The capacity of the human iliotibial band to store elastic energy during running

Full text

The capacity of the human iliotibial band to store elastic energy during
running
Carolyn M. Eng
a,b,
n
, Allison S. Arnold
a
, Daniel E. Lieberman
b
, Andrew A. Biewener
a
a
Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, MA, USA
b
Department of Human Evolutionary Biology, Harvard University, Cambridge, MA, USA
article info
Article history:
Accepted 15 June 2015
Keywords:
Elastic energy storage
Iliotibial band
Fascia
Musculoskeletal modeling
abstract
The human iliotibial band (ITB) is a poorly understood fascial structure that may contribute to energy
savings during locomotion. This study evaluated the capacity of the ITB to store and release elastic energy
during running, at speeds ranging from 2–5 m/s, using a model that characterizes the three-dimensional
musculoskeletal geometry of the human lower limb and the force–length properties of the ITB, tensor
fascia lata (TFL), and gluteus maximus (GMax). The model was based on detailed analyses of muscle
architecture, dissections of 3-D anatomy, and measurements of the muscles' moment arms about the hip
and knee in five cadaveric specimens. The model was used, in combination with measured joint kine-
matics and published EMG recordings, to estimate the forces and corresponding strains in the ITB during
running. We found that forces generated by TFL and GMax during running stretch the ITB substantially,
resulting in energy storage. Anterior and posterior regions of the ITB muscle–tendon units (MTUs) show
distinct length change patterns, in part due to different moment arms at the hip and knee. The posterior
ITB MTU likely stores more energy than the anterior ITB MTU because it transmits larger muscle forces.
We estimate that the ITB stores about 1 J of energy per stride during slow running and 7 J during fast
running, which represents approximately 14% of the energy stored in the Achilles tendon at a compar-
able speed. This previously unrecognized mechanism for storing elastic energy may be an adaptation to
increase human locomotor economy.
&2015 Elsevier Ltd. All rights reserved.
1. Introduction
Because bipedalism is a fundamental derived feature of homi-
nins (species more closely related to humans than chimpanzees),
many distinctive features of the human spine and lower extremity
are adaptations to improve bipedal locomotor performance. Many
adaptations for standing and walking, for example, appear early in
hominin evolution including a inferiorly-oriented foramen mag-
num, a lordotic lumbar spine, and a sagittally-oriented ilium (see
Aiello and Dean (1990) and Zollikofer et al. (2005)). Additional
features that first appear later in the genus Homo may reflect
selection for endurance running, including a stabilized sacroiliac
joint, an expanded attachment of gluteus maximus, and shorter
toes (Bramble and Lieberman, 2004;Lieberman et al., 2006;Rolian
et al., 2009). Although the selective factors underlying the evolu-
tion of both walking and running are debated, it is likely that
locomotor economy played a key role. Hypothesized energy-
saving features for walking include long legs and dorsally oriented
ischia (Crompton et al., 1998;Pontzer et al., 2009;Robinson, 1972;
Sockol et al., 2007). Energy saving features for running in the
genus Homo include a long, compliant Achilles tendon and a
spring-like median longitudinal arch, which are known to store
and recover elastic energy during running in other vertebrates
(Biewener, 2003;Ker et al., 1987;Roberts, 2002). In addition, the
human lower extremity has a number of fascial structures with
elastic properties that are not present in apes, but whether these
structures store energy or serve another function remains poorly
understood.
One of the most interesting of these structures is the iliotibial
band (ITB). The ITB is a thickening of the lateral fascia of the thigh
that originates on the pelvis and inserts on the tibia; it receives
muscle fibers from the tensor fascia lata (TFL) anteriorly and from
the gluteus maximus (GMax) posteriorly (Gottschalk et al., 1989;
Gray et al., 1995;Kaplan, 1958;Ober, 1936;Stern, 1972). The ITB is
traditionally considered to function as a “strut”during walking,
stabilizing the hip in the frontal plane (Gottschalk et al., 1989;
Inman, 1947;Kaplan, 1958). However, the high compliance of the
ITB (Butler et al., 1984;Derwin et al., 2008;Gratz, 1931), the fact
that it crosses both the hip and knee, and the presence of in-series
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/jbiomech
www.JBiomech.com
Journal of Biomechanics
http://dx.doi.org/10.1016/j.jbiomech.2015.06.017
0021-9290/&2015 Elsevier Ltd. All rights reserved.
n
Correspondence to: Department of Ecology and Evolutionary Biology, Brown
University, 34 Olive St., Box G-B204, Providence, RI 02912, USA.
E-mail address: carolyn_eng@brown.edu (C.M. Eng).
Journal of Biomechanics 48 (2015) 3341–3348

muscles suggest that the ITB may play other roles. If the ITB
stretches substantially while transmitting muscle forces, storing
elastic energy, then it may decrease the metabolic cost of loco-
motion. Prior studies have estimated that energy recovered from
the Achilles tendon during running reduces muscle work by as
much as 35% (Alexander and Bennet-Clark, 1977;Ker et al., 1987).
Whether the ITB also stores and recovers elastic energy, and how
this compares to Achilles tendon energy recovery, is unknown.
As a first step toward evaluating the ITB's role in locomotor
economy, this study examined the capacity of the ITB to store
elastic energy at running speeds ranging from 2 to 5 m/s. We
hypothesized that forces generated by TFL and GMax stretch the
ITB during running, storing elastic energy that may be recovered
later in the stride. We tested this hypothesis by developing a
musculoskeletal model of the ITB and inserting muscles. Our
model characterizes the 3-D skeletal geometry, the hip and knee
kinematics, and the attachments and force–length (F–L) properties
of the ITB, TFL and GMax for an average-sized adult male (femur
length: 39.8 cm; tibia length: 36.2 cm). Because existing repre-
sentations of TFL and GMax were not sufficiently accurate for this
study, we performed detailed analyses of these muscles' archi-
tecture and measured their moment arms (MAs) about the hip and
knee in cadaveric specimens. The TFL has largely been neglected in
previous studies of muscle architecture (e.g., Ward et al. (2009)
and Wickiewicz et al. (1983)) and locomotor function (e.g., Dorn
et al. (2012) and Sasaki and Neptune (2006)), despite being active
during running (Andersson et al., 1997;Mann et al., 1986;Mon-
tgomery et al., 1994;Paré et al., 1981). GMax is routinely modeled
as a uniarticular hip extensor that inserts on the femur (e.g.,
Arnold et al. (2010) and Delp et al. (1990)), despite evidence that a
substantial portion of GMax inserts on the ITB (Gray et al., 1995;
Stern, 1972). Our refined musculoskeletal model, which addresses
these limitations, is available on SimTK (simtk.org). Using this
model, we estimated the forces transmitted to anterior and pos-
terior regions of the ITB at body positions corresponding to run-
ning, predicted the length changes of each region, and calculated
the corresponding ITB strain energies over the course of a stride
based on published measurements of the tissue's elastic modulus
(Butler et al., 1984;Derwin et al., 2008).
