherwise stated. A one-way repeated ANOVA on walking lengthened until the middle of the push-off phase
rank was used to analyze the significant difference of the
EMG activity between the different phases during the
movements. Where ANOVA was significant, a least Table 1
significant difference post hoc test was used to look for Measured and calculated parameters (n = 7).
significant differences between phases. The Wilcoxon Running Walking
signed-rank test was used to test for a difference between Contact time (ms) 296  28.4 629  33.0**
the parameters (EMG, length changes and ground reaction Averaged EMG (mV)
force) of walking and running. We considered p < 0.05 as a Pre-activation
statistically significant difference. Braking phase
Push-off phase
DLMTU (cm)
Contact-peak 2.2  0.9 1.2  0.4*
3. Results Peak-toe-off 4.0  0.3 3.7  0.5*
DLfascicle (cm)
Both running and walking showed the stretching and Contact-peak 0.4  0.3 0.2  0.3
Peak-toe-off 2.0  0.6 1.0  0.4*
subsequent shortening of the MTU and TT during the ground DLTT (cm)
contact period in all subjects (Fig. 2D and F). LMTU and LTT at Contact-peak 3.0  1.2 1.4  0.5*
the initial ground contact and at the toe-off points did not show Peak-toe-off 3.7  0.6 3.2  0.8
any significant difference between running and walking. Values are expressed as means  S.D. *, ** Significantly different between
However, LMTU and LTT at the peak lengthening point were running and walking at p < 0.05, p < 0.01, respectively. #, ## Significantly
longer in running (50.4  2.7 and 45.7  3.9 cm, respec- different between phases at p < 0.05, p < 0.01, respectively.
 M. Ishikawa et al. / Gait & Posture 25 (2007) 380–384
(70% of the contact period) after the initial shortening due to
the sudden plantar flexion (10% of the contact period). They
then shortened during the late push-off phase (70–100% of the
contact period). In comparison between running and walking,
Lfa at the initial ground contact and at the peak lengthening
point did not show any significant difference between running
(5.9  1.8 and 6.3  1.5 cm, respectively) and walking
(6.1  1.4 and 6.3  1.5 cm, respectively). Lfa at the toe-
off point was greater in walking (5.3  1.3 cm) than in
running (4.3  1.3 cm) ( p < 0.05). The magnitudes of the
fascicle shortening from the peak Lfa until toe-off were greater
in running than in walking ( p < 0.05, Table 1), but the
stretching magnitudes did not show any difference. The
magnitudes of the TT shortening were not significantly
different between tasks but in running the TT stretching was
greater than in walking (Table 1).
The peak vertical ground reaction force (Fz) was greater
in running (1805  282 N) than in walking (978  215 N)
( p < 0.05). When compared to walking, the MG aEMGs
were greater in the preactivation and braking phases of
running ( p < 0.05). This was the opposite in the push-off
phase ( p < 0.05) (Table 1). Thus in running EMG activity of
MG occurred primarily during the braking phase (Fig. 2C,
Table 1). In contrast, the greater muscle activation in
walking was observed during the early push-off phase.
4. Discussion
We investigated the modulation of the MG fascicle-TT
interaction in response to changes in mechanical demands
associated with tasks. The TT stretching magnitudes of MG
from the contact to peak length was greater in running
(7.4  2.5%) than in walking (4.1  0.9%). A similar
difference was observed for the MTU stretching (running
4.3  1.7%, walking 2.5  0.6%). In running, however, the
MG fascicles behaved differently from the MTU and TT.
The MG fascicles in running shortened throughout the
ground contact after the rapid fascicle stretching (0–10% of
the contact period). These results can indicate the existence
of the fascicle-TT interaction, which is movement specific in
both human walking and running.
Roberts et al. [8] have shown that in the turkey running,
the tendon stretched and recoiled as the ankle flexed and
extended while the foot was on the ground, but the fascicles
remained almost constant in length. The present results also
demonstrated that TT was stretched during the braking
phase, followed by shortening during the push-off phase in
running. Thus, we confirmed the existence of the elastic
behavior of TT, which allows for storage of elastic energy in
the first half of the step and its release in the second half of
the contact phase [13]. In this TT behavior, the MG fascicles
shorten throughout the ground contact phase. This MG
fascicle shortening during the MTU stretching accompa-
nied by the higher EMG activation (Fig. 2) can increase the
TT stretch as compared to the fascicle lengthening or the
383
constant length. However, it has been proposed that the
concerted fascicle action, where activation can be matched
to the imposed load in order to keep the length of the
contractile component constant [14], can favor the effective
storage of TT elastic energy during the ground contact.
