BIOMECHANICAL RESPONSE OF THE ANKLE TO
EXCESSIVE EXTERNAL ROTATION
By
Keith D. Button

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Engineering Mechanics-Doctor of Philosophy
2015

ABSTRACT
BIOMECHANICAL RESPONSE OF THE ANKLE TO EXCESSIVE EXTERNAL
ROTATION
By
Keith D. Button
Acute ankle damage is one of the most commonly observed athletic injuries, accounting for 1030% of sports-related injuries in young athletes. Most frequently, damage occurs to the lateral
ligamentous complex, involving the anterior talofibular ligament and calcaneofibular ligament.
Medial ankle injuries involve damage to the anterior deltoid ligament, while syndesmotic (or
high) ankle injuries are defined as damage to the anterior tibiofibular ligament. While less
common than lateral ankle injuries, both medial and high ankle injuries are of particular interest
to researchers due to their longer recovery time and potential for long-term ankle dysfunction.
While combinations of eversion, dorsiflexion, and external rotation are often implicated in
medial and high ankle sprains, the exact mechanisms are still unclear. Additionally, although
studies have investigated the effect of shoe-surface interface on injury risk, few researchers have
examined the effect of shoe rotational stiffness on motion of the ankle, torque generated, and
subsequent injury location and severity during external rotation. One hypothesis of this
dissertation was that the motion of the ankle joint and the location of injury during external
rotation are functions of both the position of the ankle prior to rotation and the amount of
constraint on the ankle. A second hypothesis was that specimen-specific, rigid body modeling
could be utilized to simulate injury-level ankle rotation in addition to modeling the effect of
different footwear on motion of the ankle. Computational modeling, in addition to cadaveric and
in vivo human subject testing was utilized to test these hypotheses.

Specimen-specific computational models simulating injuries observed experimentally revealed
that external foot rotation primarily strains the medial ankle ligaments, but pre-everting the
ankles prior to rotation puts the primary strain on the syndesmotic ankle ligaments. Additionally,
there was a significant difference in the amount of strain in simulations modelling complete
ruptures versus partial tears. A follow-up study involving the measurement of cadaver joint
kinematics during external foot rotation concluded there was significantly more talocrural joint
rotation, but significantly less subtalar joint rotation in a neutral versus pre-everted ankle,
potentially explaining the location of injury during external rotation. In a separate cadaver study
investigating the effect of foot constraint, ankles constrained by a ‘stiff’ football shoe
experienced more ankle joint rotation, but less eversion than ankles constrained by a more
‘flexible’ shoe. As a result, ankles in stiff shoes experienced combination syndesmotic and
medial ankle injuries in addition to high strains in these ligaments. Ankles in flexible shoes saw
less combination injuries, but higher strains in the subtalar ligaments. This effect of ankle
constraint on joint motion was then quantified by measuring joint kinematics in human subjects,
and computational modelling provided an estimation of injury risk as a function of foot rotation
and constraint. A final human subject study involved calculating the stiffness of the
unconstrained ankle joint during voluntary external foot rotation in order to more accurately
model the ankle in the future. The information from these studies may aid in the implementation
of preventative measures in order to mitigate the risk of future ankle injury resulting from
excessive levels of external rotation.

Thank you Mom, Dad, Rachael, and Erin. You are my heroes and your support means
everything.

iv

ACKNOWLEDGEMENTS
I would like to thank my mentor, Dr. Roger C. Haut, for his support, expertise, and leadership
throughout my research at the Orthopaedic Biomechanics Laboratories (OBL). I would like to
acknowledge Dr. Dashin Liu, Dr. Eric Meyer, Dr. John Powell, Dr. Feng Wei, and Dr. Neil
Wright for their insight when serving on my committee. I would also like to thank Mr. Jerrod
Braman for his expertise in gait analysis, Mr. Cliff Beckett for his technical assistance, Mrs. Jean
Atkinson for help with specimen preparation, and Dr. Maureen Schaefer for clinical advice. I
would like to thank my co-investigators at Colorado State University: Garrett Coatney, Dr.
Tammy Haut Donahue, Kristine Fischenich, Hannah Pauly, and Benjamin Wheatley. I would
lastly like to thank all my coworkers in the OBL for their help and friendship: Benjamin
Carruthers, Mark Davison, Trevor Deland, Kathleen Fitzsimons, Mike Klein, Katie Landwehr,
Kevin Leikert, Patrick Vaughn, and Brian Weaver.

v

TABLE OF CONTENTS

LIST OF TABLES ......................................................................................................................... ix
LIST OF FIGURES .........................................................................................................................x

CHAPTER 1: INTRODUCTION AND BACKGROUND .............................................................1
Ankle Anatomy and Kinematics ..........................................................................................1
Ankle Injury in Sports ..........................................................................................................3
Inversion Injuries .....................................................................................................4
External Rotation Injuries ........................................................................................4
Risk Factors for External Rotation Injuries .........................................................................6
Shoe Surface Interface .............................................................................................6
Shoe Stiffness...........................................................................................................7
Computational Modelling ....................................................................................................8
Summary and Objectives .....................................................................................................9

CHAPTER 2: SPECIMEN-SPECIFIC COMPUTATIONAL MODELS OF ANKLE
SPRAINS PRODUCED IN A LABORATORY SETTING .............................................11
ABSTRACT .......................................................................................................................11
INTRODUCTION .............................................................................................................13
METHODS ........................................................................................................................15
Cadaver Experiment...............................................................................................15
Model Development...............................................................................................17
Simulation ..............................................................................................................19
RESULTS ..........................................................................................................................20
DISCUSSION ....................................................................................................................26

CHAPTER 3: EFFECT OF PRE-EVERSION ON TALOCRURAL AND SUBTALAR
JOINT MOTION DURING ANKLE EXTERNAL ROTATION ....................................31
ABSTRACT .......................................................................................................................31
INTRODUCTION .............................................................................................................32
METHODS ........................................................................................................................33
RESULTS ..........................................................................................................................36
DISCUSSION ....................................................................................................................37

vi

CHAPTER 4: ROTATIONAL STIFFNESS OF AMERICAN FOOTBALL SHOES AFFECTS
ANKLE BIOMECHANICS AND INJURY SEVERITY .................................................42
ABSTRACT .......................................................................................................................42
INTRODUCTION .............................................................................................................43
METHODS ........................................................................................................................46
Shoe Selection ........................................................................................................46
Cadaver Tests .........................................................................................................47
Motion Analysis .....................................................................................................51
Computational Modelling ......................................................................................52
Statistical Analysis .................................................................................................52
RESULTS ..........................................................................................................................53
Experimental ..........................................................................................................53
Computational Modelling ......................................................................................56
DISCUSSION ....................................................................................................................59

CHAPTER 5: THE EFFECT OF ROTATIONAL STIFFNESS ON ANKLE JOINT MOTION
AND LIGAMENT STRAIN DURING EXTERNAL ROTATION .....................65
ABSTRACT .......................................................................................................................65
INTRODUCTION .............................................................................................................66
METHODS ........................................................................................................................69
Tape Stiffness Tests ...............................................................................................69
Human Subject Tests .............................................................................................71
Computational Modelling ......................................................................................74
Statistical Analysis .................................................................................................76
RESULTS ..........................................................................................................................76
DISCUSSION ....................................................................................................................82

CHAPTER 6: A METHOD OF DETERMINING IN VIVO DYNAMIC HUMAN
ANKLE STIFFNESS DURING EXTERNAL ROTATION .............................................87
ABSTRACT .......................................................................................................................87
INTRODUCTION .............................................................................................................87
METHODS ........................................................................................................................89
RESULTS ..........................................................................................................................93
DISCUSSION ....................................................................................................................93

vii

CHAPTER 7: CONCLUSIONS ...................................................................................................97
Overview ............................................................................................................................97
Chapter 2 ............................................................................................................................98
Chapter 3 ............................................................................................................................99
Chapter 4 ..........................................................................................................................100
Chapter 5 ..........................................................................................................................102
Chapter 6 ..........................................................................................................................104
Contributions to the Literature .........................................................................................106
Influence of shoe design on ankle movement and injury ...................................106
Determination of failure-level ligament strains ..................................................106
Use of specimen-specific computational models ................................................107

APPENDICES .............................................................................................................................108
RESEARCH PUBLICATIONS .......................................................................................109
SOP (Standard Operating Procedure) for Building Specimen-Specific Computational
Ankle Models ...................................................................................................................112
SOP (Standard Operating Procedure) For Calculating Ankle Joint Motion Given Vicon
Data ..................................................................................................................................118

REFERENCES ............................................................................................................................132

viii

LIST OF TABLES
Table 2-1 Results of the cadaver experiment for the neutral and everted ankle ...........................25
Table 2-2 Comparison of the talus external rotations with respect to the tibia (deg) from the
cadaver study with the computational talus rotations ..................................................................25
Table 3-1 Motions of the subtalar and talocrural joint for 20 degrees of external rotation ........37
Table 4-1 Lateral (a), posterior (b), and medial (c) views of one of the specimen-specific models
in SolidWorks with linear springs representing ligaments. ...........................................................54
Table 4-2 Strains at failure rotation for the anterior tibiofibular ligament (ATiFL), anterior
tibiotalar ligament (ATiTL), posterior tibiotalar ligament (PTiTL), interosseous talocalcaneal
ligament (ITaCL), lateral talocalcaneal ligament (LTaCL), medial talocalcaneal ligament
(MTaCL), and posterior talocalcaneal ligament (PTaCL). (Underlining represents ligaments
that ruptured or avulsed) ...............................................................................................................55
Table 6-1 Ankle stiffnesses in three directions obtained from subjects performing a single legged
external rotation of the foot with standard deviations in parenthesis. Note: the absolute value of
the inversion-eversion slope was taken so that the reported number was positive. ......................92
Table 6-2 Stiffness values from previous studies ...........................................................................94
Table 7-1 Motions of the subtalar and talocrural joint for 20 degrees of external rotation .......100
Table 7-2 Summary of injury mechanisms and failure metrics from cadaver experiment ..........101
Table 7-3 Ankle stiffness in three directions obtained from subjects performing a single legged
external rotation of the foot. . .....................................................................................................105
Table A-1 Raw Data from Cadaver Failure Tests (Chapter 4) ...................................................124
Table A-2 Raw Data from Tape Stiffness Tests (Chapter 5) .......................................................127
Table A-3 Input Motions for Computational Model (Chapter 5) ................................................128

ix

LIST OF FIGURES
Figure 1-1 Medial view of the bony anatomy of the ankle ..............................................................1
Figure 1-2 Lateral (left) and medial (right) view of the osteoligamentous ankle anatomy ............2
Figure 1-3 The three motions of the ankle illustrated about a Cartesian coordinate system .........3
Figure 2-1 Setup for the cadaver experiment by Wei et al. ............................................................16
Figure 2-2 The cadaver taping pattern (a) and the model taping pattern represented by springs
(b). Note: for clarity all ligament springs and the plate springs on the medial side of the ankle
were set invisible. ...........................................................................................................................18
Figure 2-3 Injury mechanisms reported from the cadaver study by Wei et al. (Wei et al., 2012b).
Partial tear of the ATiFL (a); Rupture of the ATiFL (b); Partial tear of the ATiTL (c); Rupture
of the ATiTL (d); Tibial avulsion of the ATiTL (e); Partial tear of the TiNL(f). Note: the location
of injury is shown with a hemostat.................................................................................................21
Figure 2-4 Ligaments with average model predicted strains greater than 2% for the neutral case
in 3 of 4 cases in which ligament injury occurred .........................................................................22
Figure 2-5 Strains in the ATiFL and the two deltoid ligaments (ATiTL and TiNL) for the neutral
case for each specimen modeled ....................................................................................................23
Figure 2-6 Ligaments with average model predicted strains greater than 2% for the everted
simulations .....................................................................................................................................24
Figure 2-7 Strains in the ATiFL and the two deltoid ligaments (ATiTL and TiNL) for the everted
simulations .....................................................................................................................................24
Figure 3-1 Cadaver foot taped to a polycarbonate plate (a). Ankle with reflective marker array
in the tibia, talus, and calcaneus in the testing fixture (b). ............................................................34
Figure 3-2 Orthogonal coordinate axes on the tibia, talus, and calcaneus. .................................36
Figure 3-3 The sustentaculum tali of the calcaneus locks with the articulating surface of the talus
in a neutral position shown from the posterior (a) and medial (b) view. When everted, these
surfaces unlock, shown in the posterior (c) and medial (d) view. ................................................40
Figure 4-1 Posterior view of a neutral (left) and everted foot (right). .........................................44
Figure 4-2 Football shoes used for the current study: Nike Flyposite (a) and Nike Zoom Air (b).
........................................................................................................................................................47
Figure 4-3 Football cleat mold made of epoxy resin used to fix the shoe to the testing fixture ....48
x

Figure 4-4 The experiments were performed on a biaxial testing machine with the motions of the
talus and tibia tracked by a 5-camera motion capture system. .....................................................50
Figure 4-5 An example of the torque data from a sub-failure (a) and a failure (b) test. The
failure occurs when there is a steep drop in torque, as indicated by the arrow............................50
Figure 4-6 Lateral (a), posterior (b), and medial (c) views of one of the specimen-specific models
in SolidWorks with linear springs representing ligaments. ...........................................................51
Figure 4-7 Fibular fracture (a), medial malleolus fracture (b), ATiFL avulsion (c), ATiTL
avulsion (d), ATiFL tear (e), and ATiTL tear (f)............................................................................57
Figure 4-8 Representative plot of the ATiFL and ATiTL strains versus tibia rotation in a
simulation of the combination ATiFL and ATiTL injury scenario (specimen 2, stiff shoe). ..........58
Figure 5-1 The surrogate ankle taped to the polycarbonate plate using Elastikon (a), PowerFlex
(b), and Sher-Light (c). ..................................................................................................................69
Figure 5-2 The surrogate limb was attached to the testing machine .............................................70
Figure 5-3 Marker arrays were placed on the subject’s anterior tibia, about 10 cm distal to the
patella, and the calcaneal tuberosity. ............................................................................................72
Figure 5-4 Three orthogonal axes of the ankle. Rotations about these axes are described as
plantarflexion/dorsiflexion, eversion/inversion, and internal/external rotation. ..........................74
Figure 5-5 Torque-rotation curves for the three tape designs and two shoe designs from
rotational stiffness tests (a), and rotational stiffness values determined from the slopes of the
torque-rotation curves (b). Curves with different symbols (* # ƚ ¥) indicate statistically different
stiffness between designs................................................................................................................77
Figure 5-6 Mean ankle joint dorsiflexion-tibia rotation curves for the three tape (and no-tape)
designs from an in vivo external foot rotation (n=6). There were no significant differences
between tape designs......................................................................................................................78
Figure 5-7 Mean ankle joint dorsiflexion-tibia rotation curves for the three tape (and no-tape)
designs from an in vivo external foot rotation (n=6). There were no significant differences
between tape designs. ...................................................................................................................78
Figure 5-8 Mean ankle joint eversion-tibia rotation curves for the three tape (and no-tape)
designs from an in vivo external foot rotation (n=6). Different symbols (* #) indicate
statistically different curves. ..........................................................................................................79

xi

Figure 5-9 Predicted anterior tibiofibular ligament (ATiFL) (a) and anterior tibiotalar ligament
(ATiTL) (b) strain versus tibia rotation for an ankle constrained with tape stiffness ranging from
0 to 30 Nm/deg. Inputs for ankle joint motion were based on predictive equations 3a-3c. .........81
Figure 6-1 Marker set of the Oxford Foot Model (Carson et al., 2001). The subjects stood on
one foot and performed an internal rotation of the body (external rotation of the foot)...............89
Figure 6-2 Sample of a moment versus time plot (left) and rotation versus time plot (right) for all
three movements: internal-external rotation, inversion-eversion, and plantarflexion-dorsiflexion.
Data from the loading phase (shaded in the figure) for moment vs. time were extracted and
plotted against the extracted rotation vs. time data, as shown in Figure 6-3................................90
Figure 6-3 Samples of moment versus rotation plots in the loading phase for internal-external
rotation (left), inversion-eversion (right), and plantarflexion-dorsiflexion (bottom). ...................91
Figure 7-1 Strains in the ATiFL and two deltoid ligaments (ATiTL and TiNL) for neutral and
everted simulations. .....................................................................................................................98
Figure 7-2 Representative plot of the ATiFL and ATiTL strains versus tibia rotation ................101
Figure 7-3 Mean ankle joint dorsiflexion, external rotation, and eversion vs. tibia rotation for
three tapes and a tapeless condition. ..........................................................................................103
Figure 7-4 Predicted anterior tibiofibular ligament (ATiFL) (a) and anterior tibiotalar ligament
(ATiTL) (b) strain versus tibia rotation for an ankle constrained with tape stiffness ranging from
0 to 30 Nm/deg. Inputs for ankle joint motion were based on predictive equations 3a-3c. .......104
Figure A-1 Typical MIMICS interface ........................................................................................112
Figure A-2 Orthogonal coordinate axes on the tibia, talus, and calcaneus. ...............................118
Figure A-3 Marker arrays in a cadaver ankle (a) and human subject (b) ..................................119

xii

CHAPTER 1:
INTRODUCTION AND BACKGROUND
Ankle Anatomy and Kinematics
The ankle plays an important role in both athletic performance and everyday movement. The
ankle bones include the tibia, fibula, talus, and calcaneus and contacts between these bones
define the distal syndesmotic joint, the talocrural joint, and the subtalar joint (Figure 1-1). The
ligaments that connect the ankle bones can be divided into three groups (Figure 1-2). The lateral
collateral ligaments consist of the anterior talofibular ligament (ATaFL), interosseous
talocalcaneal ligament (ITaCL), calcaneofibular ligament (CaFL), and posterior talofibular
ligament (PTaFL). The medial ligaments consist of the anterior deltoid—the anterior tibiotalar
ligament (ATiTL) and tibionavicular ligament (TiNL)—and the posterior deltoid—the posterior
tibiotalar ligament (PTiTL) and tibiocalcaneal ligament (TiCL). The distal syndesmotic
ligaments include the anterior tibiofibular ligament (ATiFL), interosseous ligament (IOL), and
posterior tibiofibular ligament PTiFL) (Golano et al., 2010).

Figure 1-1 Medial view of the bony anatomy of the ankle

1

Figure 1-2 Lateral (left) and medial (right) view of the osteoligamentous ankle anatomy

Motion of the ankle is complex and involves coupled motion across three axes (Figure 1-3).
Rotation about the y-axis (medial-lateral) is called plantarflexion when the toes rotate away from
the body and dorsiflexion when the toes rotate towards the body. Rotation about the x-axis (long
axis) is called eversion when the bottom of the foot faces away from the body center and
inversion when the bottom of the foot points towards the body center. Rotation about the z-axis
(inferior-superior) is called internal rotation when the long axis points towards body center and
external rotation when the long axis points away from the body center. Motion of the talocrural
joint is primarily responsible for plantarflexion/dorsiflexion and internal/external rotation while
motion of the subtalar joint is primarily responsible for inversion/eversion.

Motion of the ankle rarely isolates one of these three rotations, resulting in coupled motions due
to the bony and ligamentous constraints of the ankle. Eversion and external rotation are strongly
coupled and both weakly coupled with dorsiflexion. Inversion and internal rotation are also

2

strongly coupled and weakly coupled with plantarflexion. Forced rotation about one axis usually
results in a ‘transferal of rotation’ in which rotations occur about the other axes in order to
dissipate torsion (Hicks, 1953, Lundberg et al., 1989b). When this natural response of the ankle
is compromised by restricting movement about one or more axes, increased strain is generated in
ligaments or bones resulting in injury.

Figure 1-3 The three motions of the ankle illustrated about a Cartesian coordinate system

Ankle Injury in Sports
Damage to the ankle is the most frequently observed injury in the emergency room (Boruta et al.,
1990) and accounts for 10-30% of all sports injuries (Waterman et al., 2010). In some sports,
these percentages can be even higher: Garrick and Requa report that the rate could be as high as
74% in softball, 76% in racquet sports and football, 77% in weight lifting and dancing, 79% in
basketball, and 82% in volleyball (Garrick and Requa, 1989). These injuries can also be
debilitating with up to 40% of individuals with a history of ankle injury having complaints
interfering with daily life (Golano et al., 2010, Gerber et al., 1998). These interferences with

3

daily life include pain, instability, crepitus, weakness, stiffness, and swelling (Yeung et al.,
1994).

In professional sports, absence of key players due to ankle injury may result in defeats

in major games, as well as tremendous economic losses (Fong et al., 2007). Because of these
detrimental effects, researchers are continuously working on ankle injury prevention.
Inversion Injuries
Most commonly, damage occurs to the lateral ligamentous complex and the primary mechanism
of these injuries is inversion (Renstrom and Konradsen, 1997, Beynnon et al., 2002). If the ankle
is in plantarflexion, the first ligament injured during inversion is the ATaFL (Dias, 1979). When
the ankle is in neutral flexion, some studies report that the first ligament injured is the CaFL
(Parenteau et al., 1998), while others report a roughly even distribution between CaFL and
ATaFL injury (Funk et al., 2002, Rasmussen and Kromannandersen, 1983). Continued inversion
has also been shown to injure the PTaFL (Dias, 1979). Many cases of lateral ankle sprain involve
additional injury to other soft tissue structures (Fallat et al., 1998) and, in severe cases, avulsion
fractures to the lateral malleolus common in children and adults over 40 years old.

External Rotation Injuries
Excessive external rotation has been shown to result in both medial and high ankle injuries
(Waterman et al., 2010, Gerber et al., 1998). Medial ankle injuries involve damage to the TiNL
and the ATiTL, while high (or syndesmotic) ankle injuries are defined as damage to the ATiFL
(Figures 1,2). High ankle sprains are mostly prevalent in football, team handball, basketball, and
soccer while medial ankle sprains occur most commonly in rugby, gymnastics, and soccer
(Waterman et al., 2011).

While less common than inversion injuries, medial and high ankle

injuries are of particular interest to researchers due to their longer recovery time (Boytim et al.,

4

1991), and potential for long-term ankle dysfunction (Gerber et al., 1998, Waterman et al.,
2010), especially in high-intensity sports.

Often the above ankle injuries occur in conjunction with one another (Williams et al., 2007,
Miller et al., 1995, Edwards and Delee, 1984). The various mechanisms of these injuries have
been described in previous studies (Hughes et al., 1979, Laugehansen, 1950), but all occur as a
result of external rotation. According to the Weber-C classification, external rotation to an
everted ankle puts maximum tension on the medial side of the ankle, first causing rupture of the
ATiTL, or fracture of the medial malleolus. The injury may stop here, referred to as the
pronation-exorotation stage [25]. Once the ATiTL ruptures, the talus may continue to externally
rotate and move laterally causing rupture of the ATiFL However, another mechanism of failure
has been proposed in which the ATiFL fails first: axial loading of a pre-everted ankle causes
lateral translation of the talus and pre-strain in the ATiFL, followed by external rotation which
fails the ATiFL (Wei et al., 2012b).

Previous laboratory studies have produced medial and syndesmotic injuries in cadavers, but the
majority of these injuries were fractures. A recent study by Haraguchi and Armiger show that
external rotation to an everted foot with 700 N of axial load results in the classic Lauge-Hansesn
supination-external rotation injury pattern: a short oblique fibular fracture starting at the level of
the tibial plafond and running in a posterio-superior direction, disruption of the ATiFL and
PTiFL complexes, and either a medial malleoular fracture or ATiTL tear (Haraguchi and
Armiger, 2009). These findings have been confirmed by an analysis of injury videos posted on
YouTube.com which show that five out of seven ankles subjected to pronation-external rotation

5

loading show a fracture pattern that is consistent with the pattern described by Haraguchi (Kwon
et al., 2010). Other studies which have rotated ankles to failure in a neutral position result in a
nearly even distribution of medial and lateral injuries with a low incidence of syndesmotic
injuries and a high incidence of fractures (Rasmussen and Kromannandersen, 1983, Hirsch and
Lewis, 1965). A recent study by Wei, however, produced both medial and syndesmotic
ligamentous injuries in cadaver ankles (Wei et al., 2012b). All ankles were pre-dorsiflexed prior
to rotation, but left ankles were pre-everted as well. Wei observed high ankle sprain in the preeverted left ankles and medial ankle sprain in the neutral right ankles.
Risk Factors for External Rotation Injuries
Shoe-Surface Interface
Traction, the resistance to relative motion between a shoe outsole and sports surface (Villwock et
al., 2009b), can be quantified in either the linear or rotational motions (Kent et al., 2012). While
linear traction is necessary for performance, it is generally accepted that excessive rotational
traction may increase the risk of ankle injury (Nigg and Yeadon, 1987, Lambson et al., 1996).
Torg and Quedenfeld (Torg and Quedenfe.Tc, 1973) were among the first researchers to
document the interaction between shoes and surfaces as an injury risk factor. They noted that the
size and number of cleats are correlated with the frequency of ankle injury, with less aggressive
cleats generating fewer injuries. Livesay et al. (Livesay et al., 2006) later noted that shoe-surface
rotational stiffness, defined as the rate at which moment is developed under shoe rotation on the
surface, may act as another risk factor for external rotation ankle injuries. The Livesay study
involved measuring the rotational traction and shoe-surface rotational stiffness between 5
playing surfaces and 2 types of shoes. The results show that differences in shoe-surface
rotational stiffness were greater than differences in traction, suggesting that rotational stiffness

6

may be a more sensitive indicator of interaction between shoes and surfaces than rotational
traction. A more recent study by Villwock et al. (Villwock et al., 2009c) involved measuring
both the traction and shoe-surface rotational stiffness between 10 football shoe models and 4
playing surfaces. The study concludes than artificial surfaces yield significantly higher peak
moments and shoe-surface rotational stiffness than natural grass, indicating that natural grass
may be a better surface for mitigating rotational injury risk.
Shoe Stiffness
While shoe-surface interface has been studied as an ankle injury risk factor for decades, it has
only recently been proposed that the constraint on the ankle has an effect on the type of injury
generated under external rotation (Wei et al., 2011d). A cadaver study by Wei et al. has
compared the effect of two different foot constraints on motion of a cadaver ankle during
external rotation: a potted constraint in which the calcaneus was fixed and a taped constraint in
which the calcaneus was free to collapse (Wei et al., 2010). When the potted foot was externally
rotated, the talus experienced dorsiflexion and inversion, stretching the posterior-lateral aspect of
the ankle and generating PTaFL injury. However, external rotation of the taped foot resulted in
talus eversion and plantarflexion, stretching the anterior-medial aspect of the ankle and
generating ATiTL failure.