2. Materials and methods
2.1. Muscle architecture measurements
We characterized the isometric force-generating capacity of TFL
and GMax based on measurements of muscle architecture in three
formalin-fixed human cadavers (Table 1). Specimens were dis-
sected and the muscles isolated and removed. Total mass (M)of
each muscle was measured; in addition, the masses of four regions
of the GMax were measured separately. A muscle fascicle was
carefully dissected from each region of GMax and from two
regions of TFL and the fascicle lengths (L
f
’) measured. Surface
pennation angles between the fascicles and ITB were also mea-
sured. Under magnification, muscle fiber bundles were isolated
from each fascicle and mounted on slides. Following Lieber et al.
(1990), bundle sarcomere length (L
s
') was determined by laser
diffraction and used to calculate optimal fascicle length (L
f
)
⎛
⎝
⎜⎞
⎠
⎟
LL m
L
2.7
1
ff
s
μ
=‵ ‵()
where 2.7
μ
m is the optimal sarcomere length for human muscle
(Lieber et al., 1994). Physiological cross-sectional area (PCSA)was
calculated from muscle mass and optimal fascicle length (Powell
et al., 1984)
PCSA M
L2
f
ρ
=⋅()
where
ρ
is muscle density (1.056 g/cm
3
;Mendez and Keys, 1960).
Our architecture data for GMax are consistent with data reported
by Ward et al. (2009).
2.2. Muscle moment arm measurements
We measured MAs of the muscle-ITB paths in five fresh frozen
cadaveric hemi-pelvises obtained from MedCure (Portland, OR).
MAs were determined for hip flex/extension, rotation, ab/adduc-
tion, and knee flex/extension using the tendon excursion method
(An et al., 1984;Brand et al., 1975). We approximated TFL with two
Kevlar thread paths (Fig. 1A and B) and GMax with four paths
(Fig. 1A and C). The ITB was left intact during these measurements.
Each thread was anchored to a screw eye at the path's insertion,
routed over the ITB through plastic tubing to a screw eye at the
path's origin, and attached to one of two cable-extension position
transducers (PTX101, Celesco, Canoga Park, CA) that measured
length changes with an accuracy of 70.32 mm while applying a
tension of 1.4 or 2.8 N. The tubing ensured a repeatable path along
the surface of the ITB and decreased friction. Detailed procedures
for defining each path are described in Supplementary materials.
Hip and knee angles were measured simultaneously with
muscle-ITB length changes using a motion tracking system (Pol-
hemus Fastrak, Colchester, VT) and custom software (LabView,
National Instruments Corporation, Austin, TX). Receivers were
rigidly attached to the pelvis, femur, and tibia to track the seg-
ments' positions and orientations. Segment coordinate systems
were defined along anatomical axes by digitizing bony landmarks
and determining the hip center (Fig. S1), as described in the
Supplementary materials. For each muscle-ITB path, we digitized
the origin, insertion and key “via”points that constrained the path
with hip or knee motion. We also tracked the relative motions of
nine marker pairs sutured along the ITB using high-speed video.
These data guided development of the model and were analyzed
to determine the hip and knee angles at which the anterior and
posterior ITB began to stretch.
Each specimen was mounted in a custom frame (Fig. 2) that
allowed independent control of hip flex/extension, rotation, ab/
adduction, and knee flexion following Arnold et al. (2000).
Alignment and mounting of the specimen comprised four main
steps, each performed with real-time feedback to ensure that the
pelvis, femur, and tibia were secured to within 5 mm and 2°of the
desired alignment. First, the pelvis was secured to a table and
aligned with either its medial-lateral axis (for flex/extension MAs)
or anterior–posterior axis (for ab/adduction MAs) perpendicular to
the table. Second, the femur was mounted on a cart equipped with
two concentric rings. The femur was secured to the inner ring so
that the femur's long axis (from hip center to the midpoint
between femoral epicondyles) was centered perpendicular to the
plane of the rings. Third, the base of the cart was adjusted so that
Table 1
Muscle architecture of tensor fascia lata (TFL) and gluteus maximus (GMax).
Muscle Mass (g) Optimal fascicle
length (cm)
Pennation angle
(deg.)
PCSA
a
(cm
2
)
TFL 35.579.6 9.8 70.7 1.171.1 3.271. 0
GMax 412.1769.7 14.4 70.7 26.375.0 30.6 75.1
Data from 3 elderly cadaveric specimens (2 male, 1 female; mean age: 7876 years)
are expressed as mean7s.e.m.
a
Pennation angle is not included in the PCSA calculation since our SIMM model
multiplies PCSA, specific tension, and pennation angle to determine a muscle's
maximum isometric force.
C.M. Eng et al. / Journal of Biomechanics 48 (2015) 3341–33483342

its wheels rolled in an arc about the specimen's hip center. Fourth,
the tibia was secured to a locking hinge attached to the inner ring.
When measuring knee MAs, the hinge was removed and the tibia
was flexed and extended. When measuring hip rotation MAs, the
inner ring was rotated relative to the outer ring, which internally
and externally rotated the hip. When measuring hip flex/extension
or ab/adduction MAs, the cart was rotated about the specimen's
hip center, thereby flex/extending or ab/adducting the hip. When
measuring MAs about one axis, the other axes were locked in a
neutral position (hip flexion¼0°, hip rotation ¼5°, hip
adduction¼0°, knee flexion¼0°). To verify alignment, we mon-
itored coupling of hip angles and ensured that hip adduction
varied o2°and hip rotationo4°over a 75°range of flexion. When
the specimen was aligned for hip ab/adduction, we ensured that
hip flexion varied o2°and hip rotation o4°over a 50°range of
ab/adduction.
To measure muscle-ITB MAs, the excursion of each thread path
was recorded while slowly moving the specimen through its range
of hip and knee motion. Excursion and joint angle data were
sampled at 10 Hz (National Instruments BNC-2090 A/D converter).
