EFFECTS OF TWO MARKER PLACEMENT AND DATA ANALYSIS METHODS
ON RUNNING GAIT ANALYSIS
by
JAMES N.M. BECKER
A THESIS
Presented to the Department ofHuman Physiology
and the Graduate School of the University of Oregon
in partial fulfillment of the requirements
for the degree of
Master of Science
March 2010
"Effects of Two Marker Placement and Data Analysis Methods on Running Gait
Analysis," a thesis prepared by James N.M. Becker in partial fulfillment of the
requirements for the Master of Science degree in the Department of Human Physiology.
This thesis has been approved and accepted by:
Dr. Li-Shan Chou, Chair of the Examining Committee
ii
Date
Committee in Charge:
Accepted by:
Dr. Li-Shan Chou, Chair
Dr. Louis Ostemig
Dr. Stanley James
Dean of the Graduate School
© 2010 James N.M. Becker
iii
IV
An Abstract of the Thesis of
James N.M. Becker for the degree of Master of Science
in the Department of Human Physiology to be taken March 2010
Title: EFFECTS OF TWO MARKER PLACEMENT AND DATA ANALYSIS
METHODS ON RUNNING GAIT ANALYSIS
Approved:
Dr. Li-Shan Chou
This study evaluated the effects of two marker placement methods and two data
analysis methods on running gait analysis. Markers placed on the shoe heel counter were
compared with markers placed directly on the calcaneous and visible through heel
windows cut into the shoe. When analyzed using a traditional group design no
significant differences were found between marker conditions for rear foot eversion
excursion, percent stance at which peak eversion occurred, maximal instantaneous
eversion velocity, or maximal instantaneous vertical loading rate. Ankle frontal plane
variability was significantly different between conditions. When analyzed with a single
subject design some individuals demonstrated significant differences between conditions
while others did not. In some individuals the heel windows condition revealed previously
masked coupling parameters thought to be related to injury. The results of this study
suggest the heel windows method and single subject analysis should be used for a
longitudinal study of runners.
CURRICULUM VrTAE
NAME OF AUTHOR: James Nicholas Macksoud Becker
PLACE OF BIRTH: Boston, MA
DATE OF BIRTH: 07/17/1979
GRADUATE AND UNDERGRADUATE SCHOOLS ATTENDED:
University of Oregon, Eugene, Oregon
Middlebury College, Middlebury, Vermont
DEGREES AWARDED:
Master of Science, Human Physiology, March 2010, University of Oregon
Bachelor of Arts, Geography, June 2002, Middlebury College
AREAS OF SPECIAL INTEREST:
Biomechanical contributors to running related injuries
Applying biomechanical analyses to improve sports performance
PROFESSIONAL EXPERIENCE:
Graduate Teaching Fellow, University of Oregon, Eugene, Oregon, 2009-2010
High School Science Teacher, Vergennes Union High School, Vergennes,
Vermont, 2003-2007
Head Coach Track & Field and Cross Country Teams, Vergennes Union High
School, Vergennes, Vermont, 2002-2007
v
VI
ACKNOWLEDGMENTS
This work would not have been possible without the support of the three members
ofmy committee. I would like to thank Dr. Li-Shan Chou for providing the support,
guidance, and freedom in design which allowed this project to come to fruition. I would
also like to thank Dr. Louis Osternig for his feedback while developing the study and his
help in the preparation of this manuscript. Lastly, I would especially like to thank Dr. Stan
James for donating his time to perform the clinical exams in this study, for his feedback
and ideas, and for his help in imagining future directions following this work.
I would also like to acknowledge the contributions and help of my labmates Scott
Breloff, Betty Chen, Shiu-Ling Chiu, Masahiro Fujimoto, and Vipul Lugade. Thank you
all for your help, insight, support, and general friendship. Thank you to Steve Laurie and
Colin Wallace for their help in developing and testing the protocols used in this study.
Finally, I wish to extend a big thank you to Hilary Senesac. Without her help and
support this project never would have gotten off the ground. From protocol testing to data
collection to preparing for the oral defense, this project would not have happened without
your constant support, encouragement, and understanding. Thank you.
Vll
TABLE OF CONTENTS
Chapter Page
I. INTRODUCTION 1
Running Injuries: Incidence and Costs 1
Retrospective Studies on Possible Causes of Running Injuries.......... 3
Lower Limb Kinematics During Running............................................................. 6
Relationship Between Rear Foot Kinematics and Running Injuries 8
Measuring Rear Foot Motion................................................................................. 10
Purposes and Hypotheses of the Study.................................................................. 16
II. METHODS.............................................................................................................. 19
Subjects.................................................................................................................. 19
Experimental Instruments...................................................................................... 19
Dynamometer................................................................................................... 19
Force Plates 20
Motion Capture System.................. 20
Data Collection and Experimental Procedures 20
Muscle Strength Measurements............................ 20
Clinical Exam.... 23
Motion Analysis............................................................................................... 24
Data Analysis.... 27
Statistical Analysis..................................... 29
III. RESULTS 32
Subject Data........................................................................................................... 32
Heel Windows Dimensions.................................................................................... 33
Chapter
viii
Page
Running Speed 35
Kinematic Variables Results.................................................................................. 37
IV. DISCUSSION........................................................................................................ 60
Variability ofAnkle Joint Inversion Eversion Angles........................................... 62
The Subject Sample as a Potential Source ofNon-Significance 64
Kinematic and Kinetic Parameters......................................................................... 66
Uncontrolled Factors.............................................................................................. 67
Data Analysis Methods 70
V. SINGLE SUBJECT DESIGN 75
Background and Rational....................................................................................... 75
Single Subject Design Methods......... 81
VI. SINGLE SUBJECT RESULTS 87
Subjects 87
Graphical Results................................................................................................... 89
Statistical Analysis Results... 92
VII. SINGLE SUBJECT ANALYSIS DISCUSSION 100
Appropriateness of Using Statistical Analysis....................................................... 101
Kinematic Differences Revealed with Heel Windows and Single Subject
Analysis...... 103
Differences in Maximal Instantaneous Vertical Loading Rates............................ 106
VIII. CONCLUSIONS, LIMITATIONS, FUTURE DIRECTIONS........................... 109
Conclusions............................................................................................................ 109
Limitations 111
Chapter
IX
Page
Future Directions...... 113
APPENDICES 114
A. SUBJECT INFORMED CONSENT FORM...... 114
B. CLINICAL EVALUATION FORM 117
C. METHODOLOGY FOR ESTABLISHING THE ANATOMIC AND
TRACKING MARKER COORDINATE SySTEMS 121
BIBLIOGRAPHY 133
LIST OF FIGURES
Figure Page
1. Schematic illustration ofwhere the measurements were taken for the
arch height index............................................................................................... 24
2. The markers used to develop the anatomic coordinate system 30
3. Picture of the heel windows marker placement and set up of the heel
windows............................................................................................................. 31
4. Example left sagittal plane ankle angles 41
5. Example right sagittal plane ankle angles 41
6. Example frontal plane ankle angles................................................................... 42
7. Example right frontal plane ankle angles 42
8. Example left transverse plane ankle angles....................................................... 43
9. Example right transverse plane ankle angles 43
10. Example left sagittal plane knee angles 44
11. Example right sagittal plane knee angles. 44
12. Example left frontal plane knee angles 45
13. Example right frontal plane knee angles........ 45
14. Example left transverse plane knee angles 46
15. Example right transverse plane knee angles...................................................... 46
16. Example left sagittal plane hip angles........... 47
17. Example right sagittal plane hip angles............................................................. 47
18. Example left frontal plane hip angles 48
19. Example right frontal plane hip angles.............................................................. 48
20. Example left transverse plane hip angles 49
21. Example right transverse plane hip angles 49
22. Example left limb vertical ground reaction forces normalized to body
weight................................................................................................................ 50
23. Example right limb vertical ground reaction forces normalized to body
weight................................................................................................................ 50
x
xi
Figure Page
24. Changes in eversion excursion from shoe markers to heel windows markers
condition for each individual subject 52
25. Changes in time to peak eversion from shoe markers to heel windows
condition for each individual subject 52
26. Changes in percent stance at which peak eversion occurred between shoe
markers and heel windows markers 55
27. Change in maximal instantaneous eversion velocity from shoe markers to
heel windows markers 55
28. Changes in maximal instantaneous vertical loading rate from shoe markers
to heel windows marker conditions................................................................... 57
29. Changes in CMC values from shoe markers to heel windows markers............ 57
30. Rear foot eversion curves from a study by Reinschmidt et al. (1997) 72
31. Rear foot eversion curves from a study by Stacoff et al. (2000)....... 73
32. Example of a graphical analysis which demonstrated significant differences
between shoe markers and heel windows markers for the percent stance at
maximum eversion excursion............................................................. 90
33. Example of a graphical analysis which demonstrated significant differences
between shoe markers and heel windows markers for eversion excursion....... 90
34. A graphical example where there mayor may not be a significant difference
between the two marker conditions................................................................... 91
35. A second graphical example where there mayor may not be a significant
difference between the two marker conditions.................................................. 91
36. Sample rear foot eversion and knee flexion curves for both the shoe markers and
heel windows marker conditions.. 105
37. A second example of rear foot eversion and knee flexion curves for both the shoe
markers and heel windows marker conditions 106
Xll
LIST OF TABLES
Table Page
1. Age, number of years running, approximate weekly mileage, arch height
ratios, and injury history for the original thirteen recruited subjects.. 34
2. Summary ofheel windows holes for the final feet used in the study.............. 35
3. Heel windows areas, shoe brand, and shoe type for the subjects used in the
final analysis 36
4. Running speed for shoe markers and heel windows conditions 37
5. Eversion excursion results for both the shoe markers and heel windows
markers conditions........................................................................................... 51
6. Time to peak eversion results for both the shoe markers and heel windows
markers conditions 53
7. Percent stance at which maximal eversion occurred for both the shoe markers
and heel windows markers conditions.. 54
8. Maximal instantaneous eversion velocity for shoe markers and heel windows
marker 56
9. Maximal instantaneous vertical loading rates for shoe markers and heel
windows markers............................................................................................. 58
10. CMC values for the shoe markers and heel windows markers 59
11. Representative kinematic and kinetic results reported from other studies..... 68
12. Critical values for the Model Statistics analysis 86
13. Number of trials used in the single subject analysis for both the kinematic
and kinetic variables 88
14. Autocorrelation coefficients for the three kinematic variables for both the
shoe markers and the heel windows markers................................................... 94
15. Autocorrelation coefficients for the maximal instantaneous vertical loading
rates for both the shoe markers and heel windows markers 95
16. Independent t test and Model Statistics results for eversion excursion 96
17. Independent t test and Model Statistics results for percent stance at which
maximal eversion occurs.......................... 97
X111
Table Page
18. Independent t test and Model Statistics results for the maximal instantaneous
eversion velocity... 98
19. Independent t test and Model Statistics results for the maximal instantaneous
vertical loading rate... 99
CHAPTER I
INTRODUCTION
Running Injuries: Incidence and Costs
Runners get injured at an incredibly high rate. Several review articles have reported
injury rates of 24% to 75% for recreational and competitive runners over a one year
period (Hreljac, 2005; Jacobs & Berson, 1986; Marti, Vader, Minder, & Abelin, 1988;
Milani & Hennig, 2000; Van Ghent et aI., 2007; Van Mechelen, 1992; Walter, Hart,
McIntosh, & Sutton, 1989). A 1986 study of running injury incidence suggested there
were approximately ten million regular runners in the U.S. (Jacobs & Berson, 1986). A
similar number was reported in a 2004 sports participation report by American Sports
Data, Inc. which concluded there were approximately 10.3 million runners in the U.S.,
defining a runner as someone who ran one hundred or more days per year. With injury
rates between 24% - 75%, this suggests anywhere from approximately 2.5 to 7.5 million
individuals suffer a running injury each year.
A study on the etiology of almost two thousand running injuries found that 31 % of
injured runners sought medical treatment, and that most cases required an average of 3.8
medical consultations (Marti et aI., 1988). The same study also reported that 5% of
injuries were serious enough to lead to an absence from work, with the average duration
of absence being 10.1 days. Another prospective cohort study on running injuries found
that 40.6% of injured runners sought medical treatment, and that 25% of these individuals
2had persistent symptoms up to three months after their first consultation (van
Middelkoop, Kolkman, van Ochten, Bierma-Zeinstra, & Koes, 2006). Jacobs and Burton
(1986) reported that 70% of the injured runners in their study sought medical attention for
their injuries, with treatments ranging from muscle strengthening to orthotics to surgical
intervention in several cases. Based on this epidemiologic data, while the exact cost of
running injuries is unknown, the sheer number of individuals involved suggests it is not
insignificant.
One ofthe more troubling conclusions from the epidemiologic studies study is that
previously injured runners carried a high risk of sustaining another injury. For instance,
Marti et al. (1988) found previously injured runners had 74% risk of sustaining a second
running injury while Walter et al. (1989) found that following an initial injury, men and
women were 1.69 and 2.35 times more likely to sustain another injury, respectively.
Even with all the advancements in sports medicine over the intervening years these
numbers have not changed as a prospective study on injury incidence among 844
recreational runners conducted by Taunton et al. (2003) found that 50% of subjects who
reported an injury during their study also had sustained some form ofprior running
injury. Powell, Kohl, Caspersen, & Blair (1986) suggested three main reasons why a
previous history of injury may increase the likely hood ofa second injury. These
included the previous injury not healing completely, the repaired tissue not functioning
as well or having the strength of the original tissue, or the original fundamental cause of
the injury was not addressed leading to re-injury upon resumption of activity. It is in
3addressing this third reason where a biomechanical analysis may prove particularly useful
for the clinician and patient.
Retrospective Studies on Possible Causes of Running Injuries
In 1978 James, Bates, and Osternig (1978) published what has become one of the
landmark papers on overuse injuries in runners. Based on injuries observed in Dr.
James's clinic, the authors concluded that the causes of overuse running injuries fell into
three broad categories: training errors, anatomic and biomechanical factors, and external
factors such as shoe selection or training surface. Training errors and external errors are,
in theory, fairly easy to control or modify since they can be identified through a careful
examination of an individual's training plan. However, the anatomical and biomechanical
factors contributing to injuries are much more difficult to identify. Often these injuries
result from interactions between numerous parameters with no one injury directly caused
by a specific anatomic or biomechanical factor. In the thirty years following James et
al. ' s article there has been a huge volume of literature produced investigating the
contributions of anatomic and biomechanical factors to overuse running injuries. These
studies have generally focused on three areas: anatomic and anthropometric variables,
kinematic variables, and kinetic variables.
Anatomic and anthropometric variables that have been linked to running injuries
include leg length discrepancies, femoral neck anteversion, varus or valgus alignment of
the calcaneous relative to the fore foot and tibia, pes planus or pes cavus foot structure
4under static or dynamic conditions, squinting patellae, Q angle at the knee, and genu
varus or valgus alignment at the knee (Bandholm, Boysen, Haugaard, Kreutzfeldt-Zebis,
& Beneke, 2008; Bennett et ai., 2001; James et ai., 1978; Kaufman, Brodine, Shaffer,
Johnson, & Cullison, 1999; Korpelainen, Orava, Karpakka, Siira, & Hulkko, 2001;
Lysholm & Wiklander, 1987; Rauh, Koepsell, & Rivara, 2006; Reinking, 2006; Ryan,
MacLean, & Taunton, 2006; Van Mechelen, 1992). However, there is no clear
agreement in the literature since other studies have found no link between femoral
anteversion, patella alignment, or a rear foot valgus alignment and running injuries
(Walter et ai., 1989). Further complicating the issue are several studies which have
suggested there may not be any relationship between measures of static lower limb
alignment and running injuries (Lun, Meeuwisse, Stergiou, & Stefanyshyn, 2003; Wen,
Puffer, & Schmalzried, 1997, 1998). The differing findings in these studies leave no
clear picture on exactly how specific anatomic or anthropometric variables may
predispose and individual to overuse running injuries.
A lack of flexibility or range of motion has been suggested to contribute to
overuse running injuries (James et ai., 1978; Kaufman et ai., 1999). However, as with the
anthropometric variables, this too does not find unanimous support in the literaure. Some
authors have found that runners who stretch regularly actually experience more injuries
than those who do not (Jacobs & Berson, 1986). Other authors have found that
individuals who stretch intermittently or infrequently to be at a higher risk than those who
either stretch regularly or never (Walter et ai., 1989). Still other authors have found that
individuals in the most and least flexible quartiles to be more injury prone than those in
5the middle quartiles (Jones & Knapik, 1999). Even though stretching before and after
running is commonly suggested for injury prevention, at this time there exists no
empirical experimental studies to support this claim (Hreljac, 2005).
In addition to anatomic and anthropometric factors, numerous studies have
focused on the potential relationship between kinetic variables overuse injuries in
running. These studies have generally focused on the impact forces imparted to the
runner as the foot makes contact with the ground. For a traditional heel striking runner, a
plot of the vertical ground reaction force contains two peaks, an impact peak
corresponding to the impact of the foot and ground, and an absolute peak generated by
muscular activity during push off. The impact peak is usually 1.5 to 2 times body weight
while the active peak can range up to 3 or more times body weight, depending on the
velocity of the runner (Cavanagh & Lafortune, 1980). Despite being the smaller of the
two, the impact peak has received the bulk of the attention in studies relating kinetic
parameters and overuse injuries.