These different fascicle length patterns (shortening and/or
constant) may also be related to the intensity of the
movement [9,15]. It can be speculated that in the MG
muscle the elastic energy in TT can be utilized due to the
fascicle shortening during the braking phase of the present
running task.
According to the previous human walking studies
[10,11], the interaction between the fascicles and TT can
play an important role in the storage and release of elastic
energy. In walking, however, the elastic energy cannot be
provided directly from initial impact due to the low Achilles
tendon force [11,16] but mainly from the muscle action.
Consequently, the MG muscle in walking was not strongly
activated during the preactivation and braking phases while
the MG fascicles were lengthened during the braking phase
(Fig. 2E). Both in running and jumping, however, the
Achilles tendon force (ATF) increases during the braking
phase [11]. The elastic energy can come mainly from the
initial negative work in running due to the strong muscle
activation during the preactivation and braking phases. In the
present running condition, the MG fascicles contracted
together with high muscle activation during the preactiva-
tion and braking phases (Fig. 2). Thus, the present results
suggest that the fascicle-TT interaction is movement specific
in both human walking and running.
It has been suggested that the muscle fiber contraction
occurs around the plateau region of the sarcomere force–
length curve during human movements [17]. Supporting
their suggestion, the estimation of the force–length relation
of the sarcomeres (Fig. 3) shows that the working range of
Fig. 3. The proposed working area of sarcomere length–force relation of
human gastrocnemius muscle during human locomotion ((a): Ref. [10], (b):
Ref. [21], (c): Ref. [9]). Average sarcomere lengths, which were estimated
by dividing the fascicle length by the average number of sarcomeres in
series in fascicles [22], have been estimated during human locomotion and
superimposed on a force–length relationship for human muscle derived
from the data of Walker and Schrodt (see Refs. [17] and [20] for details).
The shaded areas refer to the estimation which is based on the present
results.
384 M. Ishikawa et al. / Gait & Posture 25 (2007) 380–384
active muscle fibers in the force–length relationship can shift
more to a plateau (optimal) phase in the present walking
speed (1.5  0.1 m/s) than in a slower walking speed (Ref.
[10]; 0.8 m/s) with the advantages of the relatively larger
force generation. This possible fascicle behavior can support
the early suggestions that in walking elastic energy can be
provided mainly from muscle action. In running, however,
the working ranges of sarcomeres may shift more to an
ascending limb (shorter length). This shift in running may
favor for the storage of elastic energy in TT. These
assumptions suggest further that the modulation of the MG
fascicle-TT interaction takes place in response to changes in
mechanical demands associated with locomotory tasks and
need to be verified in future studies.
Previous studies addressing extensor muscle activities in
human hopping and running have identified a short-latency
stretch-reflex (SLR) response following touch-down
[18,19]. In line with these studies, a quick fascicle stretch
(Fig. 2E) was observed immediately after the ground contact
of running followed by the rapidly increasing EMG activity
within 30–75 ms after the initial contact. This stretch
induced reflex together with length changes of muscle fibers
has not been previously reported during in vivo human
locomotion. This is methodologically important for the
identification of the beginning of the stretch reflex loop
activation in the muscle.
There are certain points in the methods that need to be
addressed. In the present study as in many other similar
studies earlier, only one muscle (MG) was examined. In a
recent study [9] it was suggested that the fascicles in mono-
and bi-articular muscles cannot necessarily behave similarly
during SSC exercises. Thus, the present results must be
interpreted to apply only to the MG muscle and only for
running and walking. Secondly, there are differences in the
literatures (e.g. Ref. [20]) between the models for the MTU
length calculation, which can influence the estimation of TT
length. In our understanding, these differences deal mainly
with the calculation of the moment arms. However, we
believe that the task comparison can be made successfully in
the same subject group provided that the same MTU
calculation method has been applied. In conclusion, our data
showed that the MG fascicles behave differently between
walking and running during ground contact, although TT
can be stretched prior to shortening in human running and
walking. These task specific fascicle-TT interactions during
the ground contact are expected to have implications for
storage and release of elastic energy.
Acknowledgements
The authors gratefully acknowledge T. Rantalainen, H.