The lengthening excursion versus joint angle data were fit with a
fourth order polynomial, and the derivative of the polynomial was
averaged across trials to estimate the MA. A minimum of five trials
was collected for each condition.
Following MA measurements, muscles were freed, cleaned of
fat and connective tissue, and weighed (Table 2). In two speci-
mens, the regions of GMax were carefully dissected to determine
the relative masses of the portions that insert on the ITB versus the
femur.
2.3. Model of TFL, Gmax, and ITB F-L properties
We modified the paths of TFL and GMax muscle–tendon units
(MTUs) in the model reported by Arnold et al. (2010) to match our
digitized attachments and MA data (Fig. 3). Using SIMM (Software
Fig. 1. (A) Lateral view of the human ITB showing paths of the inserting muscles, TFL and GMax, as characterized during the moment arm measurements, (B) Lateral diagram
showing the anterior and posterior paths of TFL and (C) Posterior diagram showing the superior (Gmax1,2) and inferior (GMax3,4) regions of GMax. Muscle-ITB paths are
described in the Supplementary materials.
Fig. 2. Hardware and procedure for measuring hip ab/adduction and rotation
moment arms. The hardware consisted of a fixed table for aligning and securing the
pelvis, an adjustable cart for moving the femur through a range of hip ab/adduction
angles, and a set of concentric rings for rotating the femur about its mechanical
axis, following Arnold et al. (2000). Receivers (shown in gray) were rigidly attached
to the pelvis, femur, and tibia to track their motions in real time. (1) The pelvis was
secured to a fixed table with its anterior–posterior axis perpendicular to the table,
(2) the femur was secured to the inner of the two rings so that the femur's long axis
was centered perpendicular to the plane of the rings, (3) the bases of the cart were
adjusted so that the cart's wheels rolled in an arc about the specimen's hip joint
center and (4) the tibia was secured to a locking hinge. Hip flex/extension moment
arms were measured by re-orienting the pelvis on the table so that its medial-
lateral axis was perpendicular to the table. More details are provided in the Sup-
plementary materials.
C.M. Eng et al. / Journal of Biomechanics 48 (2015) 3341–3348 3343

for Interactive Musculoskeletal Modeling, v7.0, MusculoGraphics,
Santa Rosa, CA), we initially created two paths for TFL and eight for
GMax (four to the ITB and four to the femur). Via points and
wrapping surfaces were iteratively adjusted so that the model's
paths reproduced the three-dimensional paths digitized during
the experiments and the model's MAs matched the experimentally
determined MAs. Because the model's MAs are extremely con-
sistent with our experimental data (Fig. 4 and S2), we are con-
fident that the model accurately predicts length changes of these
MTUs.
To estimate strain in regions of the ITB during running, we
created three additional MTUs, representative of the major paths
of ITB force transmission determined from our experiments
(Fig. 3B and C). One path accounts for force transmitted by the
anterior ITB when TFL is active (TFL-ITB
ant
). The other paths
account for the cumulative force transmitted by the posterior ITB
when superior (GMax1,2-ITB
post1
) or inferior (GMax3,4-ITB
post2
)
portions of GMax are active. Attachments and via points of each
path were iteratively adjusted to yield average MAs of the com-
bined MTUs (Fig. 4). This model represents the muscles as inde-
pendent, proximal-to-distal MTUs, even though the ITB is multi-
layered and loaded from different directions, based on detailed
dissections and on biaxial testing of goat fascia lata, which showed
that the fascia's material properties are not strongly influenced by
its biaxial strain environment (Eng et al., 2014).
We used a Hill-type muscle model (Delp et al., 1990;Zajac,
1989) to estimate isometric forces generated by TFL-ITB
ant
,
GMax1,2-ITB
post1,
and GMax3,4-ITB
post2
at different activation
levels. Two parameters, maximum isometric force (F
max
) and
optimal fiber length (L
opt
), scaled normalized active and passive F–
Lcurves to each muscle (Table S1). F
max
and an additional para-
meter, tendon slack length (L
TS
), scaled a normalized “tendon”F–L
curve to each MTU. We specified parameters for each MTU based
on our architecture measurements and data reported by Ward
et al. (2009). We adjusted L
TS
such that ITB
ant
and ITB
post
began to
stretch passively at hip and knee angles consistent with our
experimental data.
For each MTU, we created a normalized F–Lcurve for the ITB
(Fig. 5) based on published material properties of the human ITB
(Butler et al., 1984;Derwin et al., 2008). We assumed a transition
strain of 3% based on data from goat fascia lata (Eng et al., 2014).
Above 3% strain, we assumed a linear relationship between force
and strain with a normalized stiffness
k
()
~
determined using the
ITB's elastic modulus (E), the muscle's F
max
, and the ITB's effective
cross-sectional area (a):
kE
F3
max
α
~=·
()
The effective cross-sectional area of the ITB was calculated from
measurements of thickness and width in cadaveric specimens (see
Table S2). The width of each ITB region was measured while pla-
cing tension on the inserting muscle and visually assessing ITB
strain. We used an elastic modulus of 400 MPa, which is consistent
with values reported in the literature (Butler et al., 1984;Derwin
et al., 2008;Hammer et al., 2012;Steinke et al., 2012). Below 3%
strain, in the toe region, we decreased stiffness by a factor of 2/3.
At F
max,
the ITB strains 5–11% in our model, which seems plausible
given the range of yield strains reported in the literature (10–27%;
(Butler et al., 1984;Hammer et al., 2012;Hinton et al., 1992). The
regional variation in strain at F
max
is consistent with our mea-
surements of ITB thickness, which are relatively uniform in ante-
rior and posterior regions despite the fact that the inserting
muscles differ substantially in force-generating capacity.
2.4. Assessment of ITB energy storage
We used our model in combination with published joint
kinematics and EMG recordings to examine the capacity of the ITB
to store elastic energy during running. First, we calculated the
lengths of the MTUs at hip and knee angles corresponding to
running using data from 10 experienced runners, at speeds of 2, 3,
4, and 5 m/s (Hamner and Delp, 2013). Next, we divided MTU
lengths into muscle fiber lengths and ITB lengths by indepen-
dently activating each MTU in the model and solving for the
lengths at which the muscle and ITB forces were equivalent,
accounting for pennation angle. Maximum activation levels for
running were assumed to range between 20% and 65% of the EMG
activation measured during a maximum voluntary contraction
(MVC). In particular, we set each muscle's maximum activation to
20%, 35%, 50%, or 65% to estimate ITB strains during running at 2,
Table 2
Muscle regional masses of tensor fascia lata (TFL) and gluteus maximus (GMax).