As with the anthropometric and alignment parameters, there is disagreement over
exactly how, or if, specific kinetic parameters are related to running injuries. Some
studies have reported higher instantaneous vertical loading rates in individuals who have
previously sustained at least one tibial stress fracuture (Hreljac, Marshall, & Hume, 2000;
Milner, Ferber, Pollard, Hamill, & Davis, 2006). Other studies have implicated the
twisting torque between the foot and the ground, a parameter which is reportedly higher
in individuals who have sustained at least one stress fracture as a potential mechanism.
(Holden & Cavanagh, 1991; Milner, Davis, & Hamill, 2006). This higher twisting torque
in previously injured individuals was reported in a study examining retrospective
predictors of tibial stress fracture in female runners (Pohl, Mullineaux, Milner, Hamill, &
Davis, 2008). However, this same study also found that maximal instantaneous vertical
loading rates were not a significant predictor for previous stress fracture history. A
retrospective study by Pohl, Hamill, & Davis (2009) found individuals with a history of
plantar fasciitis demonstrated greater peak vertical loading rates than individuals who had
never experienced that injury, suggesting kinetic factors can affect soft tissues as well as
bone.
However, there are also studies suggesting impact forces are not related to
running injuries. For instance two retrospective studies, one in men and one in women,
found that higher peak impact forces was not predictive of a history of stress fracture for
either sex (Bennell et aI., 2004; Crossley, Bennell, Wrigley, & Oakes, 1999). Further
complicating the issue are two studies by Nigg (1997; 2001) which found that not only
were impact forces not related to the incidence ofrunning injuries, but that runners with
larger vertical loading rates tended to have fewer injuries than those with lower vertical
loading rates. So, as with the anthropometric parameters, the exact relationship of impact
forces to injury in running is not clear.
Lower Limb Kinematics During Running
Even though anthropometries and kinetics in running have received significant
attention, it seems as if the volume of studies on kinematics during running is
7considerably larger. Out of all the studies on limb kinematics, by far the most heavily
investigated parameter has been rear foot motion. During running, the typical impact
force with the ground is around 1.5 to 2 times body weight, though this increases with
'increased running speed. (Cavanagh & Lafortune, 1980). One of the functions of the foot
during this time is to absorb and dissipate some of that impact force. A runner generally
strikes the ground with the rear foot in a slightly inverted position. The calcaneous then
everts, a motion which "unlocks" the transverse tarsal joints allowing the foot to become
flexible and act as a shock absorber (Hintermann & Nigg, 1998; Novacheck, 1988). Due
to the tight articulation between bones of the foot this motion cannot happen in isolation
and is therefore accompanied by dorsiflexion and abduction of the foot. This
combination of rear foot eversion, forefoot abduction, and talocrural dorsi flexion is
known as pronation.
While playing an important role, the foot is not the only shock absorbing
mechanism in the body. The knees also flex during weight acceptance to help absorb
some of the impact shock. However, the shape of the femoral condyles means the tibia
must rotate internally with knee flexion and externally with knee extension. Since the
foot is fixed to the ground and cannot abduct relative to the line ofprogression, the
abduction component of pronation is accomplished through tibia internal rotation.
In theory therefore, maximal rear foot eversion and maximal knee flexion should
occur at the same time. This synchrony appears to be supported by numerous studies of
lower limb kinematics during running, a summary of which can be found in review
article by DeLeo, Dierks, Ferber, & Davis (2004). The authors cited twelve studies
8which examined the timing of rear foot eversion and knee flexion, among other variables,
and found maximum rear foot eversion occuring between 39.3% and 53.9% of stance and
maximal knee flexion occuring between 36.0% and 45.3% of stance (DeLeo et aI., 2004).
Relationship Between Rear Foot Kinematics and Running Injuries
Disruptions to this natural foot motion are thought to be related to numerous
overuse running injuries. Generally, studies relating foot motion to injury have focused
on three different aspects of rear foot motion including the total amount of rear foot
motion, the velocity over which this motion takes place, and the relative timing of the
motion in relation to other limb segments. Different overuse running injuries appear to
be sensitive to different combinations ofthese three aspects. For instance, too much
eversion has been cited as a contributing factor to Achilles tendon injuries due to the
whipping nature imposed on the tendon by excessive calcaneal eversion (Donoghue,
Harrison, Laxton, & Jones, 2008; Paavola et aI., 2002). Too much eversion has also been
cited as a contributing factor to plantar fasciitis for the stress it places on that tissue
(Warren, 1984, 1990). Perhaps most commonly, excessive eversion is often cited as a
contributing factor in the development ofmedial tibial stress syndrome, as it is thought
that greater eversion places increased strain on the soft tissue structures of the lower limb.
(Messier & Pittala, 1988; Reinking & Hayes, 2006; Tweed, Campbell, & Avil, 2008;
Willems, Witvrouw, De Cock, & De Clercq, 2007).
9The injuries described above occur in the soft tissue structures of the limb, with
excessive eversion often cited as a contributing factor. Traditionally higher arches and
smaller amounts of rear foot eversion have been associated with bony injuries such as
stress fractures (Korpelainen et aI., 2001; Williams, McClay, & Hamill, 2001). However,
this too is not always true. For instance, a retrospective study on tibial stress fractures in
female runners found rear foot eversion excursion to be one of the strongest predictors of
injury, with injured subjects having greater eversion excursions than non-injured subjects
(Pohl et aI., 2008). Based on this body of literature it appears excessive rear foot
eversion can contribute to injuries to both the soft tissue and bony structures of the lower
limb.
Other studies investigating the relationship between rear foot eversion and injury
have concluded the actual amount of rear foot eversion may not be as important as the
velocity of rear foot eversion. For instance, a study examining potential biomechanical
factors contributing to patellar tendiopathy in female runners found that, compared to
controls, injured runners had a higher maximal eversion velocity, but similar overall
eversion excursions (Grau et aI., 2008).
Other authors have suggested that it is not just the amount or velocity of rear foot
eversion which matters, but how rear foot motion is coupled with motion at other joints
and segments, specifically the mid-tarsal joints and tibial rotation. As the rear foot everts
the mid-tarsal joints unlock and the foot becomes more flexible. However, beyond mid-
stance through the push offphase the foot needs to be a rigid level for an effective push
off. If rear foot eversion is prolonged then this will not happen and the result will be a
10
push off with a soft and flexible foot. Achieving an effective push off in this scenario
requires more effort from the musculature of the lower limb placing greater stresses on
the soft tissue structures, and potentially cause injury (James et aI., 1978; McClay &
Manal, 1998a; Novacheck, 1988).
An alternative hypothesis involves the relative timing of maximal eversion and
maximal knee flexion. Due to the tight coupling between the rear foot, talus, and tibia,
prolonged rearfoot eversion leads to prolonged internal rotation of the tibia. If the knee
starts extending while the rear foot is still everted then a torsion stress will be placed on
the tibia. This mistiming also increases stress on the soft tissue structures at the knee and
has been hypothesized as a potential mechanism for knee injuries in runners (Stergiou,
Bates, & James, 1999; Tiberio, 1987).
Measuring Rear Foot Motion
While it is not clear whether it is too much rear foot eversion, too great an
eversion velocity, or the timing of rear foot eversion relative to the movement of other
joints and segments, that contributes the most to injury, it is clear that rear foot eversion
is an important component of any running gait analysis. Therefore, there have been
numerous attempts to determine the most optimal measures to quantify rear foot eversion.
Historically, with two dimensional analysis methods, rear foot eversion was quantified by
two markers on the vertical bisection of the shoe and two markers on the vertical
bisection of the lower limb (Edington, Frederick, & Cavanagh, 1990; McClay, 1995;
11
Novacheck, 1988). The relative movements and angles created between these lines
provided an idea of rear foot motion. However, a two dimensional approach may not be
optimal since it has been shown that camera alignment can seriously affect the resulting
measurements (Areblad, Nigg, Ekstrand, Olsson, & Ekstron, 1990). For instance,
changing the camera alignment only 2° can result in a 1° change in the measured
parameter.
Additionally, a two dimensional approach would be susceptible to errors resulting
from out ofplane motions, which is especially concerning given the tri-planar nature of
foot motion during running. One study mathematically estimated that inclining a
segment 9° from the projected plane ofmotion could change measured joint angles by as
much as 40% (Soutas-Little, Beavis, Verstraete, & Markus, 1987). One study on joint
angles during walking suggested out of plane motion was only an issue during toe off, as
rear foot motion was not significantly different during the first 60% of stance when
measured with two dimensional or three dimensional approaches (Cornwall & McPoil,
1995). However, this is not true in all cases. For instance, some individuals may lack
dorsiflexion at the ankle joint. In order to place their foot flat on the ground these
individuals must have compensatory pronation. The main component of compensatory
pronation will be forefoot abduction, resulting in increased out ofplane motion. One
study specifically examining differences in joint angles resulting from different levels of
foot abduction found that as the abduction angles increased so did the differences
between the two dimensional and three dimensional joint angles (McClay & Manal,
1998b). Taken as a whole these studies suggest that, when possible, a three dimensional
12
approach should be used to minimize errors associated with out of plane motions
occurring at the ankle joint.
Even when using a three dimensional approach, researchers are still left with a
large obstacle in that the foot is enclosed within the shoe and therefore directly measuring
the motion of the rear foot is difficult. There are numerous ways researchers have
attempted to deal with this problem. One option is to place the markers directly on the
heel counter of the shoe, and indeed this is what many authors have done. (McClay &
Manal, 1998a, 1998b; Noehren, Davis, & Hamill, 2007; Pohl et aI., 2008; Souza &
Powers, 2009; Stefanyshyn, Stergiou, Lun, Meeuwisse, & Worobets, 2006; Walter et aI.,
1989). However, it is recognized that this marker set up is, by necessity, tracking the
motion ofthe shoe, which mayor may not be representative of foot bone motion. To
explore these differences several studies have used intracortical bone pins to compare
differences between rear foot motions as measured with shoe markers and with the bone
pins (Reinschmidt, van den Bogert, Murphy, Lundberg, & Nigg, 1997; Stacoff et aI.,
2001). The authors found that while the patterns ofmotion were similar between the two
marker methods, the magnitudes of motion were slightly different. While the differences
were unique for each individual, the authors concluded that, in general, the shoe markers
overestimated the true skeletal motion.
While intracortical bone pins may provide the truest measures of bone motion,
they are not a practical method for conducting large studies on running kinematics.
Therefore, researchers have explored other non-invasive methods which may provide
reasonable estimates of rear foot motions. One such method would be using
13
electrogoniometers to record changes injoint angles during running. While each study
may use slight variations, the basic premise of this method involves fashioning a heel cup
over the fat pad of the calcaneous and attaching it to a fixed anchor point on the lower
limb. Eversion and inversion of the calcaneous turns a potentiometer located between the
two attachment points, yielding the rear foot angle across the gait cycle. Several studies
have used this method to examine rear foot kinematics during running. (Derrick, Dereu,
& McLean, 2002; Milani & Hennig, 2000). While the results of these studies suggest this
may be a reliable method for measuring rear foot eversion, the method cannot
simultaneously measure angles in other planes of motion. Recording motion other than
eversion-inversion requires additional electrogoniometers, which can quickly become
bulky and cumbersome for the subject, potentially affecting their normal running
mechanics.
In an effort to avoid external attachments yet still record accurate foot motion,
some authors have used modified running shoes or sandals in their studies. For instance,
O'Connor and Hamill (2004) used a specially constructed shoe with an intact sole and no
upper or heel counter to examine the role of various foot muscles during running. The
shoe was held in place with strong elastic over the dorsum of the foot and an attachment
to a band around the tibia which ran over the medial and lateral malleoli. This custom
shoe design has been used in other studies as well (MacLean, McClay Davis, & Hamill,
2006; Snyder, Earl, O'Connor, & Ebersole, 2009). Other authors have used specially
designed running sandals which allow marker attachment directly the foot to examine
kinematics, kinetics, and forefoot-rearfoot coupling patterns during running gait (Eslami,
14
Begon, Farahpour, & Allard, 2007; Morio, Lake, Guegen, Rao, & Baly, 2009;
Mundermann, Nigg, Humble, & Stefanyshyn, 2003; Nawoczenski, Saltzman, & Cook,
1998).
While these methods allow easy placement of markers directly on the foot, there
are several issues preventing their widespread adoption. Firstly, custom shoes are both
costly and not readily available, and their specificity precludes their adoption on a wide
scale, especially for a laboratory which may work with a variety of subjects. Given the
variation in foot types, foot sizes, and running shoe history, a complete set of custom
shoes seems prohibitive. Secondly, there are issues concerning the external validity of
any study which examines the mechanics of running without using actual running shoes,
specifically the extent to which the observed mechanics in these studies are representative
ofthe subject's actual mechanics. The authors of all these studies reported that subjects
ran "comfortably" or "normally" in the sandals or specialized shoes however, no actual
assessments were made of the subjects running mechanics in standard running shoes.
This is an especially important point considering both the custom shoes and lacked a heel
counter. Though a study by Van Gheluwe, Tielemans, & Roosen (1995) suggested rear
foot motion may be independent of heel counter stiffness, it is traditionally thought the
heel counter is an important component of running shoes which aids in rear foot
movement control (Edington et ai., 1990; Stacoff & Luethi, 1986).
In an attempt at trying to match standard enclosed running shoes with a more
accurate marker placement, some authors have cut holes in the heel counter of the shoe,
effectively enabling them to use a traditional gait analysis marker set up but place the
15
markers directly on the rear foot. It is thought this method is one way to get more
accurate kinematic data. Initially, this method was done first with two dimensional
measuring techniques (Nigg, 1986; Reinschmidt, Stacoff, & Stussi, 1992; Stacoff,
Reinchmidt, & Stussi, 1992). More recently, it is has been adopted for use with three
dimensional analysis and has been used extensively in the Motion Analysis Laboratory
and Running Injury Clinic at the University of Delaware (Butler, Davis, & Hamill, 2006;
Dierks, Manal, Hamill, & Davis, 2008; Stackhouse, McClay Davis, & Hamill, 2004;
Williams, McClay Davis, & Baitch, 2003; Williams, McClay Davis, Scholz, Hamill, &
Buchanan, 2004; Williams, McClay, Hamill, & Buchanan, 2001). Given the relative
importance of the heel counter in stabilizing the shoe and controlling foot motion, it is
unclear exactly how holes in the shoe would affect the measured parameters. Ifthe holes
do not unduly influence rear foot motion, then this may be a method which yields a more
accurate estimate of rear foot motion during running and potentially helps clarify the
differences between shoe motion and rear foot motion during running.
So far only two studies have examined this issue in any depth. Both studies
compared rear foot eversion as measured with both heel windows and shoe mounted
markers (Nigg, 1986; Stacoff et aI., 1992). Their results were similar to the bone pin
studies, and the authors concluded that while the patterns of motion are similar, the shoe
markers overestimate rear foot motion compared to the heel markers. Interestingly, the
difference in motion between the two marker conditions was related to the size of the
holes. As hole size increased, the differences between shoe motion and rear foot motion
increased as well. The only three dimensional study examining the effects of hole
16
windows looked at their effect on heel counter stability, concluding that small windows
resulted in only a 10% decrease in heel counter stiffness (Butler et aI., 2006). However,
to date, no study has used a three dimensional approach to examine the effects of heel
counter windows on rear foot kinematics and kinetics.
Purposes and Hypotheses of the Study
The current disagreement in the literature over the relationship between various
anthropometric, kinematic, and kinetic parameters and running injuries presents several
important considerations for designing a prospective, longitudinal study. First, it is clear
anyone piece of information by itself most likely will yield an incomplete picture or fail
as an assessment of injury potential. Therefore, a prospective, longitudinal study should
attempt to include as many of these puzzle pieces as possible. Secondly, the
disagreement among studies suggests it may be important to consider each runner as an
individual, rather than as a member of a group. The variability inherent in the group may
wash out the significance ofpotential injury markers which can be identified in a single
individual. Lastly, the lack of agreement reinforces the importance of using methods
which provide valid, reliable measures of the parameters of interest.
In light of these suggestions, the purpose of this study was to use three
dimensional analysis techniques to examine the effects of heel counter windows on
kinematic and kinetic measurements during running gait. Specifically, this study sought
to examine differences between shoe mounted markers and heel window markers on
17
common measurements thought to be related to overuse running injuries. The
measurements chosen for analysis were the time to maximum rear foot eversion, the
percent stance at which maximum rear foot eversion occurs, maximal rear foot eversion
excursion, maximal instantaneous eversion velocity, and the maximal vertical loading
rate during heel impact in the stance phase of the gait cycle. An additional measurement
included the variability ofjoint angle curves under each marker condition. A secondary
purpose of this study was to be able to make a conclusive recommendation on which
marker system should be used for a planned future longitudinal study on overuse running
injuries.