Another study compared the motion of the talus when externally rotated in football shoes (Wei et
al., 2012a). The study first measured the rotational stiffness (the rate at which moment was
developed under external foot rotation in the shoe) of four different shoes. The rotational
stiffness in each of the four designs was significantly different from one another, indicating that
the material constructing the upper of the shoe may be a unique property. The stiffest and most

7

flexible shoe designs were selected for the cadaver study. Twelve (six pairs) cadaver ankles
were externally rotated 30 degrees in either the stiff or flexible shoe while motion capture
tracked the rotations of the talocrural joint using reflective marker arrays screwed into the talus
and tibia. Results of the motion capture analysis have shown that ankles in the stiff shoe yielded
more talocrural rotation, but less talocrural eversion than the ankles in a flexible shoe. This
suggests that ankles in stiff shoes experience higher ATiFL injury risk, but lower ATiTL injury
risk than those in flexible shoes, but this had not yet been tested at failure levels.
Computational Modelling
Validated computational models can be used to understand joint function and, clinically, to
understand and prevent sports injuries. Researchers generally develop models that are either
finite element analysis-based (FEA) or multibody kinematic- or dynamic- based (Chao, 2003,
Cheung et al., 2006, Konradsen and Voigt, 2002, Kwak et al., 2000, Iaquinto and Wayne, 2010).
Since both approaches have their advantages, the method chosen depends on the information
sought. While FEA models have the advantage of solving for small deformations, rigid body
models are able to solve the mechanics of large structures using highly efficient algorithms,
which execute much faster than FEA models(Kwak et al., 2000). These models are driven by
experimental measurements of bone movements, assumptions about joint degrees of freedom,
and assumptions regarding muscle forces that drive the motion of bones (Holzbaur et al., 2005,
Kitaoka et al., 1997, Konradsen and Voigt, 2002, Kwak et al., 2000). Liacouras and Wayne have
developed a 3D computational approach to model the lower leg in order to simulate cadaver
ligament sectioning studies of syndesmotic injury and ankle inversion stability (Liacouras and
Waynel, 2007). A more recent model by Wei et al. (Wei et al., 2011c) has been developed and
validated against two cadaver studies of ankle ligament strains (Colville et al., 1990a) and ankle

8

joint torques (Wei et al., 2010). This model has proven its ability to simulate footwear behavior
in the ankle (Wei et al., 2012a), predict ligament strains during an in vivo inversion ankle sprain
(Fong et al., 2011), and estimate ligament strains in human subjects during external rotation (Wei
et al., 2011c). However, fewer models incorporate ‘subject-specific’ parameters such as
articulation geometry, ligaments, and other anatomical features. Specifically in the ankle, the
curvature of the tronchlea tali (top surface of the talus) determines the stability of the ankle in the
anterior-posterior direction (Kleipool and Blankevoort, 2010) and the talar facets (contacting
surfaces) have a significant impact on the position of the axis for movements between the talus
and calcaneus (Barbaix et al., 2000). Variations in these parameters could have a significant
impact on the results of a rigid-body simulation.

Summary and Objectives
Two key phases in preventing injury include understanding the mechanism responsible for the
injury and decreasing the probability of this mechanism occurring. This can be accomplished
through investigations of clinical injuries, cadaveric testing, understanding in vivo kinematics,
and computational modelling. Cadaveric testing offers advantages in the control that can be
taken with experimental design, precision when measuring the forces and motions during injury,
and in-depth pathology afforded by post-test dissection. However, these tests often neglect
muscle action, which occurs during in vivo injuries. Yet, researchers can supplement
information drawn from cadaver studies with that measured from in vivo human subjects
performing similar motions in order to gain a fuller understanding of human ankle joint
movement. Additionally, specimen-specific rigid body modelling provides a way to costeffectively assess injury risk using boundary conditions from both cadaveric and in vivo tests.

9

One of the hypotheses of the current research is that the movement of the talocrural joint and
location of injury during ankle external rotation depends on the position of the ankle prior to
rotation and the amount of constraint on the ankle. Reporting the specific injury generated as a
result of foot constraint during external ankle rotation could aid in designing athletic shoes that
mitigate injury risk without compromising performance and comfort. Another hypothesis of the
research is that specimen-specific, rigid body modelling can be utilized to simulate injury-level
ankle rotation in addition to modelling the effect of various shoe rotational stiffnesses on
talocrural and subtalar joint movement. The information from these models could help quantify
failure properties of various structures of the ankle and may help mitigate the risk of rotational
ankle injury through footwear design.

10

CHAPTER 2:
SPECIMEN-SPECIFIC COMPUTATIONAL MODELS OF ANKLE SPRAINS PRODUCED
IN A LABORATORY SETTING

ABSTRACT
The use of computational modeling to predict injury mechanisms and severity has recently been
investigated, but few models report failure level ligament strains. The hypothesis of the study
was that models built off neutral ankle experimental studies would generate the highest ligament
strain at failure in the anterior deltoid ligament, comprised of the anterior tibiotalar ligament
(ATiTL) and tibionavicular ligament (TiNL). For models built off everted ankle experimental
studies the highest strain at failure would be developed in the anterior tibiofibular ligament
(ATiFL). An additional objective of the study was to show that in these computational models
ligament strain would be lower when modeling a partial versus complete ligament rupture
experiment. To simulate a prior cadaver study in which six pairs of cadaver ankles underwent
external rotation until gross failure, six specimen-specific models were built based on computed
tomography (CT) scans from each specimen. The models were initially positioned with 20°
dorsiflexion and either everted 20° or maintained at neutral to simulate the cadaver experiments.
Then each model underwent dynamic external rotation up to the maximum angle at failure in the
experiments, at which point the peak strains in the ligaments were calculated. Neutral ankle
models predicted the average of highest strain in the ATiTL (29.1 ± 5.3%), correlating with the
medial ankle sprains in the neutral cadaver experiments. Everted ankle models predicted the
average of highest strain in the ATiFL (31.2 ± 4.3%) correlating with the high ankle sprains
documented in everted experiments. Strains predicted for ligaments that suffered gross injuries
were significantly higher than the strains in ligaments suffering only a partial tear. The

11

correlation between strain and ligament damage demonstrates the potential for modeling to
provide important information for the study of injury mechanisms and for aiding in treatment
procedure.

12

INTRODUCTION
Ankle injury is the most common injury in sports and accounts for 10-30% of sports injuries
(Waterman et al., 2010). Injury to the lateral ligamentous complex occurs under excessive foot
inversion and is known as a “lateral ankle sprain” (Colville et al., 1990b). Injury to the anterior
deltoid ligament (ADL), which consists of the tibionavicular ligament (TiNL) and the anterior
tibiotalar ligament (ATiTL), is known as a “medial ankle sprain” (Wolfe et al., 2001). High
ankle sprains occur in the distal tibiofibular syndesmosis, which is comprised of the anterior and
posterior tibiofibular ligaments (ATiFL and PTiFL) and the interosseous ligament (IOL) (Dattani
et al., 2008). While most ankle sprains are lateral ankle injuries, high and medial ankle injuries
are typically more severe and result in more recovery time (Wolfe et al., 2001). High ankle
sprains, in particular, occur in 11-20% of ankle injuries, depending on the level of competition
(Hopkinson et al., 1990, Guise, 1976). Historically these injuries have been underdiagnosed and
assessment in terms of severity and optimal treatment has not been determined. More recently, a
heightened awareness in sports medicine has resulted in more frequent diagnoses of high ankle
sprains (Williams et al., 2007).

The clinical literature has mostly attributed both high and medial ankle sprains to external foot
rotation (Waterman et al., 2010, Boytim et al., 1991). Other studies, however, have indicated
that the mechanism of high ankle sprain involves a combination of dorsiflexion, eversion, and
external rotation (Wolfe et al., 2001, Boytim et al., 1991, Waterman et al., 2011).

A

computational model of the ankle was developed in our laboratory in order to further investigate
ankle injuries occurring in cadaver studies or a clinical setting. The model was first built to
study the effect of constraints of the foot on ankle injury using human cadaver studies (Wei et

13

al., 2010, Wei et al., 2011d). The model generated ankle torques, rotational stiffnesses, and
ligament strains that were consistent with the cadaver data.

The model was later used to

parametrically investigate the roles of dorsiflexion, eversion, and external rotation on high ankle
sprains. Since the damage experienced in ankle sprains is primarily due to mechanical disruption
from excessive deformation of ligaments (Villwock et al., 2009c), ligament strain was used in
the model to indicate ankle sprain risk. In a Wei et al. study (Wei et al., 2012b), the simulations
suggest that an everted, externally rotated foot generates the highest strains in the ATiFL, while a
neutral (non-everted), externally rotated foot generates the highest strains in the ADL. Despite
this useful information provided by the previous model, the input motions used in those
simulations correlate with subfailure level of sprains, and therefore injurious ligament strains are
unknown.

Recently our laboratory conducted external rotation, failure level experiments on pairs of
dorsiflexed, human cadaver ankles that were either everted 20 degrees or kept neutral prior to
failure (Wei et al., 2012b). Neutral ankles failed by injury to the ADL. The injuries involved
partial and complete ruptures, tibial avulsion in one case, and, in another case, a combination of
partial tears in the ADL and ATiFL. In all everted ankles taken to failure under external rotation,
cases of complete and partial ruptures of the ATiFL were reported. This was the first study to
produce isolated ATiFL injury, representative of a high ankle sprain, in a laboratory setting.

In the current study, specimen-specific computational models were developed to investigate
ankle ligament strains generated in the eversion study of Wei et al. (Wei et al., 2012b). Knowing
the relative motion of the tibia and talus during that study, commercial software was used to

14

determine the ligament strains at failure in the current study. Our hypothesis was that in the
neutral ankle externally rotated to failure, the highest ligament strain would be generated in the
ADL. For everted ankles the highest strain would be developed in the ATiFL. An additional
objective of the study was to show that in these computational models ligament strain would be
lower in specimens with a partial ligament rupture.

METHODS
No new experimental work was performed for this study. However, details of the previous study
from which the specimen-specific computational models were based will be presented, in brief,
in order to adequately describe the current study.

Cadaver Experiment
Six male cadaver lower limbs(aged 56 ± 12 years) were thawed for 24 hours and were transected
approximately 15 cm distal to the center of the knee (Wei et al., 2012b). The foot was taped to a
polycarbonate plate and inserted into a rectangular tray (Figure 2-1). Both the right and the left
limbs were dorsiflexed 20 degrees. The left limbs were positioned at 20 degrees of eversion,
while the right limbs were initially positioned in neutral with regard to the inversion/eversion
rotation prior to dynamic external foot rotation.

A compressive preload of 1500 N was applied axially and internal tibia rotation (external foot
rotation) was input in position control at a frequency of 1 Hz (0.5 seconds to peak rotation for
each test). The tests started with 20 degrees of rotation and then repeated in 10 degree
increments successively until an observed injury to the ankle. A steep drop in the magnitude of

15

Figure 2-1 Setup for the cadaver experiment by Wei et al..

the torque signal indicated failure, at which point the foot rotation was documented and the ankle
was examined for injury after each experiment (Table 2-2-1). Prior to the test, reflective marker
arrays were screwed into the talus and tibia, allowing an analysis of the motion of the talus with
respect to the tibia using a Vicon motion capture system (Oxford Metrics Ltd., Oxford, United
Kingdom) (Figure 2-1 and Table 2-2). While these arrays were randomly oriented on the tibia
and talus, coordinate transformations were set-up to determine the orientation of the tibia marker
set with respect to the long axis of the tibia and the talus marker set with respect to the mediallateral axis of the talus, based on the initial orientation of a line between the medial and lateral
malleoli. This provided an anatomical axis for each bone. The third joint coordinate axis
(floating axis) was determined by the cross product of these anatomical-based vectors. Clinical
angles were determined as previously described by Soutas-Little et al. (Soutas-Little et al., 1987)
(Table 2-2).

16

Model Development
Computed tomography (CT) scans were performed on the right limb from each specimen pair to
obtain three-dimensional joint anatomy with 0.6 mm spacing between images and in-plane
spatial resolution of 0.508 mm. Scans were performed with the foot taped to a polycarbonate
plate and placed in a fixture to maintain neutral flexion during scanning. The CT scans were
transferred from Digital Imaging and Communications in Medicine (DICOM) files into
Materialise’s Interactive Imaging Control System (MIMICS, Leuven, Belgium) software. The
individual bones were separated and meshed as solid bodies. A “mask” was created for each
bone, accurately reflecting the surface geometry of the bone. In order to reduce the size of the
files, the bones were remeshed using the auto remesh feature. Each bone was then exported as a
stereolithography (STL) file and imported into Solidworks 2009 (Trimech Solutions, LLC,
Columbia, MD) as mesh files where they were simplified using the ScanTo3D package.

SolidWorks Motion was used to add ligaments (modeled as tension-only linear springs) and
assemble the bones in their anatomically correct orientation based on the CT scans. Insertion
locations of the ligaments were determined from both dissection and an anatomical atlas (Netter
and JT, 2003).

A pre-load was induced on each ligament by reducing its length by 2% from the

scanned lengths, as in Wei et al. (Wei et al., 2011d). Three-dimensional contacts were
implemented, similar to the procedure outlined by Iaquinto and Wayne(Iaquinto and Wayne,
2010), between adjacent bones in order to prevent overlap during the simulation. This was done
by calculating the interference at each time step and applying an outward force if any overlap
was detected.

The force was a function of the material stiffness k (10,000 N/mm), the

penetration depth g (0.001 mm), the exponent e (1.75), the damping coefficient cmax (50 N/mm-

17

s), penetration depth at maximum damping dmax, and penetration velocity

dg
. Friction and the
dt

effects of gravity were considered negligible.
 dg 
Fn = kg e + 
 f (c max , d max )
 dt 

Since toe involvement was likely minimal in the cadaver study, the phalanges were excluded
from the model.

a

c

b

d

e

Figure 2-2 The cadaver taping pattern (a) and the model taping pattern represented by springs
(b). Note: for clarity all ligament springs and the plate springs on the medial side of the ankle
were set invisible.

18

The neutral position of the ankle from the CT scan was maintained. The tibia was fixed in space
while the fibula, calcaneus, and talus were free to move. Due to minimal relative motion among
the forefoot bones and in order to reduce run-time, the rest of the bones were fused together and
moved as a rigid body. The model included 22 linear spring elements (Figure 2-2) to represent
ligaments with stiffness values taken from the literature, as referenced by Wei et al. (Wei et al.,
2011d).

Simulation
To simulate the cadaver ankle rotation, a plate was created in SolidWorks with dimensions from
the experiment. The plate was attached to the foot with 14 springs representing the athletic
taping pattern used on the cadaver feet (Figure 2-2b).

Motors were used to drive the motion of the plate, with reference to the fixed tibia (as per the
cadaver experiment), with one motor for plate dorsiflexion, one for plate eversion, and one for
plate external rotation. For each ankle model in SolidWorks Motion, the plate was dorsiflexed
20 degrees, everted 20 degrees (for models based on everted foot tests), and an axial pre-load of
1500 N was applied to the proximal end of the model (distributing the load between tibia and
fibula with one-sixth of the loading on the fibula). The plate was then externally rotated to the
angle at which injury was documented in the cadaver experiments. All motions were relative to
the fixed tibia. Since the stiffness and preload of the athletic tape were not quantified in the
experiment, stiffness of all spring elements representing the athletic taping pattern (Figure 2-2b)
was iteratively and simultaneously adjusted in 10 N/mm increments to help match the model
external talus rotation to the experimental talus rotation. Thus one stiffness was applied to all

19

those springs that represented the effect of tape. Ligament strains, defined in percentage as the
relative elongations of ligaments, were determined at the plate rotation needed to generate gross
ankle injury in the cadaver experiments (Table 2-1).

Two-sample t-tests were performed to compare failure strains for partial tears and total ruptures,
for comparing the model generated talus rotations with experimental talus rotations, for
comparing ATiFL failure strains with ADL failure strains, and for comparing experimental plate
rotations with experimental talus rotations. P values less than 0.05 were considered significant in
all tests.

RESULTS
Injuries reported in the cadaver experiments of Wei et al. (Wei et al., 2012b) were used in this
study (Table 2-1). The injuries consisted of partial tears and complete ruptures of the ATiFL
(Figure 2-3a and 2-3b, respectively), partial tears and ruptures of the ATiTL (Figures 3c and 3d,
respectively), tibial avulsion of the ATiTL (Figure 2-3e), partial tear of the TiNL (Figure 2-3f),
and a spiral fracture of the fibula. At failure, comparisons between the rotations of the plate and
talus showed that in the neutral case, talus rotations were significantly less than plate rotations
(p<0.001), but not so in the everted case (p=0.11) (Tables 2-1 and 2-2) [13].

20

a

b

Lateral

Medial

Fibula
Anterior

Posterior

Tibia
Posterior

Anterior

Fibula
Tibia

Lateral
Medial

c

d
Tibia

Tibia
Anterior

Posterior
Anterior

Posterior

Talus

e

Medial

f
Posterior
Tibia

Tibia

Calcaneus
Lateral
Medial
Lateral
Anterior

Figure 2-3 Injury mechanisms reported from the cadaver study by Wei et al. (Wei et al., 2012b).
Partial tear of the ATiFL (a); Rupture of the ATiFL (b); Partial tear of the ATiTL (c); Rupture
of the ATiTL (d); Tibial avulsion of the ATiTL (e); Partial tear of the TiNL(f). Note: the location
of injury is shown with a hemostat.

21

The stiffness of the simulated athletic tape was variable between specimens and between 300
N/mm to 400 N/mm. Comparisons of the talus rotations showed that the model generated talus
rotations were not statistically different than the experimental rotations (p>0.88) (Table 2-2).
Models developed for the neutral experiments showed the highest strain generated in the ATiTL
(29.1 ± 5.3%) followed by the ATiFL (22.3 ± 0.8%) and the TiNL (16.1 ± 4.2%) in the three of
four cases in which ligament failure occurred (Figure 2-4). These results are in agreement with
observations of ATiTL failure from the cadaveric experiment that occurred in all but one of the
neutral cadaver experiments in which ligament failure occurred (Table 2-1). In specimen 3,
failures were reported in the TiNL and ATiFL (Table 2-1). In this model the ATiFL strain was
comparable with that of the TiNL, and it was the highest among all ATiFL strains for the neutral
simulations (Figure 2-5).

Figure 2-4 Ligaments with average model predicted strains greater than 2% for the neutral case
in 3 of 4 cases in which ligament injury occurred

22

Figure 2-5 Strains in the ATiFL and the two deltoid ligaments (ATiTL and TiNL) for the neutral
case for each specimen modeled

Computational models of the everted ankle experiments showed the highest strains in the ATiFL
(31.2 ± 4.3%) followed by the TiNL (18.5 ± 5.6%) and the ATiTL (18.1 ± 5.0%) (Figure 2-6).
These data coincided with the observation of ATiFL injury in all of these cadaver experiments
(Table 2-1). All but two of the specimens from the cadaver study suffered complete ruptures
(specimens 2 and 5 suffered partial tears) and the corresponding strains can be seen in Figure 27. For ruptured ligament cases there was no significant difference between ATiFL failure strain
and ATiTL failure strain (p>0.7). For partial tears, there was no significant difference between
ATiFL failure strain and ADL failure strain (combining cases of ATiTL failure and TiNL
failure) (p>0.45).

23

Figure 2-6 Ligaments with average model predicted strains greater than 2% for the everted
simulations

Figure 2-7 Strains in the ATiFL and the two deltoid ligaments (ATiTL and TiNL) for the everted
simulations

24

Table 2-1 Results of the cadaver experiment for the neutral and everted ankle

Specimen

Plate failure
rotation (deg)

Failure mode

Neutral

Everted

Neutral (R)

Everted (L)

1
2

36.0
47.1

43.4
55.3

3

35.1

50.1

4
5
6

39.8
31.1
37.0

50.0
43.4
38.6

Tibial avulsion of the ATiTL
Partial tear of the ATiTL
Partial tear of the TiNL and partial tear
of the ATiFL
Rupture of the ATiTL
Spiral fracture of the fibula
Rupture of the ATiTL

Rupture of the ATiFL
Partial tear of the ATiFL
Rupture of the ATiFL
Rupture of the ATiFL
Partial tear of the ATiFL
Rupture of the ATiFL

Table 2-2 Comparison of the talus external rotations with respect to the tibia (deg) from the
cadaver study with the computational talus rotations
Specimen

1
2
3
4
5
6

Experimental talus failure rotation (deg)

Computational talus failure rotation (deg)

Neutral

Everted

Neutral

Everted

20.9
29.7
23.4
25.2
17.9
22.5

37.7
46.4
45.7
44.7
39.7
36.5

18.7
31.3
21.0
24.2
18.1
20.9

38.2
44.8
47.0
43.1
37.6
37.9

Combining both neutral and everted simulations, strains of 31.8 ± 3.8% (n=6) computed from the
models of experiments that showed grossly ruptured ligaments were statistically greater than
25.5 ± 4.2% (n=4) for the partial tear simulations (p=0.046). Most of the strains for partial tear
simulations were less than those for complete ruptures except for the ATiFL strain for specimen
5 in the everted case and for the ATiTL strain for specimen 6 in the neutral case. Additionally
the ATiTL strain for specimen 4 in the everted case suggests failure, but this did not occur
experimentally. This suggests a slightly mixed correlation between strains and injury modes.

25

Furthermore, in the neutral model that suffered a spiral fracture experimentally (specimen 5), the
ATiFL, ATiTL and TiNL strains (23.6%, 10.4%, and 13.0%, respectively) were below failure
levels documented in other experiments. In the ATiTL avulsion case (specimen 1), the ATiFL
and TiNL strains (21.2% and 14.9%, respectively) were also below the average failure levels of
other simulations. The ATiTL strain, on the other hand, from the model was 32.2%, well above
the failure levels given in other simulations.

DISCUSSION
The study supported the hypothesis that the models would generate the highest strains in the
ATiFL for the everted foot and in the deltoid ligaments for the neutral foot. As suggested by
Wei et al. (Wei et al., 2012b), the lateral talar translation resulting from an everted foot puts
strain on the ATiFL by pushing the fibula away from the tibia, since the ligament is oriented
along the coronal plane. In the neutral foot, lateral talar translation was minimal during axial
load and external rotation, and thus external rotation resulted in straining the deltoid, as this
ligament is oriented along the sagittal plane.

The study also confirmed that there was a

significant difference in strains for cases of complete versus partially ruptured ligaments. In the
neutral case, specimen 3 revealed the correct prediction of injury, despite yielding a different
result than the other simulations. In all other cases of a neutral foot ligament failure the ATiTL
experienced the highest strain, yet the highest strain in specimen 3 was found in the TiNL
followed by the ATiFL. This was consistent with the injuries generated in the TiNL and ATiFL
in the cadaver study (Table 2-1 and Figure 2-5). Furthermore, the rotations to failure for the
neutral foot in specimens 1, 3, and 6 were similar, yet corresponded to different computational
ligament strains (Figure 2-5) and this was supported by the observation of different injuries
(Table 2-1). This information indicates that specimen-specific modeling may, in some cases,

26

predict injuries that may not be predictable with a generic model, such as used previously by this
laboratory. While differences in the geometry of the trochlea tali and talar facets have both been
identified as potential explanations for differences between specimens, a detailed presentation of
these differences was beyond the scope of the current experiment.

While little data exists on the failure strains of the ankle ligaments of interest in this study, these
failure data are consistent with experimental failure calculations from other ligaments. A study
by Beumer et al. (Beumer et al., 2003), for example, reported the force required to rupture the
ATiFL, PTiFL, and posterior tibiotalar ligament (PTiTL) to be 500±105 N, 708 ±91 N, and
446±51 N, respectively.

Using the stiffnesses from the literature (Wei et al., 2011d) and

ligament lengths from the simulations, the failure forces of the ATiFL, TiNL, respectively, were
610±98 N and

750±190 N, comparable to Beumer’s values, while the ATiTL prediction,

170±0N, was considerably lower. Attarian et al. (Attarian et al., 1985) measured failure strains
of bone-ligament-bone preparations in tensile experiments. They documented values of, on
average, 53% and 38% for the anterior talofibular ligament (ATaFL) and calcaneofibular
ligament (CaFL), respectively, which are higher than the failure strains reported in the current
study.
Given that ligament collagen fibers start to tear at approximately half of the failure strain (a socalled grade I ligament sprain) (Yahia et al., 1990), the partial tears might be more representative
of grade II sprain injuries, as they occurred at approximately 70% of the total rupture strain.
Interestingly, the strains for non-injured ligaments fell below this partial-tear failure strain
average for all ligaments in the neutral simulation and all but one ligament (ATiTL for specimen

27

4) in the everted simulations. This may indicate a low risk of falsely predicting ligament failure
for non-injured ligaments in future injury simulation studies.