Muscle Total mass of region (g)
(n¼5)
Percentage of mass inserting on ITB
(%) (n¼2)
TFL1
a
26.477. 2 10 0
TFL2 21.4 75.5 100
GMax1
b
110.6726.2 44.6 74.9
GMax2 109.4724.8 52.7 77.8
GMax3 121.9719.8 47.7711.3
GMax4 104.7729.7 71.7728.3
Data from 5 adult males (mean age: 62 710 years) are expressed as mean7s.e.m.
a
TFL was divided into two anterior–posterior regions based on origin and
fascicle orientation.
b
GMax was separated into four superior–inferior regions.
Fig. 3. Lower extremity model modified from Arnold et al. (2010). (A) Lateral view
showing the two TFL-ITB paths that reproduce our experimental data, (B) posterior
view showing the four GMax-ITB paths originating on the ilium, sacrum, and
coccyx and inserting on the ITB and (C) lateral view of the combined MTU paths
used to estimate energy storage. TFL-ITB
ant
, GMax1,2-ITB
post1
, and GMax3,4-ITB
post2
paths are shown at touchdown, midstance, toeoff, and midswing during running at
5 m/s. The TFL-ITB
ant
MTU is maximally stretched in early swing, while the GMax-
ITB
post
MTUs are most stretched during late swing.
C.M. Eng et al. / Journal of Biomechanics 48 (2015) 3341–33483344

3, 4, and 5 m/s, respectively. These values are based on reported
maximum activations of 20% MVC in GMax and TFL during level
walking (Perry, 1992), 40% MVC in GMax and TFL during level
running at 4.0 m/s (Montgomery et al., 1994), and 65% MVC in
GMax during level running at 4.5 m/s (Swanson and Caldwell,
2000). Time-varying patterns of activity were derived from EMG
recordings reported for GMax and TFL (Jönhagen et al., 1996;
Montgomery et al., 1994;Paré et al., 1981;Swanson and Caldwell,
2000), which we scaled to the maximum activation at each speed
(Fig. 7). Lastly, we estimated energy storage capacity at each speed
by integrating the ITB F–Lcurves from L
TS
to peak ITB length
during running. Length changes of the ITB were determined
relative to slack length. Total elastic energy stored in the posterior
ITB was calculated as the sum of the energies stored in GMax1,2-
ITB
post1
and GMax3,4-ITB
post2
. We assessed the sensitivity of our
analysis to the F–Lproperties by varying normalized stiffness by
720% and transition strain by 72% (Fig. 5) and re-calculating
energy storage.
3. Results
The TFL and GMax MTUs in our model undergo substantial
excursions during running (Fig. 6). Because of its hip flexion and
knee extension MAs, TFL-ITB
ant
is maximally stretched during
early swing, when the hip is extended and the knee flexed
(Fig. 3C). EMG recordings show that TFL is highly activated during
this time (Figs. 6 and 7)(Montgomery et al., 1994;Paré et al.,
1981). In contrast, because of their role in hip extension and knee
flexion MAs, GMax-ITB
post1
and GMax-ITB
post2
are maximally
stretched during late swing, when the hip flexes and the knee
extends (Fig. 3C). EMG recordings show that GMax is highly acti-
vated during this time (Figs. 6 and 7)(Jönhagen et al., 1996;
Swanson and Caldwell, 2000). Inferior portions of GMax lengthen
about 7% more than proximal portions, due to larger hip extension
MAs when the hip is flexed.
The largest strains in ITB
ant
occur in early swing (Fig. 7A), with
ITB
ant
stretching 0.9–1.7 cm beyond slack length in our model. TFL
muscle fiber length is longer than optimal when it begins gen-
erating force in late stance, and near optimal when it is maximally
activated in early swing. Peak strains in ITB
post
occur in late swing
(Fig. 7B), with ITB
post
stretching 1.4–3.0 cm beyond slack length in
our model. GMax3,4 is shorter than optimal length when it begins
generating force in mid swing; however, it is stretched beyond
optimal length as the hip flexes in swing. In late swing, when
GMax3,4 is maximally activated, it operates near optimal length
and generates forces that stretch ITB
post
in our model. A similar
pattern occurs in GMax1,2-ITB
post2
. Passive strains in the ITB,
Fig. 4. Hip and knee moment arms of TFL-ITB
ant
and GMax-ITB
post
compared with experimental data. (A) TFL has a large hip flexion MA that increases as the hip flexes,
(B) TFL has a large hip abduction MA that increases with hip abduction, (C) the most posterior part of TFL has a small knee extension MA that decreases with knee flexion,
(D) and (G) all portions of GMax that insert on the ITB have large hip extension MAs, (E) the superior portions of GMax have hip abduction MAs, (H) the inferior portions of
GMax have hip adduction MAs, (F) and (I) the portions of GMax that insert on the ITB have small knee flexion MAs. Solid lines and shaded regions indicate the means and
standard deviations of experimentally determined MAs from five cadaveric limbs. (A)–(C) Dashed lines show the MAs of TFL1-ITB (dark gray), TFL2-ITB (light gray), and the
combined TFL1,2-ITB
ant
(black) predicted by our model. D–I: Dashed lines show the MAs (from superior to inferior) of GMax1 (dark gray), GMax2 (light gray), GMax3 (dark
gray), GMax4 (light gray), and the combined paths for GMax1,2-ITB
post1
(black, D-F) and GMax3,4-ITB
post2
(black, G-I) as predicted by our model. Note the y-axes have
different scales.
C.M. Eng et al. / Journal of Biomechanics 48 (2015) 3341–3348 3345

without muscle activation, are relatively small in our model. ITB
ant
strains 1.7% and ITB
post
strains 1.1% over the stride cycle when the
muscles are not activated.
Because the TFL and GMax MTUs are stretched to relatively
long lengths when the muscles are active during running, both
anterior and posterior regions of the ITB in our model have the
capacity to store elastic energy. We estimate that ITB
ant
strains
about 4% during running at 5 m/s, which means that the ITB
ant
may store nearly 1 J of energy per stride during early swing
(Fig. 8A). ITB
post1
and ITB
post2
strain about 4% during slow running
and 7% during faster running in late swing when GMax is maxi-
mally active in our model. These data suggest that the ITB
post
may
store as much as 6 J per stride during late swing (Fig. 8B).