Given the bone pin studies and previous two dimensional studies which indicated
shoe markers overestimated true rear foot motion, it was hypothesized in this study that
the maximal rear foot eversion excursion would be less with the heel windows marker set
than with the shoe mounted marker set. It was also hypothesized that the time to
maximum eversion excursion and the percent stance at which maximum eversion
excursion occurred would remain unchanged between the two marker conditions. If the
amount of eversion excursion decreases, but the amount of time and percent stance at
maximal eversion does not change, then rear foot eversion velocity should be smaller in
the heel windows condition. Therefore it was hypothesized that the heel windows marker
condition would indicate a higher rear foot eversion velocity compared to the shoe based
marker condition. Since the amount of rear foot eversion excursion should be less in the
heel windows condition, it was hypothesized that the maximal instantaneous vertical
loading rate would be larger in the heel windows condition compared to the shoe
18
mounted marker condition. Lastly, it was hypothesized that foot motion would be more
variable in the heel windows condition compared to the shoe markers condition.
19
CHAPTER II
METHODS
Subjects
Thirteen recreational runners from the University and local communities were
recruited for this study. Individuals were recruited through personal correspondence, or
referral from individuals familiar with the goals of the study. All subjects reviewed and
signed the informed consent form which had been approved by the University of Oregon
institutional review board prior to participating in the study. A copy of this form can be
found in Appendix A.
The main inclusion criterion for this study was status as an active runner, which
for this study was classified as running 20 or more miles per week. The second main
inclusion criterion was that the individual was currently healthy and not suffering from
any musculoskeletal injuries. A previous history ofmusculoskeletal injury did not
preclude a subject from participating in the study.
Experimental Instruments
Dynamometer
Isometric maximal voluntary torque generation was measured using a BioDex
System 3 dynamometer (Biodex Medical Systems, Inc., Shirley, NY).
20
Force Plates
Three AMTI (Advanced Mechanical Technologies Inc., Watertown, MA) force
plates were used to collect ground reaction forces and moments. The force plates were
located in series in the center ofthe 5 meter runway, sampling at a rate of 1000 Hz.
Motion Capture System
An eight camera motion capture system was used to record three dimensional
(3D) marker trajectories, with the sampling rate set to 200 Hz. The motion capture
system was calibrated prior to each subject's testing session as per the manufacturer's
instructions.
Data Collection and Experimental Procedures
Muscle Strength Measurements
Subjects were tested bilaterally for maximal isometric torque generated in hip
flexion, hip abduction, ankle plantar and dorsi flexion, and ankle inversion and eversion.
It is recognized that there were multiple muscles involved in each contraction and no
attempt was made to separate out individual muscles since it was thought that the muscles
would act as a group during running scenarios. These specific muscles and motions were
selected based on their importance for running and their potential link to running injuries.
For instance, several studies have highlighted the importance of the ankle musculature in
controlling the motion of the ankle and foot, their role in stabilizing the foot during
21
stance, and their contributions to generating force. (Christina, White, & Gilchrist, 2001;
Kibler, Goldber, & Chandler, 1991; Lun et aI., 2003; Reber, Perry, & Pink, 1993; Scott &
Winter, 1990). There has also been a large volume of literature published investigating
the relationship between hip muscle strength and overuse injuries, especially at the knee
(Grau et aI., 2008; Ireland, Willson, Ballantyne, & McClay Davis, 2003; Noehren et aI.,
2007; Powers, 2003; Souza & Powers, 2009).
Isometric torque of the hip flexors was measured with the subject standing next to
the dynamometer. The resistance pad was placed approximately in the middle of the
thigh and center of rotation on the dynamometer was aligned with the greater trochanter
of the test leg when the subject was standing in an upright position. This upright position
was considered to be 0° of hip flexion and the isometric test was performed at 300 ofhip
flexion. The subject was allowed to place their hands on the dynamometer and an
additional supporting structure for balance but was instructed to maintain an upright
posture, not bend at the waist, and attempt to isolate the hip flexors.
Isometric torque of the hip abductors was measured with the subject standing
facing the dynamometer. The resistance pad was placed over the lateral aspect of the
thigh, approximately mid-way between the lateral femoral condyle and the greater
trochanter. The center of rotation on the dynamometer was aligned with the hip joint
center estimated to be several centimeters lower than the anterior superior iliac spine in
the sagittal plane. This upright position was considered to be 0° of hip abduction and the
isometric test was performed at 10° of hip abduction. The subject was allowed to place
their hands on the dynamometer for balance but was instructed to maintain an upright
22
posture, not internally or externally rotate the leg, and attempt to isolate the hip
abductors.
Isometric torque of the ankle dorsi and plantar flexors was measured with the
subject in a seated position and the knee flexed approximately 200 • A supporting pad was
placed under the popliteal area and a restraining strap was placed across the subject's
waist. The isometric torque was tested with the ankle joint in a neutral position, as
indicated by an estimated 900 angle between the foot and tibia. The center of rotation of
the dynamometer was aligned with the talocrural joint as estimated by the center of the
lateral malleolus.
Isometric torque of the ankle invertors and evertors was measured with the subject
in a seated position and the knee flexed approximately 200 • A supporting pad was placed
under the popliteal area and restraining straps were placed across the subject's waist and
across the thigh. The subject was instructed to perform the motions only at the ankle
joint and avoid any rotation at the knee. The center of rotation of the dynamometer was
aligned with the approximated subtalar joint axis.
For all tests the subject was allowed to perform a familiarization trial. For the
actual tests the subjects were asked to push as hard as possible against the resistance pad
for 5 seconds. This procedure was repeated two more times for a total of three trials.
The subject rested for 5 seconds between the trials.
23
Clinical Exam
A clinical exam was performed by Dr. Stan James, an orthopedic surgeon with
extensive experience in evaluating running injuries. The exam is a modified version of
one he uses in current practice as part ofhis examination of an injured runner, and has
also been previously used in research studies on runners (Stergiou et aI., 1999). The exam
focuses on general structural alignment, joint mobility, and flexibility of the lower
extremity. The complete examination can be found in Appendix B.
One portion of the exam classified the arch structure of the subject using a ratio of the
height of the dorsum ofthe foot at 50% ofthe full foot length divided by the truncated
foot length. Figure 1 illustrates where these measurements were taken. This method has
been shown to have high reliability and validity in both 10% and 90% weight bearing
positions (Williams & McClay, 2000). In a previous study sampling 200 individuals, the
mean arch ratio was 0.316. Ratios lower than 0.274 are considered "low" arches, and
ratios higher than 0.356 are considered "high" arches (Williams, McClay, Hamill et aI.,
2001).
.<
FL~ ~_-t
TFL\
24
Figure 1. Schematic illustration of where the measurements were taken for the arch
height index. The ratio is the height of the dorsum ofthe foot at 50% full foot length
(FL) divided by the truncated foot length (TFL). The truncated foot length is measured
from the most posterior aspect ofthe calcaneous to the 1st metatarsal-phalangeal joint.
Image adapted from Willams & McClay (2000).
Motion Analysis
A total of 34 reflective markers were placed on bony landmarks of the subject,
using a modified Helen Hayes marker set (Kadaba, Ramakrishnan, & Wootten, 1990).
For the pelvis segment markers were placed on the sacrum midway between the posterior
superior iliac spines, and bilaterally on the anterior superior iliac spines. Marker
placement for the thigh, shank, and foot were similar bilaterally. For the thigh markers
were placed on the medial and lateral femoral epicondyles and on the thigh collinear
between the lateral femoral epicondyle and the greater trochanter. The anatomic
coordinate systems for the thigh and shank were defined as per International Society of
Biomechanics (lSB) recommendations (Wu, 2002). The thigh anatomic coordinate
system was defined using the two femoral epicondyle markers, the hip joint center and
25
the thigh marker, as per the ISB recommendations. The hip joint center was defined
based on the anthropometric measurements of the subject (Vaughan, Davis, & O'Connor,
1999). The anatomic coordinate system for the shank shared the medial and lateral
femoral epicondyle markers and had additional markers on the medial and lateral malleoli
and a marker on the medial shank collinear with the medial malleolus and the medial
femoral epicondyle markers. A static calibration trial was collected and the medial
malleoli and medial femoral condyle markers were removed for the actual gait trials.
Since the subject wore shoes for both conditions, the foot markers were placed on
the shoe. Subjects wore their own running shoes that they ran in every day. A toe
marker was placed approximately between the first and second metatarsal shafts. For the
markers on the shoe condition, three markers were placed on the heel counter of the shoe,
one on the lateral aspect, and the other two along the vertical bisection of the heel
counter, with the midpoint of these two markers level with the toe marker. This marker
placement on the heel has previously been used to collect rear foot kinematic data
(McClay & Manal, 1997, 1998a, 1998b; Noehren et aI., 2007). The anatomic coordinate
system was defined using the midpoint ofthe two heel counter markers, the toe marker,
and the ankle joint center, as defined as the midpoint between the medial and lateral
malleoli.
For the heel windows condition holes were cut into the heel counter ofthe shoe
and the marker bases were attached directly to the shoe, in the same locations as the
markers on the shoe condition. The markers were then screwed into the based from
outside the shoe. The size of the holes was kept small, approximately 1.5 - 2.0 em. in
26
diameter since holes larger than this have been shown to affect rear foot movement
(Stacoff et aI., 1992). The same drill bit was used to cut each hole in an attempt to make
them as similar as possible, however, small adjustments were made on a subject by
subject basis. Previous studies using similar methodology have used an Instron materials
testing device (Instron Corp., Norwood, MA) to examine the effect on heel counter
stability and have found that the holes resulted in only a 10% decrement in heel counter
rigidity (Butler et aI., 2006; Stackhouse et aI., 2004).
The placement of the anatomic and tracking markers are shown in Figure 2 and
the heel windows marker set up can be seen in Figure 3. A detailed description of the
methods used to define the anatomic and tracking coordinate systems for each segment
can be found in Appendix C.
The running protocol involved the subject running laps in the Motion Analysis
laboratory. Each lap was approximately 25 meters long. Data were collected as subjects
passed through a 5 meter region in the center of the capture volume. Each subject ran at
a self selected speed, though their speed was recorded using two photocells placed 5
meters apart. The three force plates were located in series in the center of the capture
volume. The subject was instructed to return to approximately the same starting position
after each lap to maintain consistency, however they were also instructed not to alter their
stride to hit the force plates. Therefore, passes resulting in a clean force plate strike were
used for both kinematic and kinetic data analysis, while those without a clean force plate
strike were only used for kinematic analysis. Each subject completed between 30 and 40
passes under both conditions.
27
Data Analysis
Raw marker trajectories were identified using the EvaRT 5.0 motion capture
software (Motion Analysis Corp., Santa Rosa, CA). The raw trajectories were low pass
filtered using a Butterworth filter with a cutoff of 8 Hz. All trials were analyzed only for
the stance portion of the gait cycle from heel strike to toe off. In cases where the trial
was a clean force place strike these frames were identified by looking at ground reaction
force curves. In cases where the trial was not a clean force plate strike these frames were
identified based on visual analysis and the vertical coordinates of the heel and toe
markers. The heel strike and toe off frames, and which force plate was cleanly hit, were
manually recorded for each trial. The filtered marker trajectories and the analog
photocell and force plate data were then exported as ASCII files which were used to
calculated the joint angles.
Joint angles during the stance phase of gait were calculated using a custom
LabView program (National Instruments, Austin TX). A detailed description of the
methods used can be found in chapter seven of David Winter's biomechanics book
(Winter, 2005). Three dimensional angles for the ankle, knee, and hip joints were
determined using ajoint coordinate system (Grood & Suntay, 1983). Cardan angles were
used to define the joint angles across stance, referencing the movement of the distal
segment to the proximal segment. For the hip and knee joints the order of rotations was
ZXY, corresponding to flexion-extension, ab/adduction, and internal rotation. For the
ankle joint the order of rotations was ZYX, corresponding to dorsi/plantar flexion,
28
inversion-eversion, and ab/adduction. Zero degrees for all joint angles was assumed to
be when the coordinate systems of the proximal and distal segments were aligned.
Once the joint angles across stance were calculated the discreete variables of
interest were then identified, again using a custom LabView program (National
Instruments, Austin, TX). These included several parameters which are often reported in
studies on running kinematics including maximum rear foot eversion excursion, time to
and percent stance at which maximum rear foot eversion occurs, the maximal
instantaneous eversion velocity, the maximal instantaneous vertical loading rate from the
ground reaction force, and a measure of the variability or repeatability of the joint angle
curves between trials. Maximum rear foot eversion excursion was defined as the absolute
difference between the rear foot angle at heel strike and the point of maximum rear foot
eversion. The time to peak eversion referred to the time, in seconds, from heel strike to
peak rear foot eversion while the percent stance of peak eversion referred to the percent
stance at which peak eversion occurred. Maximal instantaneous rear foot eversion
velocity was identified as the maximal point on the plot of the first derivative ofthe rear
foot eversion curve. Maximal vertical instantaneous loading rate was determined as the
maximal loading rate between ten and ninety percent of the vertical ground reaction force
impact peak (Milner, Ferber et aI., 2006; Williams et al., 2004; Williams, McClay,
Hamill et aI., 2001). The variability and repeatability of the joint angle curves were
assessed using a coefficient of multiple correlation (CMC) (Kadaba et aI., 1989). A
CMC value of one indicates the joint angle curves are identical each trial and the lower
the CMC value the more variability in the kinematic patterns.
29
For each subject, the left and right feet were analyzed separately. Though the
limbs belong to the same subject, several studies have demonstrated some level of
kinematic and kinetic asymmetry during running is common in most individuals
(Vagenas & Hoshizake, 1992; Zifchock, Davis, & Hamill, 2006). Thus, it was thought
each the left and right feet of each subject may not respond in the same manner to the two
marker conditions. Additionally, since overuse injuries tend to occur in one limb, not
both at the same time, it has been hypothesized larger levels of asymmetry might predict
or predispose an individual to injuries on one side of their body and not the other
(Zifchock, Davis, Higginson, McCaw, & Royer, 2008). Therefore, this study choose to
examine analyze each subject's left and right foot independently.
Statistical Analysis
A minimum of five trials per subject were averaged to create an ensemble average
for that individual for all the parameters of interest. Dependent observations t tests were
used to compare differences between marker conditions for all dependent variables. To
investigate the possibility of changes being running speed dependent a dependent
observations t test was used to compare average running speed from the shoe markers
condition and the heel windows markers condition. The significance level for all
statistical tests was set to a = 0.05 a priori. All statistical tests were performed in
Statistical Package for the Social Sciences (SPSS) (SPSS Inc., Chicago IL).
30
• •
.- •
~
ea···
• •
Figure 2. The markers used to develop the anatomic coordinate system. For the tracking
marker set the only difference was the medial femoral condyle and medial malleolus
markers were removed.
31
Figure 3. Picture of the heel windows marker placement and set up of the heel windows.
The marker bases were attached directly to the ca1caneous and the markers were screwed
in from the outside once the subject had put the shoe on their foot.
32
CHAPTER III
RESULTS
Subject Data
Subject data including age, number of years running, weekly mileage, arch height
ratios, and injury histories can be seen in Table 1. The average age of the subjects was
25.46 years (± 8.86 years) and the average number of years running was 7 years (± 3.92
years). All runners met the inclusion criteria for weekly mileage, with an average weekly
mileage of 40.31 miles (± 11.03 miles). The average arch height ratios for the runner's
feet were 0.321 (± 0.022) for the left and 0.317 (± 0.425) for the right, however, it should
be noted that there was a wide range of arch heights. Most ofthe subjects had sustained
at least one of the more common running injuries (Jacobs & Berson, 1986; James et ai.,
1978; Rauh et ai., 2006; Taunton, 2002, 2003; Van Ghent et ai., 2007).
Though thirteen subjects (twenty six feet) initially participated, only eleven
subjects were used in the final analysis. One subject had to leave in the middle of the
testing session and did not return to complete the testing. During testing it was
discovered that a second subject was currently injured so their data were not used in the
study. Additionally, there were errors and difficulties with the marker tracking and data
collection on two subjects. For these two only one of feet was used for data analysis.
Therefore the final number of feet analyzed for this study was 20 (n = 20).
33
Heel Windows Dimensions
Table 2 shows the average size of the heel windows for the entire group of
subjects. Individual sizes of heel window holes, brand, and type of shoe for each subject
can be seen in Table 3. The average heel window dimensions were 2.13 cm.2 (± 0.31
cm.) for the bottom hole and 1.87 cm? (± 0.28 cm.) for the top hole. Since the subjects
wore their own shoes there was a wide variety of shoe brands and types. The type of
shoe was a general classification based on the sole construction. Shoes with dual density
midsoles or visible plastic stabilizing devices were classified as stability shoes. Shoes
without these devices were classified as cushioning shoes. One subject used a light
weight training or racing flat as their everyday training shoe, and this is reflected in the
shoe classification.
34
Table 1. Age, number of years running, approximate weekly mileage, arch height ratios,
and injury history for the original thirteen recruited subjects.