One limitation of the current study was that we were not able to accurately quantify the ankletape structural stiffness of the cadaver study, and thus had to estimate it in our simulations. We
attempted to compensate for this aspect by adjusting the spring stiffness that represented the
taping of the foot to the plate for each specimen so that the motion of the model talus was similar
to that of the cadaver talus for a given level of plate rotation at failure. These values varied
between specimens, as the structural stiffness due to the amount of tape stretch in the cadaver
study may have actually varied between specimens. Despite potential errors in matching tape
stiffness, we believe that as long as the talar motions in the simulations closely matched those of
the experiments, the model’s ligament strains would provide reasonable estimates of strain.
Another limitation of the study was that it assumed simple, linear elastic behavior of the ankle
ligaments. While ligaments may behave linearly for small deformations, they do behave nonlinearly for failure strains. Because of this, non-linear computational models should be used for
failure-level studies in the future. The model also did not account for the viscoelastic properties
of ligaments. Funk et al. (Funk et al., 2000) produced non-linear models for eight major
ligaments of the foot-ankle complex, which accounted for both their elastic and viscoelastic
properties. They concluded that a viscoelastic assumption could also be neglected for very slow
(<0.0001/s) or very fast (>1/s) strain rates, but substantial effects seem to exist on ligament
behavior for intermediate strain rates. In the cadaver experiment by Wei et al. (Wei et al.,
2012b), the tibia rotation was input at 1 Hz, so the ligaments reached their maximum strain at
0.5s. This means that the failure strain rates were, on average, 1.000/s for partial tears and 0.933

28

/s for ruptures, indicating that there could be some viscoelastic effect. However, while the model
may not currently predict the exact mechanical behavior of ligaments, quantifying failure
ligament strains, rather than forces, may be a better indicator of injury tolerance due to the
viscoelastic nature of ankle ligaments (Edwards and Delee, 1984, Boytim et al., 1991). This may
be supported by earlier studies that show that the ultimate tensile strength of ligaments and
tendons increases significantly with strain rate, while ultimate strain does not (France et al.,
1987), (Ng et al., 2004). Yet, in order to improve upon the model in future predictions of ankle
ligament failure strains from in vitro and in vivo studies, more accurate constitutive modeling
would still seem to be warranted.

The use of specimen-specific ankle modeling could be expanded to investigate in vivo injury
cases simulating situations where the geometry and motion of the foot/ankle complex from a
specimen are known. For example, in the 2008 Beijing Olympic Games two ankle inversion
injuries were captured on film and analyzed using a model-based image matching (MBIM)
motion analysis technique from calibrated video sequences. The results of the study indicated a
disparity in injury mechanics from previous studies. Previous studies of lateral ankle injuries
suggest that the injurious motion involves inversion plus an internal rotation of the foot (Safran
et al., 1999) with plantarflexion (Vitale and Fallat, 1988). However, the MBIM study revealed
that plantarflexion was not involved in these ankle sprains, indicating that the talocrural joint was
less involved in these inversion ankle sprains (Mok et al., 2011). By combining the MBIM with
computational foot and ankle models developed from CT or magnetic resonance images (MRI),
ankle injury mechanisms may be better understood. Such studies could aid in the design of
intervention strategies and new equipment designs that might help limit the severity and

29

frequency of athletic ankle injuries. Based on the results of the current study, specimen-specific
modeling may help provide even better data for the study of ankle injury mechanics on the
athletic field.

30

CHAPTER 3:
EFFECT OF PRE-EVERSION ON TALOCRURAL AND SUBTALAR JOINT MOTION
DURING ANKLE EXTERNAL ROTATION

ABSTRACT
Eversion prior to excessive external foot rotation has been shown to predispose the anterior
tibiofibular ligament (ATiFL) to failure, yet protect the anterior deltoid ligament (ADL) from
failure despite high levels of foot rotation. The purpose of the current study was to measure the
rotations of both the subtalar and talocrural joints during foot external rotation in either a neutral
or a pre-everted position in order to better understand injury generated during rotation. Fourteen
(seven pairs) cadaver lower extremities were externally rotated 20 degrees in either a pre-everted
or neutral configuration. Marker arrays were screwed into the tibia, talus, and calcaneus and
motion capture was performed to track the motion of these bones. A joint coordinate system was
used to analyze motions of the two joints. While talocrural joint rotation was greater in the
neutral ankle (13.33 ± 2.09 deg versus 10.51 ± 2.75 deg, p=0.006), subtalar joint rotation was
greater in the pre-everted ankle (2.38 ± 1.89 deg versus 1.12 ± 1.04 deg, p=0.014). Overall, the
talocrural joint rotated more than the subtalar joint (11.92 ± 2.77 deg versus 1.75 ± 1.61 degrees,
p<0.001). It was proposed that the calcaneus and talus ‘lock’ in a neutral position, but ‘unlock’
when the ankle is everted prior to rotation. This locking/unlocking mechanism could be
responsible for an increased subtalar rotation, but decreased talocrural rotation when the ankle is
pre-everted, protecting the ADL from failure. This study may provide information valuable to
the study of external rotation kinematics and injury risk.

31

INTRODUCTION
Positioning of the ankle prior to external rotation affects the behavior of the talus and therefore
location of injury (Wei et al., 2012b, Button et al., 2013, Haraguchi and Armiger, 2009). A
previous study in this laboratory has shown that the combination of ankle eversion and
dorsiflexion followed by axial load and external rotation produces isolated anterior tibiofibular
ligament (ATiFL) injury, also known as high (or syndesmotic) ankle sprain. In a similar set of
experiments, removing the eversion component mostly results in anterior deltoid ligament (ADL)
injury, or a medial ankle sprain. Since the ATiFL restricts both fibular movement and external
talus rotation(Sarsam and Hughes, 1988), the mechanism of high ankle sprain is attributed to a
pre-tensioning of the ATiFL, as a result of the eversion and axial load, following by external
rotation of the talus(Wei et al., 2012b). Oriented approximately in the antero-posterior direction
in the sagittal plane (Wei et al., 2012b), the ADL is ruptured due to external rotation of the foot,
supported by a clinical review which states that external rotation of the foot will “first rupture the
deltoid ligament, with subsequent injury to the ATiFL” (Dattani et al., 2008, Wei et al., 2011d).
Yet, despite no observation of ADL injury in the pre-everted ankles, the amount of foot rotation
at failure was significantly higher than the failure rotation in neutral ankles (Wei et al., 2012b),
seemingly contradicting the notion that external foot rotation primarily strains the ADL.
However, while there is more overall foot rotation in the pre-everted ankles, ligament strain is
dependent on relative bone motion which was not documented at the point of failure in this
previous study.

While strain in the ADL is strongly correlated with talocrural joint rotation, it has been suggested
that the subtalar joint equally contributes to overall foot rotation (Siegler et al., 1988). Excessive

32

subtalar joint rotation could also be problematic due to increased risk of subtalar ligament injury,
which has been implicated in subtalar joint instability and calcaneal tilt (Kato, 1995, Heilman et
al., 1990).

Previous studies have measured subtalar and talocrural joint rotation during

plantarflexion/dorsiflexion, inversion/eversion (Wong et al., 2005, Fassbind et al., 2011), and
passive rotation (Siegler et al., 1988). While these studies noted that the contribution of both the
subtalar and talocrural joints to foot rotation was dependent on the positioning of the ankle, they
did not document the effect of pre-eversion and axial load on joint motion during rotation. The
objectives of the current study were 1) to measure the contributions from the talocrural and
subtalar joints to the overall rotation of the foot-shank in an axially loaded foot and 2) to identify
the effect of pre-eversion on these contributions.

METHODS
Fourteen (seven pairs) fresh-frozen human cadaver (age 58.5 ± 9.2 years) lower-extremities were
used in this study. The limbs were stored at -20°C and thawed to room temperature for 24 hours
prior to testing. The tibia and fibula were transected 15 cm distal to the center of the knee, and
the proximal 10 cm of the bones was removed of tissue and cleaned with 70% alcohol. The
cleaned portion of the bone was potted in an aluminum box with room-temperature curing epoxy
(Fiber Strand, Martin Senior Corp., Cleveland, OH). Two screws were placed in the anterior and
medial tibia before potting to help restrain the tibia in the potting material. Reflective marker
arrays were screwed in the anterior medial aspect of the talus, posterior calcaneus, and on the
tibia approximately 20 cm proximal to the articular surface. CT scans verified the placement of
the marker arrays.

33

a

b

Figure 3-1 Cadaver foot taped to a polycarbonate plate (a). Ankle with reflective marker array
in the tibia, talus, and calcaneus in the testing fixture (b).

Elastic athletic tape (Elastikon, J&J, New Brunswick, NJ) was used to constrain the foot onto a
polycarbonate plate (32 cm x 8 cm) (Figure 3-1a). Experiments were performed with a
customized hydraulic biaxial testing machine with a 244 Nm rotary actuator (Model SS-001-1C,

34

Micromatic, Berne, IN) and a vertically oriented linear actuator (Model 204.52, MTS Corp.,
Eden Prairie, MN). The proximal (potted) end of the foot was attached to the rotary actuator
through a biaxial load cell (Model 1216CEW-2K, Interface, Scottsdale, AZ) with a capacity of
8896 N of axial force and 113 Nm of torsion (Figure 3-1b). Right ankles were pre-everted 20
degrees prior to rotation and left ankles were neutral in the coronal plane. A 1500 N
compressive preload was applied to simulate dynamic weight bearing prior to rotation and 20
degrees of internal tibia rotation (external foot rotation) was input in position control at 1Hz.

A Vicon motion capture system (Oxford Metrics Ltd, Oxford, UK) was used to document the
position of the marker arrays during motion. Coordinate transformations were performed on the
tibia, talus, and calcaneus marker arrays in order to establish orthogonal coordinate systems in
line with the three axes of the ankle (Figure 3-2). Joint coordinate systems (JCS) were
established between the tibia and talus (talocrural joint) and between the talus and calcaneus
(subtalar joint), similar to previously defined systems (Soutas-Little et al., 1987, Wu et al.,
2002). Calculations for flexion, rotation, and inversion/eversion of the talocrural joint are shown
below (Equations 1-2) (note: the same procedure is used for the subtalar joint).
=

[1a]

=

[1b]

=|

=

=

∙

=

∙
∙

×
×

[1c]

|

(flexion)
(rotation)

(inversion/eversion)

35

[2a]
[2b]
[2c]

Figure 3-2 Orthogonal coordinate axes on the tibia, talus, and calcaneus.

Paired t-tests were used to compare the amount of joint flexion, rotation, and inversion/eversion
between neutral and everted ankles. Two-way repeated measures ANOVA (pre-positioning,
joint) was used to compare the amount of flexion, rotation, and inversion/eversion between the
two joints. P-values less than 0.05 were considered significant in all tests.

RESULTS
Motions of the subtalar and talocrural joints during 20 degrees of internal tibia rotation were
shown in Table 3-1 (Note: talocrural motion refers to talus motion with respect to the tibia, while
subtalar motion refers to calcaneus motion with respect to the talus). There was significantly
more external rotation in the talocrural joint compared with the subtalar joint (p<0.001).
Additionally there was significantly more talocrural external rotation (p=0.006), but significantly
36

less subtalar external rotation (p=0.014) in the neutral foot when compared with the pre-everted
foot.
Table 3-1 Motions of the subtalar and talocrural joint for 20 degrees of external rotation

Specimen

1
2
3
4
5
6
7
AVG
STDEV

Subtalar
Eversion (deg)

Talocrural
Eversion (deg)

Subtalar
Dorsiflexion
(deg)

Talocrural
Dorsiflexion
(deg)

Subtalar
Rotation (deg)

Talocrural
Rotation (deg)

Neutral Everted

Neutral Everted

Neutral Everted

Neutral Everted

Neutral Everted

Neutral Everted

0.59
1.66
0.50
0.25
0.97
1.04
1.71
0.96
0.56

p-value

0.29
2.43
0.21
0.32
0.41
1.99
0.84
0.93
0.91

0.39
2.68
1.53
1.35
1.30
1.50
0.58
1.33
0.74

0.39

0.19
3.80
1.03
0.30
1.34
1.31
0.85
1.26
1.21
0.79

0.787

1.46
0.16
0.03
0.34
2.24
1.43
0.98
4.09
1.29
0.75
3.11
3.50
0.97
0.53
0.55
0.52
0.84
0.16
0.79
0.02
1.23
1.02
2.91
0.78
0.04
0.52
0.72
0.48
1.15
0.65
1.30
1.39
0.67
0.46
1.21
1.67
0.054
0.882
0.197

0.04
0.69
15.81 11.08
1.11
1.70
12.30 10.97
2.53
5.78
11.84
6.41
0.23
0.88
15.67 12.87
0.02
0.94
14.93 14.20
1.98
2.88
12.09 10.53
1.90
3.81
10.65
7.52
1.12
2.38
13.33 10.51
1.04
1.89
2.09
2.75
0.014
0.006
<0.001

DISCUSSION
While previous studies have investigated the effects of ankle pre-eversion on location of injury
and strain generated in key ankle ligaments during external rotation of an axially-loaded foot
(Button et al., 2013, Wei et al., 2012b), the current study examined the effects that pre-eversion
had on talocrural and subtalar joint movement. Specifically, a key aim of the current study was
to investigate why pre-everted ankles saw no ADL injury despite high levels of foot rotation at
failure. The study involved using a JCS to calculate joint motion during 20 degrees of external
ankle rotation in either a pre-everted or a neutral foot. Results of the study indicated that when
the foot was axially loaded and externally rotated most of the foot-shank rotation comes from the
talocrural joint. However, pre-eversion lowered the amount of talocrural rotation, potentially
lowering ADL injury risk.

37

The geometry of the subtalar joint may explain the differences in rotation between the subtalar
and talocrural joints. The talus contacts the calcaneus on two articular surfaces: the subtalar
posterior joint and anterior subtalar joint (Viladot et al., 1984). The calcaneal surface on the
medial section of the anterior subtalar joint is called the “sustenaculum tali”, which the
articulating surface of the talus fits into (Figure 3-3) (Viladot et al., 1984), and the movement of
this complex is fundamental to the transmission of rotations from the foot to leg (Leardini et al.,
2001). In a neutral ankle, the sustentaculum tali may ‘lock’ with the articulating surface of the
talus, restricting its rotation (Figure 3a,b). As a result, when the tibia rotates, this ‘locking’
restricts the talus from rotating with the tibia, resulting in a high talocrural rotation between the
tibia and talus. In contrast, when the ankle is everted, (Figure 3-3c,3-d) the contact between the
sustentaculum tali and articulating surface of the talus ‘unlocks’. Consequently, this ‘unlocking’
enables the tibiotalar ligaments to pull the talus with the rotation of the tibia, resulting in lower
relative motion.

This phenomenon can possibly explain the injuries observed in the Wei et al. study (Wei et al.,
2012b). In that study, the neutral ankles experienced ADL injury in 5 of 6 cases while the preeverted ankles experienced no instances of ADL injury, despite failing at significantly higher
foot rotations. Since the ADL is the primary ligament resisting ankle rotation (Wei et al.,
2011d), one would have expected the significantly higher foot rotation in the everted ankles to
fail the ADL as well. However, results of the current study may imply that even though the preeverted ankles experienced more foot rotation at failure in Wei et al. (2012), the ‘unlocking’ of
the subtalar joint (due to pre-eversion) might have protected the ADL from failure by limiting
talocrural joint rotation. As suggested by Wei et al. (2012), axial loading may have translated

38

the talus laterally, thereby pushing the fibula away from the tibia, producing pre-strain in the
ATiFL, and potentially predisposing the foot to a high ankle sprain. Increased subtalar joint
rotation in the pre-everted ankles of the current study potentially could increase the risk of
subtalar ligament failure, resulting in chronic instability of the joint. On the other hand, the
subtalar joint remained ‘locked’ in the neutral scenario, explaining ADL injury for these ankle
(Wei et al., 2012b).

39

a

b

d

c

Figure 3-3 The sustentaculum tali of the calcaneus locks with the articulating surface of the
talus in a neutral position shown from the posterior (a) and medial (b) view. When everted,
these surfaces unlock, shown in the posterior (c) and medial (d) view.

Results of the current study appear to contradict conclusions from another study measuring
subtalar and talocrural rotations using a JCS (Siegler et al., 1988). The Siegler study concluded
that the talocrural and subtalar joints contributed nearly equally to overall foot rotation.

40

However, that study involved applying rotation to an unloaded ankle, while the current study
involved preloading the joint with 1500 N. The difference in experimental results of the two
studies may suggest that compressing the subtalar joint prior to rotation may be essential for the
‘locking’ mechanism to occur—suggesting that axial load may increase the risk of ADL injury.

A limitation of the current study was that the tests were run at sub-failure levels. While the
motions of the joints at failure levels could be estimated based on extrapolations of the motions
from this study, a future failure-level study is warranted. Additionally, the Wei 2012 failurelevel study involved pre-dorsiflexion to both neutral and everted ankles (Wei et al., 2012b),
while the current study did not. While pre-dorsiflexing the ankles in the current study may have
better replicated the previous experiment, a previous parametric study concluded that predorsiflexion has little effect on ADL strain during external rotation (Wei et al., 2011b) indicating
that pre-dorsiflexion may have changed the results of this study very little.

In conclusion, everting the foot prior to external rotation affects the kinematics of the subtalar
and talocrural joints. When the ankle is in neutral, the subtalar joint ‘locks’, restricting rotation
of the joint and straining the ADL. Everting the ankle prior to rotation ‘unlocks’ the subtalar
joint which allows the talus to rotate with the tibia, potentially limiting ADL strain and putting
the primary strain on the ATiFL. This study may improve our understanding of the various
mechanisms of ankle injury during external rotation of the foot.

41

CHAPTER 4:
ROTATIONAL STIFFNESS OF AMERICAN FOOTBALL SHOES AFFECTS ANKLE
BIOMECHANICS AND INJURY SEVERITY

ABSTRACT
While previous studies have investigated the effect of shoe-surface interaction on injury risk, few
studies have examined the effect of shoe rotational stiffness. The hypothesis of the current
study was that ankles externally rotated to failure in shoes with low rotational stiffness would
allow more talus eversion than those in shoes with a higher rotational stiffness, resulting in less
severe injury. Twelve (six pairs) cadaver lower extremities were externally rotated to gross
failure while positioned in 20 degrees of pre-eversion and 20 degrees of pre-dorsiflexion by
fixing the distal end of the foot, axially loading the proximal tibia, and internally rotating the
tibia. One ankle in each pair was constrained by an American football shoe with a stiff upper,
while the other was constrained by an American football shoe with a flexible upper.
Experimental bone motions were input into specimen-specific computational models to measure
ligament strains at failure. Ankles in flexible shoes failed at 97.4 ± 47.0 Nm which was greater
than the failure torque for ankles in stiff shoes (89.2 ± 44.9 Nm, p=0.049) in addition to allowing
6.7 ± 2.4 degrees of talus eversion during rotation, significantly greater than the 1.7 ± 1.0
degrees for ankles in stiff shoes (p=0.01).

The significantly greater eversion in flexible shoes

was potentially due to a more natural response of the ankle during rotation, possibly affecting the
injuries that were produced. All ankles failed by either medial ankle injury or syndesmotic
injury, or a combination of both. Complex (more than one ligament or bone) injuries were noted
in 4 of 6 ankles in stiff shoes and 1 of 6 ankles in flexible shoes. Models of ankles in flexible
shoes showed significantly less strain in the deltoid at failure (p=0.023), but significantly more

42

strain in three of four subtalar ligaments (p=0.025, 0.03, and 0.047). The current study suggests
flexibility (or rotational stiffness) of the shoe may play an important role in both the risk and
severity of ankle injuries for athletes.

INTRODUCTION
Ankle injury can result from a variety of loading situations and accounts for 10-30% of all sports
injuries (Waterman et al., 2011). In American football, ankle injuries can occur in contact and
noncontact situations, but frequently the mechanism of injury involves a foot planted on the
playing surface with internal rotation of the upper body (Williams et al., 2007). This external
rotation of the foot with respect to the body is known to result in both medial and syndesmotic
ankle injuries (Waterman et al., 2011), as well as acute bone fractures (Haraguchi and Armiger,
2009, Laugehansen, 1950), and makes up 10-15% of all ankle injuries (Gerber et al., 1998, Fallat
et al., 1998). While less common than other types of ankle injuries, syndesmotic and medial
ankle injuries are of particular interest to researchers due to their long recovery time (Boytim et
al., 1991), and the potential for chronic ankle dysfunction (Gerber et al., 1998, Waterman et al.,
2011, Wolfe et al., 2001).

Syndesmotic ankles sprains (also known as high ankle sprains) involve injury to the anterior
tibiofibular ligament (ATiFL). While most clinical literature attribute high ankle sprains to
external rotation (Norkus and Floyd, 2001), a recent study with cadavers has produced high
ankle sprain when the ankles are everted (rotated about the long axis of the foot (Figure 4-1))
prior to external rotation (Wei et al., 2012b). The same study indicates that when keeping the
ankle neutral in the coronal plane (no eversion), external rotation results in medial ankle injury,

43

involving injury to the anterior deltoid, also known as the anterior tibiotalar ligament (ATiTL).
Another study also indicates that a neutral foot undergoing external rotation will result in rupture
of the deltoid ligament with subsequent injury to the ATiFL and fibula [11]. These studies
suggest that pre-eversion of the talus is an important factor in the type of injury produced during
external rotation of the ankle.

Figure 4-1 Posterior view of a neutral (left) and everted foot (right).

While previous studies have examined the effects of running shoe structure on joint kinematics
(Peltz et al., 2014, Bonacci et al., 2013, Williams et al., 2012) as well as surface type (Meyers,
2010) and shoe-surface interface (Villwock et al., 2009a, Kuhlman et al., 2010) on injury
potential, few studies have investigated the effect of a shoe’s rotational stiffness on motion of the

44

talus and the subsequent type and risk of ankle injury. One recent study, however, investigated
the role of an American football shoe’s rotational stiffness on talus motion at sub-injurious levels
(Wei et al., 2012a). The study involved rotating cadaver ankles in either stiff or flexible shoes,
while tracking motion of the talus with respect to the tibia. The results show that ankles in
flexible shoes experience more eversion of the talus with respect to the tibia, but less external
rotation. In addition to this sub-failure study, another cadaver study (Haraguchi and Armiger,
2009), using a rigid foot constraint, shows that external rotation of the ankle results in deltoid
injury prior to ATiFL injury with fibular fracture. Other studies (Wei et al., 2011d) suggest that
more subtalar motion and arch collapse occur with a more flexible constraint of the ankle, thus
allowing a more ‘natural’ response of the foot during external rotation of the ankle. As
suggested by Lundberg et al., these ‘natural’ motions of the ankle bones may act as a torsion
dissipating devices to increase the ankle’s resistance to injury (Lundberg et al., 1989a).

Computational models of the ankle have previously been utilized to understand how different
boundary conditions affect the biomechanics and ligament strains in the ankle under external
rotation (Button et al., 2013, Wei et al., 2011d). Given the complicated manner in which the
various ankle structures develop strain in the three planes of motion (Lin et al., 2006), these
models have the advantage of quickly determining ligament strains and reaction forces during
solid-body motion of the ankle. By inputting motions from experimental studies, the simulations
can provide information about the foot’s ability to respond to a given motion with various
constraints. The current plan was to develop a model to help in the assessment of different
footwear during injurious scenarios.

45

The purpose of the current study was to utilize a setup similar to a previous study (Wei et al.,
2012a), but rotate the ankles to failure in order to investigate the effect of shoe stiffness on injury
risk and type of injury. In order to produce conditions that result in high ankle sprain, the ankles
were pre-dorsiflexed and pre-everted, per a previous study (Wei et al., 2012b). Additionally, the
bone motions were input into specimen-specific computational models in order to study ankle
ligament strains and compare these to experimental injuries. The hypothesis of the study was
that ankles in flexible shoes would allow more talus eversion during external rotation of the foot
than those in stiff shoes, thus allowing the subtalar ligaments to carry some of the applied torque
and result in a less severe ankle injury.

METHODS
Shoe Selection
Shoes were selected based on a previous experimental study (Wei et al., 2012a). Briefly, four
styles of American football shoes were tested in order to determine rotational stiffness using a
surrogate (rigid) ankle. The four shoes were subjected to a compressive pre-load of 1500 N, a
rotational pre-torque of 2 Nm, and a dynamic torque of 60 Nm input in load control at a
frequency of 1 Hz. The rotational stiffness of the shoes was defined as the slope of the loading
portion of the torque-rotation curve. The study found that the shoe with the highest rotational
stiffness (50.0 ± 1.7 Nm/deg) was the Nike Flyposite and the lowest rotational stiffness (21.9 ±
2.8 Nm/deg) was the Nike Air (Wei et al., 2012a) (Figure 4-2). These two styles of shoe were
selected for the current study.

46

Cadaver Tests
Twelve (six pairs) human cadaver (age 60 ± 6 years) lower-extremities were used in the current
study. The limbs were stored at -20°C and thawed to room temperature for 24 hours prior to
testing. The tibia and fibula were transected approximately 15 cm distal to the center of the
knee, and the proximal ends of the bones were removed of tissue and cleaned with 70% alcohol.
The proximal tibia and fibula were potted in an aluminum box with room-temperature curing
epoxy (Fiber Strand, Martin Senour Corp., Cleveland, OH). Two screws were placed in the
anterior and medial tibia before potting to help restrain the tibia in the potting material.

Figure 4-2 Football shoes used for the current study: Nike Flyposite (a) and Nike Zoom Air (b).