4. Discussion
This is the first study to quantitatively characterize the 3-D
musculoskeletal geometry of the human ITB and its inserting
muscles. Dissections confirmed that all fibers of TFL insert into the
anterior ITB and a large fraction of GMax (40–70% by mass) inserts
into the posterior ITB. Thus, the ITB likely transmits substantial
force. Additionally, our MA measurements confirmed that the
inserting muscles have relatively large MAs about the hip, and
thus undergo large MTU excursions, with hip flexion and exten-
sion. In combination, the ITB's high compliance and its potential to
transmit force while changing length, suggest a plausible, pre-
viously unrecognized mechanism for storing elastic energy during
running.
We created a model that characterizes the geometry and F–L
properties of the ITB, TFL, and GMax to test the hypothesis that
forces generated by TFL and GMax stretch the ITB during running,
Fig. 5. Normalized force-length curves for anterior and posterior regions of the ITB
derived from experimental data (Butler et al., 1984;Derwin et al., 2008;Eng et al.,
2014). Curves are shown for TFL-ITB
ant
(short dash), GMax1,2-ITB
post1
(long dash),
and GMax3,4-ITB
post2
(dot-dash) with shaded regions indicating 720% stiffness
used in the sensitivity analysis. The ITB is more compliant than tendon (solid line),
as shown by the typical tendon force–length curve generated by Millard et al.
(2013) and by ultrasound-based measures of tendon force–length properties
reported by Magnusson et al. (2001), light gray shaded region. This figure is
adapted from Fig. 3 in Millard et al. (2013).
Fig. 6. Length changes of TFL-ITB and GMax-ITB MTUs during fast running (5 m/s).
TFL-ITB
ant
(light gray) stretches during stance phase, while GMax1,2-ITB
post1
(dark
gray) and GMax3,4-ITB
post2
(black) stretch during the swing phase. Regions of
intermediate muscle activity (10-30% activation; intermediately thickened portion
of each curve) and maximal muscle activity (31-65% activation; thickened portion
of each curve) demonstrate that these muscles are maximally active when the MTU
is near maximum length.
Fig. 7. TFL-ITB
ant
and GMax3,4-ITB
post
activation patterns over the stride cycle. For
each muscle, the time-varying pattern was scaled to an activation level of 20%, 35%,
50%, or 65% during running at 2, 3, 4, and 5 m/s, respectively (increasingly darker
lines). Filled squares on each line indicate toe-off. Superimposed are plots of force
versus normalized fiber length and force versus ITB strain at key points in the gait
cycle during running at 5 m/s. Circles on the curves show where the muscle or ITB
is acting at that point in the gait cycle. A: TFL is longer than optimal length (L
opt
)
prior to toe-off when the muscle begins to generate force. When TFL is maximally
activated in early swing, it operates near optimal length and stretches ITB
ant
to its
longest length in our model. B: GMax3,4 is shorter than optimal length when it
begins generating force in mid swing and is stretched beyond optimal length in
swing. When GMax3,4 is maximally activated in late swing, it operates near opti-
mal length and stretches ITB
post
to its longest length in our model.
C.M. Eng et al. / Journal of Biomechanics 48 (2015) 3341–33483346

storing elastic energy. Analysis of the model revealed that the ITB
has the capacity to store 7 J per stride during running at 5 m/s. The
posterior ITB stores substantially more energy than the anterior
ITB because it transmits larger muscle forces.
How does the amount of energy stored in the ITB during run-
ning compare to energy stored in the Achilles tendon? Using a 3-D
model and static optimization, Lai et al. (2014) estimated that the
soleus and gastrocnemius store 40–50 J per stride in the Achilles
tendon at slow to fast running speeds (3.5–5 m/s), an estimate
consistent with previous experimental studies (Alexander and
Bennet-Clark, 1977;Hof et al., 2002;Ker et al., 1987). We therefore
calculate that the combined anterior and posterior ITB stores 14%
as much energy as the Achilles tendon at a 5 m/s pace.
To provide additional context, we compared energy stored in
the ITB to hip muscle work during running. Sasaki and Neptune
(2006) used a muscle-driven dynamic simulation to estimate the
mechanical work performed by hip muscles and series elastic
elements during running at 2.4 m/s. They reported that the hip
extensors do 40 J of work per stride during stance, while the hip
flexors do 6 J of work during swing. Recovery of 2 J from ITB
post
during slow running could account for 5% of the work done by hip
extensors in stance, while recovery of 0.3 J from ITB
ant
could
contribute 5% of the work done by hip flexors in swing. Although
the extent to which energy recovery would drive selection for
endurance running is unknown, these comparisons suggest that
energy storage in the ITB is not negligible.
This analysis has several limitations. First, although our data
confirm that forces generated by TFL and GMax stretch the ITB
during running, storing useful energy, our study did not test
whether the human ITB reduces muscle work or enhances loco-
motor economy. Second, uncertainty exists in the parameters used
to derive the F–Lcurves. For example, our measures of ITB width
and thickness in cadaveric specimens may not be representative of
healthy young subjects, thus potentially underestimating ITB
stiffness. However, varying normalized stiffness by 720% and
transition strain by 72% altered our estimates of energy storage
by only 0.1 J in the anterior ITB and by 1.2 J in the posterior ITB at
the fastest running speed. At the slowest running speed, varying
stiffness and transition strain altered our estimates of energy
storage by about 0.05 J. Thus, we are reasonably confident in our
model of the ITB's force–strain behavior and that the ITB con-
tributes to energy storage at all running speeds. Third, we esti-
mated the peak forces generated by TFL and GMax ignoring the
muscles' force–velocity (F–V) properties and assuming the mus-
cles' activation patterns during running. If the muscles shorten
substantially during running, or if we overestimated activation,
then we likely overestimated ITB energy storage. It is plausible,
however, that the ITB's length and compliance allow GMax to
operate nearly isometrically when generating maximum force in
late stance, mitigating the effects of F–V properties on muscle-ITB
mechanics. In the running simulations described by Lai et al.
(2014), muscles inserting on the Achilles tendon contracted nearly
isometrically across a range of running speeds. Lastly, we esti-
mated the capacity of the ITB to store elastic energy during run-
ning but not walking. It is likely that the ITB transmits smaller
forces, and thus stores less energy, during walking than reported
here.