Arch Arch
Years Weekly Height Height
Subject Age Running Mileage (L) (R) Injury History
25.46
7.0 40.38 .322 .318
Mean (± 3.92) (± 11.17) (± 0.02) (± 0.04)
R03 25 7 35 - 40 .352 .337 None
L femoral neck stress
R04 20 5 60 .323 .313 fracture, L fibula stress
fracture
R05 27 15 50 .330 .360 R ankle sprain, Lhamstring strain
R06 42 5 35 .352 .385 none
R07 46 8 40 .353 .373 L piriformis strain
R08 28 6 30 .300 .297 R patellar tendonitis, Lgroin strain
R09 25 5 40 .306 .292 R tibia stress fracture, Rfoot generic sprain
RIO 21 7 35 - 40 .321 .301 R tibia stress fracture
Rll 20 3 25 - 30 .301 .256 R IT band strain, Rpatellar tendonitis
R12 19 8 60 .280 .244 R IT band strain, L 3
rd
metatarsal stress fracture
R13 18 4 25 .325 .325 L IT band syndrome
R14 19 3 30 - 40 .313 .300 L tibia stress fracture
2 x R tibia stress fracture,
R 15 21 15 45 - 50 .324 .349 R hamstring strain, R IT
band syndrome
Note. Subjects 4 and 9 were not used in the final analysis. Only the right foot was used
for subject 10 and only the left foot was used for subject 12.
35
Running Speed
Due to unknown alignment issues, the photocells did not record running speed for
every subject. In these situations the average forward velocity of the point representing
the center ofmass for each subject's pelvic segment was used to indicate the running
velocity for that trial. This point was selected since it was thought to approximate the
location of the subject's center of mass. Appendix C specifies how this point was
determined.
Running speeds for both the shoe markers condition and heel windows marker
conditions are shown in Table 4. Group mean speeds and speeds for each individual
subject are shown. There was no significant difference in running speeds between the
shoe markers and heel windows marker conditions, t(10) = 0.717,p = .490.
Table 2. Summary of heel window holes for the final feet used in the study.
Variable
Upper Heel Window Area (cm2)
Lower Heel Window Area
(cm2)
N
20
20
M
1.87
2.13
SD
± 0.28
± 0.31
36
Table 3. Heel window areas, shoe brand, and shoe type for the subjects used in the final
analysis.
Subject Bottom Hole (cm) Top Hole (cm) Brand Type
R03 L 2.00 1.77
Asics Stability
R03R 2.00 1.77
R05 L 1.88 1.77
Nike Cushioning
R05R 1.88 2.14
R06L 1.77 1.77
Brooks Cushioning
R06R 1.65 1.77
R07L 2.00 2.00
Nike Racing Flat
R07R 2.36 2.12
R08L 2.59 2.36
Addidas Stability
R08R 2.27 2.51
RIOR 2.01 2.12 Nike Cushioning
R11 L 2.27 1.47
Asks Stability
R11 R 2.20 1.76
R12 L 2.83 1.65 Mizuno Stability
R13R 2.51 1.77
Nike Cushioning
R13 L 2.51 1.77
R14L 2.14 2.00
R14R 1.77 1.32
Saucony Stability
R15 L 1.88 1.77
Nike Stability
R15R 1.98 1.76
37
Table 4. Running speed for shoe markers and heel windows conditions trials.
Shoe Markers Heel Windows
Subject n M SD n M SD
Mean 11 7.37 0.62 11 7.32 0.82
Running Velocity Running Velocity
R03 7.17 6.73
R05 6.85 6.80
R06 7.75 8.23
R07 7.12 7.12
R08 6.59 6.16
RIO 6.81 6.61
R11 6.77 6.52
Rl2 7.58 7.58
R13 8.05 7.94
R14 8.13 8.46
R15 8.32 8.32
Note: Running speed presented as minutes per mile pace.
Kinematic Variables Results
Figures 4 through 21 show an example ofjoint angles across the stance phase for
one subject (R06) for the shoe marker condition. The general movement patterns were
similar across subjects and between conditions, with slight variations. All individuals in
the study demonstrated some asynchrony between the joint angle patterns on their left
and right feet, which supports the decision to analyze each foot individually. Figures 22
and 23 show an example of one subject's (R06) vertical ground reaction forces for the
38
left and right limbs, respectively. All subjects demonstrated similar magnitude vertical
ground reaction forces with impact peaks between 1.5 and 2 times body weight and
active force peaks of2.5 to 3 times body weight.
Mean eversion excursion for all feet under both conditions, eversion excursion for
each individual foot under both conditions, and the percent change between the shoe
markers and heel windows markers conditions is shown in Table 5. There was not a
significant difference in eversion excursion from the shoe markers condition to the heel
windows condition, t(19) = -0.296,p = .770. Some feet demonstrated greater eversion
excursion in the heel windows condition, while some feet demonstrated less eversion
excursion. To more clearly illustrate this mixed response, a graph showing the changes
in eversion excursion from the shoe markers to the heel windows markers can be seen in
Figure 22.
Mean time to peak eversion for all feet under both conditions, time to peak
eversion for each individual foot under both conditions, and the percent change in time to
peak eversion between the shoe markers and heel windows markers conditions is shown
in Table 6. There was not a significant difference in time to peak eversion between the
two marker conditions, t(19):::: -0.291,p = .774. Some feet demonstrated a shorter time
to peak eversion in the heel windows conditions while others had longer times to peak
excursion. The difference in times from the shoe markers to the heel windows markers
conditions can be seen in Figure 25.
The mean percent stance at peak eversion for all feet under both conditions, the
percent stance at peak eversion for each individual foot for both conditions, and the
39
percent change in percent stance at which peak eversion occurred from the shoe markers
to the heel windows marker conditions can be seen in Table 7. The difference in the
percent stance at which peak eversion occurred in the shoe markers and the heel windows
conditions was not significant, t(l9) == - 1.089, p == .290. For some feet peak eversion
occurred earlier in stance with the shoe markers compared to the heel windows while
other subjects had peak eversion occur later with the shoe markers compared to the heel
windows. The changes in percent stance at which peak eversion occurred between the
two conditions are shown in Figure 26.
The mean maximal instantaneous eversion velocity for all feet under both
conditions, the maximal instantaneous eversion velocity for each individual foot under
both conditions, and the percent change in maximal instantaneous eversion velocity
between the two conditions can be seen in Table 8. There was not a significant
difference in the mean maximal instantaneous eversion velocities between the shoe
markers and heel windows conditions, t(l9) == 0.837,p == .413. For some feet the
maximal instantaneous eversion velocity increased in the heel windows conditions while
for other subjects it decreased. Changes in maximal instantaneous eversion velocity
between the shoe markers and heel windows markers are shown in Figure 27.
The mean maximal instantaneous vertical loading rate for all feet under both
conditions, the maximal instantaneous vertical loading rate for each individual foot under
both conditions, and the percent change in maximal instantaneous vertical loading rate
between the shoe markers and heel windows markers conditions is shown in Table 9.
There was not a significant difference in the mean maximal instantaneous loading rates
40
between conditions, t(16) = 0.780,p = .447. Some feet experienced higher maximal
instantaneous loading rates in the heel windows condition compared to the shoe markers
condition while other experienced lower ones. Changes in maximal instantaneous
loading rate from the shoe markers condition to the heel windows conditions are shown
in Figure 28.
The mean ankle frontal plane CMC for all feet under both conditions, the ankle
frontal plane CMC for each individual foot under both conditions, and the percent change
in ankle frontal plane CMC values from the shoe markers to the heel windows markers
for each subject is shown in Table 10. Overall, the heel windows marker condition
resulted in significantly lower CMC values than the shoe marker conditions, indicating
higher variability in the ankle joint frontal plane angle curves, t(19) = 3.56,p = 0.002.
While almost all feet demonstrated greater variability in the heel windows markers
condition the magnitude of the change varied from foot to foot. A graphic illustration of
the changes in CMC values from the shoe markers condition to the heel windows marker
conditions is shown in Figure 29.
41
L-11 (. ,- .tt,-,' F'l'- - - ,'. -, ':1- 6 . '"'Tl- ,~,e ,.':.dgl .•:u .:ille ...u.JJ.r. e ... ~lt. e.:,
+
"'-'
.~I,",
25,
0I====k::;~::ti~~~J~===~===~-- ......::": ......../<~¢.~:-- '~~\ ..
10,0E~~..~r·""-::'--I--""f\~~\:'~~"_~-\~...:.,' -r---l~.. -----_.-.- . ~
~\... ····1.
0, [I+----+---+-----+--~,~~..0.~,l'\~:...,....·+-.....----1
\'\~ ""
~~ \~~~.\
.~.. -10, (I +------+-----+-------ir----""f-~~ .....o::__-__1
,..... .::~....\.;"
-::: ..... 'J..~ •. ""\
.3 -20. (I -+-----+----4-----+----...p........: ;;,,;';1.,:~~>-,~:i'e...::--...•:-1..,
~ ....... ,~... '.
"-
'--
100.060,040.020.0
-30.0+------+-----+-------ir------+---__1
0.0
Figure 4. Example left sagittal plane ankle angles. Dorsiflexion is positive, plantar
flexion is negative
:37.('
+ 30.0
'-'
.~
20,0'.",M
C'
0
10.0
.-.
0.0
-'
....
...
-10.0~
.....
~
-22.0
0.0 20.0 40.0 60.0 eo,o 100.0
Figure 5. Example right sagittal plane ankle angles. Dorsiflexion is positive, plantar
flexion is negative.
42
Left- Frontal Plane 1.61.nkle ,Angles
11.0
+ 7 c-.--,
--'
~. 5.0...
...
>-4
.-:' ~
.;;.. ,_I
0.0
-
.-:- co
.::... I ,_I
.-
-5.0~.
~
-9.0
0.0 20.0 40.0 60.0 :::0.0 100,0
Pell:ent Shl\,:e
Figure 6. Example frontal plane ankle angles. Rear foot eversion is negative, inversion
is positive.
100,0:::0.060.040.020.0
5, 0-ii8~---+-----+-----t---_M~~---__l
o. 0+-~~-__+-----+-------1f_~~+-_+_---___l
-5. 0-t---~~t_----+_---_:'!1~~:...-----1t----__l
1c- -_'. U-t-----+-----+-----t-------1t-""-: 'Vt"-----t
10,0+-----+----+-----+------,o'~~-__
'-',
~'jght Frontal Plane ,Ankle Angles
25.0 -r------r------r-----r------,r------,
~.
=>-4
r-, 20.0 +-----t-------+-----t-----+---':!"7f1l8lOl
+
~ -10, 0-+-----P~~E!:S~;;:=~~I-----t---I
1c- -- -_', U+-----+-----+_------1t-------jf-----l
0,0
Pell:ent Shl\':e
Figure 7. Example right frontal plane ankle angles. Rear foot eversion is negative,
inversion is positive.
43
L.::.ft TI··:'t·l·:'T~et··:,f. F'l·:. t ·1f. 6t·.1;1f. 6t·1(Tlf.·~. 1_,. I:J.J. ~I I) _' ~. _' I::J.J. _' ~ 1.l.Lr:..l _. ~ I..l 1::- _ ..
.:;. 24. 0-r-------,-----,-----~---__r----_,
...
I·~:: 22.04------+------+-------1-----1-------1
.- ..--...-
... ...,., .....-..,:::.... 8:.
~.~ .:•...~~~_ .. Pl;:;~'---.
W 20. 0 +---=------l~---_+----_b~..,..._=_=:_!~~~'a!=:'.,
.....- --- ....... ...::, ;r:?'4A'.....- -~~.~~~
-;::t;:::::.:===--.L:= _.--- ",::: ...... ~",:. '.. ,
...: --.::... .......:::. .~ ,-"" . \."x'.~:=:-- '''-;'' - - •.~~ v.;- '. ~"2:
I 1::: .0 v·~::.::.=--.::;:: . .::It" -~:••, '''..'-,
'._.' ~~:- - - --= --7;' ""-'"
... ....
~ 16.0 +...'-----+-----_+-------I-----t-------i
100.0:::0.060.040.020.0
"E
- 14.0-+-----+----+-----+-----+------1
0.0
F'elcent St.?1n:::e
Figure 8. Example left transverse plane ankle angles. Internal rotation is
negative,external rotation is positive.
.'-', -1 7 n..,..-------r-----,----r-----r------,+ .. -
+------f------+------+-----+------I
-'.
...I:' ~ "'"
~ -20. 0+-----l------+---.....,.4--~--+_---~
... ~:::.- "-" .
~.~ --~-- '-. '.
r.' -22.0r--I-:~~~~~~~~I--l....... ..----;..- ~ ~ \.. \ .
..... ~._ .....~ ;:';~i. t ...... \.i~~~~-.~~..~~':A:'!:·~~~ ~.~'~""'~.~.~...~\~;.:==~~-~·4 n ~. - . .::~:~-t.:. '~"""'"'" ~.;;,
.-:.-...:..:..lI=: I.-" ••••.-....... ":~::~~i. _ ___~.
._.' -26. 0 +:::-.--!.-~~--t-•.-...:='-,...:...---+------+-~::lo,_........~:-.;:.~"-.~-;;:::~.:~ .."'-.--i
0 _----... r..= ::::::=:_.
~ -2:::.0 +0--==---+-----_+_----_+------+------1
100.0:::0.060.040.020.0
...
~
- -30.0+------+------+------+------+-----1
0.0
Figure 9. Example right transverse plane ankle angles. Internal rotation is negative and
external rotation is positive.
44
Left. ~~agittal Plane Knee ~6~ngles
46,0
,-, 40.0+
,_."
~.: 35.0
'1'~ 30.0
-
25.0
I
'-' 20.0
-:.~
~ 15.0 .'
9.0
0.0 20.0 40.0 60.0 100.0
P~ll:~nt Shll':€:
Figure 10. Example left sagittal plane knee angles. Flexion is positive and extension is
negative.
F' ' ~1· t ,-. - a' tt,-.1 F1l- . - l,:""· - - ... - al- .,d:='. 1, .-:=' d,...1 , dl a!le r·Jlee H!l,... e:.)
- -' -'C4 -,_,' .U
50.0
,-,
+ 45.0-'
;.~
40.0'1'r;::
35.0
,-, 30.0I
'-'
-
25.0..
r',
~ 20.0
15.0
0.0 20.0 40.0 60.0 :::0.0 100.0
P€: I'': €: n t S t ,:S.no: €:
Figure 11. Example right sagittal plane knee angles. Flexion is positive and extension is
negative.
45
L-ft F .- 0 t -.1 F'l- . - t,:"' - - - .', • -r1- ..,e ' I UU ~:1l a,Ile Lllee ..uJ1e:. e::·
2,0....------r-------,-----,----.-------,
100.0ao.o60.040.020,0
-10, 0 +-----+~~~""~-~....-:,....~;,g.;..;.-.--+-----{~
-=( -12,0 4-----1----...:=f=--O;-=---+------+-----\
-14.0+-----+-----+-----+------+-----1
0.0
0.04-----1------+-----+-----+-------1
,-,
I
'T;:1 -2. 0 +-----t-----+-----+-----+--~li"":;:~
..c,
.=( -4. 0~~~__,_.,.-+-----+----+------t-7I5+::~---i
Pe:rcent St,:altce
Figure 12. Example left frontal plane knee angles. Abduction is postitive and adduction
is negative.
D ' -rl- t F 0 - - t.-.1 F'l- . - t,:"'· - - ,', -r1- ~F1e:. 1 ' I uU ~ a,Ile f·..Uee ..uJ1e:, e::.
---==-.--
2.0
+ 0.0
'-'
oTj
-2,0
..c'
o=(
-4.0
-6.0
'-'
-a.o
'T;:1
-10,0'T;:1
-=(
-13.0
0,0 20.0 40.0 60,0 eo.o 100,0
F'·:t· .. ·:ttt ·-::t"'~t .. ·~ I_~ ~" I.I. I_~
Figure 13. Example right frontal plane knee angles. Abduction is positive and adduction
is negative.
46
L-ft T·- . ~ T T - • ~ - F'l- . - t,:"" • • - t. • "1- ~'t', t au~, 1.1 ex ~,e aUe f':..tlee "uJ1E, e~·
:::0.0 100,060.040.020.0
,-, 49.0 ..,.------,----..,------r-------,-----.,
~.. 45. (I +====~==.=,,:-:~>~;.~>~.. ~ i;.:>'i-~.--i_?-~~__..;~~~2=+.:...;..=...'"==~
42, 5+-----+--.""7...ff".~'Jr. ..,e;;."':>/'~----+--~-......::::::II.....~~~;:--.;),'"~•..-.--t~ 4 - - +--__~f-""'-§'S!.:.JIr~ ----.-+----+-----._f'~~:-...~~.: ~. .....~"':-""'''''rl
- 'u, U /-;i ."- ",.,.~,:- .....
:37, 5+---~_M~~-L---+-----_+_---_+-4~~l
,_....~:,~ ~ ,/ 'S~~
'~" 35.0 .-' ""':1'/ ... "'...~::
"-' ':1'--:1 t:' ~--~.,~~ulT.~r-.. -..-"l-----+----+----_+---~
- ..'0::... '.' l'~~~ -:"" ,0""~:~ _. • ~~.JE;,.-,...:-/_+----~----t_---_+---_"1
-. :.:iU.U V, ,· '
.... ~)7 ,co.' ~-=-..------i----_+_----t_---__+---_1
:.~ -
W 25,0+-----+----+-----+----+-----1
0,0
F'. l' ~ . l\t '::' t :. ..\ ~ .':' I_,=, ...,.~ I_t:::::
Figure 14. Example left transverse plane knee angles. Internal rotation is positive and
external rotation is negative.