Piitulainen and M. Puttonen for assistance. This study was
supported by a grant #88/627/2005 (P.V. Komi) from the
Ministry of Education, Finland.
References
[1] Alexander R. McN, Bennet-Clark HC. Storage of elastic strain energy
in muscle and other tissues. Nature 1977;265:114–7.
[2] Cavagna GA, Saibene FP, Margaria. R effect of negative work on the
amount of positive work performed by an isolated muscle. J Appl
Physiol 1965;20:157–8.
[3] Komi PV. Stretch-shortening cycle: a powerful model to study normal
and fatigue muscle. J Biomech 2000;33:1197–206.
[4] Biewener AA, Roberts TJ. Muscle and tendon contributions to force,
work, and elastic energy savings: a comparative perspective. Exerc
Sport Sci Rev 2000;28:99–107.
[5] Fukunaga T, Kawakami Y, Kuno S, Funato K, Fukashiro S. Muscle
architecture and function in humans. J Biomech 1997;30:457–63.
[6] Narici MV, Binzoni T, Hiltbrand E, Fasel J, Terrier F, Cerretelli P. In
vivo human gastrocnemius architecture with changing joint angle at
rest and during graded isometric contraction. J Physiol 1996;496:287–
97.
[7] Biewener AA, Gillis GB. Dynamics of muscle function during loco-
motion accommodating variable conditions. J Exp Biol 1999;202:
3387–96.
[8] Roberts TJ, Marsh RL, Weyand PG, Taylor CR. Muscular force in
running turkeys: the economy of minimizing work. Science 1997;275:
1113–5.
[9] Ishikawa M, Niemela E, Komi PV. The interaction between fascicle
and tendinous tissues in short contact stretch-shortening cycle exercise
with varying eccentric intensities. J Appl Physiol 2005;99:217–23.
[10] Fukunaga T, Kubo K, Kawakami Y, Fukashiro S, Kanehisa H,
Maganaris CN. In vivo behavior of human muscle tendon during
walking. Proc R Soc Lond 2001;B268:229–33.
[11] Ishikawa M, Komi PV, Grey MJ, Lepola V, Bruggemann GP. Muscle-
tendon interaction and elastic energy usage in human walking. J Appl
Physiol 2005;99:603–8.
[12] Hawkins D, Hull ML. A method for determining lower extremity
muscle-tendon lengths during flexion/extension movements. J Bio-
mech 1990;23:487–94.
[13] Alexander RM. Tendon elasticity and muscle function. Comp Bio-
chem Physiol A Mol Integr Physiol 2002;133:1001–11.
[14] Hof AL, Geelen BA. Van den Berg J. Calf muscle moment, work and
efficiency in level walking; role of series elasticity. J Biomech
1983;16:523–37.
[15] Ishikawa M, Komi PV, Finni T, Kuitunen S. Contribution of the
tendinous tissue to force enhancement during stretch-shortening cycle
exercise depends on the pre stretch and concentric phase intensities. J
Electromyogr Kinesiol 2005;3 [Epub ahead of print].
[16] Komi. PV Relevance of in vivo force measurements to human
biomechanics. J Biomech 1990;23:23–34.
[17] Fukunaga T, Kawakami Y, Kubo K, Kanehisa H. Muscle and tendon
interaction during human movements. Exerc Sport Sci Rev 2002;30:
106–10.
[18] Dietz V, Schmidtbleicher D, Noth J. Neural mechanisms of human
locomotion. J Neurophysiol 1979;42:1212–22.
[19] Melvill-Jones G, Watt DGD. Observation on the control of stepping
and hopping movements in man. J Physiol 1971;219:729–37.
[20] Yuen TJ, Orendurff MS. A comparison of gastrocnemius muscle-
tendon unit length during gait using anatomic, cadaveric and MRI
models. Gait Posture 2005;23:112–7.
[21] Kurokawa S, Fukunaga T, Fukashiro S. Behavior of fascicles and
tendinous structures of human gastrocnemius during vertical jumping.
J Appl Physiol 2001;90:1349–58.
[22] Bobbert MF, Huijing PA, van Ingen Schenau GA. A model of the
human triceps surae muscle-tendon complex applied to jumping. J
Biomech 1986;19:887–98.
[23] Walker SM, Schrodt GRI. Segment length and thin filament period in
skeletal muscle fibers of the Rhesus monkey and the human. Anat Rec
1974;178:63–82.