47

For each pair of lower extremities, one foot was randomly fit into the stiff shoe and the other
foot was fit into a flexible shoe and equal numbers of right and left feet were assigned to the
different shoe types. Shoes were selected to fit the size of each cadaver ankle. Regular-length
cotton socks were used on each ankle. The base of the shoe was held by an epoxy cleat mold
(Figure 4-3), which was attached to the testing fixture (Figure 4-4). Each shoe, with an attached
mold in an aluminum tray, was dorsiflexed 20 degrees and everted 20 degrees to replicate
conditions that previously produced high ankle sprains during external rotation of the foot (Wei
et al., 2012b).

Figure 4-3 Football cleat mold made of epoxy resin used to fix the shoe to the testing fixture

Experiments were performed with a customized hydraulic biaxial testing machine with a 244 Nm
rotary actuator (Model SS-001-1C, Micromatic, Berne, IN) and a vertically oriented linear
actuator (Model 204.52, MTS Corp., Eden Prairie, MN). The proximal (potted) end of the foot
48

was attached to the rotary actuator through a biaxial load cell (Model 1216CEW-2K, Interface,
Scottsdale, AZ) with a capacity of 8896 N of axial force and 113 Nm of torsion (Note: because
of the high resistive torque for specimen 6, another load cell (Model 6467-107, serial #229,
Lowell) with a capacity of 20kN of axial force and 1130 Nm of torsion was used). The distal
end of the extremity was then attached to the aluminum tray via the cleat mold. The tray was
attached to a plate that allowed for x-y adjustments to align the tibia with the linear and torsional
actuators. A 1500 N compressive preload was applied to simulate dynamic weight bearing prior
to rotation, as used in previous studies (Wei et al., 2012b). Internal tibia rotation (external foot
rotation) was input in position control at 1 Hz (0.5s to peak rotation). The first experiment was
to 20° rotation and subsequent experiments were repeated on the same ankle in increments of 10°
until the torque signal indicated an injury (Figure 4-5). At failure, tibia rotation and torque were
documented. Dissections of the ankle (by MCS) were performed after joint failure.

49

Figure 4-4 The experiments were performed on a biaxial testing machine with the motions of the
talus and tibia tracked by a 5-camera motion capture system.

Figure 4-5 An example of the torque data from a sub-failure (a) and a failure (b) test. The
failure occurs when there is a steep drop in torque, as indicated by the arrow.

50

Motion Analysis
Reflective marker arrays were screwed into the anteriomedial aspect of the talus and on the tibia
approximately 20 cm proximal to the articular surface. Care was taken not to damage any
ligaments during marker insertion. CT scans verified the placement of the marker arrays. A
Vicon motion capture system (Oxford Metrics Ltd, Oxford, UK) was used to document the
position of the markers during the experiments. A joint coordinate system (JCS), as defined by
the International Society of Biomechanics (ISB) (Wu et al., 2002), was established between the
talus and tibia to determine the talocrural motions (talus motion with respect to the tibia) up to
ankle failure. This involved aligning the local coordinate systems (based on the reflective
marker triad) on the talus and tibia (Figure 4-4) with their respective anatomical axes through
coordinate transformations. The JCS took one coordinate axis from each of these local
coordinate systems and a final floating coordinate axis perpendicular to the other two. By
calculating the dot product between the coordinate axes, as described by the ISB, the amount of
ankle flexion, eversion, and rotation was determined at joint failure.

Figure 4-6 Lateral (a), posterior (b), and medial (c) views of one of the specimen-specific
models in SolidWorks with linear springs representing ligaments.

51

Computational Modelling
Model development has been previously described in depth (Button et al., 2013). Briefly,
specimen-specific computational models were created based on CT scans (0.6 mm per slice) of
each ankle specimen. Digital Imaging and Communications in Medicine (DICOM) files of each
specimen were imported into Materialise’s Interactive Medical Imaging Control System
(MIMICS)(Materialise, Ann Arbor, MI), yielding three-dimensional surface models of the bones
as Stereolithography (STL) files. Motion data from the JCS analysis were used to drive each
specimen-specific computational model. Nineteen linear springs were used to model ankle
ligaments and nine linear springs were used to model the tibiofibular interosseous ligament
(Figure 4-6); with the initial length reduced 2% prior to motion in order to induce a small preload in each ligament to represent in situ pre-tensions (Liacouras and Waynel, 2007). Ligament
stiffness values were taken from the literature, as described by Wei et al. (Wei et al., 2011d).
Three-dimensional contacts were implemented between adjacent bones in order to prevent
overlap during simulation, with the parameters for the contact described previously by Button et
al. (Button et al., 2013). The forefoot was fixed and motors simulating rotation of the tibia and
talus, eversion of the talus, and the bone displacement as a result of axial pre-load were used to
drive each model. The calcaneus was fixed in the inferior-superior direction simulating contact
between the hind foot and the shoe. Strains were documented at the experimental point of gross
ankle failure.

Statistical Analysis
Paired t-tests were used to compare the differences between shoe constraints in failure torque,
talocrural eversion at failure, talocrural eversion per degree tibia rotation, talocrural rotation at

52

failure, talocrural rotation per degree of tibia rotation, and ligament strain. A one-way chi-square
test was used to test for a difference in complex injury rates in the two different shoe types. A
multiple regression analysis with categorical factors (subject, shoe) was used to analyze the
correlation between torque and eversion. P-values less than 0.05 were considered significant.

RESULTS
Experimental
Table 4-1 summarizes the injury mechanisms and failure metrics. The ankles in flexible shoes
failed at an average torque of 97.4 ± 47.0 Nm, significantly higher than the failure torque in stiff
shoes, 89.2 ± 44.9 Nm (p=0.049). While there was no difference in tibia failure rotation
(p=0.53) or talocrural failure rotation (p=0.51) between flexible and stiff shoes, there was
significantly more talocrural eversion at failure for ankles in flexible shoes (6.7 ± 2.4 versus 1.7
± 1.0, p=0.01).

53

Table 4-1 Summary of the injury mechanisms and failure metrics (* indicates significant difference between stiff and flexible shoes)

Specimen

Failure Torque
(Nm)

Tibia Failure
Rotation (deg)

Talocrural
Talocrural
Failure Rotation Failure Eversion
(deg)
(deg)

Injury Mechanism

Stiff

Flexible

Stiff

Flexible

Stiff

Flexible

Stiff

Flexible

1

85.6

83.4

44.7

46.5

19.3

17.5

3.5

3.9

2

75.8

93.4

37.0

39.4

27.9

21.5

0.9

7.4

3

44.9

51.0

44.6

31.1

28.4

18.5

0.8

5.2

Medial malleolus fracture

1) ATiFL avulsion
2) ATiTL avulsion

4

55.8

59.0

29.4

33.2

19.3

13.9

1.7

5.1

1) ATiFL rupture
2) ATiTL rupture

ATiTL rupture

5

102.4

117.9

37.4

43.0

13.9

15.1

2.0

8.2

ATiFL avulsion

Fibular head fracture

6

170.7

179.7

48.1

63.2

20.3

27.2

1.1

10.4

1) ATiFL rupture
2) ATiTL rupture

ATiTL rupture

Average

89.2

97.4

40.2

42.7

21.5

19.0

1.7

6.7*

Standard
Deviation

44.9

47.0

6.9

11.6

5.6

4.9

1.0

2.4

54

Stiff
1) Medial malleolus fracture
2) Spiral fracture of the fibula
1) ATiFL rupture
2 )ATiTL rupture

Flexible
ATiFL avulsion
ATiFL rupture

Table 4-2 Strains at failure rotation for the anterior tibiofibular ligament (ATiFL), anterior tibiotalar ligament (ATiTL), posterior
tibiotalar ligament (PTiTL), interosseous talocalcaneal ligament (ITaCL), lateral talocalcaneal ligament (LTaCL), medial
talocalcaneal ligament (MTaCL), and posterior talocalcaneal ligament (PTaCL). (Underlining represents ligaments that ruptured or
avulsed)

Percent Elgonation at Failure Rotation (%)
ATiFL

ATiTL

PTiTL

ITaCL

LTaCL

MTaCL

PTaCL

Specimen
Stiff

Flexible

Stiff

Flexible

Stiff

Flexible

Stiff

Flexible

Stiff

Flexible

Stiff

Flexible

Stiff

Flexible

1

19.9

21.2

18.9

14.6

14.9

17.2

11.4

14.3

16.0

16.7

0.1

2.9

9.1

12.0

2

35.5

34.4

37.2

28.3

1.8

10.6

3.0

18.8

1.8

18.1

1.9

2.0

1.9

17.4

3

19.1

21.1

34.6

20.7

10.3

12.3

3.8

1.3

1.9

11.1

11.9

1.7

1.9

1.8

4

18.3

7.3

26.8

20.0

5.5

8.0

1.9

11.7

1.9

1.4

2.0

17.3

7.2

18.1

5

33.4

32.8

4.2

2.0

17.4

19.8

3.7

10.8

20.0

31.1

2.7

1.8

18.3

22.8

6

29.8

15.9

32.9

31.6

7.7

23.3

1.9

7.5

2.0

15.3

20.9

30.7

9.0

12.5

Average

26.0

22.1

25.7

19.5

9.6

15.2

3.3

10.7

7.3

15.6

6.6

9.4

7.9

14.1

Standard
Deviation

8.5

10.9

13.3

9.8

6.4

4.8

4.4

6.4

8.9

10.8

4.7

16.2

6.8

8.0

55

Additionally, the amount of talocrural eversion per degree tibia rotation in the flexible shoe (0.16
± 0.04 degree/degree) was significantly greater than that for the stiff shoe (0.04 ± 0.02
degree/degree (p=0.023)). There was slightly more talocrural rotation per degree of tibia rotation
in the stiff shoe versus the flexible shoe (0.55 ± 0.16 degree/degree and 0.45 ± 0.10
degree/degree, respectively), although this was not quite significant (p=0.082). There was no
correlation between torque and eversion at failure (p=0.189).

The injury mechanisms also differed between the two types of shoes (Figure 4-7 and Table 4-1).
Medial injuries were observed in 5 of 6 stiff-shoe ankles and 3 of 6 flexible-shoe ankles.
Syndesmotic injuries were observed in 5 of 6 stiff-shoe ankles and 4 of 6 flexible-shoe ankles.
When the specimens were compared with a chi-square analysis, the effect of shoe type on the
frequency of complex injuries was not quite significant (p=0.079): there were combination
(complex) medial and syndesmotic injuries in 4 of 6 stiff-shoe ankles while only 1 of 6 ankles in
the flexible shoes suffered combination injuries. Bone injuries were also more prevalent in stiffshoe ankles, occurring three times compared with only one occurrence in flexible-shoe ankles.

Computational Modelling
Results of the SolidWorks simulations are shown in Table 4-2 for the seven ligaments that
developed the highest levels of strain during external rotation of the ankle: the ATiFL, ATiTL,
posterior tibiotalar ligament (PTiTL), interosseous talocalcaneal ligament (ITaCL), lateral
talocalcaneal ligament (LTaCL), medial talocalcaneal ligament (MTaCL), and posterior
talocalcaneal ligament (PTaCL). Models of the stiff and flexible shoe cases showed no
difference in ligament strains for the ATiFL (p=0.221) and MTaCL (p=0.264) at ankle failure.

56

Figure 4-7 Fibular fracture (a), medial malleolus fracture (b), ATiFL avulsion (c), ATiTL
avulsion (d), ATiFL tear (e), and ATiTL tear (f).

57

However, the stiff shoe models showed significantly higher levels of strain in the ATiTL (25.7 ±
13.3% versus 19.5 ± 9.8% in the flexible models, p=0.023). Stiff shoe models also showed
significantly lower levels of strain in the ITaCL (3.3 ± 4.4% versus 10.7 ± 6.4%), LTaCL (7.3 ±
8.9% versus 15.6 ± 10.8%), and PTaCL (7.9 ± 6.8% versus 14.1 ± 8.0%) at ankle failure,
(p=0.035, 0.03, and 0.047, respectively). Additionally, there was a decrease in PTiFL strain in
the stiff versus flexible shoe models (9.6 ± 6.4% versus 15.2 ± 4.8%), but not significantly
(p=0.056).

Figure 4-8 Representative plot of the ATiFL and ATiTL strains versus tibia rotation in a
simulation of the combination ATiFL and ATiTL injury scenario (specimen 2, stiff shoe).

58

In all four cases of complex ligamentous injury, the combination of pre-dorsiflexion, preeversion, and axial load pre-strained the ATiFL more than the ATiTL. This resulted in higher
ATiFL strains for low levels of rotation, but an increasing ATiTL strain at near-failure levels of
rotation, while the ATiFL strain leveled off. In models 2, 4, and 6, the ATiTL strain level
surpassed the ATiFL strain in the last 3-5° of rotation before failure. In the other case
(specimen 3 flexible) the ATiTL strain nearly equaled the ATiFL strain, but did not surpass it. A
representative plot of simulations that showed combined ATiFL/ATiTL failure is shown in
Figure 4-8 (specimen 2, stiff shoe).

DISCUSSION
The rotational stiffness of American football shoes may affect high and medial ankle injury risks.
The current study investigated this premise by rotating pairs of cadaver ankles to failure in either
a stiff or flexible shoe. All ankles showed high or medial ankle injury, or a combination of each.
The ankles in flexible shoes failed at statistically higher torques and at significantly higher levels
of talocrural eversion than the ankles in stiff shoes. Computational models simulating flexible
shoe experiments exhibited statistically higher strains in subtalar ligaments during experimental
rotations, but lower strains in the ATiTL. Additionally, there was a higher frequency of complex
(combination) injuries in the stiff shoes, but not quite at a significant level. These results
indicate that shoes with a flexible upper may result in less severe ankle injuries than those with a
stiff upper.

59

The increased ankle eversion in the flexible shoe may be due to less restriction of the subtalar
joint to allow more motion between the talus and calcaneus, which may more closely simulate
the ankle’s natural, unrestrained response to external rotation. Previous investigations on
anatomical specimens have concluded that external rotation of the unrestrained ankle results in
pronation (the combination of eversion, abduction, and dorsiflexion) of the ankle (Lundberg et
al., 1989b). Rotations about the anteroposterior and vertical axes of the talonavicular joint have
been shown larger than the external rotation input about the tibia (Vanlangelaan, 1983). This is
described as the natural ‘transferal of rotation’ (Lundberg et al., 1989b). In the current study, a
lower talus constraint may have allowed the ankles in flexible shoes to undergo this ‘transferal of
rotation’ resulting in significantly more eversion than in a more stiff constraining type of shoe.
This natural sequence of rotations may help the ankle absorb applied rotational torques, offering
a possible explanation as to why ankles in flexible shoes failed at slightly higher torques, but
similar levels of shoe rotation.

The modelling aspect of this paper provided insight into what occurs at a ligamentous level in the
stiff versus flexibly constrained ankle. In the stiff shoe simulations, the subtalar joint remained
relatively closed during external rotation and as a result the applied torque primarily stretched the
ATiTL and ATiFL, with the ATiTL straining significantly more than in the flexible shoe
simulations (Table 4-2). However, the flexible shoe constraint allowed the subtalar ligaments to
deform as well, potentially increasing the amount of load the ankle could withstand before gross
failure. While an applied torque to the ankle in a stiff shoe resulted in an almost pure talocrural
rotation with high ATiTL and ATiFL strains, an applied torque to the ankle in a flexible shoe
allowed a combination of ankle rotation and eversion. This combination of rotation and eversion

60

strained more ligaments, but at more moderate, sub-failure levels. This difference in talus
response and corresponding ligament strain may explain the different injury mechanisms in the
stiff versus flexible shoes.

To help explain some of these injuries, other studies have documented the order and mechanisms
of pronation-external rotation injuries (Norkus and Floyd, 2001). The combination of medial
and syndesmotic injury (specimens 1, 2, 4, and 6 in the stiff shoe and specimen 3 in the flexible
shoe) closely follow the mechanism that is described in a Weber C injury (Harper, 1992). In
these injuries, the medial side of the ankle is in maximum tension due to the combination of
eversion and external rotation, first causing rupture of the ATiTL, or fracture of the medial
malleolus. The injury may stop here, possibly explaining the medial malleolus fracture in
specimen 3 with the stiff shoe and isolated ATiTL failure in specimens 4 and 6 with the flexible
shoe. Once the ATiTL ruptures, the talus may continue to externally rotate and move laterally
causing rupture of the ATiFL. This mechanism is supported by the computational study (Figure
4-8), as the ATiTL strain rate increases more rapidly and overtakes the ATiFL strain level near
the point of failure. This supports the claim that the mechanism of injury may follow that of a
Weber C classification: the ATiTL fails first, causing lateral translation of the talus and a
subsequent increase in ATiFL strain to cause failure of the ATiFL.

Interestingly, 4 of the 5 combination medial/syndesmotic injuries occurred in stiff-shoe ankles,
possibly suggesting that restricting motion of the talus may also lead to a more severe ankle
injury. This may be supported by other studies: Haraguchi, performing eversion-external
rotation tests on ankles with a rigid constraint, noted combination injuries in 19 of 22 specimens;

61

while Wei, performing eversion-external rotation tests on ankles with a compliant constraint,
documented isolated injuries in all six specimens (Wei et al., 2012b). Isolated ATiFL tear or
avulsion (specimens 1 and 2 in the flexible shoe and specimen 5 in the stiff shoe) most likely
occurred as described by Wei et al. (Wei et al., 2012b): the combination of pre-eversion, predorsiflexion, and axial load causing pre-tension in the ATiFL, making it the first ligament to
rupture under external rotation. This mechanism of injury was also supported by the
computational study, as specimens with ATiFL failure showed the highest strain in the ATiFL
(Table 4-2).

One limitation of the current study was the use of older-aged cadavers. Fractures have been
shown to occur more in patients with osteoporotic bone, most likely the result of advanced age
(Haraguchi and Armiger, 2009). The occurrence of osseous injury instead of ligamentous injury
could be related to the specimen’s age and bone quality. However, these osseous injuries did
not seem to bias the results. Three specimens (1, 3, and 5) suffered osseous injuries and in each
case these injuries were prevalent in both the stiff and flexible shoe ankles. However, there were
more fractures in the stiff shoe ankles and more avulsions in the flexible shoe ankles, potentially
indicating that more rigidly constraining ankles by a stiff shoe upper may stress the bones more
during external rotation of the ankle. This may be supported by a recent study analyzing injury
videos posted on Youtube.com where a fracture pattern similar to that observed from the current
study was noted under pronation-external rotation (Kwon et al., 2010), suggesting that bone
fracture may be a common occurrence in young athletes as well. Another limitation was the
limited number of shoes tested, as well as the lack of documented technical information about
the shoes. A recent study has compared the rotational characteristics of the two shoes from the

62

current studies with other shoes (Wei et al., 2012a), but future studies may want to look at more
shoes (it is important to note that a constant 1500 N axial load was applied to a surrogate ankle
during this stiffness testing, and different axial loads may result in different rotational
stiffnesses). Furthermore, technical characteristics of the shoes, besides rotational stiffness,
were not documented, and these could have an impact on talocrural joint motion and subsequent
injuries that were generated in these experiments. Future studies may want to document
additional differences in shoe construction that may have an effect on talocrural joint motion.

Additionally, the effect of muscle action was not included, which, on top of the specimen age,
may have been a limitation when attempting to reproduce the physiological behavior during an
actual sports injury. Yet, while it has been previously suggested that human subjects generate
10-20% more ankle joint torque than cadavers during sub-injurious external rotation (Wei et al.,
2011c), the magnitude of torque at failure was of little importance to the current study.
Furthermore, the proprioceptive muscle response is often too slow to have an effect on joint
motion during injury (Dubin et al., 2011, Delahunt et al., 2006), so it is likely that biomechanics
of the joint would differ only slightly during in vivo loading. Lastly, the small number of
specimens has an effect on reducing the statistical power of the study. Increasing the sample size
may have improved the robustness of the conclusions: more noted differences could possibly be
seen in talocrural rotation/degree of tibia rotation as well as differences in complex injury
frequency. However, the sample size was adequate to see differences in eversion, as well as key
ligament strains at failure.

63

In conclusion, the current study investigated ankle injury risk and its severity under combined
dorsiflexion, eversion, and external rotation while in either an American football shoe with a
relatively stiff or flexible upper design. Results of the study indicated that more talus eversion
occurred in flexible shoes under external rotation of the shoe, potentially allowing for a more
‘natural’ response of the ankle to rotation. As a result, the ankles in flexible shoes required
slightly more torque to fail and had fewer cases of complex ankle injury. Computational models
of the experiments showed that the response of the ankle to external rotation in a more flexible
shoe strained more ligaments than in a stiff shoe, but at more moderate levels. However, as
conclusions from this study only apply to external rotation of the ankle future studies may want
to investigate the performance characteristics of stiff versus flexible shoes in a variety of
circumstances in order to best recommend footwear that optimizes injury mitigation. In
addition, future studies may want to investigate the effect of shoe stiffness on joint kinematics in
regards to performance as well as injury, as previously studied with soccer shoes (Hennig, 2011).
The current study may help drive research and design of future American football shoes in order
to help limit injury risk without compromising the athlete’s comfort and performance.

64

CHAPTER 5:
THE EFFECT OF ROTATIONAL STIFFNESS ON ANKLE JOINT MOTION AND
LIGAMENT STRAIN DURING EXTERNAL ROTATION
ABSTRACT
The rotational stiffness of footwear has been previously shown to have an effect on ankle joint
kinematics and injury risk, but this relationship has not yet been quantified. The purpose of the
current study was to derive equations that could be used to predict ankle joint kinematics based
on the amount of ankle constraint. Three athletic tapes were tested for their ability to constrain
the ankle joint during external rotation. Six subjects then performed a voluntary external foot
rotation using the selected tape designs to constrain the ankle, as well as an unconstrained case.
Motion capture was performed to track the motion of the hindfoot with respect to the tibia (ankle
joint) up to fifteen degrees of tibia rotation and predictive equations were determined to establish
ankle joint rotation, eversion, and flexion as a function of gross tibia motion and tape stiffness.
These predictive equations were then used to drive a computational model in which ankle
ligament strains were determined at 15 degrees of tibia rotation and for ankle constraint stiffness
ranging from 0 Nm/deg to 30 Nm/deg. The three tapes provided significantly different
constraint stiffness during external foot rotation. There was no statistical effect of ankle joint
constraint on the dorsiflexion response of the ankle (p=0.461). In contrast, there was an effect of
constraint stiffness on ankle joint external rotation (p<0.001) and ankle joint eversion (p<0.001).
Results of the model simulation revealed the highest ligament strains in the anterior tibiotalar
ligament (ATiTL) and anterior tibiofibular ligament (ATiFL). ATiTL strain increased with
increasing constraint stiffness, while there was little effect of constraint stiffness on ATiFL
strain. Results from this study could aid in the design of footwear, as well as the analysis of
clinical injuries.

65

INTRODUCTION
External rotation of the foot with respect to the body is known to result in both medial and
syndesmotic ankle sprains and fractures (Waterman et al., 2011, Boytim et al., 1991, Haraguchi
and Armiger, 2009). Medial ankle injury involves damage to the anterior tibiotalar ligament
(ATiTL) (Wolfe et al., 2001), while syndesmotic (or high) ankle injury involves damage
primarily to the anterior tibiofibular ligament (ATiFL) (Dattani et al., 2008). These injuries are
most prevalent in contact sports where the injury mechanism is frequently attributed to a foot
being planted on the playing surface with internal rotation of the upper body (Guise, 1976).
While both injuries have been shown to involve some combination of external rotation,
dorsiflexion, and eversion (Waterman et al., 2010, Gerber et al., 1998, Norkus and Floyd, 2001,
Wei et al., 2012b), understanding more precisely the joint kinematics at the point of failure is an
important aspect for preventing injury.

Previous researchers have examined joint motion during medial and high ankle sprains using a
variety of techniques. One study used a rigid-body ankle model to investigate the effect of
different levels of joint eversion and dorsiflexion on ATiFL and ATiTL injury risk (Wei et al.,
2011a). Conclusions from this computational study motivated a cadaver study in which pairs of
ankles were rotated to failure in either an everted or neutral configuration, producing
syndesmotic and medial ankle sprains, respectively (Wei et al., 2012b). Results of both studies
concluded that the amount of eversion is critical in producing a syndesmotic versus a medial
ankle sprain. Yet, to provide a more complete understanding of external rotation injuries
occurring on the field, an in depth analysis of in vivo injuries may be warranted.

66

Krosshaug and colleagues have developed a model-based image matching (MBIM) motion
analysis technique to analyze three-dimensional human motion in an un-calibrated environment
using panning cameras (Krosshaug and Bahr, 2005). This technique was used to investigate two
cases of inversion ankle injury from the 2008 Beijing Olympics (Mok et al., 2011) and five cases
of inversion sprains in tennis (Fong et al., 2012). Results from these studies have provided new
information regarding the mechanism of lateral ankle sprain and have led to recommendations
for prevention of such sprains. However, accurate estimations of actual foot motion inside shoes
remains challenging.

In external rotation injuries motion of the ankle joint has been considered to play an important
role in generating ligament strains, particularly to the ATiTL and ATiFL (Sarsam and Hughes,
1988, Wei et al., 2011d). Therefore, in order to use MBIM to study external rotation injuries the
analysis would need to accurately quantify ankle joint motion. Previous studies from this lab,
however, have shown that ankle joint motion during upper body (tibial) rotation and the
subsequently generated injuries are highly dependent on rotational stiffness of the shoe (Wei et
al., 2012a). In order to analyze injurious scenarios or model the effect of different shoe
rotational stiffness, the relationship between constraint stiffness and ankle joint motion needs to
be investigated in more detail.