100.0:::0.060.040.020.0
Co' "l·t T·· . ~TT··~· F'l- . - t·:""· - - .'.• "1- ~r'·.lE. 1 ' t au::· vex::, e aUe f':..tlee "uJ1E. e::·
.-
----..
-
,
-
-
- ".
~...- ~>..,.. -..::::
,.. ~~._-../7~ -- , ..
~ ""~~.,.~~J:O'"
'. ~..
~ .
ti7
~'
4.0
0,0
-5,0
-215
1co •. '_"U
",_.'
"-'.
+
-~
-19.0
0.0
- -1'-::' co(. .;..., ,_I
~
.-;-.. -10.0
"._,'
P~l'l:~nt St'll\,:e
Figure 15. Example right transverse plane knee angles. Internal rotation is positive and
external rotation is negative.
47
L-ft ,-. - " tt -, 1 F'1- -- Hi - 6· ,,1- .'·e .•-':'dE)·~:1l alle 1-' ... ~le:. e.:·
42.0 +=====:::+======+=====::::t========t======i
10.0+-----+-----j.----+---:.~~..,-_t_---__i
20. 0+-----I-------t--~~ ~-----+-------1
,
'-'
.'-',
..
p',
'1'r;::
-:.; O. 0+-----+-----+------i'----~~~~-____i
f,z.J
100.0:::0.060.040.020.0
-12. 0:t======+======::t======1=======+=====~
0.0
Figure 16. Example left sagittal plane hip angles. Flexion is positive and extension is
negative.
F" 'lot '-'- 'tt-,1 F'1- -- Hi - ,'.. n1-'..j~. l· .-::'dgl.~:1l a,l1e 1.1 HIle:, e::,
40.0 -r------r-----r-------r-------,..---__
100.0eo.o60.040.020.0
o. 0 +-----+-----+----~f__----:;;~~~.,...._____l
10.0+-----+- +- _+_......=:o~~~+_---__1
20. 0+----~f__----+--....:IiI~l+.~---+_---__l
1·... - +------11-------+-----+------+-------1- ·.:o.U
0.0
..p',
.1,1
r;::
.'-'.,
t--~·-~·~-~-~~~~;:::;.;L_=__--~L---J----J:;.. 30.0
',-,'
Pel".::ent Shl'll::e
Figure 17. Example right sagittal plane hip angles. Flexion is positive and extension is
negative.
48
Left, Frontal Plane Hip i~Jlgles
2.0
+ 0.0'-'
'-r:1
..c,
-2,0.:(
-4.0
-6,0
~
-8,0
-:(
-10.0
0.0 20,0 40,0 60.0 80,0 100.0
Figure 18. Example left frontal plane hip angles. Abduction is positive and adduction is
negative.
Pight Frontal Plane Hip ~6Jlg1es
100.080,060,040,020.0
7.0 -r--------,-----r------,-----r-----.,
'-:' c:: 4------+----+-----!-----+~;§i;2~
.:.., '-'
-5, 0~e~,..-__t----_t_------:;
o. 0 +-------+-----+-------If---.......,,,,,..~:..,,...:--~
-2. 5+----+-----+-----+---..~x;....,..:....t---__l
-7I 5+----'=-:
-14.0+-----+-----+-----+------+-------1
0,0
"_0"
"-"+
~ -10, 0+----~~~~~~~~::::=~r----+----i
-:(
..-.
F' -1'" -l\t ·~t "' ..\ .. -~ I_~ ~,.~ ,_~
Figure 19. Example right frontal plane hip angles. Abduction is positive and adduction
is negative.
49
L.. ft T '.- . ,'TT" ...... F'1'- . - U~· 6- (T1-'~e. I alL:. I) ex .:. e a.He ! ill.! .. U1E;. e.:·
100,0eo,o60.0
--- ;;-. ~-.-----~--.-
40.020.0
.-. 14.0
+
.... 10.0,-,
~ ... c
.... .' , ,_f
...
... 5.0......
.-;. C
~ •._I
- 0.0,
....
-
..... C
(I ~,._I
~
-5,0
....
..
p',
~
-9.0
0.0
F'· ~.... ~lt ·~t :'~1'"~ ... I_t::... ..I .... I_~
Figure 20. Example left transverse plane hip angles. Internal rotation is positive and
external rotation is negative.
..- -.:=-'--:;i -- . -
-.... .- .
~.;;:;r-. - -~- -~~-.Jt..,.e -- ~~~~ .... .., •. :~:~.~'~
....•:~...:~:~~
" '.
~I •.... '~~-'I •
. •.... '\
....-.
"-'. 17.0
-0 14.0
~ 12.0
....~ 10.0
......
100.0eo.o60.040.020,0
F' ."Tl- t T -.- ..~T T" -.~" F'1'" ... U~· .... 1"T1-·"·.le:. 1· I all.:. '.I ex .:. e alle rill.! ..u"UE;. e.:.
2,0
6.0
4,0
',-,'
X [Ilel
~ -2.0
0.0
F'· ~.~. ~lt ·~t :'~1~'"~ ... I_~... ...,.~ 1_,=
Figure 21. Example right transverse plane hip angles. Internal rotation is positive and
external rotation is negative.
50
Left Ground Re3.etlon Fore es
100,0eo,o60.040,0
....
...-.-- ---..
20.0
3,0
.-.
E~ .-. c
,.::..
.;: , "--
~·i'
~ 2,0z
_.'
~ 1.5e, 1.0
'-:;:1
'-'....
1:: 0.5
'1'
..-
..-"
0,0
0,0
F'-::1"c-::nt Shnc-::
Figure 22. Example left limb vertical ground reaction forces normalized to body weight.
f,",}-t (4 -- - -1 f,1_- -t'- - F- ---~dE:. 1· -_TI ullIn ·_e:iC .luI! UI t: 1:'::.
3.0
.-.
E~ .-:. c
,.::.. .:.:.. I "_I
~'.I)
~ 2.0.....
..:-.
-
'.', 1.5~e, 1,0
'-:;:1
'-'....
1:: 0.5
'1'
==- 0.0
0,0 20.0 40.0 60,0 100,0
F'-::1"c-::nt St.:t"nc-::
Figure 23. Example right limb vertical ground reaction forces normalized to body weight.
51
Table 5. Eversion excursion results for both the shoe markers and heel window markers
conditions.
Shoe Markers Heel Windows Percent Change
Subject n M SD n M SD n M SD
Mean 20 13.66 ±6.76 20 13.83 ± 6.91 20 3.37 ±6.91
Eversion Excursion Eversion Excursion Percent Change
R03 L 9.42 9.81 4.09
R03R 19.94 20.06 0.60
R05 L 9.56 6.87 -28.15
R05R 12.32 11.41 -7.33
R06L 10.87 11.36 4.54
R06R 17.92 19.87 10.88
R07L 10.33 8.57 -17.06
R07R 15.96 15.28 -4.22
R08L 7.73 11.09 43.44
R08R 9.34 10.48 12.21
R10R 17.18 21.23 23.57
R11 L 10.89 10.05 -7.71
R11 R 17.42 13.43 -22.95
R12L 10.25 13.55 32.26
R13 L 14.82 12.89 -13.05
R13R 33.14 35.56 7.32
R14L 8.05 7.09 -11.87
R14R 25.21 20.97 -16.85
R15 L 5.88 6.08 3.38
R15 R 7.06 10.90 54.31
Note: Eversion excursion measured in degrees CO), from heel strike to point of maximal
eversion. The mean row shows mean for the group, while each individual foot's data are
displayed below. Positive percent change indicates greater eversion excursion in the heel
windows condition while a negative percent change indicates less eversion excursion in
the heel windows condition.
-~
c:a:i
o (])
.~ ..r.
:l 0 ~u+J1i~ ~ ...(]) ttl
c: ~ E
o ~.~ E ~
~ (]) 0
(]) 0 -g
c: ..r. .-
.- VI :i:(]) E
~e
ttl'+-
..r.
U
5.0
4.0
3.0
2.0
1.0
0.0
-1.0
-2.0
-3.0
-4.0
-5.0
Changes in Eversion Excursion
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
m m ~ ~ ~ ~ ~ ~ ~ ~ 0 ~ ~ N m m ~ ~ ~ ~
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
52
Figure 24. Changes in eversion excursion from shoe markers to heel windows markers
condition for each individual subject. Eversion excursion measured in degrees from heel
strike to maximum rear foot eversion.
0.03
0.02
0.01
0.00
-0.01
-0.02
-0.03
Changes in Time to Peak Eversion
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
m m ~ ~ ~ ~ ~ ~ ~ ~ 0 ~ ~ N m m ~ ~ ~ ~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
Figure 25. Changes in time to peak eversion from shoe markers to heel windows
condition for each individual subject. Time measured in seconds.
53
Table 6. Time to peak eversion results for both the shoe markers and heel windows
markers conditions.
Shoe Markers Heel Windows Percent Change
Subject n M SD n M SD n M SD
Mean 20 0.088 ± 0.029 20 0.089 ± 0.032 20 0.45 ± 13.11
Time to Peak Eversion Time to Peak Eversion Percent Change
R03 L 0.075 0.063 -16.00
R03R 0.108 0.110 1.85
R05 L 0.061 0.050 -18.03
R05R 0.114 0.132 15.79
R06L 0.052 0.063 21.15
R06R 0.096 0.121 26.04
R07L 0.076 0.062 -18.42
R07R 0.119 0.118 -0.84
R08L 0.068 0.072 5.88
R08R 0.117 0.125 6.84
RlOR 0.122 0.130 6.56
R11 L 0.053 0.054 1.89
R11 R 0.110 0.088 -20.00
R12L 0.073 0.073 0.00
R13 L 0.055 0.061 10.91
R13R 0.125 0.111 -11.20
R14L 0.045 0.042 -6.67
R14R 0.101 0.107 5.94
R15 L 0.056 0.051
-8.93
R15 R 0.129 0.137 6.20
Note: Time to peak eversion measured in seconds from heel strike to maximal eversion.
A positive percent change indicates a longer time to peak eversion in the heel windows
condition while a negative percent change indicates a shorter time to peak eversion in the
heel windows condition.
54
Table 7. Percent stance at which peak eversion occurred for both the shoe markers and
heel windows markers conditions.
Shoe Markers Heel Windows Percent Change
Subject n M SD n M SD n M SD
Mean 20 36.76 ± 12.0 20 36.77 ± 14.87 20 - 1.63 ± 13.11
Percent Stance at Percent Stance at Peak
Peak Eversion Eversion Percent Change
R03 L 28.83 28.17 -2.31
R03R 46.69 47.46 1.63
R05 L 25.80 21.88 -15.21
R05R 49.95 55.64 11.40
R06L 23.70 27.60 16.47
R06R 41.26 49.90 20.93
R07L 30.21 24.85 -17.75
R07R 47.14 46.21 -1.97
R08L 26.74 29.08 8.76
R08R 46.14 49.83 8.00
RlOR 51.20 54.63 6.69
Rll L 23.08 25.27 9.49
Rll R 48.56 42.67 -12.13
R12L 29.46 28.78 -2.30
R13 L 22.45 24.81 10.51
R13R 49.89 43.96 -11.88
R14L 20.68 19.15
-7.43
R14R 46.10 49.03 6.38
R15 L 23.92 25.92 8.36
R15R 53.40 60.50 13.30
Note: Percent stance indicates the percent of the stance phase of the gait cycle at which
peak eversion occurred. A positive percent change indicates peak eversion occurred later
in stance in the heel windows condition while a negative percent change indicates peak
eversion occurred earlier in stance in the heel windows condition.
10.00
8.00
6.00
4.00
vi
.... 2.00<IJ
..:Jt:.
....
ttl 0.00E
-2.00
-4.00
-6.00
-8.00
Changes in Percent Stance at Peak Eversion
55
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
M M ~ ~ ~ ~ ~ ~ 00 00 0 ~ ~ N M M ~ ~ ~ ~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
Figure 26. Changes in percent stance at which peak eversion occurred between shoe
markers and heel windows markers.
Vl 80.00
:::l <IJ ~~ ,g ~ 60.00
c: Vl ....
oS E ttl 40.00
c: 0 Etl ~ ~ 20.00
c:lii"o
.- ......... "'C 0.00
E~c:
:::l >'§ -20.00
.§ -0 Qj -40.00
>< 0 <IJ
ttlE Qj.r:; -60.00> 0
c: g -:;; -80.00
<IJ .- ....~ ~ ~ -100.00
is <IJ E -120.00
-140.00
Chan es in Maximum Instantaneous Eversion Velocit
Figure 27. Change in maximum instantaneous eversion velocity from shoe markers to
heel windows markers. Velocity measured in degrees per second (o/s).
56
Table 8. Maximal instantaneous eversion velocity for shoe markers and heel windows
markers conditions.
Shoe Markers Heel Windows Percent Change
Subject n M SD n M SD n M SD
Mean 20 269.5 ± 117.6 20 260.6 ± 115.1 20 -0.28 ± 20.47
Max.Instant. Max. Instant.
Eversion Velocity Eversion Velocity Percent Change
R03 L 225.50 216.85 -3.83
R03R 285.62 290.44 1.69
R05 L 252.82 190.59 -24.61
R05R 178.59 168.38 -5.71
R06L 296.59 283.60 -4.38
R06R 315.45 246.57 -21.84
R07L 209.39 199.51 -4.72
R07R 246.69 240.69 -2.43
R08L 179.29 235.58 31.39
R08R 152.03 129.43 -14.86
RI0R 257.99 316.92 22.84
Rll L 296.23 268.70 -9.29
Rll R 267.96 285.31 6.48
R12L 247.65 295.09 19.16
R13 L 381.55 330.56 -13.36
R13R 645.29 681.20 5.56
R14L 276.14 210.39 -23.81
R14R 424.24 304.21 -28.29
R15 L 159.28 176.59 10.87
R15R 91.42 140.34 53.51
Note: Maximal instantaneous eversion velocity measured in degrees per second Cis). A
positive percent change indicates a greater eversion maximal instantaneous eversion
velocity in the heel windows condition while a negative percent change indicates smaller
maximal instantaneous eversion velocity in the heel windows condition.
57
Changes in Maximal Instantaneous Vertical Loading Rate
15.00
VI ~;:,0 3:-Q) 10.00c: CO Q)
ttl * Q)
... oo..c:
c:
..:¥. 0 5.00ttl .............
t;; z VI
c: - ~ viQ)~ ~ 0.00ttl ... ~ 0ttl ttlE ~ E "0
'x c:
ttl c: Q) .~ -5.00
E "0 0ttl..c:
c:
.2 VI
Q) E -10.00ttl00 U 0c: ~
'f .....ttl
-15.00
..c: Q)
U > ....J 0::: ....J 0::: ....J 0::: ....J 0::: 0::: ....J 0::: ....J 0::: ....J 0::: ....J 0:::
m m IJ) IJ) c..o c..o r-. r-. 00 .-l .-l m m
"" ""
IJ) IJ)
0 0 0 0 0 0 0 0 0 .-l ..-I ..-I .-l ..-I .-l ..-I .-l0::: 0::: 0::: 0::: 0::: 0::: 0::: 0::: 0::: 0::: 0::: 0::: 0::: 0::: 0::: 0::: 0:::
Figure 28. Changes in maximal instantaneous vertical loading rate from shoe markers to
heel windows marker conditions. Loading rate measured in Newtons per kilogram of
body weight per second (Nlkg*BW/s).
Q) ~
o Q)~~
§ E
... \II~ 3
Q) 0
;:'-0
- c:co ._
> 3
u
~ Q)u~
.S .s
to ~~~
co ...
~ co
u E
0.10
0.05
0.00
-0.05
-0.10
-0.15
-0.20
-0.25
Chan es in CMC Values
Figure 29. Changes in CMC values from shoe markers to heel windows markers.
58
Table 9. Maximal instantaneous vertical loading rates for shoe markers and heel
windows markers.
Shoe Markers Heel Windows Percent Change
Subject
Mean
n M SD
20 84.47 ± 18.48
n
20
M
82.23
SD
± 17.16
n
20
M
- 2.01
SD
± 7.52
R03 L
R03R
R05 L
R05R
R06L
R06R
R07L
R07R
R08L
R08R
RlOR
R11 L
R11 R
R12L
R13 L
R13R
R14L
R14R
R15 L
R15 R
Max. Instant. Vertical
Loading Rate
108.41
102.76
97.91
81.37
84.16
97.34
76.15
65.42
112.92
74.96
80.88
54.16
41.98
84.78
91.18
89.38
92.25
Max. Instant Vertical
Loading Rate
96.96
96.84
92.85
71.79
92.05
94.73
71.13
62.99
115.35
83.54
76.56
51.73
48.41
79.80
90.44
86.98
85.78
Percent Change
-10.56
-5.76
-5.17
-11.78
9.38
-2.68
-6.59
-3.71
2.15
11.45
-5.34
-4.49
15.31
-5.87
-0.81
-2.69
-7.02
Note: Maximal instantaneous vertical loading rate is measured in Newtons per kilogram
of body weight per second (N/Kg BW/s). A positive percent change indicates a higher
loading rate in the heel windows markers condition compared to the shoe markers
condition while a negative percent change indicates a lower loading rate in the heel
windows condition.
59
Table 10. CMC values for the shoe markers and heel windows markers conditions.