Our study has implications for understanding the evolution of
human bipedalism. While these data do not exclude the possibility
that the ITB stores substantial energy during walking, selection for
the capacity to run long distances would have presented unique
demands on the anatomy and physiology of Homo (see Bramble
and Lieberman (2004) for review). Among these demands is the
need to efficiently accelerate the swing limb, which is long and
massive in humans (14% body mass) compared to chimpanzees
(9% body mass; Zihlman and Brunker, 1979). The human ITB is
stretched substantially just before swing, when the TFL is active
and the hip is extending (Figs. 6 and 7). Subsequent recoil of the
ITB may help accelerate the swing limb. Although the energetic
cost of running is primarily determined by muscle forces that
support the body during stance (Kram and Taylor, 1990), the cost
of accelerating the swing limb may be as much as 27% of total
metabolic cost (Marsh et al., 2004;Modica and Kram, 2005;Myers
and Steudel, 1985). Thus, selection for increased running economy
may have favored traits that increase swing phase energy recovery
in Homo. The need to decrease locomotor costs may also help
explain the expansion of GMax evident in Homo. This adaptation is
thought to play a role in trunk stabilization during endurance
running (Lieberman et al., 2006), but it may also facilitate elastic
energy storage by increasing the forces transmitted to the ITB as it
is stretched in late swing.
Conflict of Interest Statement
The authors have no conflicts of interest to disclose.
Acknowledgments
The authors fondly remember Farish A. Jenkins Jr. (1940–2012)
for many stimulating and insightful discussions. Professor Jenkins
helped guide C.M.E.'s dissertation research, which provided the
basis for this study, and he deserves much credit. The authors
thank two anonymous reviewers for constructive comments that
significantly improved this manuscript. We gratefully acknowl-
edge Delande Justinvil and Zachary Lewis for technical assistance
during the moment arm experiments. We thank Casey Boyle and
Yasmin Rawlins for assistance during pilot studies, and we thank
Andrew Mountcastle and Glenna Clifton for help with video-
graphy. We are grateful to Tom Roberts for helpful comments on a
Fig. 8. Elastic energy stored in the ITB during running at 2, 3, 4, and 5 m/s as
predicted by the model. A: Peak energy stored in ITB
ant
when TFL is activated 20,
35, 50, or 65%. B: Peak energy stored in ITB
post
when GMax is activated 20, 35, 50,
or 65%. The energy stored in GMax-ITB
post
is calculated as the sum of energies
stored in GMax1,2-ITB
post1
(gray) and GMax3,4-ITB
post2
(white).
C.M. Eng et al. / Journal of Biomechanics 48 (2015) 3341–3348 3347

preliminary version of the manuscript. This research was funded
by a Wenner-Gren Dissertation Fieldwork Grant to C.M.E. under
award no. 8588.
Appendix A. Supplementary material
Supplementary data associated with this article can be found in the
online version at http://dx.doi.org/10.1016/j.jbiomech.2015.06.017.
References
Aiello, L.C., Dean, C., 1990. An Introduction to Human Evolutionary Anatomy, Bos-
ton.. Academic Press
Alexander, R.M., Bennet-Clark, H.C., 1977. Storage of elastic strain energy in muscle
and other tissues. Nature 265, 114–117.
An, K.N., Takahashi, K., Harrigan, T.P., Chao, E.Y., 1984. Determination of muscle
orientations and moment arms. J. Biomech. Eng. 106, 280–282.
Andersson, E.A., Nilsson, J., Thorstensson, A., 1997. Intramuscular EMG from the hip
flexor muscles during human locomotion. Acta Physiol. Scand. 161, 361–370.
Arnold, A.S., Salinas, S., Asakawa, D.J., Delp, S.L., 2000. Accuracy of muscle moment
arms estimated from MRI‐based musculoskeletal models of the lower extre-
mity. Comput. Aided Surg. 5, 108–119.
Arnold, E.M., Ward, S.R., Lieber, R.L., Delp, S.L., 2010. A model of the lower limb for
analysis of human movement. Ann. Biomed. Eng. 38, 269–279.
Biewener, A.A., 2003. Animal Locomotion. Oxford University Press, Oxford.
Bramble, D., Lieberman, D., 2004. Endurance running and the evolution of Homo.
Nature 432, 345–352.
Brand, P.W., Cranor, K.C., Ellis, J.C., 1975. Tendon and pulleys at the metacarpo-
phalangeal joint of a finger. J. Bone Joint Surg. 57, 779–784.
Butler, D.L., Grood, E.S., Noyes, F.R., Zernicke, R.F., Brackett, K., 1984. Effects of
structure and strain measurement technique on the material properties of
young human tendons and fascia. J. Biomech. 17, 579–596.
Crompton, R.H., Li, Y., Wang, W., Günther, M.M., Savage, R., 1998. The mechanical
effectiveness of erect and “bent-hip, bent-knee”bipedal walking in Aus-
tralopithecus afarensis. J. Human Evol. 35, 55–74.
Delp, S.L., Loan, J.P., Hoy, M.G., Zajac, F.E., Topp, E.L., Rosen, J.M., 1990. An interactive
graphics-based model of the lower extremity to study orthopaedic surgical
procedures. IEEE Trans. Biomed. Eng. 37, 757–767.
Derwin, K.A., Baker, A.R., Spragg, R.K., Leigh, D.R., Farhat, W., Iannotti, J.P., 2008.
Regional variability, processing methods, and biophysical properties of human
fascia lata extracellular matrix. J. Biomed. Mater. Res. A 84, 500–507.
Dorn, T.W., Schache, A.G., Pandy, M.G., 2012. Muscular strategy shift in human
running: dependence of running speed on hip and ankle muscle performance. J.
Exp. Biol. 215, 1944–1956.
Eng, C.M., Pancheri, F.Q., Lieberman, D.E., Biewener, A.A., Dorfmann, L., 2014.
Directional differences in the biaxial material properties of fascia lata and the
implications for fascia function. Ann. Biomed. Eng. 42, 1224–1237.
Gottschalk, F., Kourosh, S., Leveau, B., 1989. The functional anatomy of tensor fasciae
latae and gluteus medius and minimus. J. Anat. 166, 179–189 .
Gratz, C.M., 1931. Tensile strength and elasticity tests on human fascia lata. J. Bone
Joint Surg. 13, 334–340.