The purposes of the current study were (1) to have human subjects perform ankle external
rotation while constraining the foot with athletic tapes of different rotational stiffness (as a means
of determining the potential role of shoe stiffness on ankle motion) (2) to develop equations
relating the effect of constraint stiffness on 3D ankle motion, and finally (3) to use these

67

equations to drive a computational model to estimate the range of ligament strains that are
generated over a domain of constraint stiffness representative of various levels of shoe rotational
stiffness. It was hypothesized that taped constraint of the ankle would have a significant effect
on the amount of ankle joint motion during external rotation of the foot with respect to the upper
body, and that the effect on ankle joint kinematics would generate differences in key ligament
strains during external foot rotation. Information from this study may help investigators better
estimate ankle motion within various shoe designs and improve theoretical predictions of the
potential for ankle injuries during field conditions in order to optimize athletic footwear design
for preventing external rotation injuries of the ankle.

68

Figure 5-1 The surrogate ankle taped to the polycarbonate plate using Elastikon (a), PowerFlex
(b), and Sher-Light (c).

METHODS
Tape stiffness tests

Three athletic tapes were evaluated to determine their rotational stiffness properties for the ankle.
The experiments were conducted on a custom, hydraulic, biaxial testing machine with a 244 Nm
rotary actuator (Model SS-001-1V, Micromatic, Berne, MN) and a vertically oriented linear

69

actuator (Model 204.53, MTS Corp., Eden Praire, MN). The tapes evaluated in the study were
Andover PowerFlex, Covidien Sher-Light, and Johnson & Johnson Elasikon. The tapes
constrained a surrogate lower extremity made from room temperature curing epoxy resin (Fiber
Strand, Martin Senior Corp., Cleveland, OH) to a polycarbonate plate (32 cm x 8 cm) (Figure 51). The taping was performed by an athletic training student (PT) with the goal of modelling the
constraint provided by a football shoe upper by applying two medial and two lateral heel locks.
The polycarbonate plate was clamped to a customized fixture that allowed x-y adjustments to
align the surrogate shank with the linear and torsional actuators. The proximal end of the
surrogate limb was inserted into an aluminum box which was attached to a rotary actuator
through a biaxial load cell (Model 1216CEW-2K, Interface, Scottsdale, AZ) with a capacity of
8896 N axial force and 113 Nm torsion (Figure 5-2).

Figure 5-2 The surrogate limb was attached to the testing machine

70

The axial load was zeroed and a rotational pre-torque of 2 Nm was applied to the surrogate foot
prior to internal rotation of the shank (external ankle rotation). The pre-torque was applied to
ensure tape contact on the medial side of the foot. A dynamic torque of 50 Nm was input in load
control at a frequency of 1 Hz (0.5 s to peak torque) and repeated six times for each tape. The
loading portion of the torque-rotation curve was averaged across the final five cycles of loading.
The slope of the linear portion of the curve (between 10 Nm and 50 Nm) was determined to yield
a rotational stiffness for each tape in Nm/deg.

This procedure was repeated, but with the surrogate ankle constrained by shoe/cleat molds
instead of tapes, as previously described by Wei et al. (Wei et al., 2012a). The shoes tested were
the Nike Flyposite and Nike Air, the stiffest and most flexible shoes from a previous study,
respectively (Button et al., 2015). In order to accurately mimic the tape stiffness tests, the axial
load was zeroed prior to rotation, unlike the Wei study which applied a 1500N axial pre-load
load to simulate body weight. The procedure from the Wei study was deemed ineffective for the
tape study because of potential high levels of torsional friction between the surrogate ankle and
the polycarbonate plate.

Human subject tests

Institutional Review Board approval was obtained for these experiments. Kinematic data from
six young (age: 27.3 ± 4.6 years), male subjects (body weight: 77.5 ± 8.9 kg) with no history of
lower body injury were collected after obtaining informed consent. A six-camera Vicon MX
Motion Capture System (OMG plc. Oxford, UK) was used to capture motion data at 100 Hz.
Marker arrays were placed on the anterior portion of the tibia, approximately 10 cm distal to the

71

patella, and on the calcaneal tuberosity (Figure 5-3). Each subject’s right foot was measured
prior to testing and a polycarbonate plate was cut to fit. The subject’s right foot was taped to the
1/8” thick polycarbonate plate by the athletic trainer (PT) using two medial and two lateral heel
locks and one of three athletic tapes for each experiment. The foot/plate was adhered to the
ground using a Dycem adhesive sheet (Dycem Ltd, Warwick, RI, USA). The left foot was placed
adjacent to the right foot on a raised surface so the feet were level. The subject performed an
internal rotation of the body (external rotation of the foot) at a self-determined maximal effort,
trying to minimize tibia flexion and eversion. Each subject repeated this motion until 10
successful trials were collected. The procedure was repeated for each tape and under a tapeless
condition (note: the shoes were not used in the human subject tests).

Figure 5-3 Marker arrays were placed on the subject’s anterior tibia, about 10 cm distal to the
patella, and the calcaneal tuberosity.

72

Coordinate transformations were performed on the tibia and hindfoot marker arrays in order to
establish orthogonal coordinate systems in line with the three axes of the ankle (Figure 5-4). A
joint coordinate system (JCS) was then established between the tibia and hindfoot (henceforth
referred to as the ankle joint), similar to a previously defined system (Wu et al., 2002):
=

(1a)

=

(1b)

=|

×
×

(1c)

|

=

∙

(flexion)

(2a)

=

∙

(rotation)

(2b)

=

∙

(inversion/eversion)

(2c)

Calculations of absolute tibia rotation, eversion, and flexion with respect to the laboratory were
performed by the same method, but the JCS was established between the tibia coordinate system
and a fixed laboratory coordinate system aligned with the axes shown in Figure 5-4.
Calculations for ankle joint flexion, rotation, and inversion/eversion were determined over a
range of 0 to 15 degrees of tibia rotation.

73

Figure 5-4 Three orthogonal axes of the ankle. Rotations about these axes are described as
plantarflexion/dorsiflexion, eversion/inversion, and internal/external rotation.

Predictions of ankle joint dorsiflexion, eversion, and rotation were determined as a function of
tibia rotation, tibia eversion, tibia dorsiflexion, and tape stiffness. Irrelevant predictors and
interaction terms were eliminated by using the ‘stepwisefit’ and ‘LinearModel.fit’ functions,
respectively, in MATLAB.

Computational modelling

Predicted ankle joint motion data from the curve fitting equations were used to drive a generic
computational ankle model. Details of the model development and its validation have been
described in previous studies (Wei et al., 2011d, Button et al., 2013), so they will only be
described briefly here. The ankle model was based on a computed tomography (CT) scan of a
fresh frozen, right male cadaver foot (52 years old) showing no signs of abnormal anatomy.
Digital information from the scan was transferred from Digital Imaging and Communications in

74

Medicine (DICOM) files into Materialise’s Interactive Imaging Control System (MIMICS,
Leuven, Belgium) software. Three dimensional surface models of the bones, as
Stereolithography (STL) files, were created and exported to SolidWorks 2011 (Trimech
Solutions, LLC, Columbia, MD), where the bones were simplified and assembled. SolidWorks
Motion was used to add ligaments (modelled as linear springs) and insertion locations were
determined from previous dissections in this laboratory (KB and FW) and an anatomical atlas
(Netter and JT, 2003). A pre-load was induced in each ligament by reducing its length by 2%,
inducing in situ preloads (Liacouras and Waynel, 2007). The model has been previously
validated based on ankle ligament strain (Colville et al., 1990a), ankle joint torque (Wei et al.,
2010), talus rotation (Button et al., 2013), and injuries generated during external rotation
experiments (Button et al., 2013).

The motion simulation was run in displacement control with motors driving the threedimensional rotations of the tibia and talus. The input for the tibia motor was a ramp to 15
degrees of external rotation, with no eversion or dorsiflexion (modelling an isolated external
rotation). The talus rotation, eversion, and dorsiflexion (determined from results of ‘Curve
Fitting’ section) were dependent on the amount of tibia rotation at a given time point and the
rotational stiffness parameter. The forefoot was fixed while the calcaneus, talus, tibia, and fibula
were free to move. The body weight was simulated as an axial load on the proximal end of the
model with one-sixth of the loading on the fibula and the rest on the tibia (Wei et al., 2011d).
Ligament strains were determined during rotation.

75

Statistical analysis

One-way ANOVA and Tukey post-hoc tests were used to statistically determine differences in
rotational stiffness between tapes/shoes. Two-way repeated measures ANOVA and Tukey posthoc tests were used to statistically determine differences in ankle joint dorsiflexion, external
rotation, and eversion for the three tape (and no tape) designs (factor one) at each rotation level
(factor two). In all statistical tests, p-values less than 0.05 were considered significant.

RESULTS
The 2 Nm pre-load was zeroed out by shifting the torque-rotation curves, which showed nearly
linear behavior beyond 10 Nm of input torque (Figure 5-5a and b). The rotational stiffness of the
Elastikon, Sher-Light, and PowerFlex tapes were 25.7 ± 1.6 Nm/deg, 22.6 ± 0.75 Nm/deg, and
13.5 ± 0.37 Nm/deg, respectively. The rotational stiffness of the Air and Flyposite shoes were
16.5 ± 0.44 Nm/deg and 25.8 ± 1.1 Nm/deg, respectively. There were significant differences
between all the tape/shoe designs, except between the Flyposite shoe and Elastikon tape
(p=0.99).

76

a

b

Figure 5-5 Torque-rotation curves for the three tape designs and two shoe designs from
rotational stiffness tests (a), and rotational stiffness values determined from the slopes of the
torque-rotation curves (b). Curves with different symbols (* # ƚ ¥) indicate statistically different
stiffness between designs.

77

Figure 5-6 Mean ankle joint dorsiflexion-tibia rotation curves for the three tape (and no-tape)
designs from an in vivo external foot rotation (n=6). There were no significant differences
between tape designs.

Figure 5-7 Mean ankle joint external rotation-tibia rotation curves for the three tape (and notape) designs from an in vivo external foot rotation (n=6). Different symbols (* # ƚ ǂ) indicate
statistically different curves.

78

Figure 5-8 Mean ankle joint eversion-tibia rotation curves for the three tape (and no-tape)
designs from an in vivo external foot rotation (n=6). Different symbols (* #) indicate
statistically different curves.

There were no differences in ankle joint dorsiflexion response between the three tapes or the
tapeless condition (p=0.461) (Figure 5-6). There was, however, an overall difference in ankle
joint external rotation response (p<0.001) for the four conditions. Individually, the PowerFlex
external rotation response differed from the Sher-Light response (p=0.0001), and all other curves
differed from one other (p<0.0001) (Figure 5-7). There was also an overall difference in ankle
joint eversion response (p<0.001). Individually, the Elastikon eversion response was
significantly different from the PowerFlex and tapeless responses (p<0.0001), but not from the
Sher-Light response (p=0.659). The PowerFlex response was significantly different than the
Sher-Light response (p<0.0001), but not the tapeless response (p=0.999). The Sher-Light and
tapeless curves were also significantly different (p<0.0001) (Figure 5-8).

79

Predictive equations for ankle joint dorsiflexion (AJDorsi), ankle joint rotation (AJRot), and
ankle joint eversion (AJever) as a function of tibia dorsiflexion (TIdorsi), tibia rotation (TIrot),
tibia eversion (TIever), and tape stiffness (Stiff) are shown in equations 3a-c with correlation
coefficients 0.99, 0.99, and 0.95, respectively.

AJdorsi = 0.0343 + 0.6794 ∗ ./01

(3a)

AJrot = 0.29551 + 0.6804 ∗ .107 + 87 99 0.0089 ∗ .107 − 0.0099

(3b)

AJever = 0.11525 + 0.4209 ∗ . = 1 + 0.08371 ∗ .107 − 87 99 0.0024 ∗ .107 + 0.0007 ∗
87 99 + 0.0002

(3c)

External rotation of a neutral ankle strained the ATiFL and ATiTL the most. The effects of
ankle constraint stiffness on the ATiTL strain were greater than those on the ATiFL strain
(Figure 5-9a and b). Fifteen degrees of tibial rotation strained the ATiFL 15.5-16.2%, while the
ATiTL strain was between 6.7% and 14.5%. At 15 deg of tibial rotation, each increment of 5
Nm/deg in stiffness theoretically increased ATiTL strain approximately 1.2%.

80

a

b

Figure 5-9 Predicted anterior tibiofibular ligament (ATiFL) (a) and anterior tibiotalar ligament
(ATiTL) (b) strain versus tibia rotation for an ankle constrained with tape stiffness ranging from
0 to 30 Nm/deg. Inputs for ankle joint motion were based on predictive equations 3a-3c.

81

DISCUSSION
The current study investigated the effect of ankle constraint on the three-dimensional motion of
the ankle using different athletic tapes. Three tapes (and a no-tape condition) were used to
constrain human ankles during an in vivo voluntary external ankle rotation. The rotational
stiffness of the tapes was comparable to those of a relatively stiff and a compliant football shoe
design. Predictive equations for determining ankle joint motion as a function of tibia motion and
tape stiffness were developed from human subject testing results. These predictive equations
were used to drive a generic computational ankle model for the calculation of ligament strain
during isolated external ankle rotation. Results of the experimental study indicated that for a
given tibia rotation the amount of ankle joint eversion and external rotation was dependent on the
level of ankle constraint, with a stiffer ankle constraint allowing greater ankle rotation, but less
eversion. Constraint on the ankle also affected theoretical ligament strain: increasing constraint
stiffness substantially increased ATiTL strain, but had little effect on ATiFL strain. These
results supported the hypothesis that constraint stiffness would have a significant effect on ankle
joint eversion and rotation during foot external rotation.

Previous investigations on anatomical specimens conclude that external rotation of an
unconstrained ankle results in a combination of eversion and rotation (Lundberg et al., 1989b).
This potentially explains the increased eversion measured in ankles constrained with a more
flexible tape, likely because the applied rotation can be more easily transferred about other axes
of the ankle. In particular, less constraint of the subtalar and transverse tarsal joints may allow
more motion between the talus and calcaneus, generating foot eversion (Hertel et al., 1999, Wei
et al., 2012a). With increased tape stiffness, however, this natural response was restricted,

82

resulting in nearly isolated ankle external rotation. Additionally, while eversion and external
rotation are strongly coupled, they are only weakly coupled with dorsiflexion (Lundberg et al.,
1989b), potentially explaining why ankle dorsiflexion was only a function of absolute tibia
dorsiflexion (Equation 3a). The apparent correlation between tibia rotation and ankle joint
dorsiflexion noted in Figure 5-6 was most likely due to an increased tibia flexion that occurred as
a result of loading the leg.

Foot external rotation primarily strained the ATiTL and ATiFL in the computational model.
Figure 5-9 shows that axial load put an ~8% pre-strain on the ATiFL, likely due to the
tibiofibular syndesmosis opening slightly around the talus upon contact. External rotation put
additional strain on the ligament, and constraint stiffness had little effect on strain magnitude.
However, increasing constraint stiffness correlated with considerable increases in ATiTL strain.
Unlike the ATiFL, the ATiTL strain did not increase with axial load, but increased linearly with
tibia rotation once the ligament began to tense around 5-7° of tibia rotation. The observation of
increased ATiTL strain with increasing rotational ankle constraint suggested that a stiffer
constraint may predispose the ankle to medial ankle injury during external rotation. This
potentially explains the results from a previous cadaver study in which pairs of cadaver ankles
were rotated to failure in either a stiff (Flyposite) or flexible (Air) football shoe. A high
frequency of combination medial and syndesmotic ankle injuries were noted in the stiff shoes,
while ankles in flexible shoes mostly suffered islolated syndesmotic injuries (Button et al.,
2015).

83

The information regarding the effect of ankle constraint on ankle joint kinematics and injury risk
may benefit the future design of athletic shoes. Previous studies have examined the effects of
running shoe structure (Bonacci et al., 2013, Peltz et al., 2014, Williams et al., 2012) and football
shoes (Wei et al., 2012a, Button et al., 2015), as well as the effects of shoe-surface interface
(Villwock et al., 2009a) on joint kinematics and injury risk, but no study has quantified and
modelled these effects. Two shoe designs from a previous cadaver study (Wei et al., 2012a,
Button et al., 2015) were subjected to rotational stiffness testing in order to assess whether or not
the constraint offered by the athletic tapes was a realistic simulation of athletic footwear (Figure
5-5). It was noted that the stiffer shoe (Nike Flyposite) behaved almost identically to the stiffest
tape (Elastikon), while the more flexible shoe (Nike Air) behaved between the moderate (SherLight) and most flexible tape (PowerFlex). This indicated that the tapes used in this study
reasonably matched stiff and flexible shoe designs used in previously studies by this laboratory
during external rotation, but future studies may want to constrain in vivo or cadaver ankles with
different shoes instead of tapes to further investigate this effect.

Predictive equations derived in the current study may also help in the analysis of injury
mechanisms during athletic competition. Using Krosshaug’s MBIM technique to analyze leg
and shoe kinematics in game films, Fong et al. (Fong et al., 2012) and Mok et al. (Mok et al.,
2011) documented wide variations in the amount of joint motion at ligament failure, with peak
inversion values ranging from 48o to 148°. However, much of this disparity could be due to
different constraints on the ankle joint in different footwear. Predictive equations from the
current study exhibited that 15 degrees of isolated tibia rotation could result in 10.5o-14.2° of
ankle joint rotation and 6.7-14.2% ATiTL strain, depending on ankle constraint. Future studies

84

may want to not only use MBIM to investigate the kinematics of the leg with respect to the shoe
during external rotation, but complement this information with the known rotational shoe
stiffness-ankle motion relationships to more accurately assess ankle joint kinematics inside the
shoe. Such information, combined with cadaver studies and computational modeling, could help
provide researchers with a more complete understanding of ankle kinematics within the shoes
during high and medial ankle sprains.
While the current study utilized human subjects, future studies may benefit from studying joint
motion under various constraints using cadaver specimens. While in vivo studies rely largely on
voluntary motions, using cadaver specimens would allow more precise control of input motions.
Reflective markers could also be screwed into the talus, calcaneus, and tibia. Marker arrays on
both the talus and calcaneus would enable the calculation of both subtalar and talocrural joint
kinematics. Additionally, using screw-in marker sets instead of skin marker sets could
potentially mitigate any errors associated with skin motion. Furthermore, a limitation of the
current study was that the analysis was restricted to sub-failure rotations. Using cadavers could
allow researchers to measure ankle kinematics and ligament behavior at failure-level foot
rotations, which may differ from sub-failure behavior. Comparing results from a cadaver study
with an in vivo study could also help quantify the effect of muscle action on ankle joint
kinematics during external foot rotation.

In conclusion, the current study quantified the effect of constraint on ankle joint motion during
voluntary external foot rotation. The study showed that constraint of the ankle had an effect on
the amount of ankle eversion and rotation for a given degree of tibia rotation, while dorsiflexion
was unaffected. A computational model of the ankle and its major ligamentous structures

85

showed that ATiTL strain was highly dependent on ankle constraint, while ATiFL strain was not.
The current study represents an important step in understanding the effect of different footwear
on ankle joint kinematics during external rotation of the foot relative to the upper body.

86

CHAPTER 6:
A METHOD OF DETERMINING IN VIVO DYNAMIC HUMAN ANKLE STIFFNESS
DURING EXTERNAL ROTATION

ABSTRACT
This study presents a method to determine stiffness of the human ankle in internal-external
rotation, inversion-eversion, and plantarflexion-dorsiflexion during an in vivo, single-legged,
external foot rotation experiment. Five young male subjects stood on one foot and performed an
internal rotation of the body (external rotation of the foot) on a force plate, while motions were
captured with a motion capture system. Hindfoot rotations relative to the tibia and resistive
moments in the above three directions were obtained using a commercial software package. The
resistive moments were plotted against rotation during the loading portions of the curves. Linear
approximations of these curves were used to determine stiffness of the ankle in directions of
internal-external, inversion-eversion, and plantarflexion-dorsiflexion. Future data from the above
experiment and other athletic motions may be useful for the design of surrogate devices used to
evaluate the potential performance and ankle injury risk characteristics of various shoe-surface
interfaces.

INTRODUCTION
Acute ankle trauma accounts for 10-30% of sports-related injuries (Hootman et al., 2007). While
medial and high ankle sprains are less common than lateral ankle sprains, they represent a more
disabling problem requiring longer recovery times (Wolfe et al., 2001). Preventative measures
may require the assessment of current footwear and playing surfaces.

87

The mechanism of medial and high ankle sprains has been clinically ascribed to excessive
external foot rotation (Guise, 1976, Boytim et al., 1991). Additionally, foot eversion prior to
external rotation has been shown critical in generating a high ankle sprain(Wei et al., 2012b). It
is also generally accepted that excessive rotational traction (Cawley et al., 2003) at the shoesurface interface may precipitate ankle and knee injuries (Nigg and Yeadon, 1987). Livesay et al.
(Livesay et al., 2006) suggest that rotational stiffness of the shoe-surface interface may be a more
important parameter than torque in determining ankle injury risk. Thus, evaluations of shoesurface interface with devices and development of preventative measures should use surrogate
ankles that have appropriate 3D properties and can simulate motions of the human ankle.

Previous studies documenting human ankle stiffness (or compliance) have determined ankle
stiffness using controlled input motions (Villwock et al., 2009c, Petit, 1996, Crandall et al.,
1996). However, little information exists on motions of the human ankle and its stiffness during
voluntary athletic maneuvers. The objective of the current study was to describe a method that
integrates experimentally obtained kinematic and kinetic data to produce useful structural
characteristics of the in vivo, human ankle during external foot rotation. This information may
help in the future development of surrogate devices to assess performance and ankle injury risk
characteristics of shoe-surface interfaces.

88

Figure 6-1 Marker set of the Oxford Foot Model (Carson et al., 2001). The subjects stood on
one foot and performed an internal rotation of the body (external rotation of the foot).

METHODS
Institutional Review Board approval was obtained for this project. In vivo kinematic data were
collected after obtaining informed consent from five young, male subjects with body mass of
77.9 ± 18 kg. A six-camera Vicon MX Motion Capture System (OMG plc. Oxford, UK) was
used to capture motions at 100 Hz from a marker set defined by the Oxford Foot Model (OFM)
(OMG plc., Oxford, UK) (Carson et al., 2001). An AMTI (Advanced Mechanical Technology,
Inc., Watertown, MA) force plate was used to capture ground reaction data at 1000 Hz. The
subjects stood with a flexed right leg on the force plate, balanced themselves, and performed an
internal rotation of the body (external rotation of the foot) with a self-determined maximum
effort (Figure 6-1). The subjects were instructed to internally rotate their bodies at a self-selected
speed, but within 3s, while keeping their foot on the force plate. Data were acquired over four

89

trials and processed using commercial kinetic (Plug-in Gait, OMG plc., Oxford, UK) and
kinematic (Dynamic OFM) software packages.

Figure 6-2 Sample of a moment versus time plot (left) and rotation versus time plot (right) for all
three movements: internal-external rotation, inversion-eversion, and plantarflexion-dorsiflexion.
Data from the loading phase (shaded in the figure) for moment vs. time were extracted and
plotted against the extracted rotation vs. time data, as shown in Figure 6-3.

Hindfoot rotations relative to the tibia and the corresponding resistive moments of the ankle in
three directions were obtained from the Plug-in Gait package. The moments were measured with
respect to the center of the ankle joint, as determined by the Plug-in Gait package. Moment and
rotation versus time plots were generated from the trials of each subject (Figure 6-2). The three
dimensional moments were plotted against the corresponding rotations from the beginning of
ankle loading (either a peak or sink in the response curve) to the end of loading (another peak or
sink). A linear regression was performed on the moment-rotation plot in order to determine the
slope and coefficient of determination (R2) for each trial. The average slope of the four trials
from each subject was used to represent subject ankle stiffness (Figure 6-3).

90

8

35

7

28
21

y = -2.81x - 18.276
R2 = 0.9818

6

y = 2.449x + 10.097
2

Moment (N-m)

Moment (N-m)

R = 0.9972

14
7
0

5
4
3

-7

2

-14

1

.

0

-21
-10

-7

-4

-1

2

-9

5

External Rotation (degrees)

-8.75 -8.5 -8.25
-8
Eversion (degrees)

-7.75

84
81
y = 3.8462x - 8.287

Moment (N-m)

78

R2 = 0.9888

75
72
69
66
63
60
19

20
21
22
23
Plantarflexion (degrees)

Figure 6-3 Samples of moment versus rotation plots in the loading phase for internal-external
rotation (left), inversion-eversion (right), and plantarflexion-dorsiflexion (bottom).

One-way ANOVA (Minitab Inc., State College, PA) was used to compare potential differences
of rotation speed between subjects. A linear regression analysis was used to determine the
correlation between subject body mass and ankle stiffness.

91

Table 6-1 Ankle stiffnesses in three directions obtained from subjects performing a single legged external rotation of the foot with
standard deviations in parenthesis. Note: the absolute value of the inversion-eversion slope was taken so that the reported number
was positive.