Shoe Markers Heel Windows Percent Change
Subject n M SD n M SD n M SD
Mean 20 0.924 ± 0.070 20 0.872 ± 0.120 20 - 6.05 ± 7.95
CMC Value CMC Value Percent Change
R03 L 0.846 0.698 -17.57
R03R 0.985 0.982 -0.28
R05 L 0.870 0.917 5.30
R05R 0.972 0.986 1.38
R06L 0.924 0.867 -6.12
R06R 0.992 0.987 -0.47
R07L 0.886 0.823 -7.08
R07R 0.980 0.990 1.04
R08L 0.897 0.823 -8.21
R08R 0.969 0.990 2.18
R10R 0.986 0.948 -3.83
Rll L 0.894 0.836 -6.42
Rll R 0.984 0.858 -12.77
R12L 0.877 0.739 -15.68
R13 L 0.845 0.785
-7.10
R13 R 0.979 0.971 -0.82
R14L 0.886 0.769
-13.21
R14R 0.988 0.961 -2.73
R15 L 0.754 0.550
-27.06
R15 R 0.967 0.952
-1.55
Note: CMC indicates the repeatability or variability of the angle curve across stance. A
higher CMC value indicates less variability while a lower CMC value indicates greater
variability. A positive percent change indicates a higher CMC value in the heel windows
markers condition compared to the shoe markers while a negative percent change
indicates a lower CMC value in the heel windows condition compared to the shoe
markers condition.
- --------------
60
CHAPTER IV
DISCUSSION
The purpose of this study was to examine differences in commonly measured
parameters thought to be related to overuse injury in runners when measured with
markers mounted on the shoe and with markers mounted directly on the rear foot using
heel windows. Specifically, this study examined differences between the two marker
conditions in the time to peak rear foot eversion, the percent stance at which maximal
eversion occurred, the maximal instantaneous eversion velocity, and the maximal
instantaneous vertical loading rate. Additionally, this study also measured the variability
and repeatability of the ankle j oint inversion-eversion curve. A secondary purpose of this
study was to make a recommendation on marker methods and protocols to use in a
longitudinal study.
The hypotheses for this study were that maximal rear foot eversion would
be less in the heel windows condition compared to the shoe marker condition, that time to
peak eversion excursion and the percent stance at which peak eversion excursion
occurred would not be different between the two marker systems, the maximal
instantaneous eversion velocity would increase in the heel windows condition compared
to the shoe markers condition, and that the maximal instantaneous vertical loading rate
would decrease from the shoe marker condition to the heel window marker condition. It
61
was also hypothesized that the variability of the rear foot inversion-eversion curve would
increase in the heel windows condition compared to the shoe markers conditions.
The only hypothesis which was supported in the results of this study was that the
variability of the ankle joint inversion-eversion curve was higher in the heel windows
condition compared to the shoe markers condition. For all the other variables the results
were not statistically significant, failing to provide support for most hypotheses ofthe
study.
There are numerous possible explanations for why there were no significant results
among the kinematic parameters. Firstly, the subjects themselves could be a poor
sample, not representative of an average cross section of the running population.
Secondly, even if this sample of subjects does provide a reliable estimate of the general
running population, they may have different or unique kinematic or kinetic patterns
which led to the lack of statistical significance. Thirdly, there could have been
differences between running trials under the two marker conditions which were not
adequately controlled and resulted in the lack of significance. Lastly, perhaps the
analysis methods employed in this study were not robust enough to reveal subtle
differences between the two conditions. The remainder ofthis chapter will focus on
discussing the observed changes in variability and addressing each of the potential
confounding effects among the other parameters observed.
62
Variability of Ankle Joint Inversion Eversion Angles
The only statistically significant findings of this study was that the heel windows
condition resulted in greater variability in the ankle joint inversion eversion curve when
compared to the shoe markers conditions. There are several potential explanations for
this change. Perhaps the most obvious explanation is the cutting of the heel windows
weakened the mechanical properties of the shoe in a way which allowed greater
movement of the foot. While no materials testing was performed in this experiment to
quantify the effects of cutting the heel windows, this explanation seems unlikely for
several reasons.
First, the heel windows method for placing markers on the rear foot has been
frequently used in previous studies where materials testing was performed (Butler et aI.,
2006; Ferber, 2005; Williams et al., 2003; Williams et aI., 2004). These studies indicated
that the heel windows resulted in only a 10% decrement in heel counter stability.
Whether such a small change is enough to seriously influence foot motion is unknown,
however it appears unlikely. Secondly, the heel windows employed in this study were,
on average, 2.13 cm2 for the upper hole and 1.87 cm2 for the lower hole. Stacoff et al.
(1992) found significant changes in rear foot motion did not occur until the heel windows
were 4.12 cm2 for the lower window and 3.61 cm2 for the upper window, values
substantially larger than the ones used in this study. Lastly, it is debatable the extent to
which the rigidity of the heel counter is actually responsible for controlling rear foot
motion, as Van Gheluwe et aI. (1995) demonstrated that rear foot motion occurred
63
independently from the stiffness of the heel counter. When taken together, the results of
these studies suggest it is unlikely any changes in rear foot motion were due only to the
cutting of the holes in the heel counter of the shoe. A more likely explanation for the
increased variability observed in the heel windows markers condition involves the
relationship between the number of movement patterns available within the central
nervous system and how the marker placement may reveal or mask these various options.
Though the exact mechanisms are not completely understood, it is thought that the
general motor firing patterns responsible for producing human gait are established by
central pattern generators located in the spinal cord (Duysens & Van de Crommert, 1998;
MacKay-Lyons, 2002). These generators are heavily influenced by both supraspinal and
afferent sensory nerve input, resulting in a wide range of individual movement patterns
within the context of the larger central pattern (MacKay-Lyons, 2002; Van de Crommert,
Mulder, & Duysens, 1998). It is thought that this variety of movement patterns allows
for flexibility and adaptability within the system.
When the markers are placed on the exterior of the shoe they are all attached to the
heel counter, which is essentially a rigid object. This means the markers have little
ability to move independently from each other. Additionally, in theory, the deformations
of the shoe during the running stance phase should generally be repeatable from stride to
stride. When analyzed with markers placed in this fashion, the combination of these two
conditions will suggest foot motion with very little variability between trials. However,
this is most likely not an accurate measure of the actual amount of variability present in a
given individual.
64
Using the heel windows method allows placement of the markers directly on the skin
of the rear foot. This alleviates the restrictions of placing the markers on the shoe in that
the three markers can move independently from each other, and are not constrained to
move in the same pattern each stride. As seen in the results, the measured variability is
increased when the markers are placed in this fashion. This marker placement allows a
more accurate assessment of the variety of movement patterns available to an individual,
a measurement with potentially important clinical implications
As previously discussed, the variety of available movement patterns allows for
flexibility and adaptability within the system. Additionally, for cyclical motions such as
running, this variety ofmovement patterns also helps avoid loading biological tissues in
exactly the same manner every cycle. Some authors have suggested that there is an
optimum amount ofvariability and common running injuries may be related to a decrease
in the variability ofmovement patterns available to an individual (Hamill, van Emmerik,
Heiderscheit, & Li, 1999; Heiderscheit, Hamill, & Van Emmerik, 2002; Miller, Meardon,
Derrick, & Gillette, 2008; Stergiou, Harboume, & Cavanaugh, 2006). Placing the
markers on the shoe would mask any changes in variability, thereby potentially masking
potential injury causing mechanisms.
The Subject Sample as a Potential Source ofNon-Significance
There is the potential that the lack of significant results in this study resulted from
a sample which is not representative of the general running population. Subjects were
65
recruited into this study based on their desire or ability to participate, and their
membership in local running clubs or groups, making the subjects in this study a sample
of convenience and not truly a random sample. Additionally, most of the subjects were
recruited from a university campus, thereby potentially biasing the sample in regards to
age, number of year running, or injury history. Lastly, the subjects in this study were all
recreational runners or recreationally competitive and none of them were highly trained
or elite athletes.
Though these issues raise concerns, it is felt they had little influence on the external
validity of the current study and are similar methods to what has previously been reported
in the literature. For instance, in a study validating the use of the arch height index
Williams and McClay (2000) examined fifty one recreational runners who volunteered
from the surrounding community. The mean age of these subjects was 27.1 years and
their mean arch height index was 0.316. Another study specifically using this arch height
index method to establish reference data among recreational runners found a mean arch
height index of 0.340 (Butler, Hillstrom, Song, Richards, & Davis, 2008). This method
and the reference data presented in these two studies has since been used in numerous
studies on runners as an established method and mean for comparing arch heights among
individuals (Chang, Van Emmerik, & Hamill, 2008; Molloy et aI., 2009; Pohl et aI.,
2009; Williams et aI., 2003; Williams et aI., 2004; Williams, McClay, & Hamill, 2001;
Williams, McClay, Hamill et aI., 2001; Zifchock, Davis, Higginson, McCaw, & Royer,
2009). In the current study the mean age of the subjects was 25.46 and the mean arch
height index was 0.30, suggesting these subjects were representative of the mean foot
66
structure found in the broader population of recreational runners examined by Butler et
aI. (2008) and Willams and McClay (2000).
It has been shown by Williams et aI. (2001) that there are significant kinematic and
kinetic differences between high arched and low arched runners. The subjects in this
study had arch height indices ranging from 0.244 to 0.390, suggesting they represented a
cross section of individuals with low, medium, and high arches. If only high or low
arched runners had been used in this study it would have both potentially biased the
results and limited the potential applications of the findings. However, this does not
appear to have happened and the foot structure of the subjects does not appear to have
played a role in the results observed in this study.
Kinematic and Kinetic Parameters
In general, reviews on the biomechanics of running suggest an individual lands with
their rear foot in a slightly inverted position then everts until approximately midstance.
The amount of eversion excursion usually falls within a range of approximately 10° to
20° with maximal eversion occurring between 20% and 40% of the stance phase
(Edington et aI., 1990; McClay, 1995; Novacheck, 1988). Experimental evidence
supports these numbers. Summaries of several studies which reported the same kinematic
and kinetic parameters as this study are shown in Table 11. Most of these studies
compared these parameters in an injured and a healthy control group. When that scenario
occurred, these values were taken from the healthy control group.
67
From this infonnation it appears the results reported in this study are well within the
ranges observed in other studies. Therefore, the lack of statistical significance between
the shoe markers and heel windows marker methods is not due to kinematic or kinetic
values unique to the subjects in this study and other reasons must be considered.
Uncontrolled Factors
As a result of the methods used in this particular study there are some uncontrolled
factors which could potentially affect the internal validity of the study and help explain
the lack of statistical significance between the shoe markers and heel windows marker
conditions. For instance, the subjects ran at a self selected pace for both trials. Most of
the parameters examined in this study have been shown to vary with running speed
(Cavanagh & Lafortune, 1980; Pohl, Messenger, & Buckley, 2007). Therefore, large
differences in running speed between the shoe markers and heel windows markers
conditions might cause significant differences in the kinematic and kinetic parameters.
However, as shown in Table 4 running speeds were not significantly different between
the two marker conditions, suggesting differences or lack of differences between the two
conditions cannot be simply attributed to running speed.
68
Table 11. Representative kinematic and kinetic results reported from several studies
Maximal
Time to Peak Maximal Instantaneous
Eversion Eversion (s) Instantaneous Vertical
Excursion or Eversion Loading Rate
Authors N C) (% stance) Velocity Cis) (BW/s)
(McC1ay& 12.70 0.09 s9Manal, 1998a) (± 4.1) (± 0.026)
(Pohl et aI., 14.90 52.6% 82.9252009) (± 4.0) (± 6.0) (± 18.7)
(Willems et aI., 18.06 47.18% 447.27 133.303342007) (± 4.53) (± 12.14) (±131.71) (± 45.98)
(Pohl et aI., 8.80 83.80302008) (± 4.1) (± 23.20)
(Milner, Ferber 79.6520
et aI., 2006) (± 18.81)
(Mundermann et 16.0 464.720
aI.,2003) (± 2.3) (± 155.2)
(McClay & 12.7 0.11 s9Manal, 1997) (± 3.5) (± 0.05)
Note: Some studies report percent stance at maximal eversion in seconds (s) and some
report it as a percent of stance phase (%). This convention has been followed in the table
above.
69
Another uncontrolled factor in this study was the fact subjects wore their own running
shoes instead ofa standardized laboratory shoe. Some of the subjects in this study were
wearing stability shoes while others wore cushioning shoes. In theory stability shoes are
designed to control rear foot motion during running, suggesting the integrity of these
shoes could potentially be more susceptible to modifications such as cutting holes in the
heel windows. While this may affect the internal validity of the study it was felt this was
an important point since the shoes worn for this study are the shoes the individual wears
on a daily basis while running.
It has been shown that due to the repetitive compressive forces over time the foam
used to construct a running shoe slowly breaks down, losing its cushioning properties and
changing the pressure distribution under the foot (Verdejo & Mills, 2004). Additionally,
it has been shown that hardness of the shoe midsole has significant effects on the
kinematics and kinetics of the foot during the stance phase of running gait (De Wit, De
Clercq, & Lenoir, 1995; Hamill, Bates, & Holt, 1992; Stergiou & Bates, 1997). Given
this information it was felt that placing a subject in a controlled laboratory shoe with
different foam characteristics and hardness then what they were used to had the potential
to artificially alter their kinematics and kinetics and lead to artificial conclusions about
differences in foot motion between the two marker conditions.
However, it appears having subjects wear their own shoes did not have any effect on
the results of this study. There was no consistent pattern between changes in any of the
parameters measured in this study and the type of shoe the subjects wore. For instance,
all subjects wearing stability shoes did not respond the same for any the parameters
70
measured in this study when switching from the shoe markers to the heel windows
markers. This further supports the argument presented earlier in this chapter that the
increase in variability observed in the ankle joint inversion eversion curve in the heel
windows condition was not a result of the modifications to the shoe.
The last uncontrolled factor between the subjects in this study was the actual size of
the heel windows cut into the shoes. While every attempt was made to make the heel
windows the same size on each shoe, the windows were cut by hand resulting in small,
unavoidable variations between shoes. However, the impact of these small variations
appears negligible. As previously mentioned, the heel windows used in this study were
substantially smaller than the large windows used by Stacoffet al. (1992) and therefore
should not have affected the motion of the rear foot. Though it was not tested
statistically, a visual examination of the data presented in Tables 3,5,6, 7, 8, and 9
shows no clear relationship between either the magnitude or direction of changes in any
of the measured parameters and the size of the heel windows. Therefore it is unlikely
that any small variations in heel window sizes resulted in the lack of statistical
significance observed in this study.
Data Analysis Methods
The original hypothesis of this study suggested that the changes from the shoe
markers to heel windows marker conditions should be similar across both subjects and
71
feet within a subject. Based on bone pin studies by Stacoff et ai. (2000; 2001) and
Reinschmidt et aI. (1997) it was hypothesized the shoe markers would overestimate the
amount of rear foot motion compared to the heel windows markers, a claim commonly
found in studies citing differences between shoe and foot motion (Butler et aI., 2006;
Butler, Hamill, & Davis, 2007; Dierks et aI., 2008; Milani & Hennig, 2000; Pohl,
Messenger, & Buckley, 2006). However, as the information in Tables 4 through 9 and
Figures 24 through 28 illustrate, this clearly was not the case in this study. For most
variables there is an even split with approximately half the feet responding one way and
the other half responding in the opposite direction. For instance, Figure 24 shows that
half the feet demonstrated an increase in eversion excursion with the heel windows
compared to the shoe markers while the other half of the feet demonstrated a decrease.
Similar patterns are shown in Figures 25,26,27, and 28 for time to peak eversion,
percent stance at peak eversion, maximal instantaneous eversion velocity, and maximal
instantaneous vertical loading rates.
This individualized response to the two marker conditions can also be seen in the
results of the original bone pin studies. For example, the eversion results from the
Reinschmidt (1997) study are shown in Figure 30. While most of the subjects do
demonstrate a decrease in eversion when measured with the bone pins, subject three does
not. For subject three the bone pins actually resulted in greater eversion. The authors
have suggested this may be due to the shoe markers being placed in a more inverted
position, thereby artificially shifting the shoe marker curve, however there is the
possibility that this subject simply responded differently than the other subjects.
72
Regardless of the marker alignment issues, a more detailed examination of Figure 30
shows that the differences between the shoe markers and bone pin markers were unique
to each individual and not standardized across subjects. Similar individualized
differences were reported in a study which also used bone pins to compare differences
between barefoot and shod running, whose rear foot eversion curves can be seen in
Figure 31 (Stacoff et aI., 2000). While not directly comparing the differences in rear foot
motion between shoe markers and bone pin markers, these results also reinforce the idea
that each subject or foot may respond differently to the different marker conditions.
SUBJECT 1
20- - --.----- ..
SUBJECT 2 SUBJECT 3 SUBJECT 4 SUBJECTS
o 50 100 0 50 100 0 50 50 100 0 50 100
Figure 30. Rear foot eversion curves from a study by Reinschmidt et ai. (1997)
examining differences in rear foot motion during running as measured with intracortical
bone pins and external shoe based markers. The dashes lines represent the shoe based
markers and the solid lines represent the bone pin markers.