Gray, H., Williams, P.L., Bannister, L.H., 1995. Gray's Anatomy: The Anatomical Basis
of Medicine and Surgery. Churchill Livingstone, Edinburgh.
Hammer, N., Lingslebe, U., Aust, G., Milani, T.L., Hädrich, C., Steinke, H., 2012.
Ultimate stress and age-dependent deformation characteristics of the iliotibial
tract. J. Mech. Behav. Biomed. Mater. 16, 81–86.
Hamner, S.R., Delp, S.L., 2013. Muscle contributions to fore-aft and vertical body
mass center accelerations over a range of running speeds. J. Biomech. 46,
780–787.
Hinton, R., Jinnah, R.H., Johnson, C., Warden, K., Clarke, H.J., 1992. A biomechanical
analysis of solvent-dehydrated and freeze-dried human fascia lata allografts: a
preliminary report. Am. J. Sports Med. 20, 607–612.
Hof, A., Van Zandwijk, J., Bobbert, M., 2002. Mechanics of human triceps surae
muscle in walking, running and jumping. Acta Physiol. Scand. 174, 17–30.
Inman, V.T., 1947. Functional aspects of the abductor muscles of the hip. J. Bone
Joint Surg. 29, 607–619.
Jönhagen, S., Ericson, M., Nemeth, G., Eriksson, E., 1996. Amplitude and timing of
electromyographic activity during sprinting. Scand. J. Med. Sci. Sports 6, 15–21.
Kaplan, E.B., 1958. The iliotibial tract: clinical and morphological significance. J.
Bone Joint Surg. 40, 817–832.
Ker, R.F., Bennett, M.B., Bibby, S.R., Kester, R.C., Alexander, R.M., 1987. The spring in
the arch of the human foot. Nature 325, 147–149.
Kram, R., Taylor, C.R., 1990. Energetics of running: a new perspective. Nature 346,
265–267.
Lai, A., Schache, A.G., Lin, Y.-C., Pandy, M.G., 2014. Tendon elastic strain energy in
the human ankle plantar-flexors and its role with increased running speed. J.
Exp. Biol. 217, 3159–3168.
Lieber, R.L., Fazeli, B.M., Botte, M.J., 1990. Architecture of selected wrist flexor and
extensor muscles. J. Hand Surg. 15A, 244–250.
Lieber, R.L., Loren, G.J., Fridén, J., 1994. In vivo measurement of human wrist
extensor muscle sarcomere length changes. J. Neurophysiol. 71, 874–881.
Lieberman, D.E., Raichlen, D.A., Pontzer, H., Bramble, D.M., Cutright-Smith, E., 2006.
The human gluteus maximus and its role in running. J. Exp. Biol. 209,
2143–2155.
Magnusson, S.P., Aagaard, P., Rosager, S., Dyre-Poulsen, P., Kjaer, M., 2001. Load-
displacement properties of human triceps surae aponeurosis in vivo. J. Physiol.
531, 277–288.
Mann, R., Moran, G., Dougherty, S., 1986. Comparative electromyography of the
lower extremity in jogging, running, and sprinting. Am. J. Sports Med. 14,
501–510.
Marsh, R.L., Ellerby, D.J., Carr, J.A., Henry, H.T., Buchanan, C.I., 2004. Partitioning the
energetics of walking and running: swinging the limbs is expensive. Science
303, 80–83.
Mendez, J., Keys, A., 1960. Density and composition of mammalian muscle. Metab.:
Clin. Exp. 9, 184–188.
Millard, M., Uchida, T., Seth, A., Delp, S.L., 2013. Flexing computational muscle:
modeling and simulation of musculotendon dynamics. J. Biomech. Eng. 135,
1–11.
Modica, J.R., Kram, R., 2005. Metabolic energy and muscular activity required for leg
swing in running. J. Appl. Physiol. 98, 2126–2131.
Montgomery, W.H., Pink, M., Perry, J., 1994. Electromyographic analysis of hip and
knee musculature during running. Am. J. Sports Med. 22, 272–278.
Myers, M., Steudel, K., 1985. Effect of limb mass and its distribution on the energetic
cost of running. J. Exp. Biol. 116, 363–373.
Ober, F.R., 1936. The role of the iliotibial band and fascia lata as a factor in the
causation of low-back disabilities and sciatica. J Bone Joint Surg 18, 105–110.
Paré, E., Stern, J., Schwartz, J., 1981. Functional differentiation within the tensor
fasciae latae. J Bone Joint Surg. 63, 1457–1471.
Perry, J., 1992. Gait Analysis: Normal and Pathological Function. Slack Inc., Thoro-
fare, NJ.
Pontzer, H., Raichlen, D.A., Sockol, M.D., 2009. The metabolic cost of walking in
humans, chimpanzees, and early hominins. J. Human Evol. 56, 43–54.
Powell, P.L., Roy, R.R., Kanim, P., Bello, M., Edgerton, V.R., 1984. Predictability of
skeletal muscle tension from architectural determinations in guinea pig hin-
dlimbs. J. Appl. Physiol. 57, 1715–1721 .
Roberts, T., 2002. The integrated function of muscles and tendons during locomo-
tion. Comp. Biochem. Physiol.-A: Mol. Integr. Physiol. 133, 1087–1099.
Robinson, J.T., 1972. Early Hominid Posture and Locomotion. University of Chicago
Press, Chicago.
Rolian, C., Lieberman, D.E., Hamill, J., Scott, J.W., Werbel, W., 2009. Walking, running
and the evolution of short toes in humans. J. Exp. Biol. 212, 713–721.
Sasaki, K., Neptune, R.R., 2006. Muscle mechanical work and elastic energy utili-
zation during walking and running near the preferred gait transition speed.
Gait Posture 23, 383–390.
Sockol, M.D., Raichlen, D.A., Pontzer, H., 2007. Chimpanzee locomotion energetics
and the origin of human bipedalism. Proc. Natl. Acad. Sci. 104, 12265–12269.
Steinke, H., Lingslebe, U., Böhme, J., Slowik, V., Shim, V., Hädrich, C., Hammer, N.,
2012. Deformation behavior of the iliotibial tract under different states of
fixation. Med. Eng. Phys. 34, 1221–1227.
Stern Jr., J.T., 1972. Anatomical and functional specializations of the human gluteus
maximus. Am. J. Phys. Anthropol. 36, 315–339.