Subject
Rotation Speed (deg/s)

Internal/external rotation

86.21 (16.51)

63.75

17.3

3

4

5

52.47 (23.63)

75.20 (19.07)

61.83 (15.65)

43.00 (11.49)

2.67

1.80 (0.86)

1.52

(1.03)

2.65 (0.65)

2.31

(0.04)

2.19

0.52

Range (deg)

11.87 (6.69)

18.41 (5.78)

14.69 (4.98)

5.93 (1.00)

23.69 (4.84)

14.92

6.69

R

0.98 (0.00)

0.99

(0.01)

0.99

(0.00)

0.98 (0.02)

0.99

(0.00)

0.99

0.01

Stiffness (Nm/deg)

0.71 (0.35)

1.67

(1.05)

2.13

(0.92)

2.08 †

2.64

(.28)

1.85

0.72

Range (deg)

6.72 (3.78)

2.77

(1.56)

1.03

(0.50)

4.14 †

1.37 (0.43)

3.2

2.32

R

0.99 (0.00)

0.96

(0.05)

0.95

(0.05)

0.98 †

0.98 (0.01)

0.97

0.02

Stiffness (Nm/deg)

12.13 (0.08)

3.47 (1.13)

8.05

(1.60)

2.20 (0.53)

2.50 (1.07)

5.68

4.31

Range (deg)

0.72 (0.35)

4.54

(1.28)

3.27

(0.23)

2.77 (2.21)

2.11

(0.95)

2.68

1.41

0.89 (0.14)

0.89 (0.08)

0.99

(0.01)

0.92 (0.04)

0.97

(0.04)

0.93

0.04

2

Plantarflexion/dorsiflexion

SD

2

Stiffness (Nm/deg)

2

Inversion/eversion

Mean

1

2

R

92

(0.23)

RESULTS
The average rotation speed utilized by subjects was 63.5 ± 17.3 deg/sec and showed no
significant difference between subjects (p=0.11) (Table 6-1). The ranges of internal-external
rotational motion, inversion-eversion motion, and plantarflexion-dorsiflexion motion for these
experiments were 14.92 ± 6.69, 3.20 ± 2.32, and 2.68 ± 1.41 (deg), respectively. The level of
internal-external motion was significantly different than other motions (p<0.001). The
coefficients of determination (R2), assuming linearity, of the moment-rotation responses for
internal-external, inversion-eversion and plantarflexion-dorsiflexion rotation were 0.99 ± 0.01,
0.97 ± 0.02 and 0.93 ± 0.04, respectively (Table 6-1). The stiffness values of the ankle in
internal-external rotation, inversion-eversion, and plantarflexion-dorsiflexion were 2.19 ± 0.52,
1.85 ± 0.72 and 5.68 ± 4.31 (Nm/deg), respectively. A linear regression analysis revealed no
correlation between subject mass and stiffness for internal-external rotational, inversioneversion, and plantarflexion-dorsiflexion (p=0.73, 0.85, and 0.38, respectively).

DISCUSSION
The current study documented a method to extract in vivo ankle stiffness in internal-external,
inversion-eversion and plantarflexion-dorsiflexion during a single legged, external foot rotation
experiment. Results from the ATMI force plate and OFM-based motion analysis provided
moments and angular rotations in the three directions. These experimentally collected data were
utilized to calculate in vivo ankle stiffnesses in the three directions. Despite a small sample size,
the method proposed in this study may provide more realistic ankle stiffness values than previous
studies using either human cadavers or human subjects with passive motions (Crandall et al.,

93

1996, Petit, 1996, Villwock et al., 2009c) due to the fact that those muscles that were specifically
needed to cause internal body rotation were being utilized.
Table 6-2 Stiffness values obtained from previous studies

Internal/External Rotation

Stiffness (Nm/deg)
1.24-1.96

Reference
[9], [13]

Inversion/Eversion

0.83-1.7

[11], [10]

Dorsiflexion/ Plantar Flexion

3.7-6.4

[14, 19]

As opposed to previous studies that either applied controlled motions to human subjects or
utilized cadavers, the current study, by allowing voluntary motions, showed a higher stiffness in
internal-external and inversion-eversion directions, and a comparable stiffness in dorsiflexionplantar flexion direction (Table 6-2). A study by Villwock et al. (Villwock et al., 2009c) applied
an external rotation to cadaver ankles while measuring the resistive torque, showing an average
internal/external rotation stiffness of 1.24 Nm/deg. Other external foot rotation experiments
using cadavers document a rotational stiffness of 1.65 Nm/deg and 1.96 Nm/deg with the foot in
a flexible and rigid football shoe, respectively (Wei et al., 2012a). Applying an isolated motion
to human volunteers, Crandall et al. (Crandall et al., 1996) documents the inversion/eversion
stiffness to be 0.83 Nm/deg while Petit el al. (Petit, 1996) measures a stiffness of 1.7 Nm/deg by
applying the same motion to cadavers. Loram and Lakie (Loram and Lakie, 2002) document a
stiffness in dorsiflexion/plantar flexion between 3.7 and 4.8 Nm/deg by strapping human subjects
to a fixed vertical support and having them balance an inverted pendulum. Casadio et al.
document a higher stiffness (6.4 Nm/ deg) by mounting a motorized footplate onto a strain gauge
force platform.

94

One limitation of the current study was that subjects generated internal rotation of their bodies
with only a small amount of initial plantar-dorsiflexion or inversion-eversion rotation, resulting
in fewer data points. A previous study by Smith and Wong (Smith and Wong, 2012) has noted
that, when calculating ankle stiffness, more data points result in a higher reliability. Future
studies may have to adjust sampling rate in order to increase data reliability. Another limitation
of the current study was that ankle stiffness was determined in only the voluntary range of
loading. Previous cadaver studies (Villwock et al., 2009c, Petit, 1996, Crandall et al., 1996) have
noted a bilinear stiffness response of the ankle, with the stiffness increasing near failure levels.
It, of course, was not possible to conduct the current in vivo studies at high levels of internal
body rotation due to human subject safety issues. However, by comparing data from subfailure
in vivo and cadaver studies, the role of ankle musculature may be estimated by the use of a

mathematical model, similar to that utilized in a glenohumeral joint study by Elmore and Wayne
(Elmore KA, 2012) and the ankle study by Wei et al. (Wei et al., 2011d) (without musculature
effects). Based on such approximations, a surrogate ankle may be constructed that might yield a
better tool to determine the roles of various shoe-surface-interfaces on athletic performance and
their potential in producing high ankle sprains. Finally, a limitation of all studies involving
human subjects is the potential for variances in subject performance, particularly evident in the
ranges of motion during external rotation This is likely a result of the complex anatomy of the
talocrural joint which can result in differences in range of motion (and thus loading time) due to
differences in bone shape and spacing. As a result, while we believe in vivo data are most
appropriate for evaluations of shoe-surface interface characteristics, future studies may need to
provide better control of subject motions and very likely involve a large number of human
participants to generate reliable ankle stiffness information for the development of a biofidelic

95

surrogate device. Additionally, more trials from each subject could be taken in order to test the
repeatability of the current data.

While the current study was limited to internal rotation of the body with the foot planted on a
force plate, the methodology for determining ankle stiffness could be utilized in studies of other
athletic maneuvers, such as external rotation of the body with a planted foot or for various
cutting maneuvers. The current study utilized motion capture and a force plate to accomplish
this, but future studies might also utilize pressure insoles to help determine ankle stiffness during
more complex motions outside the laboratory (Weaver et al., 2013). Such data could more
completely characterize human ankle stiffness under a variety of situations on the athletic field.
These data may then be useful for the development of future surrogate test devices to evaluate
potential athletic performance and ankle injury risk characteristics of shoe-surface interfaces.

96

CHAPTER 7:
CONCLUSIONS
Overview
The first hypothesis of the current research was that the movement of the talocrural joint and the
location of injury during external rotation is a function of both the position of the ankle prior to
rotation and the amount of constraint on the ankle. This was investigated through both in vivo
and cadaveric testing. In vivo testing allowed for the measurement of ankle joint motion with
various constraints on the foot while accounting for muscle force and dynamic weight bearing.
These conditions closely replicated injurious conditions in real-life, but at sub-failure levels.
Cadaveric studies, despite not accounting for muscle-force, were advantageous because they
allowed for precision in experimental design and measurement, an accurate calculation of joint
motion by using screw-in marker sets, and an in-depth diagnosis of injury via post-test
dissection.

The second hypothesis of the current research was the that specimen-specific, rigid-body
modelling can be utilized to simulate injury-level ankle rotation in addition to modelling the
effect of various shoe rotational stiffnesses on talocrural and subtalar joint kinematics.
Computational modelling was an efficient, cost-effective, and non-invasive way to supplement
information measured from cadaveric and in vivo tests. Computational modelling enabled the
calculation of failure properties of ligaments given different boundary conditions. This was
useful in explaining experimental results and subsequently drawing conclusions regarding
external rotation ankle injury risk factors. Together, computational modelling, in vivo testing,
and cadaveric testing helped explain the functional roles of various structures of the ankle and
how those roles contribute to injury.

97

Chapter 2
Chapter 2 was a computational study which utilized CT scans to build specimen-specific models
and replicate injurious motions. Per previous laboratory testing, pairs of cadaver ankle models
were pre-everted 20 degrees or neutral in the coronal plane prior to external rotation up to the
angle of failure in the experiments, at which point peak strains in the ligaments were calculated.
The neutral ankle models showed highest strain in the ATiTL (29.1 ± 5.3%), correlating with the
medial ankle sprains in the neutral cadaver experiments. Everted ankle models showed highest
strain in the ATiFL (31.2 ± 4.3%) correlating with the high ankle sprains documented in everted
experiments. Combining both neutral and everted simulations, strains of 31.8 ± 3.8% computed
from the models of experiments that showed grossly ruptured ligaments were statistically greater
than 25.5 ± 4.2% for the partial tear simulations. The correlation between strain and ligament
damage demonstrated the potential for modelling to provide important information for the study
of injury mechanisms along with aiding in prevention of future injury.

Figure 7-1 Strains in the ATiFL and two deltoid ligaments (ATiTL and TiNL) for neutral and
everted simulations.

98

Chapter 3
Chapter 3 was an experimental study that investigated the effect of pre-eversion on talocrural and
subtalar joint motion during external rotation. Fourteen (seven pairs) cadaver lower extremities
were externally rotated 20 degrees in either a pre-everted or neutral configuration. Vicon motion
capture was used to track the motion of marker arrays on the tibia, talus, and calcaneus. Joint
coordinate systems between the tibia and talus (talocrural) and talus and calcaneus (subtalar)
were established to measure rotation, flexion, and in-eversion. While talocrural joint rotation
was greater in the neutral ankle (13.33 ± 2.09 deg versus 10.51 ± 2.75 deg, p=0.006), subtalar
joint rotation was greater in the pre-everted ankle (2.38 ± 1.89 deg versus 1.12 ± 1.04 deg,
p=0.014). Overall, the talocrural joint rotated more than the subtalar joint (11.92 ± 2.77 deg
versus 1.75 ± 1.61 degrees, p<0.001).

It was proposed that the calcaneus and talus ‘lock’ when axial load is applied to a neutral joint,
limiting subtalar motion, but maximizing talocrural rotation. When the foot is everted prior to
axial load, the subtalar joint may ‘unlock’ enabling the tibiotalar ligaments to pull the talus with
the rotation of the tibia, resulting in lower relative motion. This ‘unlocking’ of the subtalar joint
puts less strain on the medial ligaments potentially explaining why the combination of eversion
and external rotation does not always damage the anterior deltoid. This study may improve our
understanding of various mechanisms of ankle injury during external rotation of the foot.

99

Table 7-1 Motions of the subtalar and talocrural joint for 20 degrees of external rotation

Specimen

1
2
3
4
5
6
7
AVG
STDEV

p-value

Subtalar
Eversion (deg)

Talocrural
Eversion (deg)

Subtalar
Dorsiflexion
(deg)

Talocrural
Dorsiflexion
(deg)

Subtalar
Rotation (deg)

Talocrural
Rotation (deg)

Neutral Everted

Neutral Everted

Neutral Everted

Neutral Everted

Neutral Everted

Neutral Everted

0.59
1.66
0.50
0.25
0.97
1.04
1.71
0.96
0.56

0.29
2.43
0.21
0.32
0.41
1.99
0.84
0.93
0.91

0.39
2.68
1.53
1.35
1.30
1.50
0.58
1.33
0.74

0.39

0.19
3.80
1.03
0.30
1.34
1.31
0.85
1.26
1.21
0.79

0.787

1.46
0.16
0.03
0.34
2.24
1.43
0.98
4.09
1.29
0.75
3.11
3.50
0.97
0.53
0.55
0.52
0.84
0.16
0.79
0.02
1.23
1.02
2.91
0.78
0.04
0.52
0.72
0.48
1.15
0.65
1.30
1.39
0.67
0.46
1.21
1.67
0.054
0.882
0.197

0.04
0.69
15.81 11.08
1.11
1.70
12.30 10.97
2.53
5.78
11.84
6.41
0.23
0.88
15.67 12.87
0.02
0.94
14.93 14.20
1.98
2.88
12.09 10.53
1.90
3.81
10.65
7.52
1.12
2.38
13.33 10.51
1.04
1.89
2.09
2.75
0.014
0.006
<0.001

Chapter 4
Chapter 4 was a two-part study with the goal of investigating the effect of shoe rotational
stiffness on talocrural joint motion and injury during external rotation. The first part of the study
involved externally rotating pairs of cadaver ankles to failure in either a stiff or flexible shoe
while tracking the motion of the talus and tibia using motion capture. The second part of the
study involved modelling the experiments using specimen specific computational models.
Ankles in flexible shoes allowed 6.7 ± 2.4 degrees of talus eversion during rotation, significantly
greater than the 1.7 ± 1.0 degrees for ankles in stiff shoes. All ankles failed by either medial
ankle injury or syndesmotic injury, or a combination of both. Complex (more than one ligament
or bone) injuries were noted in 4 of 6 ankles in stiff shoes and 1 of 6 ankles in flexible shoes
(Table 7-2). The computational study provided information regarding the mechanisms of
complex injury (Figure 7-2). Despite a ~10% ATiFL pre-strain, the ATiTL strain rate increased
more rapidly and overtook the ATiFL strain level near the point of failure. This supported the Weber

C classification of failure: the ATiTL failed first, causing lateral translation of the talus and a
subsequent increase in ATiFL strain to cause failure of the ATiFL. In the flexible shoes, limited

100

talocrural rotation often resulted in lower ATiTL strain, potentially explaining the lower
incidence of complex injuries. Information from this study may help drive research and design
of athletic footwear in order to limit injury severity during external foot rotation in sport.
Table 7-2 Summary of injury mechanisms and failure metrics from cadaver experiment

Specimen

Failure Torque
(Nm)

Tibia Failure
Rotation (deg)

Talocrural
Talocrural
Failure Rotation Failure Eversion
(deg)
(deg)

Stiff

Flexible

Stiff

Flexible

Stiff

Flexible

Stiff

Flexible

1

85.6

83.4

44.7

46.5

19.3

17.5

3.5

3.9

2

75.8

93.4

37.0

39.4

27.9

21.5

0.9

7.4

Injury Mechanism

Stiff
1) Medial malleolus fracture
2) Spiral fracture of the fibula
1) ATiFL rupture
2 )ATiTL rupture

Flexible
ATiFL avulsion
ATiFL rupture

3

44.9

51.0

44.6

31.1

28.4

18.5

0.8

5.2

Medial malleolus fracture

1) ATiFL avulsion
2) ATiTL avulsion

4

55.8

59.0

29.4

33.2

19.3

13.9

1.7

5.1

1) ATiFL rupture
2) ATiTL rupture

ATiTL rupture

5

102.4

117.9

37.4

43.0

13.9

15.1

2.0

8.2

ATiFL avulsion

Fibular head fracture

6

170.7

179.7

48.1

63.2

20.3

27.2

1.1

10.4

1) ATiFL rupture
2) ATiTL rupture

ATiTL rupture

Average

89.2

97.4

40.2

42.7

21.5

19.0

1.7

6.7*

Standard
Deviation

44.9

47.0

6.9

11.6

5.6

4.9

1.0

2.4

Figure 7-2 Representative plot of the ATiFL and ATiTL strains versus tibia rotation

101

Chapter 5
Chapter 5 was a two-part study that involved quantifying the amount of ankle joint motion as a
function of tibia rotation and constraint rotational stiffness. The first part was an in vivo
experiment where subjects performed external ankle rotation with one of four foot constraints of
varying rotational stiffnesses while tracking the tibia and hindfoot using motion capture. Ankle
joint rotations in three directions (flexion, in-eversion, and rotation) were determined for each
subject up to 15 degrees of tibia rotation and the values were averaged to give a general curve
(Figure 7-3). Using the LinearModel.fit function in MATLAB, an equation was determined for
ankle joint dorsiflexion versus tape stiffness and tibia motion, ankle joint eversion versus tape
stiffness and tibia motion, and ankle joint external rotation versus tape stiffness and tibia motion
.
The second part of the study involved inputting predicted ankle joint motion for tibia rotations 015 degrees and tape rotational stiffnesses from 0 to 30 Nm/deg. Ligament strain was determined
during rotation to provide a range of strains for key ligaments over a domain of stiffnesses. It
was determined that constraint has a substantial effect on ATiTL strain during rotation, and a
small effect on ATiFL strain (Figure 7-4). Information from this study may help the design of
footwear in the future, as well as the analysis of game-time injuries given leg rotation and
footwear properties.

102

Figure 7-3 Mean ankle joint dorsiflexion, external rotation, and eversion vs. tibia rotation for
three tapes and a tapeless condition.

103

a

b

Figure 7-4 Predicted anterior tibiofibular ligament (ATiFL) (a) and anterior tibiotalar ligament
(ATiTL) (b) strain versus tibia rotation for an ankle constrained with tape stiffness ranging from
0 to 30 Nm/deg. Inputs for ankle joint motion were based on predictive equations 3a-3c.

Chapter 6
Chapter 6 was an in vivo study which involved providing a method to determine stiffness of the
human ankle in internal-external rotation, inversion-eversion, and plantarflexion-dorsiflexion

104

during a single-legged external foot rotation experiment. Five male subjects stood on one foot
and performed an internal rotation of the body (external rotation of the foot) on a force plate,
while motions were captured with a motion capture system. Hindfoot rotations relative to the
tibia and resistive moments in the above three directions were obtained using a commercial
software package. The resistive moments were plotted against rotation during the loading
portions of the curves. Linear approximations of these curves were used to determine stiffness of
the ankle in directions of internal-external, inversion-eversion, and plantarflexion-dorsiflexion.
This method can be used in the future for a variety of athletic motions in order to more fully
characterize human ankle stiffness under different loading situations on the athletic field. The
data can then be useful for the development of surrogate test devices to evaluate athletic
performance and ankle injury risk characteristics of shoe-surface interfaces.
Table 7-3 Ankle stiffness in three directions obtained from subjects performing a single legged
external rotation of the foot.
Subject
Rotation Speed (deg/s)

Internal/external rotation

3

4

5

Mean

SD

52.47 (23.63)

75.20 (19.07)

61.83 (15.65)

43.00 (11.49)

86.21 (16.51)

63.75

17.3

2.67

1.80 (0.86)

1.52

(1.03)

2.65 (0.65)

2.31

(0.04)

2.19

0.52

Range (deg)

11.87 (6.69)

18.41 (5.78)

14.69 (4.98)

5.93 (1.00)

23.69 (4.84)

14.92

6.69

R

0.98 (0.00)

0.99

(0.01)

0.99

(0.00)

0.98 (0.02)

0.99

(0.00)

0.99

0.01

Stiffness (Nm/deg)

0.71 (0.35)

1.67

(1.05)

2.13

(0.92)

2.08 †

2.64

(.28)

1.85

0.72

Range (deg)

6.72 (3.78)

2.77

(1.56)

1.03

(0.50)

4.14 †

1.37 (0.43)

3.2

2.32

R

0.99 (0.00)

0.96

(0.05)

0.95

(0.05)

0.98 †

0.98 (0.01)

0.97

0.02

Stiffness (Nm/deg)

12.13 (0.08)

3.47 (1.13)

8.05

(1.60)

2.20 (0.53)

2.50 (1.07)

5.68

4.31

Range (deg)

0.72 (0.35)

4.54

(1.28)

3.27

(0.23)

2.77 (2.21)

2.11

(0.95)

2.68

1.41

0.89 (0.14)

0.89 (0.08)

0.99

(0.01)

0.92 (0.04)

0.97

(0.04)

0.93

0.04

2

Plantarflexion/dorsiflexion

2

Stiffness (Nm/deg)

2

Inversion/eversion

1

2

R

105

(0.23)

Contributions to the Literature
This dissertation has made three major contributions to the current literature:

Influence of shoe design on ankle movement and injury
Previous studies have investigated shoe-surface interaction as an indicator of rotational injury
risk, but few studies have studied the effect of a shoe’s rotational stiffness on injury. A primary
contribution of this dissertation was the investigation of the effect shoe rotational stiffness has on
ankle joint motion, ligament strain, and injury generated. This was first looked at in Chapter 4,
which showed that shoe rotational stiffness not only affected the amount of talocrural joint
eversion and rotation during external foot rotation, but also the severity of injury generated and
distribution of ligament strains. It was concluded that stiff shoes limit talocrural joint eversion
during ankle external rotation, but maximize talocrural joint rotation. This resulted in high
anterior deltoid strains and complex injuries. The interaction between rotational stiffness and
talocrural joint motion was quantified in Chapter 5. The equations presented in this chapter will
enable future researchers to more accurately assess ankle joint motion given gross foot rotation
and shoe rotational stiffness. Additionally, this may aid in the design of footwear with the goal
of limiting time-loss due to severe ankle injury.

Determination of failure-level ligament strains
While previous studies have performed mechanical testing on ex vivo ligaments in order to
determine failure properties, little information exists regarding the amount of strain ligaments
undergo at failure in situ. A second major contribution of this dissertation was the correlation of
ligament strain level with severity of damage. This was accomplished first in Chapter 2, which

106

involved using specimen-specific computational models to simulate failure-level cadaver studies.
The strain magnitudes in grade II vs. grade III sprains were compared and were determined to be
significantly different (25.5 ± 4.2% vs. 31.8 ± 3.8%, respectively). This was verified in Chapter
4, in which simulations modelling complete ruptures showed ligament failure at 29.1 ± 6.7%.
These magnitudes could be used to assess potential injury risk (such as in Chapter 6), in order to
design shoes, or determine the effect of different loading conditions on injury likelihood.

Use of specimen-specific computational models
While a generic model provides useful information regarding relationships between motion and
injury mechanisms, it is believed that specimen-specific modeling provides more exact strain
information. A third contribution of this dissertation to the literature was the use of specimenspecific computational modelling to analyze ankle kinematics and injury. The curvature of the
tronchlea tali (top surface of the talus) determines the stability of the ankle in the anterior
posterior direction and the talar facets (contacting surfaces) have a significant impact on the
position of the axis of movements between the talus and calcaneus. The differences in ankle
mechanics can significantly affect injury risk as talar motion plays a significant role in both
medial and high ankle sprains. The use of specimen-specific ankle modeling can be expanded to
investigate in vivo injuries by looking at scenarios where the geometry and motion of a subject
are known. By looking at these scenarios, the specifics of the injury can be analyzed, thus aiding
in the design of intervention strategies and equipment to limit the severity and frequency of
athletic ankle injuries.

107

APPENDICES

108

APPENDIX 1: RESEARCH PUBLICATIONS
Peer Reviewed Manuscripts

1. Dutcheshen N, Maerz T, Rabban P, Haut RC, Button KD, Baker KC, Guettler J. (2012) The
acute effect of bipolar radiofrequency energy thermal chondroplasty on intrinsic biomechanical
properties and thickness of chondromalacic human articular cartilage. Journal of Biomechanical
Engineering. 134(8):081007.
2. Meyer EG, Wei F, Button KD, Haut RC. (2012) Determination of ankle ligament strain using
a rigid body computational model for sports injury scenarios. Proceedings-International Research
Council on the Biomechanics of Injury. 277-288.
3. Button KD, Wei F, Meyer EG, Haut RC. (2012) Specimen-specific computational models of
ankle sprains produced in a laboratory setting. Journal of Biomechanical Engineering. 135(4),
041001.
4. Button KD, Braman JE, Wei F, Haut RC. (2013) A method of determining in vivo dynamic
human ankle stiffness under external foot rotation. Journal of Sports Engineering and
Technology. 228(2), 120-124.
5. Fischenich KM, Button KD, Decamp CD, Haut RC, Donahue TL. (2014) Evaluation of
meniscal mechanics and proteoglycan content in a modified ACL transection model for
osteoarthritis. Journal of Biomechanical Engineering. 136(7), 071001. *Editors' Choice Paper
for 2014*
6. Fischenich KM, Button KD, Coatney GA, Fajardo RS, Leikert KM, Haut RC, Donahue TL.
(2014) Chronic changes in the articular cartilage and meniscus following traumatic impact to the
lapine knee. Journal of Biomechanics. doi:10.1016/j.jbiomech.2014.11.038
7. Wheatley BB, Fischenich KM, Button KD, Haut RC, Haut Donahue TL. (2015) An
optimized transversely isotropic, hyper-poro-viscoelastic finite element model of the meniscus to
evaluate mechanical degradation. Journal of Biomechanics. (In Press).
8. Button KD, Braman JE, Davison MA, Wei F, Schaeffer MC, Haut RC. (2014) Rotational
stiffness of American football shoes affects ankle biomechanics and injury severity. Journal of
Biomechanical Engineering. doi: 10.1115/1.4029979
9. Coatney GA, Abraham AC, Fischenich KM, Button KD, Haut RC, Haut Donahue TL. (2014)
Efficacy of P188 on lapine meniscus preservation following blunt trauma. Journal of the
Mechanical Behavior of Biomedical Materials. (In Publication).
10. Button KD, Braman JE, Davison MA, Wei F, Haut RC. (2014) Effect of pre-eversion on
talocrural and subtalar joint motion during ankle external rotation. Journal of Biomechanics. (In
Revision).