73
_.. _~_......J
Subject 5
I~:----;"j
i: : :jL._...... . ....
\. . ,
I: :1
t' 'j1'- :.' :
Subjec14SUbject J
'----'- --
Subject ~
!:
"
L_~__-.J
SUbje~t 1
10,..----..
Figure 31. Rear foot eversion curves from a study by Stacoffet al. (2000) examining
differences in rear foot motion between barefoot and shod running. The dotted lines
represent the shod condition while the solid lines represent the barefoot condition.
The results of the current study are similar to the two bone pin studies in that each
foot demonstrated a unique, individualized change in the heel windows marker condition
compared to the shoe markers. For instance, subject R05's left foot demonstrated a
28.5% decrease in eversion excursion between the two conditions while subject R12's
left foot demonstrated a 32% increase in the same parameter. Similar magnitude percent
changes can be seen in other subjects and parameters such as R06's right foot
demonstrating a 26% increase in the time to peak eversion while subject Rll's right foot
showed a 20% decrease in the same parameter. However, the dependent observations t
test used for statistical analysis in this study assesses differences between the mean scores
of the group in the shoe markers and heel windows markers conditions. With some feet
demonstrating increases in the parameters measured while others demonstrated decreases,
the overall differences wash out and the result indicates there are not statistically
significant differences between the two marker conditions.
74
Yet, as discussed above, some feet demonstrated changes of up to 32% between the
two marker conditions, a number which suggests significant differences are present
between the two marker conditions. If there are individually significant changes which
get washed out with a group analysis then the particular method of data analysis
employed in this study may not be suitable for this purpose. A more useful method
would examine each subject individually and determine on a case by case basis whether
there were statistically significant differences between the two marker conditions. This
approach is called a single subject analysis and the theory and methods behind its
application are presented in the next chapter.
75
CHAPTER V
SINGLE SUBJECT DESIGN
Background and Rationale
The results and discussion detailed in Chapters III and IV indicated that besides
variability in the ankle frontal plane joint angles, there were no significant differences
between the shoe markers and the heel windows markers. As discussed in Chapter IV,
this lack of statistically significant differences may be due to the methods utilized to
analyze the data. The methods used in the previous analysis were a traditional group
analysis where the data from multiple trials of each subject was averaged to indicate an
average performance for that individual. These representative individual responses were
then averaged to yield an average group response. The statistical analysis was performed
on these average group responses to the two marker conditions and found no significant
differences between conditions.
This approach is wide spread and common in both running studies and
biomechanics research in general, and a prime example of a traditional group approach.
Review articles and book chapters suggest traditional group design evolved from a desire
to compare individuals to some "average" level, which was assumed to be desirable
(Bates, 1996; Bates, James, & Dufek, 2000). From a research perspective this is enticing
since it is thought to allow the generalization of the results to a larger population.
76
Furthermore, any errors in measurement or individual variation can more or less be
overcome with a large enough sample size.
However, as Bates (1996) points out, with the wide range of natural variation
present in each individual, rarely does anyone subject ever perfectly conform to the
"average" parameters. Over the years, many studies on running have confirmed this,
finding wide interindividual variations in the various parameters they measured (Bates,
Osternig, Mason, & James, 1979; De Wit, De Clercq, & Aerts, 2000; McClay & Manal,
1999; Van Gheluwe et aI., 1995). This same analysis holds true within the individual
since the numerous movement patterns available to an individual's neuromuscular system
mean rarely will any single trial perfectly match their "average" performance. As with
the interindividual variability, several studies have also cited high levels of
intraindividual variability in runners for both kinematic and kinetic parameters (Bates,
Osternig, Sawhill, & James, 1983; Cavanagh & Lafortune, 1980; Devita & Skelly, 1990).
With a group statistics approach variability in a sample can be dealt with by
incorporating a larger sample size. However, this approach will not work when the
variability is a result of each individual using different, unique neuromuscular strategies
to accomplish the given task (Bates et aI., 2000). These different strategies are a result of
the numerous degrees of freedom in the neuromuscular system and how the body
responds to any constraints which may influence the movement (Bates et aI., 2000). The
constraints could be from environmental sources, due to anatomical variations, or
influenced by sensory feedback during the movement task, however the end result is that
rarely will any two trials be identical, both within and between individuals (Bates et aI.,
77
2000). Evidence of the usage of different movement strategies have been previously
observed in athletic tasks such as running and landing (Caster & Bates, 1995; Dufek,
Bates, Stergiou, & James, 1995; Reinschmidt & Nigg, 1995). In these cases, when
individuals are using different strategies to solve similar movement problems, the
variability will remain, no matter how large the sample size.
This interindividual and intraindividual variability is of particular concern from
the viewpoint of trying to assess factors related to overuse injuries in runners. For
instance, imagine a prospective scenario where one wants to predict which individuals
are at risk for different injuries. If an individual resembles the normal "average" one
might conclude they are healthy and have a relatively low risk for and injury. A similar
scenario could be envisioned for a retrospective approach where one wants to identify
biomechanical factors which may have caused a particular injury. If all the subject' s
kinematic and kinetic measurements are within "average" ranges one might not be able to
make any conclusions about contributing factors to the injury. In both scenarios, when
using a group average approach, interindividual and intraindividual variability would
mask any subtle variations in the subject's anatomic alignment or changes in their
kinematics and kinetics which may be related to an injury. In these situations a single
subject design approach with its focus on changes and variability within the individual
rather than the group average would be a more effective approach.
Single subject designs traditionally have been used in the social sciences to
examine topics such as behavioral interventions in education settings or effects of
different teaching strategies (Richards, Taylor, Ramasamy, & Richards, 1999). In these
78
studies multiple baseline measurements are taken on a single subject, some intervention
is then imposed, and multiple measures are recorded after the intervention. The
researchers treat each measurement as an independent observation and examine if there
were significant changes in the pre and post intervention observations. This particular
approach is commonly referred to as an AB design, however there are numerous other
approaches which are also commonly used such as ABA, or ABABA (Richards et ai.,
1999). For behavioral studies these repeated measurements may take place over the
course over several days, however, for application to biomechanics studies, one could
construe multiple consecutive foot strikes during running as the repeated measurements
(Bates et ai., 2000). For instance, if one was evaluating the effects of an orthotic on rear
foot motion during running, a subject could perform twenty trials in a normal shoe, have
the orthotic inserted, and then perform twenty additional trials. The researchers would
then examine differences between the trials to determine the effects of the orthotic
intervention.
As mentioned above, variability is an important consideration in any study,
whether a traditional group or a single subject design. However, one of the appeals of the
single subject methodology is how it accommodates intersubject and intrasubject
variability compared to traditional group designs. Bates et al. (1992) and Dufek et al.
(1995) have performed several computer modeling studies to explore how variability
affects the results of a statistical analysis with both group and single subject designs.
Their results suggest that, compared to single subject designs, group design methods are
more susceptible to interindividual and intraindividual variability. When such variability
79
is present there is a greater chance of failing to appropriately reject the null hypothesis.
However, their results also indicate that, compared to single subject analysis methods,
group designs require far fewer trials per subject to achieve high levels of statistical
power. This has important implications for the design of any study considering the use of
single subject analysis methods.
One of the biggest considerations, and arguments against the use of single subject
methods, is that it violates several important guiding principals of statistical analysis,
including the independence of observations and the normality of the distribution of the
sample. In theory, for traditional group statistical analysis the observations of the
dependent variable are considered to be independent. In single subject analysis the
observations are repeatedly taken from the same individual. In a situation such as
running they are repeated foot strikes that occurred within a close time span. Being from
one individual, potentially in a short time span, there is the potential that one observation
influences the next and they are not truly independent observations.
Bates et al. (1996; 2000) argue, based on both computer simulations and actual
data collected in their lab, that the assumption of independence is not usually violated in
single subject analysis and these authors suggest calculating an autocorrelation
coefficient for the data to demonstrate this fact. In the autocorrelation coefficient one
calculates the correlation coefficient between consecutive observations in a sample, such
as between 1-2,2-3,3-4, etc. If the autocorrelation coefficient is high, or significant,
then the data should not be considered as independent observations (Bates, 1996; Bates et
aI., 2000; Richards et al., 1999). The auto correlation coefficient can easily be calculated
80
by most commercially available statistical programs, or can be calculated by hand using
the proper formulas and spread sheet software such as Microsoft Excel.
The second argument against the use of statistical analysis with single subject
design, that it violates assumptions of normality, is also handled in a fairly easy fashion.
Bates(l996; 2000) suggests this assumption is not an issue since many statistical tests,
such as the t test are robust to violations of normality. However, a researcher can easily
get an idea of the normality of the data set using a statistical test such as the Shapiro-Wilk
test (Bates et aI., 2000).
When the two conditions of independence of observations and normality of data
have been confirmed, most of the standard statistical tests can then be applied in a single
subject design setting. In addition to statistical analysis, single subject design also lends
itself to graphical analysis ofthe data (Richards et aI., 1999). In this approach a
researcher might plot the data and note any apparent changes in magnitude or trends
between the conditions tested. This can be an especially valid method when there are
numerous trials to examine and the intraindividual variability is low (Richards et aI.,
1999).
As this discussion has indicated, single subject analysis can be a powerful tool
when examining data where subjects may present with individualized motor strategies to
a common movement task. The group analysis discussed in Chapters II, III, and IV
suggested there were not significant differences between the shoe markers and heel
windows marker conditions. However, as previously mentioned, several individuals
81
demonstrated rather large percent changes between the two conditions. Therefore, the
purpose of this next chapter is to reanalyze the data presented in Chapter III using a
single subject approach. Again, each foot was analyzed separately. The hypothesis for
this single subject approach is that feet with large percent changes from the shoe markers
condition to the heel windows marker condition in the group analysis will show
significant differences under a single subject approach. Furthermore, it is hypothesized
that, under a single subject approach each foot would respond uniquely to the two marker
conditions, with some demonstrating significant differences between conditions while
others demonstrate no significant differences between conditions.
Single Subject Design Methods
As the name implies, a single subject design will consider all the trials done by
one individual. Therefore, a large number of trials are needed. Bates et al. (1979)
suggested that, for a traditional group analysis, a minimum of eight trials were required to
produce a reasonable picture of an individuals kinematics and kinetics while running.
Given the statistical concerns raised by Bates et al. (1992) and Dufek et al. (1995), it
seems the number of trials is potential concern for single subject analysis, however in a
discussion about single subject analysis methods Richards et al. (1999) indicate adequate
results can be achieved with a minimum of eight trials per condition. Therefore, only feet
82
which had at least eight trials for both the shoe markers and heel windows markers
conditions were included in the single subject analysis.
Richards et al. (1999) suggest graphical analysis as a potentially powerful tool for
single subject design. In this method the results from every individual trial both
conditions are plotted against time. For instance, if the particular value being examined
was eversion excursion, then the eversion excursion would be plotted for each
consecutive trial the subject completed. This scatter plot will have eversion excursion on
the Y axis with the trial number on the X axis. Any significant differences between the
two conditions might be observed as changes in the level or trend of the trials from one
condition to the next (Richards et al., 1999). This may allow researchers to identify
changes which may be meaningful, yet not statistically significant, and can be a valuable
option when certain conditions required for statistical analysis are not met in the data
(Richards et al., 1999).
Graphical evaluation was undertaken for all trials for all feet meeting the
minimum number of trials condition. The variables analyzed for the single subject design
were eversion excursion, percent stance at peak eversion, maximal instantaneous eversion
velocity, and maximal instantaneous vertical loading rate. For each foot, all these values
from each trial, under both conditions, were plotted on the Y axis, with the trial number
of on the X axis. The graphs were inspected visually for apparent changes in trend or
levels of the particular variable between the shoe markers and the heel windows markers.
83
However, graphical analysis does have several limitations, with the foremost
limitation being the subjective nature of the analysis (Richards et aI., 1999). Two
different researchers may not agree on what level of difference between the two
conditions constitutes a meaningful change. Potentially worse, they may not even agree
if there is a difference between the two conditions, especially if there is substantial
variability present in the data, as the variability may mask changes in trend or level
between conditions (Richards et aI., 1999). Therefore, in this study, all the data were also
analyzed using statistical analysis.
Before statistical analyses were carried out the data were examined to ensure
adherence to standard assumptions of normality and independence of observations.
Evaluation of normality of the distributions for both the shoe markers trials and heel
windows trials were done with a Shapiro-Wilk test (Bates et aI., 2000). Distributions
were assumed to be normal when the resulting W value was greater than 0.05, as this
indicates a 95% confidence that the data is normally distributed (Bates et aI., 2000).
Independence of observations was examined by calculating autocorrelation coefficients
for all the trials of each foot for each variable of interest. Coefficients for the shoe
markers trials and heel windows marker trials were calculated separately. Autocorrelation
coefficients were calculated using Statistical Package for the Social Sciences (SPSS),
version 18.0 (SPSS, Inc., Chicago IL), with (l = .05. In this manner ap value less than
.05 indicates statistically significant autocorrelation was present. For all variables only
autocorrelation of the first lag was calculated as it was suggested this should be sufficient
to detect autocorrelation in the data (Bates et aI., 2000; Richards et aI., 1999).
84
An independent samples t test was used to compare the trials with the shoe
markers to the trials with the heel windows markers. Though these measurements were
repeated measurements on the same foot, the whole theoretical basis for single subject
design assumes they were independent observations, thereby contraindicating the use of
repeated measures statistical methods such as a paired observations t test (Bates, 1996).
All independent t tests were conducted using SPSS (SPSS, Inc., Chicago, IL) statistical
analysis software and were performed with a = .05, set a priori.
Statistical significance of results was also evaluated using a method specifically
designed to work with single subject analysis called Model Statistics. This method was
developed in the University of Oregon Biomechanics Laboratory by Dr. Barry Bates and
has been frequently used in analysis of single subject design (Bates et aI., 1992; Bates et
aI., 2000; Dufek & Bates, 1991; Dufek, Bates, & Davis, 1995; Dufek, Bates, Stergiou et
aI., 1995; Stergiou & Bates, 1997). In this method the absolute difference of the means
of the trials under both conditions are compared to a critical value. If the absolute
difference in means is larger than the critical value then there is a statistically significant
difference between the two conditions and if the absolute difference in means is less than
the critical value there is not a statistically significant difference between the two
conditions. The critical value is calculated based on a test statistic multiplied by the
weighted standard deviation of the trials from the two conditions. The weighted standard
deviation is calculated as:
SDw = (1)
85
SDI is the standard deviation of all the trials from the shoe markers condition and
SD2 is the standard deviation of all the trials from the heel windows markers condition.
The critical statistic was selected based on the table of critical statistics found in Bates et
al. (1992). This table is reproduced in Table 12. Critical statistics are provided for
various sample sizes and for a levels of 0.1, 0.05, and 0.01, respectively. The sample size
was selected based on the smallest number of trials for either the shoe markers or the heel
windows markers. For instance if a foot had fifteen good trials in the shoe markers and
only eleven good trials in the heel windows condition, then the sample size used for the
Model Statistics analysis was eleven. This statistical analysis was carried out for each of
the four variables which had enough trials under both conditions to identify statistically
significant differences between the shoe markers condition and heel windows marker
conditions on an individual subject basis.
86
Table 12. Critical statistic values for the Model Statistics analysis.
Sample Size u = 0.10 u = 0.05 u = 0.01
3 1.3733 1.6533 2.2133
4 1.2643 1.5058 1.9867
5 1.1597 1.3662 1.7788
6 1.0629 1.2408 1.6044
7 0.9751 1.1306 1.4623
8 0.8964 1.0351 1.3473
9 0.8270 0.9536 1.2542
10 0.7673 0.8857 1.1776
11 0.7172 0.8307 1.1129
12 0.6757 0.7867 1.0581
13 0.6415 0.7516 1.0117
14 0.6132 0.7234 0.9720
15 ·0.5896 0.7001 0.9375
16 0.5695 0.6798 0.9070
17 0.5522 0.6618 0.8796
18 0.5371 0.6458 0.8548
19 0.5237 0.6311 0.8318
20 0.5114 0.6175 0.8102
Note: Table of critical statistic values for Model Statistics from sample sizes of three to
sample sizes of twenty. From Bates et al. (1992).
87
CHAPTER VI
SINGLE SUBJECT RESULTS
Subjects
A criterion of at least eight good trials for both the shoe markers and heel
windows markers conditions was required for a foot to be included in the single subject
analysis. Most of the feet achieved this for the kinematic variables of eversion excursion,
percent stance at peak eversion, and maximal instantaneous eversion velocity. Much
fewer achieved this benchmark for the kinetic parameter of maximal instantaneous
vertical loading rate. However, when this value was met, most feet had more trials than
the required minimum.
Overall, of the original twenty feet used in the group analysis, seventeen were
used for the single subject analysis for the kinematics and ten were used for the single
subject analysis for the kinetics. Seven feet had enough trials to be used to analyze the
kinematics but not enough trials to analyze the kinetics, so only the kinematics were
analyzed in these feet. These data are summarized in Table 13.
88
Table 13. Number of trials used in the single subject analysis for both the kinematic and
kinetic variables.