Swanson, S.C., Caldwell, G.E., 2000. An integrated biomechanical analysis of high
speed incline and level treadmill running. Med. Sci. Sports Exerc. 32,
114 6–115 5 .
Ward, S.R., Eng, C.M., Smallwood, L.H., Lieber, R.L., 2009. Are current measurements
of lower extremity muscle architecture accurate? Clin. Orthop. Relat. Res. 467,
1074–1082.
Wickiewicz, T.L., Roy, R.R., Powell, P.L., Edgerton, V.R., 1983. Muscle architecture of
the human lower limb. Clin. Orthop. Relat. Res. 179, 275–283.
Zajac, F.E., 1989. Muscle and tendon: Properties, models, scaling, and application to
biomechanics and motor control. Crit. Rev. Biomed. Eng. 17, 359–411.
Zihlman, A.L., Brunker, L., 1979. Hominid bipedalism: then and now. Yearbook Phys.
Anthropol. 22, 132–162.
Zollikofer, C., Ponce de Leon, M., Lieberman, D., Guy, F., Pilbeam, D., Likius, A.,
Mackaye, H., Vignaud, P., Brunet, M., 2005. Virtual reconstruction of Sahelan-
thropus tchadensis. Nature 434, 755–759.
C.M. Eng et al. / Journal of Biomechanics 48 (2015) 3341–33483348

Supplemental materials
Descriptions of muscle-ITB paths used in moment arm experiments
The anterior path (TFL1) originated on the iliac crest near the anterior superior
iliac spine (ASIS) and the posterior path (TFL2) originated on the lateral iliac crest
approximately 3 cm posterior to the ASIS. Screw eyes inserted in the ilium anchored the
origins of thread paths to these locations (Figure 1A&B). TFL1 muscle fibers were
continuous with ITB fibers inserting in the distal femur approximately 3.0 cm anterior to
the lateral epicondyle, so a screw eye was anchored at this location in the femur. TFL2
fibers were continuous with ITB fibers inserting on the tibia, so its insertion screw eye
was placed in the lateral tibial condyle ~1.0 cm posterior to the most anterior point on
the tibial tuberosity. The most superior path of GMax (GMax1) originated on the ilium
1.5 cm lateral and 2.5 cm superior to the posterior superior iliac spine (PSIS; Figure
1A&C). The origin of GMax2 was 1.0 cm inferior to PSIS on the ilium. The GMax3 origin
was placed 1.0 cm lateral to the midline of the sacrum at the level of the intervertebral
disc between the second and third sacral vertebrae. The most inferior path of GMax
(GMax4) originated 1.0 cm lateral to the coccygeal midline on the proximal coccyx.
GMax1 and GMax2 shared a screw eye placed 2.5 cm posterior to the tibial tuberosity,
while GMax3 and 4 shared a screw eye placed 4.0 cm posterior to the tibial tuberosity.
Description of the moment arm frame
The specimen was mounted in a custom frame, comprising a fixed table for securing the
pelvis, a rotating cart for moving the femur through a range of hip flex/extension and
ab/adduction angles, and a set of concentric rings for rotating the femur, following
Arnold et al. (2000; Figure 2). The cart consisted of two vertical support posts each
mounted on a wheeled baseplate with a ball caster mounted on a rod between them.
The limb is shown here with its pelvis mounted for hip ab/adduction, which is achieved
by rotating the cart.
Figure S1. Pelvis and femur coordinate systems (CSs) used to define hip
flex/extension, ab/adduction, and rotation and to align the specimen in the custom
frame. For the pelvis CS, the medial-lateral plane was defined to be parallel to the plane
formed by two vectors from the pubic symphysis (PS) to superior and inferior points on
the sacral midline. The origin of the pelvis CS was located at the PS. The right ASIS
and pubic tubercle (PT) defined the frontal plane. For the femur, the superior-inferior
axis was defined by the vector joining the midpoint between the medial and lateral
epicondyles and the hip joint center (determined as described in the methods). The hip
joint center was defined as the origin of the femur CS. The frontal plane of the femur CS
was defined to be parallel to the vector between the lateral and medial epicondyles. The
tibia CS was specified to be coincident with the femur CS at full knee extension. The
origin of the tibia CS was located at the knee joint center, which was defined as the
midpoint between the medial and lateral femoral epicondyles.
Figure S2. Hip rotation moment arms of TFL-ITBant and GMax-ITBpost compared with
experimental data. A: TFL has a hip internal rotation MA that increases as the hip
externally rotates. B,C: All portions of GMax have a large hip external rotation that

decreases as the hip externally rotates. Solid lines and shaded regions indicate the
means and standard deviations of experimentally determined MAs from five cadaveric
limbs. Dashed lines show the MAs of TFL1-ITB (dark gray), TFL2-ITB (light gray),
GMax1 (dark gray), GMax2 (light gray), GMax3 (dark gray), GMax4 (light gray), and the
combined paths for TFL1,2-ITBant (black, A), GMax1,2-ITBpost1 (black, B) and GMax3,4-
ITBpost2 (black, C ) as predicted by our model.

Table S1. Parameters used to scale a generic Hill-type muscle model to TFL- and GMax-ITB muscle-tendon units
Muscle-tendon
unit
Maximum isometric
force (N; Fmax)¥
Optimal fiber length
(cm; Lopt)
Pennation angle
(deg; θ)
Tendon slack length
(cm; LTS)
TFL1,2-ITBant$
195.2
9.8
2.5
42.6
GMax1,2-ITBpost1*
455.9
15.2
26.3
42.3
GMax3,4-ITBpost2*
558.6
16.7
26.3
41.0
$TFL PCSA, Lopt, and θ and GMax PCSA and θ from our measurements.
* GMax Lopt from Ward et al. (2009).
¥Fmax calculated as the product of PCSA and muscle specific tension of 61 N/cm2 used by Arnold et al. (2010).

Table S2. ITB thickness and width measurements used to calculate effective cross-sectional area and stiffness of each
region
ITB Region
Thickness (mm)*
Width (mm)
ITBant
0.87 ± 0.34
16.35 ± 2.03
ITBpost1
0.97 ± 0.12
16.85 ± 1.85
ITBpost2
15.33 ± 1.86
Data from 3 elderly cadaveric specimens (2 male, 1 female; mean age: 78 ± 6 years) are expressed as mean ± s.e.m.
*Thicknesses of anterior and posterior regions were measured at proximal, middle, and distal sites and averaged across sites.