109

11. Pauly HM, Larson BE, Coatney GA, Button KD, DeCamp CE, Fajardo RS, Haut RC, Haut
Donahue TL. (2015) Assessment of cortical and trabecular bone changes in two models of posttraumatic osteoarthritis. Journal of Orthopaedic Research (In Revision).
12. Fischenich KM, Button KD, Fajardo RS, DeCamp CE, Haut RC, Haut Donahue TL. (2015)
Tissue changes in surgical and traumatically-induced experimental models of knee joint injury
using magnetic resonance imaging. Magnetic Resonance Imaging (In Review).
13. Button KD, Thornton PA, Braman JE, Wei F, Haut RC. (2015) Determining ankle joint
motion and ligament strain during ankle rotation for a given rotational stiffness constraint.
Journal of Sports Engineering and Technology (In Review)

Peer Reviewed Abstracts
1. Button KD, Braman JE, Wei F, Haut RC. (2012) Determination of in vivo dynamic human
ankle stiffness under external foot rotation. ASME Summer Bioengineering Conf. Fajardo,
Puerto Rico. SBC2012-80373.
2. Button KD, Wei F, Meyer EG, Fitzsimons K, Haut RC. (2012) Determination of in situ ankle
ligament strains in cases of high and medial ankle sprains. ASME Summer Bioengineering Conf.
Fajardo, Puerto Rico. SBC2012-80378.
3. Pauly HM, Larson BE, Button KD, Haut RC, Donahue TL. (2013) Micro-Computed
tomography analysis of bone volume following traumatic closed joint injury. Rocky Mountain
American Society of Biomechanics Meeting, April, 2013, Estes Park, CO.
4. Coatney GA, Abraham AC, Button KD, Haut RC, Donahue TL. (2013) P188 efficacy on
lapine meniscus preservation following blunt trauma. Rocky Mountain American Society of
Biomechanics Meeting, April, 2013, Estes Park, CO.
5. Fischenich KM, Button KD, Fajardo RS, Decamp CD, Haut RC, Donahue TL. (2013)
Mechanical evaluation of menisci following surgical transection of both menisci and anterior
cruciate ligament. Rocky Mountain American Society of Biomechanics Meeting, April, 2013,
Estes Park, CO.
6. Fischenich KM, Button KD, Fajardo RS, Haut RC, Donahue TL. (2013) Evaluation of
menisci following a traumatic compressive tibiofemoral load. ASME Summer Bioengineering
Conf. Sunriver, Oregon. SBC2013-14193.
7. Button KD, Leikert K, Donahue TL, Haut RC. (2013) Development of a traumatic anterior
cruciate ligament and meniscal rupture model to study osteoarthritis. ASME Summer
Bioengineering Conf. Sunriver, Oregon. SBC2013-14287.

110

8. Button KD, Davison M, Braman JE, Schaefer MC, Haut RC. (2013) Effect of shoe stiffness
on injury produced under external rotation of the foot in human cadavers. ASME Summer
Bioengineering Conf. Sunriver, Oregon, 2013. SBC2013-14719.
9. Fischenich KM, Coatney GA, Haverkamp J, Button KD, DeCamp C, Haut RC, Haut
Donahue TL. (2013) Meniscal glycosaminoglycan coverage twelve weeks post injury in two
models of knee joint injury. 11th World Congress of the International Cartilage Repair Society,
Sept. 15-18, 2013, Izmir, Turkey.
10. Button KD, Steibel JP, Karcher DM, Haut RC. (2014) A non-invasive method to predict
laying hen bone mechanical properties. Poultry Science Association Annual Meeting. Christi,
Texas. PSA-161.
11. Button KD, Davison M, Braman JE, Schaefer MC, Haut RC. (2014) Investigating the effects
of shoe stiffness on ankle injury risk using computational models. 7th World Congress in
Biomechanics. Boston, MA. 14-IS-2161-WCB.
12. Button KD, Leikert KM, DeCamp CD, Donahue TL, Haut RC. (2014) Comparison of a
traumatic ACL rupture and modified transection model to study osteoarthritis. 7th World
Congress in Biomechanics. Boston, MA. 14-A-1633-WCB.
13. Donahue TL, Fischenich KM, Coatney G, Pauly H, Button KD, Haut RC. (2014) Two new
experimental models of post-traumatic knee joint injury. 7th World Congress in Biomechanics.
Boston, MA. 14-IS-2706-WCB.
14. Fischenich KM, Button KD, Fajardo RS, Decamp CD, Haut RC, Donahue TL. (2014) A
longitudinal comparison of mechanical changes of the menisci in two models of posttraumatic
osteoarthritis of the knee. 7th World Congress in Biomechanics. Boston, MA. 14-A-1696-WCB.
15. Pauly HM, Larson BE, Button KD, DeCamp CD, Haut RC, Donahue TL. (2014) Microcomputed tomography comparison of trabecular bone changes in rabbits following surgical
transection of anterior cruciate ligament and menisci or traumatic impact to the tibiofemoral
joint. 60th Annual Meeting of the Orthopaedic Research Society. March 15-18, 2014. New
Orleans, LA.

111

APPENDIX 2: SOP (Standard Operating Procedure) for Building Specimen-Specific
Computational Ankle Models
Each individual ankle is scanned following test preparation. The ankle is potted and placed in a
neutral position prior to CT scan.
The software used are MIMICS 12.1c (Materialise, Ann Arbor, MI) and SolidWorks 2011 x64
Edition (TriMech Solutions, LLC, Columbia, MD).

Mimics Procedures
Open new project wizard with CT disc inserted. Choose CT as image type

Convert.

If more than one CT is available for you to choose from the disc select the one with the most
frames/slides.
*Note: Distance between images is usually 0.6 mm
Create new mask. (this will highlight entire CT)
*can change contrast by using thresholding feature. For bone it is easiest to use default

Figure A-1 Typical MIMICS interface

112

Select Region Growing. Use cursor to select highlighted area of bone of interest in one of the 3
CT scan viewing windows. This will create a new mask; rename it for the current bone. Deselect
the original mask.
Use the edit mask in 3D feature to check that only the bone of interest is selected under the new
mask.
If there is more than one bone selected go through the CT scan frame by frame and outline the
bone of interest with eraser using edit masks feature. This will give you a large enough gap
between bone of interest and excess area to use the edit mask in 3D feature to erase remaining
excess bone.
*use Edit in 3D cautiously, it will erase anything in the plane behind the cursor
*also you can change size and shape of the erasing cursor using the options at the
bottom of the screen or in its toolbar
When excess mask is removed choose Calculate 3D and select the mask that was just cleaned up.
For quality choose high. A 3D model will appear in the lower right screen.
Right click the 3D model just created in the side bar and choose surface wrap. (For larger bones
use 4 mm for both settings, for smaller use 3 mm to 2.5 mm for both settings.)
Right click the surface wrapped model and choose calculate Mask from 3D. This will generate a
new mask which you must check for overlap with other bones in the CT scan and for hollow
spaces in the center. When this mask is sufficient, repeat step 7. You may delete all masks and
3D models except for the final mask and final model created.
Again right click the 3-D model and choose smoothing. Use a smoothing factor of 0.9 for all
bones. The Number of iterations varies, and is somewhat at your discretion for how much
important detail should be kept. Make sure that keep originals is unchecked and compensate

113

shrinkage is checked. (# of iterations used for tibia and fibula was 4, calcaneus and talus was 2,
navicular, cuneiforms and metatarsals was 1, no smoothing was done on the phalanges)
Repeat steps 4 thru 10 for all bones of interest. (Hint: it may be helpful to use the Boolean
operator after finishing each bone to delete the mask created from the general mask that you start
your region growing from)
After all bones are completed, Go to FEA/CFD>>remesh>>select all
For smoothing (1st) and remeshing (2nd) use the following settings
Fixing>>Smooth
Method: Laplacian (1st order)
Smooth Factor: 0.9
Number of Iterations: 5
Remeshing>>Auto Remesh
Shape Quality Threshold: 0.3
Max Geometrical Error: 0.5
Number of Iterations: 3
Goto File>>Export>>Binary STL file and select appropriate folder for solidworks files. (File size
should be 100KB or less, If files are too large consider reprocessing with less detail)

114

Solid Works Processing
To open files in Solid works, you must go to open, and choose file type of mesh files. Select the
STL that was exported from MIMICS
Goto Tool>>ScanTo3D>>Mesh Prep Wizard>>select bone file
Continue to hit the next arrow in the dialog box that appears until you get to global
simplification. Here choose reduction amount of 50% unless the number of surfaces is already
less than 200.
Choose no smoothing in the next dialog box and when the option is available check the box to
launch the surface wizard.
Use surface wizard to create 3D object select Automatic creation from the radial buttons in the
first window.
Under surface detail select Update Preview. Check the number of surfaces and surface errors.
If more than one surface error occurs, choose the red x in the upper left corner of the surface
wizard. Go back to mesh prep wizard and change global smoothing by moving the slider bar one
tick mark to the right. Repeat steps with surface wizard.
If there is still more than one error go back to mesh prep wizard and reduce by 10%. Repeat until
there are only one or two surface errors.
With only one or two surface errors move slider bar to increase surface detail and update the
preview again. Continue to move slider to the right and update until error is gone.
Check the number of surfaces. If over 300 continue same reduction and smoothing process as
was used for multiple errors in the preview.
If there are no errors and there are less than 300 surfaces choose the green checkmark to finish
the bone and save in a new folder of revised files.

115

Repeat process with all bones exported form MIMICS.
*Hint: for best luck with spring placement in assembly be sure that calcaneus and talus
have at least 150 surfaces, closer to 300 would be better because of the number of springs that
are placed on these bones.

SolidWorks Assembly and Motion Analysis
Choose new SolidWorks document type assembly.
Insert Components>>Choose tibia first as fixed part of the assembly.
Order of insertion for remaining components doesn’t matter. Mate components origin and align
axes to adjacent bones in order to place them in correct 3D space
Go to motion study tab; in drop down menu choose study type: Motion Analysis.
For new spring, choose the spring icon. Make it a linear spring and choose two insertion nodes
based on insertions of previous models and references. (Atlas of Human Anatomy works well)
Choose stiffness based on literature and previous models. To pre-strain the springs, take the
original free length shown in the spring tool box and divide by 1.02 to get 2% strain. Copy this
value back into the spring toolbox and check the box to update to model changes.
Add 3D surface contacts by selecting two adjacent bones and clicking on the ‘3D surface
contacts’.
Uncheck the ‘material’ and ‘friction’ box. This will allow your model to run much quicker.
Under ‘Elastic Properties’, set:
Stiffness: 10000 N/mm
Exponent: 1.75
Max. Damping: 50 N/(mm/s)

116

Penetration: 0.001 mm
If there is still overlap after running a model, decrease the penetration or increase stiffness and
run again.
Before running either motor or motion data for calculation, check that the fibula, calcaneus, and
talus are not mated and are free to move.
To create more than one simulation with same model, right click motion study tab and
‘Duplicate’, this will save time in placing a springs in a new motion study.
Under study properties, lower the accuracy and the number of frames persecond. (Frames per
second can be lowered to as little as 5 per second). Check the box to show all motion analysis
messages—this will be helpful in troubleshooting the simulation.
Under the advanced options, the integrator can be either gstiff or wstiff and the initial integrator
step size can be lowered to if an error occurs during calculation.
Close all other programs on the computer and select the calculate icon. Running background
programs will slow the calculation and sometimes cause the program to get hung up for hours.
Torques, displacements, velocities and other motion data can be retrieved after the simulation is
finished by opening the plots toolbox.
In order to calculate strain values, click on the ‘Results’ icon and select a spring. Select
‘Displacement/Velocity/Acceleration>>Linear Displacement>>Magnitude’. Right click on the
plot and select ‘Export to Spreadsheet’. This will give you the length of the spring over time.
To calculate strain, copy the length values into ‘Ligament_Strain_Template.xls’ in the SOP
folder. Enter the original ligament length. Repeat this step for all ligaments of interest.

117

APPENDIX 3: SOP (Standard Operating Procedure) For Calculating Ankle Joint Motion
Given Vicon Data

This document will outline how to calculate the amount of talocrural joint inversion/eversion,
plantarflexion/dorsiflexion, and internal/external rotation given a three dimensional, orthogonal
axis on each bone (Figure A-2). The same procedure can be followed for the subtalar joint
The raw motions can come from either cadaver data (where the marker arrays are screwed into
the bone) (Figure A-3a) or in vivo data (where the markers are taped to the skin) (Figure A-3b)

Figure A-2 Orthogonal coordinate axes on the tibia, talus, and calcaneus.

118

a

b

Figure A-3 Marker arrays in a cadaver ankle (a) and human subject (b)

119

Mathematical Background
Since the arrays are randomly oriented, transformations need to be performed to align the local x,
y, and z axes with the global Cartesian coordinate system. For the talus:
> ?
A= ? >
? ?

?
?
>

B=Initial orientation of talus in local coordinates
C=Final position of talus in local coordinates
@A ∙ BA
R1=@C ∙ BA
@D ∙ BA

@A ∙ BC
@C ∙ BC
@D ∙ BC

@A ∙ BD
@C ∙ BD
@D ∙ BC

EA ∙ BA
R2=EC ∙ BA
ED ∙ BA

EA ∙ BC
EC ∙ BC
ED ∙ BC

EA ∙ BD
EC ∙ BD
ED ∙ BC

R3=R1TR2
R4=R3T
C’= R4A (Final orientation of talus in global coordinates)
Establish Joint Coordinate System based on final positions of arrays. For talocrural (C’=final
talus, D’=final tibia):
= F′
= E′
=|

×
×

|

=

∙ F′

(flexion)

=

∙ E′

(rotation)

=

∙

(inversion/eversion)

120

Procedure

Open up ‘JCS.m’ in ‘Z:\bimgrad\SOP\JCS ANALYSIS’
The program is written for a right ankle oriented with the foot on the ground. If the foot is
upside down, or the analysis is run for a left ankle, adjustments will need to be made to the signs
of alpha, gamma, and beta in lines 80-82 and 91-93.
Line 9: type the location of your Vicon data
Line 10: specify where you want the output data to go (note: this file must be closed when
running the program)
Line 105: specify what sheet you want to write the data to (term 3) and what range within the
sheet (term 4).
Run the program by hitting the green arrow in the task bar.
The program will give 6 columns of data. The columns should be labelled accordingly:
Tibia External
Rotation
Relative to
Global

Talus
Talus External
Talus Eversion
Dorsiflexion
Rotation
Relative to Tibia
Relative to Tibia Relative to Tibia

The program is shown below.

121

Tibia
Dorsiflexion
Relative to
Global

Tibia Eversion
Relative to
Global

%**************************************************************************
%written by Keith D. Button
%**************************************************************************
%file to calculate talocrural dorsiflexion, external rotation, and eversion
%for a right ankle. For a left ankle, multiply lines 76 and 77 by -1.
%**************************************************************************
clear, clc
%**************************************************************************
file='(Vicon data file name and location)';%Input Vicon data file location
out='(Output file location)’;%specify output file location
M=xlsread(file,1,'AG1:BD10000'); %data range in file for both tibia and talus positions (should
be 24 columns)
sizem=size(M);
r=sizem(1); %number of data points in file
Result=zeros(r,6);
tibia1=M(1,1:12); %initial talus position
talus1=M(1,13:24); %initial tibia position
%**************************************************************************
origin=[0 0 0 1 0 0 0 1 0 0 0 1];
Axraw= [origin(4)-origin(1),origin(5)-origin(2),origin(6)-origin(3)];
Ax=Axraw/norm(Axraw);
Ayraw= [origin(7)-origin(1),origin(8)-origin(2),origin(9)-origin(3)];
Ay=Ayraw/norm(Ayraw);
Azraw= [origin(10)-origin(1),origin(11)-origin(2),origin(12)-origin(3)];
Az=Azraw/norm(Azraw);
A=[Ax; Ay; Az]; %matrix for orthogonal neutral axis
%**************************************************************************
i=1;
for i=1:r
tibia2=M(i,1:12); %talus position
talus2=M(i,13:24); %tibia position
Bxraw= [talus1(4)-talus1(1),talus1(5)-talus1(2),talus1(6)-talus1(3)];
Bx=Bxraw/norm(Bxraw);
Byraw= [talus1(7)-talus1(1),talus1(8)-talus1(2),talus1(9)-talus1(3)];
By=Byraw/norm(Byraw);
Bzraw= [talus1(10)-talus1(1),talus1(11)-talus1(2),talus1(12)-talus1(3)];
Bz=Bzraw/norm(Bzraw);
B=[Bx; By; Bz]; %talus initial position in local coordinates
Cxraw= [talus2(4)-talus2(1),talus2(5)-talus2(2),talus2(6)-talus2(3)];
Cx=Cxraw/norm(Cxraw);
Cyraw= [talus2(7)-talus2(1),talus2(8)-talus2(2),talus2(9)-talus2(3)];
Cy=Cyraw/norm(Cyraw);
Czraw= [talus2(10)-talus2(1),talus2(11)-talus2(2),talus2(12)-talus2(3)];
Cz=Czraw/norm(Czraw);
C=[Cx ; Cy; Cz]; %talus position after external rotation in local coordinates
%**************************************************************************
Dxraw= [tibia1(4)-tibia1(1),tibia1(5)-tibia1(2),tibia1(6)-tibia1(3)];
Dx=Dxraw/norm(Dxraw);
Dyraw= [tibia1(7)-tibia1(1),tibia1(8)-tibia1(2),tibia1(9)-tibia1(3)];
Dy=Dyraw/norm(Dyraw);
Dzraw= [tibia1(10)-tibia1(1),tibia1(11)-tibia1(2),tibia1(12)-tibia1(3)];
Dz=Dzraw/norm(Dzraw);
D=[Dx; Dy; Dz]; %tibia initial position in local coordinates
Exraw= [tibia2(4)-tibia2(1),tibia2(5)-tibia2(2),tibia2(6)-tibia2(3)];
Ex=Exraw/norm(Exraw);
Eyraw= [tibia2(7)-tibia2(1),tibia2(8)-tibia2(2),tibia2(9)-tibia2(3)];
Ey=Eyraw/norm(Eyraw);
Ezraw= [tibia2(10)-tibia2(1),tibia2(11)-tibia2(2),tibia2(12)-tibia2(3)];
Ez=Ezraw/norm(Ezraw);
E=[Ex ; Ey; Ez]; %tibia position after external rotation in local coordinates
%**************************************************************************
ROT21=[dot(Bx,Ax),dot(Bx,Ay),dot(Bx,Az);dot(By,Ax),dot(By,Ay),dot(By,Az);dot(Bz,Ax),dot(Bz,Ay),do
t(Bz,Az)];
ROT22=[dot(Cx,Ax),dot(Cx,Ay),dot(Cx,Az);dot(Cy,Ax),dot(Cy,Ay),dot(Cy,Az);dot(Cz,Ax),dot(Cz,Ay),do
t(Cz,Az)];
ROT23=ROT22*transpose(ROT21);
ROT24=transpose(ROT21)*ROT22;
ROT25=transpose(ROT24);
C2=ROT25*A; %talus final position along anatomical axis

122

%**************************************************************************
ROT26=[dot(Dx,Ax),dot(Dx,Ay),dot(Dx,Az);dot(Dy,Ax),dot(Dy,Ay),dot(Dy,Az);dot(Dz,Ax),dot(Dz,Ay),do
t(Dz,Az)];
ROT27=[dot(Ex,Ax),dot(Ex,Ay),dot(Ex,Az);dot(Ey,Ax),dot(Ey,Ay),dot(Ey,Az);dot(Ez,Ax),dot(Ez,Ay),do
t(Ez,Az)];
ROT28=ROT27*transpose(ROT26);
ROT29=transpose(ROT26)*ROT27;
ROT30=transpose(ROT29);
E2=ROT30*A; %tibia final position along anatomical axis
%**************************************************************************
%talocrural rotations
e1=E2(1,:);
e3=C2(3,:);
e2raw=cross(e3,e1);
e2=e2raw/norm(e2raw);
alpha=asind(dot(e2,E2(3,:))); %dorsiflexion
gamma=asind(dot(e2,C2(1,:))); %external rotation
beta=asind(-dot(e1,e3)); %eversion
Result(i,2)=alpha;
Result(i,3)=gamma;
Result(i,4)=beta;
%tibia rotations
e4=Ax;
e6=E2(3,:);
e5raw=cross(e6,e4);
e5=e5raw/norm(e5raw);
alpha2=-asind(dot(e5,Az)); %tib dorsiflexion
gamma2=-asind(dot(e5,E2(1,:)));%tib external rotation
beta2=-asind(-dot(e4,e6)); %tib eversion
Result(i,1)=gamma2;
Result(i,5)=alpha2;
Result(i,6)=beta2;
i=i+1;
end
Tibrot=real(Result(:,1));
TCdors=Result(:,2);
TCrot=Result(:,3);
TCev=Result(:,4);
Tibdorsi=Result(:,5);
Tibev=Result(:,6);
xlswrite(out,Result,1,'A2:F500');

123

Table A-1: Raw Data from Cadaver Failure Tests (Chapter 4)
1477 Left
Failure Point
Torque (Nm) Rotation (deg)
85.55
44.69
Incremental
4.04
5.03
11.46
10.03
18.66
15.13
26.56
20.02
36.06
25.06
46.19
30.10
57.04
35.02
68.63
40.11
77.68
45.06

1477 Right
Failure Point
Torque (Nm) Rotation (deg)
83.39
46.48
Incremental
5.47
5.05
13.85
10.12
22.75
15.15
32.63
20.14
43.07
25.12
54.88
30.16
66.50
35.14
76.56
40.03
79.56
45.02

1466 Left
Failure Point
Torque (Nm) Rotation (deg)
75.80
37.02
Incremental
6.36
5.06
15.62
10.06
26.47
15.06
38.99
20.04
52.47
25.13
65.89
30.05
73.81
35.08

1466 Right
Failure Point
Torque (Nm) Rotation (deg)
93.35
39.42
Incremental
8.33
5.07
17.82
10.05
29.44
15.08
43.13
20.10
57.66
25.10
72.82
30.03
85.30
35.05

124

Table A-1 (cont’d)
1066 Left
Failure Point
Torque (Nm) Rotation (deg)
44.94
44.60
Incremental
2.26
5.08
5.90
10.13
10.19
15.14
15.71
20.02
22.28
25.10
29.94
30.08
38.68
35.10
40.63
40.09

1066 Right
Failure Point
Torque (Nm) Rotation (deg)
50.96
31.12
Incremental
4.28
5.09
11.06
10.06
19.32
15.06
28.51
20.04
38.78
25.07
49.05
30.10

1026 Left
Failure Point
Torque (Nm) Rotation (deg)
55.76
29.40
Incremental
6.13
5.01
13.45
10.00
21.68
15.10
30.39
20.13
40.16
25.05

1026 Right
Failure Point
Torque (Nm) Rotation (deg)
58.97
33.20
Incremental
2.79
5.01
9.62
10.06
17.78
15.06
27.77
20.00
39.85
25.11
40.17
25.25
51.88
30.05

125

Table A-1 (cont’d)
1180 Left
Failure Point
Torque (Nm)
Rotation (deg)
102.44
37.40
Incremental
4.41
5.06
15.40
10.09
27.74
15.11
43.02
20.07
59.25
25.11
76.99
30.09
92.41
35.03

1180 Right
Failure Point
Torque (Nm)
Rotation (deg)
117.92
43.02
Incremental
7.75
5.01
17.80
10.02
29.40
15.01
42.33
20.05
57.42
25.10
74.01
30.02
91.99
35.08
108.64
40.05

141 Left
Failure Point
Torque (Nm)
Rotation (deg)
170.65
48.06
Incremental
8.00
5.05
19.20
10.01
34.21
15.01
52.36
20.04
74.43
25.13
99.07
30.01
125.53
35.03
145.52
40.01
161.49
45.04

141 Right
Failure Point
Torque (Nm)
Rotation (deg)
179.68
63.19
Incremental
6.14
5.01
14.10
10.01
26.02
15.15
40.44
20.19
57.66
25.07
76.87
30.05
120.25
40.18
137.73
45.04
161.67
50.09
177.05
55.02
172.32
60.12

126

Table A-2: Raw Data from Tape Stiffness Tests (Chapter 5)
Elastikon
Torque
Rotation
(Nm)
(deg)
10
0.35
20
0.80
30
1.23
40
1.62
50
2.10
10
0.24
20
0.62
30
1.03
40
1.41
50
1.83
10
0.22
20
0.58
30
0.97
40
1.32
50
1.73
10
0.20
20
0.60
30
0.97
40
1.33
50
1.74
10
0.22
20
0.58
30
0.91
40
1.27
50
1.69

PowerFlex
Torque
Rotation
(Nm)
(deg)
10
0.30
20
1.12
30
1.91
40
2.64
50
3.37
10
0.31
20
1.09
30
1.92
40
2.53
50
3.36
10
0.35
20
1.12
30
1.80
40
2.50
50
3.28
10
0.32
20
1.11
30
1.75
40
2.49
50
3.27
10
0.31
20
1.03
30
1.73
40
2.45
50
3.20

127

Sher-Light
Torque
Rotation
(Nm)
(deg)
10
0.30
20
0.79
30
1.21
40
1.68
50
2.19
10
0.32
20
0.79
30
1.23
40
1.68
50
2.10
10
0.35
20
0.78
30
1.21
40
1.68
50
2.10
10
0.31
20
0.78
30
1.22
40
1.56
50
2.07
10
0.31
20
0.74
30
1.18
40
1.58
50
2.05

Table A-3: Input Motions for Computational Model (Chapter 5)
Tibia
Tape
Talocrural
Talocrural
Rotation
Stiffness
Dorsiflexion
Rotation
(deg)
(Nm/deg)
(deg)
(deg)
0
0
0.03
0.30
1
0
0.03
0.98
2
0
0.03
1.66
3
0
0.03
2.34
4
0
0.03
3.02
5
0
0.03
3.70
6
0
0.03
4.38
7
0
0.03
5.06
8
0
0.03
5.74
9
0
0.03
6.42
10
0
0.03
7.10
11
0
0.03
7.78
12
0
0.03
8.46
13
0
0.03
9.14
14
0
0.03
9.82
15
0
0.03
10.50
0
5
0.03
0.35
1
5
0.03
1.07
2
5
0.03
1.80
3
5
0.03
2.52
4
5
0.03
3.25
5
5
0.03
3.97
6
5
0.03
4.70
7
5
0.03
5.42
8
5
0.03
6.15
9
5
0.03
6.87
10
5
0.03
7.60
11
5
0.03
8.32
12
5
0.03
9.05
13
5
0.03
9.78
14
5
0.03
10.50
15
5
0.03
11.23

128

Talocrural
Eversion
(deg)
0.12
0.20
0.28
0.37
0.45
0.53
0.62
0.70
0.78
0.87
0.95
1.04
1.12
1.20
1.29
1.37
0.11
0.19
0.26
0.33
0.40
0.47
0.54
0.62
0.69
0.76
0.83
0.90
0.97
1.04
1.12
1.19

Table A-3: (cont’d)
Tibia
Rotation
(deg)
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

Tape
Stiffness
(Nm/deg)
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15

Talocrural
Dorsiflexion
(deg)
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03

129

Talocrural
Rotation
(deg)
0.39
1.17
1.94
2.71
3.48
4.25
5.02
5.79
6.56
7.33
8.10
8.87
9.64
10.41
11.18
11.95
0.44
1.26
2.08
2.89
3.71
4.52
5.34
6.15
6.97
7.78
8.60
9.41
10.23
11.04
11.86
12.67

Talocrural
Eversion
(deg)
0.11
0.17
0.23
0.29
0.35
0.41
0.47
0.53
0.59
0.65
0.71
0.77
0.83
0.89
0.94
1.00
0.11
0.16
0.21
0.25
0.30
0.35
0.40
0.44
0.49
0.54
0.58
0.63
0.68
0.73
0.77
0.82

Table A-3: (cont’d)
Tibia
Rotation
(deg)
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

Tape
Stiffness
(Nm/deg)
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25

Talocrural
Dorsiflexion
(deg)
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03

130

Talocrural
Rotation
(deg)
0.49
1.35
2.21
3.07
3.94
4.80
5.66
6.52
7.38
8.24
9.10
9.96
10.82
11.68
12.54
13.40
0.54
1.45
2.35
3.26
4.16
5.07
5.97
6.88
7.78
8.69
9.60
10.50
11.41
12.31
13.22
14.12

Talocrural
Eversion
(deg)
0.11
0.15
0.18
0.22
0.25
0.29
0.32
0.36
0.39
0.43
0.46
0.50
0.53
0.57
0.60
0.64
0.11
0.13
0.16
0.18
0.20
0.23
0.25
0.27
0.29
0.32
0.34
0.36
0.38
0.41
0.43
0.45

Table A-3: (cont’d)
Tibia
Rotation
(deg)
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

Tape
Stiffness
(Nm/deg)
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30

Talocrural
Dorsiflexion
(deg)
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03

131

Talocrural
Rotation
(deg)
0.59
1.54
2.49
3.44
4.39
5.34
6.29
7.24
8.19
9.14
10.09
11.04
11.99
12.94
13.89
14.84

Talocrural
Eversion
(deg)
0.11
0.12
0.13
0.14
0.15
0.16
0.17
0.18
0.20
0.21
0.22
0.23
0.24
0.25
0.26
0.27

REFERENCES

132

REFERENCES

Attarian, D. E., McCrackin, H. J., Devito, D. P., McElhaney, J. H. & Garrett, W. E. 1985. A
biomechanical study of human lateral ankle ligaments and autogenous reconstructive grafts.
American Journal of Sports Medicine, 13.
Barbaix, E., Van Roy, P. & Clarys, J. P. 2000. Variations of anatomical elements contributing to
subtalar joint stability: intrinsic risk factors for post-traumatic lateral instability of the ankle.
Ergonomics, 43, 1718-1725.
Beumer, A., van Hemert, W. L. W., Swierstra, B. A., Jasper, L. E. & Belkoff, S. M. 2003. A
biomechanical evaluation of the tibiofibular and tibiotalar ligaments of the ankle. Foot & Ankle
International, 24.
Beynnon, B. D., Murphy, D. F. & Alosa, D. M. 2002. Predictive factors for lateral ankle sprains:
A literature review. Journal of Athletic Training, 37, 376-380.
Bonacci, J., Saunders, P. U., Hicks, A., Rantalainen, T., Vicenzino, B. T. & Spratford, W. 2013.
Running in a minimalist and lightweight shoe is not the same as running barefoot: a
biomechanical study. British Journal of Sports Medicine, 47, 387-392.
Boruta, P. M., Bishop, J. O., Braly, W. G. & Tullos, H. S. 1990. Acute ankle ligament injuries; a
literature review. Foot & Ankle, 11, 107-113.
Boytim, M. J., Fischer, D. A. & Neumann, L. 1991. Syndesmotic ankle sprains. American
Journal of Sports Medicine, 19.
Button, K., Braman, J., Davison, M., Wei, F., Schaeffer, M. & Haut, R. 2015. Rotational
stiffness of American football shoes affects ankle biomechanics and injury severity. Journal of
Biomechanical Engineering, 137(6), 061004.
Button, K. D., Wei, F., Meyer, E. G. & Haut, R. C. 2013. Specimen-Specific Computational
Models of Ankle Sprains Produced in a Laboratory Setting. Journal of Biomechanical
Engineering-Transactions of the ASME, 135.
Carson, M. C., Harrington, M. E., Thompson, N., O'Connor, J. J. & Theologis, T. N. 2001.
Kinematic analysis of a multi-segment foot model for research and clinical applications: a
repeatability analysis. Journal of Biomechanics, 34.
Cawley, P. W., Heidt, R. S., Scranton, P. E., Losse, G. M. & Howard, M. E. 2003. Physiologic
axial load, frictional resistance, and the football shoe-surface interface. Foot & Ankle
International, 24.

133

Chao, E. Y. S. 2003. Graphic-based musculoskeletal model for biomechanical analyses and
animation. Medical Engineering & Physics, 25, 201-212.
Cheung, J. T. M., An, K. N. & Zhang, M. 2006. Consequences of partial and total plantar fascia
release: A finite element study. Foot & Ankle International, 27, 125-132.
Colville, M. R., Marder, R. A., Boyle, J. J. & Zarins, B. 1990a. Strain measurement in lateral
ankle ligaments. American Journal of Sports Medicine, 18.
Colville, M. R., Marder, R. A., Boyle, J. J. & Zarins, B. 1990b. Strain-measurement in lateral
ankle ligaments. American Journal of Sports Medicine, 18.
Crandall, J., Portier, L., Petit, P. & Hall, G. 1996. Biomechanical response and physical
properties of the leg, foot, and ankle.) 40th Stapp Car Crash Conference.
Dattani, R., Patnaik, S., Kantak, A., Srikanth, B. & Selvan, T. P. 2008. Injuries to the tibiofibular
syndesmosis. Journal of Bone and Joint Surgery-British Volume, 90B.
Delahunt, E., Monaghan, K. & Caulfield, B. 2006. Altered neuromuscular control and ankle joint
kinematics during walking in subjects with functional instability of the ankle joint. American
Journal of Sports Medicine, 34, 1970-1976.
Dias, L. S. 1979. Lateral ankle sprain: An experimental study. Journal of Trauma-Injury
Infection and Critical Care, 19, 266-269.
Dubin, J. C., Comeau, D., McClelland, R. I., Dubin, R. A. & Ferrel, E. 2011. Lateral and
syndesmotic ankle sprain injuries: a narrative literature review. Journal of Chiropractic
Medicine, 10, 204-19.
Edwards, G. S. & Delee, J. C. 1984. Ankle diastasis without fracture. Foot & Ankle, 4.
Elmore KA 2012. Soft tissue structures resisting anterior instability in a computational
glenohumeral joint model.
Fallat, L., Grimm, D. J. & Saracco, J. A. 1998. Sprained ankle syndrome: prevalence and
analysis of 639 acute injuries. The Journal of foot and ankle surgery : Official Publication of the
American College of Foot and Ankle Surgeons, 37, 280-5.
Fassbind, M. J., Rohr, E. S., Hu, Y., Haynor, D. R., Siegler, S., Sangeorzan, B. J. & Ledoux, W.
R. 2011. Evaluating Foot Kinematics Using Magnetic Resonance Imaging: From Maximum
Plantar Flexion, Inversion, and Internal Rotation to Maximum Dorsiflexion, Eversion, and
External Rotation. Journal of Biomechanical Engineering-Transactions of the ASME, 133.
Fong, D. T.-P., Ha, S. C.-W., Mok, K.-M., Chan, C. W.-L. & Chan, K.-M. 2012. Kinematics
Analysis of Ankle Inversion Ligamentous Sprain Injuries in Sports Five Cases From Televised
Tennis Competitions. American Journal of Sports Medicine, 40, 2627-2632.

134

Fong, D. T.-P., Hong, Y., Chan, L.-K., Yung, P. S.-H. & Chan, K.-M. 2007. A systematic review
on ankle injury and ankle sprain in sports. Sports Medicine, 37, 73-94.
Fong, D. T.-P., Wei, F., Hong, Y., Krosshaug, T., Haut, R. C. & Chan, K.-M. 2011. Ankle
ligament strain during supination sprain injury–a computational biomechanics study.) ISBSConference.
France, E. P., Paulos, L. E., Abbott, P. J., Roberts, P. F., Muhic, L. A., Lemaster, J. H. &
Kazarian, L. E. 1987. Failure characteristics of the medial collateral ligament of the knee effects of high-strain rate. Aviation Space and Environmental Medicine, 58.
Funk, J. R., Hall, G. W., Crandall, J. R. & Pilkey, W. D. 2000. Linear and quasi-linear
viscoelastic characterization of ankle ligaments. Journal of Biomechanical EngineeringTransactions of the ASME, 122.
Funk, J. R., Srinivasan, S. C. M., Crandall, J. R., Khaewpong, N., Eppinger, R. H., Jaffredo, A.
S., Potier, P. & Petit, P. Y. 2002. The effects of axial preload and dorsiflexion on the tolerance of
the ankle/subtalar joint to dynamic inversion and eversion. Stapp Car Crash Journal, 46, 245-65.
Garrick, J. G. & Requa, R. K. 1989. The epidemiology of foot and ankle injuries in sports.
Clinics in Podiatric Medicine and Surgery, 6, 629-37.
Gerber, J. P., Williams, G. N., Scoville, C. R., Arciero, R. A. & Taylor, D. C. 1998. Persistent
disability associated with ankle sprains: A prospective examination of an athletic population.
Foot & Ankle International, 19.
Golano, P., Vega, J., de Leeuw, P. A. J., Malagelada, F., Cristina Manzanares, M., Gotzens, V. &
van Dijk, C. N. 2010. Anatomy of the ankle ligaments: a pictorial essay. Knee Surgery Sports
Traumatology Arthroscopy, 18, 557-569.
Guise, E. R. 1976. Rotational ligamentous injuries to the ankle in football. The American Journal
of Sports Medicine, 4.
Haraguchi, N. & Armiger, R. S. 2009. A New Interpretation of the Mechanism of Ankle
Fracture. Journal of Bone and Joint Surgery-American Volume, 91A.
Harper, M. C. 1992. Ankle fracture classification systems: a case for integration of the LaugeHansen and AO-Danis-Weber schemes. Foot & ankle, 13, 404-7.
Heilman, A. E., Braly, W. G., Bishop, J. O., Noble, P. C. & Tullos, H. S. 1990. An anatomic
study of subtalar instability. Foot & Ankle, 10, 224-228.
Hennig, E. M. 2011. The Influence of Soccer Shoe Design on Player Performance and Injuries.
Research in Sports Medicine, 19, 186-201.

135

Hertel, J., Denegar, C. R., Monroe, M. M. & Stokes, W. L. 1999. Talocrural and subtalar joint
instability after lateral ankle sprain. Medicine and Science in Sports and Exercise, 31, 15011508.
Hicks, J. H. 1953. The mechanics of the foot: 1. the joints. Journal of Anatomy, 87, 345-357.
Hirsch, C. & Lewis, J. 1965. Experimental ankle-joint fractures. Acta Orthopaedica
Scandinavica, 36, 408-&.
Holzbaur, K. R. S., Murray, W. M. & Delp, S. L. 2005. A model of the upper extremity for
simulating musculoskeletal surgery and analyzing neuromuscular control. Annals of Biomedical
Engineering, 33, 829-840.
Hootman, J. M., Dick, R. & Agel, J. 2007. Epidemiology of collegiate injuries for 15 sports:
Summary and recommendations for injury prevention initiatives. Journal of Athletic Training,
42.
Hopkinson, W. J., Stpierre, P., Ryan, J. B. & Wheeler, J. H. 1990. Syndesmosis sprains of the
ankle. Foot & Ankle, 10.
Hughes, J. L., Weber, H., Willenegger, H. & Kuner, E. H. 1979. Evaluation of ankle fractures:
non-operative and operative treatment. Clinical Orthopaedics and Related Research, 111-119.
Iaquinto, J. M. & Wayne, J. S. 2010. Computational model of the lower leg and foot/ankle
complex: application to arch stability. Journal of Biomechanical Engineering, 132.
Kato, T. 1995. The diagnosis and treatment of instability of the subtalar joint. Journal of Bone
and Joint Surgery-British Volume, 77B, 400-406.
Kent, R., Crandall, J., Forman, J., Lessley, D., Lau, A. & Garson, C. 2012. Development and
assessment of a device and method for studying the mechanical interactions between shoes and
playing surfaces in situ at loads and rates generated by elite athletes. Sports Biomechanics, 11,
414-429.
Kitaoka, H. B., Luo, Z. P. & An, K. N. 1997. Three-dimensional analysis of normal ankle and
foot mobility. American Journal of Sports Medicine, 25, 238-242.
Kleipool, R. P. & Blankevoort, L. 2010. The relation between geometry and function of the
ankle joint complex: a biomechanical review. Knee Surgery Sports Traumatology Arthroscopy,
18, 618-627.
Konradsen, L. & Voigt, N. 2002. Inversion injury biomechanics in functional ankle instability: a
cadaver study of simulated gait. Scandinavian Journal of Medicine & Science in Sports, 12.

136

Krosshaug, T. & Bahr, R. 2005. A model-based image-matching technique for three-dimensional
reconstruction of human motion from uncalibrated video sequences. Journal of Biomechanics,
38, 919-929.
Kuhlman, S., Sabick, M., Pfeiffer, R., Cooper, B. & Forhan, J. 2010. Effect of loading condition
on the traction coefficient between shoes and artificial turf surfaces. Proceedings of the
Institution of Mechanical Engineers Part P-Journal of Sports Engineering and Technology, 224,
155-165.
Kwak, S., Blakevoort, L. & Ateshian, G. 2000. A mathematical formulation for 3D quasi-static
multibody models of diarthrodial joints. Computer Methods in Biomechanics and Biomedical
Engineering, 3, 41-64.
Kwon, J. Y., Chacko, A. T., Kadzielski, J. J., Appleton, P. T. & Rodriguez, E. K. 2010. A Novel
Methodology for the Study of Injury Mechanism: Ankle Fracture Analysis Using Injury Videos
Posted on YouTube.com. Journal of Orthopaedic Trauma, 24, 477-482.
Lambson, R. B., Barhnill, B. S. & Higgins, R. W. 1996. Football cleat design and its effect on
anterior cruciate ligament injuries - A three-year prospective study. American Journal of Sports
Medicine, 24, 155-159.
Laugehansen, N. 1950. Fractures of the Ankle. Archives of Surgery, 60.
Leardini, A., Stagni, R. & O'Connor, J. J. 2001. Mobility of the subtalar joint in the intact ankle
complex. Journal of Biomechanics, 34, 805-809.
Liacouras, P. C. & Waynel, J. S. 2007. Computational modeling to predict mechanical function
of joints: Application to the lower leg with simulation of two cadaver studies. Journal of
Biomechanical Engineering-Transactions of the ASME, 129, 811-817.
Lin, C. F., Gross, M. T. & Weinhold, P. 2006. Ankle syndesmosis injuries: Anatomy,
biomechanics, mechanism of injury, and clinical guidelines for diagnosis and intervention.
Journal of Orthopaedic & Sports Physical Therapy, 36.
Livesay, G. A., Reda, D. R. & Nauman, E. A. 2006. Peak torque and rotational stiffness
developed at the shoe-surface interface - The effect of shoe type and playing surface. American
Journal of Sports Medicine, 34.
Loram, I. D. & Lakie, M. 2002. Direct measurement of human ankle stiffness during quiet
standing: the intrinsic mechanical stiffness is insufficient for stability. Journal of PhysiologyLondon, 545, 1041-1053.
Lundberg, A., Svensson, O. K., Bylund, C., Goldie, I. & Selvik, G. 1989a. Kinematics of the
ankle/foot complex--part 2: pronation and supination. Foot & Ankle, 9, 248-253.

137

Lundberg, A., Svensson, O. K., Bylund, C. & Selvik, G. 1989b. Kinematics of the foot/ankle
complex--part 3: influence of leg rotation. Foot & Ankle, 9, 304-309.
Meyers, M. C. 2010. Incidence, Mechanisms, and Severity of Game-Related College Football
Injuries on FieldTurf Versus Natural Grass A 3-Year Prospective Study. American Journal of
Sports Medicine, 38, 687-697.
Miller, C. D., Shelton, W. R., Barrett, G. R., Savoie, F. H. & Dukes, A. D. 1995. Deltoid and
syndesmotic ligament injury of the ankle without fracture. American Journal of Sports Medicine,
23.
Mok, K.-M., Fong, D. T.-P., Krosshaug, T., Engebretsen, L., Hung, A. S.-L., Yung, P. S.-H. &
Chan, K.-M. 2011. Kinematics Analysis of Ankle Inversion Ligamentous Sprain Injuries in
Sports 2 Cases During the 2008 Beijing Olympics. American Journal of Sports Medicine, 39.
Netter, F. H. & JT, H. 2003. Atlas of Human Anatomy, Teterboro, NJ, Icon Learning Systems.
Ng, B. H., Chou, S. M., Lim, B. H. & Chong, A. 2004. Strain rate effect on the failure properties
of tendons. Proceedings of the Institution of Mechanical Engineers Part H-Journal of
Engineering in Medicine, 218.
Nigg, B. M. & Yeadon, M. R. 1987. Biomechanical aspects of playing surfaces. Journal of
Sports Sciences, 5.
Norkus, S. A. & Floyd, R. T. 2001. The anatomy and mechanisms of syndesmotic ankle sprains.
Journal of Athletic Training, 36.
Parenteau, C. S., Viano, D. C. & Petit, P. Y. 1998. Biomechanical properties of human cadaveric
ankle-subtalar joints in quasi-static loading. Journal of Biomechanical Engineering-Transactions
of the ASME, 120, 105-111.
Peltz, C. D., Haladik, J. A., Hoffman, S. E., McDonald, M., Ramo, N. L., Divine, G., Nurse, M.
& Bey, M. J. 2014. Effects of footwear on three-dimensional tibiotalar and subtalar joint motion
during running. Journal of Biomechanics, 47, 2647-53.
Petit, P. 1996. Quasi-static Characterization of the human foot-ankle joints in a simulated tensed
state and updated accidentological data.) IRCOBI.
Rasmussen, O. & Kromannandersen, C. 1983. Experimental ankle injuries--Analysis of the
traumatology of the ankle ligaments. Acta Orthopaedica Scandinavica, 54, 356-362.
Renstrom, P. & Konradsen, L. 1997. Ankle ligament injuries. British Journal of Sports Medicine,
31.
Safran, M. R., Benedetti, R. S., Bartolozzi, A. R. & Mandelbaum, B. R. 1999. Lateral ankle
sprains: a comprehensive review - Part 1: etiology, pathoanatomy, histopathogenesis, and
diagnosis. Medicine and Science in Sports and Exercise, 31.

138

Sarsam, I. M. & Hughes, S. P. F. 1988. The role of the anterior tibiofibular ligament in talar
rotation: an anatomical study. Injury, 19, 62-64.
Siegler, S., Chen, J. & Schneck, C. D. 1988. The 3-dimensional kinematics and flexibility
characteristics of the human ankle and subtalar joints--part 1: kinematics. Journal of
Biomechanical Engineering-Transactions of the ASME, 110, 364-373.
Smith, A. W. & Wong, D. P. 2012. Effects of window size on ankle joint stiffness calculation
during quiet standing: How the rule changes the result. Journal of Biomechanics, 45, 2301-2305.
Soutas-Little, R. W., Beavis, G. C., Verstraete, M. C. & Markus, T. L. 1987. Analysis of Foot
Motion During Running Using a Joint-Coordinate System. Medicine and Science in Sports and
Exercise, 19.
Torg, J. S. & Quedenfe.Tc 1973. Effect of shoe type and cleat length on incidence and severity
of knee injuries among high school football players. Journal of Bone and Joint SurgeryAmerican Volume, A 55, 1319-1319.
Vanlangelaan, E. J. 1983. A kinematical analysis of the tarsal joints. Acta Orthopaedica
Scandinavica, 54, 11-269.
Viladot, A., Lorenzo, J. C., Salazar, J. & Rodriguez, A. 1984. The subtalar joint-embryology and
morphology. Foot & Ankle, 5, 54-66.
Villwock, M. R., Meyer, E. G., Powell, J. W., Fouty, A. J. & Haut, R. C. 2009a. Football Playing
Surface and Shoe Design Affect Rotational Traction.
Villwock, M. R., Meyer, E. G., Powell, J. W., Fouty, A. J. & Haut, R. C. 2009b. Football Playing
Surface and Shoe Design Affect Rotational Traction. American Journal of Sports Medicine, 37.
Villwock, M. R., Meyer, E. G., Powell, J. W. & Haut, R. C. 2009c. Development and evaluation
of a surrogate ankle for use with a rotational traction measurement apparatus. Proceedings of the
Institution of Mechanical Engineers Part P-Journal of Sports Engineering and Technology, 223.
Vitale, T. D. & Fallat, L. M. 1988. Lateral ankle sprains: evaluation and treatment. The Journal
of Foot Surgery, 27.
Waterman, B. R., Belmont, P. J., Jr., Cameron, K. L., DeBerardino, T. M. & Owens, B. D. 2010.
Epidemiology of Ankle Sprain at the United States Military Academy. American Journal of
Sports Medicine, 38.
Waterman, B. R., Belmont, P. J., Jr., Cameron, K. L., Svoboda, S. J., Alitz, C. J. & Owens, B. D.
2011. Risk Factors for Syndesmotic and Medial Ankle Sprain Role of Sex, Sport, and Level of
Competition. American Journal of Sports Medicine, 39.

139

Weaver, B., Fitzsimons, K., Braman, J. & Haut, R. 2013. Torque prediction at the shoe-surface
interface using insole pressure technology. Journal of Sports Engineering and Technology, In
Print.
Wei, F., Braman, J., Meyer, E., Powell, J. & Haut, R. 2011a. Mechanism of injury in a high
ankle sprain: A simulation study.) Summer Bioengineering Conference. Farmington, PA.
Wei, F., Braman, J. E., Weaver, B. T. & Haut, R. C. 2011c. Determination of dynamic ankle
ligament strains from a computational model driven by motion analysis based kinematic data.
Journal of Biomechanics, 44, 2636-2641.
Wei, F., Hunley, S. C., Powell, J. W. & Haut, R. C. 2011d. Development and Validation of a
Computational Model to Study the Effect of Foot Constraint on Ankle Injury due to External
Rotation. Annals of Biomedical Engineering, 39.
Wei, F., Meyer, E. G., Braman, J. E., Powell, J. W. & Haut, R. C. 2012a. Rotational Stiffness of
Football Shoes Influences Talus Motion during External Rotation of the Foot. Journal of
Biomechanical Engineering-Transactions of the ASME, 134.
Wei, F., Post, J. M., Braman, J. E., Meyer, E. G., Powell, J. W. & Haut, R. C. 2012b. Eversion
during external rotation of the human cadaver foot produces high ankle sprains. Journal of
Orthopaedic Research, 30.
Wei, F., Villwock, M. R., Meyer, E. G., Powell, J. W. & Haut, R. C. 2010. A Biomechanical
Investigation of Ankle Injury Under Excessive External Foot Rotation in the Human Cadaver.
Journal of Biomechanical Engineering-Transactions of the ASME, 132.
Williams, D. S. B., 3rd, Green, D. H. & Wurzinger, B. 2012. Changes in lower extremity
movement and power absorption during forefoot striking and barefoot running. International
Journal of Sports Physical Therapy, 7, 525-32.
Williams, G. N., Jones, M. H. & Amendola, A. 2007. Syndesmotic ankle sprains in athletes.
American Journal of Sports Medicine, 35.
Wolfe, M. W., Uhl, T. L., Mattacola, C. G. & McCluskey, L. C. 2001. Management of ankle
sprains (vol 63, pg 93, 2001). American Family Physician, 64.
Wong, Y. S., Kim, W. & Ying, N. 2005. Passive motion characteristics of the talocrural and the
subtalar joint by dual Euler angles. Journal of Biomechanics, 38, 2480-2485.
Wu, G., Siegler, S., Allard, P., Kirtley, C., Leardini, A., Rosenbaum, D., Whittle, M., D'Lima, D.
D., Cristofolini, L., Witte, H., Schmid, O. & Stokes, H. 2002. ISB recommendation on
definitions of joint coordinate system of various joints for the reporting of human joint motion part 1: ankle, hip, and spine. Journal of Biomechanics, 35, 543-548.

140

Yahia, L., Brunet, J., Labelle, S. & Rivard, C. H. 1990. A scanning electron-microscopic study
of rabbit ligaments under strain. Matrix, 10.
Yeung, M. S., Chan, K. M., So, C. H. & Yuan, W. Y. 1994. An epidemiological survey on ankle
sprain. British Journal of Sports Medicine, 28, 112-116.

141