Subject Number of Kinematic Trials
R03R 11
R05 R 14
R06L 20
R06R 19
R07L 20
R07R 19
R08L 21
R08R 21
RI0R 8
R11 L 11
R11 R 9
R12L 11
R13 L 20
R13 R 20
R14L 19
R14R 19
R15 L 11
Number of Kinetic Trials
9
8
9
11
11
10
14
18
12
16
Note: The number of trials refers to the smaller of the shoe markers or heel windows
markers conditions. For instance, if the shoe markers condition had 15 good trials and
the heel windows condition only had 11 good trials then the number of trials used for the
analysis was 11.
89
Graphical Results
As indicated above, seventeen feet had enough trials in both the shoe markers and
heel windows markers to analyze their kinematics. The three kinematic variables being
evaluated were eversion excursion, percent stance at peak eversion, and maximal
instantaneous eversion velocity. Seventeen feet with three graphs each means there were
fifty one graphs for the single subject analysis of the kinematic variables. There were
only ten feet with enough trials under both marker conditions to analyze the kinetic
parameter of maximal instantaneous vertical loading rate. This means, in total, the
graphical analysis produced sixty one graphs.
Of these sixty one, seventeen graphs indicated significant differences between the
shoe markers and the heel windows markers conditions. Two examples of these graphs
are shown in Figures 32 and 33. As discussed previously, one of the issues with a
graphical analysis is that the combination of the subjective nature of their interpretation
and variability in the parameters measured can make them difficult to interpret. An
example of this difficulty can be seen in Figures 34 and 35 which show examples of
situations where the statistical analysis indicated significant differences between
conditions while the graphical analysis mayor may not demonstrate significant
differences between conditions.
90
R06R - Percent Stance at Maximum. Eversion Excursion
b,. b,.
Mb,.b,.Mb,. b,.b,.b,. b,.b b,.b,.b,.
b,. b,.
~
0 60
.....
en
t) 55;;.-
~ 50
~ 45(1)
~-.
40~~
(1)'-" 35~ 30
r:/J
~ 25(1)
B 20(1)
~ 0 5 10 15 20 25 30 35 40 45
Trial Number
Figure 32. Example of a graphical analysis which demonstrated significant differences
between shoe markers and heel windows markers for the percent stance at maximum.
eversion excursion. Graph shows the percent stance at which peak eversion occurred for
subject R06R. The squares are the shoe markers trials and the triangles are the heel
windows trials.
R14R - Eversion Excursion
o
o
33
0-
'-' 31
'"<l)~ 29
<l)
0 27
.S 25
=:
,$2 23
'"1-0:::I 21~
~ 19
=:
,$2 17
'"1-0<l)
&5 15
0
o 0
%Cb::PDOD D
o 0 0
10 20
Trial Number
30 40 50
Figure 33, Example of a graphical analysis which demonstrated significant differences
between shoe markers and heel windows markers for eversion excursion. Graph shows
the eversion excursion for subject R14R. The squares are the shoe marker trials and the
triangles ad the heel windows trials.
91
R08R - Eversion Excursion
16
6
4
8
4540353025
of::,
o
20
o
o
o
15105
D
00 .. #0
[jJ 0 0
o 0
o
o
,-...
o
'-'
~ 14
~o 12
.9
§ 10
.~
~
c
.9
~(I)
&i
Trial Number
Figure 34. A graphical example where there mayor may not be a significant difference
between the two marker conditions. This graph shows eversion excursion in degrees for
subject R08R. The squares represent the shoe marker trials and the triangles represent
the heel windows trials.
500
450
400
!
350 '
300
250,0
200
150
100
R13L Maximal Instantaneous Eversion Velocity
o 10 20
Trial Number
30 40 50
Figure 35. A second graphical example where there mayor may not be a significant
difference between the two marker conditions. The graph shows maximal instantaneous
eversion velocity in degrees per seconds Cis) for subject R13L. The squares represent
the shoe markers trials and the triangles represent the heel windows markers trials.
92
Statistical Analysis Results
Seventeen feet had autocorrelations assessed for all trials under the shoe markers
condition and the heel windows marker condition for the three kinematic variables. This
means there were one hundred and two autocorrelation analyses performed for the
kinematic variables. Ten feet had autocorrelation analyses for both the shoe markers and
heel windows markers conditions for the one kinetic parameter. This added twenty
additional autocorrelation analyses, bringing the total to one hundred twenty two
autocorrelation analyses.
The results of the tests for autocorrelation are shown in Tables 14 and 15. Of the
one hundred twenty two autocorrelation analyses performed only five, or four percent
demonstrated statistically significant autocorrelations. Of the one hundred twenty two
Shapiro-Wilks tests for normality, eleven, or nine percent, returned results withp < .05,
suggesting they were not a normal distribution.
The results of the individual t tests and the Model Statistics Analysis can be seen
in Tables 16 through 19. According to both the independent t test and Model Statistics
methods, ten of the seventeen feet showed significant differences between the shoe
markers and heel windows markers for eversion excursion (Table 16). Ten feet also
showed significant differences between conditions for the percent stance at which peak
eversion occurred (Table 17). For maximal instantaneous eversion velocity only seven
feet demonstrated significant differences between conditions (Table 18) and for maximal
instantaneous vertical loading rate only two feet demonstrated significant differences
93
between conditions (Table 19). In all instances where the Model Statistics method
indicated there was a significant difference between conditions for that foot, the
independent t test also yielded a p value less than .05, indicating good agreement between
these two statistical analysis methods.
94
Table 14. Autocorrelation coefficients for the three kinematic variables for both the shoe
markers and the heel windows markers conditions.
Percent Stance at Peak Maximal Instantaneous
Eversion Excursion Eversion Eversion Velocity
Shoe Heel Shoe Heel Shoe Heel
Subject Markers Windows Markers Windows Markers Windows
R03 R .215 .186 -.011 -.441 .234 .117
R05R -.553 * .372 .172 .017 -.355 .306
R06L .160 -.330 -.167 .219 -.211 .321
R06R .252 .012 .100 -.296 .066 -.169
R07L -.232 -.243 .215 -.224 -.181 .047
R07R .095 -.143 -.259 -.337 .223 -.266
R08L .039 .036 .303 .201 -.072 -.118
R08R -.251 -.156 .318 -.174 .174 -.096
R10R -.080 -.355 .426 .226 -.086 .551
R11 L -.327 .267 -.180 -.011 -.316 .293
R11 R .212 -.104 -.081 .057 .048 -.099
R12L -.006 -.119 -.241 .651 * .235 .408
R13 L .392 .430 * .420 * .001 -.246 .191
R13R -.020 .050 -.159 .300 .020 .111
R14L .092 .108 -.139 -.152 -.069 -.043
R14R .014 -.244 -.074 -.001 .083 .154
R15 L -.347 .404 .263 -.253 -.352 .397
Note: Autocorrelations which were significant at the p < .05 level are indicated with an
asterisk (*).
95
Table 15. Autocorrelation coefficients for the maximal instantaneous vertical loading
rate for both shoe markers and heel windows markers conditions.
Maximal Instantaneous Vertical Loading Rate
Subject
R03R
R05 R
R06L
R06R
R07L
R07R
R08 L
R08R
R10R
Rll L
R11 R
R12L
R13 L
R13R
R14L
R14R
R15 L
Shoe Markers
.283
.078
.149
.333
.048
.351
.591 *
.119
.064
-.021
Heel Windows
.122
.349
-.380
-.177
.047
.153
.130
.116
.018
-.027
Note: Autocorrelations which were significant at the p < .05 level are marked with an
asterisks (*). Feet which did not have enough trials to analyze the kinetic parameters in a
single subject design are indicated with a dash (-).
96
Table 16. Independent t test and Model Statistics results for eversion excursion.
Percent Absolute
change from Critical Mean
Subject group design t test p value Difference Difference Significant?
R03 R 0.60 .793 1.560 0.183 NO
R05R -7.33 .256 1.685 0.904 NO
R06L 4.54 .850 1.056 0.101 NO
R06R 10.88 .005 * 1.346 2.021 YES
R07L -17.06 .001 * 0.617 2.027 YES
R07R -4.22 .367 1.184 0.541 NO
R08L 43.44 < .001 * 1.151 3.226 YES
R08R 12.21 .018 * 0.901 1.115 YES
RI0R 23.57 .001 * 2.097 4.048 YES
Rll L -7.71 .459 2.210 0.840 NO
R11R -22.95 .016 * 2.876 3.998 YES
R12L 32.26 < .001 * 1.375 3.304 YES
R13 L -13.05 .108 2.068 1.743 NO
R13R 7.32 < .001 * 2.141 2.516 YES
R14L -11.87 .026 * 0.975 1.253 YES
R14R -16.85 <.001 * 1.141 5.547 YES
R15 L 3.38 .792 0.001 0.005 NO
Note: Feet where the independent t test indicated there were significant differences
between shoe and heel windows markers are indicated with an asterisks (*). Trials where
the Model Statistics indicated a significant difference between the two conditions are
highlighted in bold text.
97
Table 17. Independent t test and Model Statistics results for percent stance at which
maximal eversion occurs.
Percent
Change Absolute
from Group Critical Mean
Subject Design t test p value Difference Difference Significant?
R03R 1.63 .318 2.551 1.764 NO
R05R 11.40 < .001 * 1.593 5.695 YES
R06L 16.47 .002 * 2.219 3.838 YES
R06R 20.93 < .001 * 1.651 8.561 YES
R07L -17.75 < .001 * 0.617 3.251 YES
R07R -1.97 .265 2.551 1.456 NO
R08L 8.76 < .001 * 19.641 53.542 YES
R08R 8.00 .003 * 17.687 28.222 YES
RI0R 6.69 .018 * 2.708 3.425 YES
Rll L 9.49 .099 2.505 2.189 NO
Rll R -12.13 .059 5.572 5.589 NO
R12L -2.30 .814 5.614 0.616 NO
R13 L 10.51 .109 2.883 2.453 NO
R13R -11.88 < .001 * 1.795 4.907 YES
R14L -7.43 .018 * 1.418 1.421 YES
R14R 6.38 .012 * 2.723 3.175 YES
R15 L 8.36 .289 2.354 1.280 NO
Note: Feet where the independent t test indicated there were significant differences
between shoe and heel windows markers are indicated with an asterisks (*). Trials where
the Model Statistics indicated a significant difference between the two conditions are
highlighted in bold text.
98
Table 18. Independent t test and Model Statistics results for maximum instantaneous
eversion velocity.
Percent
Change
from Absolute
Group Critical Mean
Subject Design t test p value Difference Difference Significant?
R03R -3.83 .960 35.941 0.777 NO
R05R -5.71 .283 19.623 10.200 NO
R06L -4.38 .431 50.103 20.803 NO
R06R -21.84 < .001 * 30.147 70.383 YES
R07L -4.72 .189 27.008 18.390 NO
R07R -2.43 .708 15.833 2.993 NO
R08L 31.39 .009 * 1.742 3.281 YES
R08R -14.86 .004 * 3.016 4.762 YES
R10R 22.84 .063 59.877 58.935 NO
Rll L -9.29 .243 44.657 22.527 NO
Rll R 6.48 .599 65.037 17.352 NO
R12L 19.16 .012 * 38.467 47.437 YES
R13L -13.36 .0146 * 41.172 53.542 YES
R13 R 5.56 .101 39.477 33.648 NO
R14L -23.81 .004 * 48.531 65.743 YES
R14R -28.29 < .001 * 21.367 122.630 YES
R15 L 10.87 .472 31.159 11.552 NO
Note: Feet where the independent t test indicated there were significant differences
between shoe and heel windows markers are indicated with an asterisks (*). Trials where
the Model Statistics indicated a significant difference between the two conditions are
highlighted in bold text.
99
Table 19. Independent t test and Model Statistics results for maximal instantaneous
vertical loading rate
Percent
Change
from Absolute
Group Critical Mean
Subject Design t testp value Difference Difference Significant?
R03R -5.76 .291 12.714 5.918 NO
R05R -11.78 <.001* 4.940 9.585 YES
R06R -2.68 .369 7.012 2.606 NO
R07L -6.59 .282 8.776 3.923 NO
R07R -3.71 .397 5.549 2.427 NO
R08R 2.15 .537 8.176 2.425 NO
Rl3 L -4.49 .536 6.543 2.249 NO
R13R 15.31 .001 * 4.059 7.767 YES
R14L -5.87 .1067 6.275 4.977 NO
R14R -0.81 .052 7.019 6.431 NO
Note: Feet where the independent t test indicated there were significant differences
between shoe and heel windows markers are indicated with an asterisks (*). Trials where
the Model Statistics indicated a significant difference between the two conditions are
highlighted in bold text.
100
CHAPTER VII
SINGLE SUBJECT ANALYSIS DISCUSSION
The purpose of this single subject analysis was to determine if individual feet
demonstrated significant differences in kinematic and kinetic parameters between the two
shoe conditions which may have been masked by the group design analysis method.
Since there were no significant differences between conditions in the group analysis, the
hypothesis for this portion of the study was that the feet which demonstrated larger
percent differences in the group design would demonstrate significant differences
between conditions for the kinematic and kinetic parameters. The results support this
hypothesis.
Due to the requirement of a higher number of trials for the single subject analysis,
only seventeen of the original twenty feet were analyzed for the kinematic variables, and
only ten were analyzed for the kinetic variable. For the kinematic variables, ten feet
demonstrated significant differences in eversion excursion between the shoe markers and
heel windows markers conditions. Ten feet also demonstrated significant differences
between conditions in the percent stance at which peak eversion occurred and seven
subjects demonstrated significant differences between conditions in maximal
instantaneous eversion velocity. Of the ten feet analyzed for the maximal instantaneous
vertical loading rate, only two demonstrated significant differences between conditions.
In all cases, the feet which demonstrated significant differences demonstrated larger
absolute percent changes in the group design. However, while large in magnitude, these
101
changes were not always in the same direction, suggesting a single subject analysis might
better detect differences between the two marker conditions, differences which were
masked when using a traditional group analysis. The remainder of this chapter will
discuss the appropriateness of using statistical analyses for the single subject design in
this particular study, differences revealed in the single subject analysis which were not
apparent in the group analysis, and discuss why the number of feet demonstrating a
significant difference between conditions decreased for the kinetic parameters.
Appropriateness of Using Statistical Analysis
When performing a statistical analysis on single subject data, there are two main
criteria which must be met, including the data not being auto-correlated and the single
subject data having normal distributions (Bates, 1996; Bates et aI., 2000; Richards et aI.,
1999). Bates (1996) indicates results of his investigations suggest that autocorrelation is
not a significant problem in running gait data, a statement supported by the findings of
this study. Only five of the one hundred twenty two autocorrelations demonstrated
statistical significance. This suggests that autocorrelation was not an issue in applying
statistical analyses to this data.
On the normality of single subject data Bates (1996) suggests the nature of data
produced when studying human gait means this assumption is often violated, even in
traditional group designs. However he also suggests that statistical analyses such as the t
test are robust against minor violations of this assumption, and therefore minor violations
102
of normalcy should not be considered grounds against using statistical analysis with
single subject data (Bates et aI., 2000). This statement was also supported by the results
of this study. Eleven of the Shapiro-Wilks tests performed to assess the normality of the
data indicated non-normal distributions, meaning the bulk of the data was approximately
normal in distribution. When examined individually, the eleven results which indicated
non-normality were most likely the result of low variability with a large cluster of scores
at a particular value, and not a distribution which was wildly different than normal.
These results suggest that a statistical analysis was appropriate for the single subject data
used in this study.
The statistical significance of the single subject results was analyzed by both a
traditional independent samples t test and a unique analysis method called Model
Statistics. Model Statistics was developed specifically for analyzing single subject data
and has been used in numerous studies (Bates et aI., 1992; Bates et aI., 2000; Dufek &
Bates, 1991; Dufek, Bates, & Davis, 1995; Dufek, Bates, Stergiou et aI., 1995; Stergiou
& Bates, 1997). While not a main focus of this particular study, the results of these two
different statistical analyses were always in agreement on the statistical significance or
lack of significance of any particular variable. This suggests Model Statistics is a robust
method for incorporating statistical analysis into single subject design data.
103
Kinematic Differences Revealed with Heel Windows and Single Subject Analysis
The single subject analysis suggested that ten of the seventeen feet analyzed
demonstrated significant differences in kinematic parameters between the shoe markers
and heel windows marker conditions. This has important implications when considering
the most optimal method to measure rear foot motion during running. For instance,
eversion excursion and maximal instantaneous eversion velocity are two of the most
commonly assessed biomechanical parameters and are widely thought to be related to a
variety of overuse running injuries (Donoghue et aI., 2008; Grau et aI., 2008; Paavola et
aI., 2002; Pohl et aI., 2008; Reinking & Hayes, 2006; Tweed et aI., 2008; Warren, 1984,
1990; T. Willems et aI., 2007). If one assessed these parameters using the shoe markers
one might conclude that the parameters are within normal physiologic ranges and not
potential sources for injury. However, as the results ofthe single subject analysis
demonstrate, the motion reported by the shoe based markers mayor may not reflect the
true motion of the rear foot, as the measurement for one variable may differ by as much
as 40% between the two conditions. In such a case a clinician might wrongly rule out
potentially injurious biomechanical markers. An example of such a situation using the
single subject data from this study is discussed below.
As the heel makes contact with the ground the rear foot everts and the knee flexes
to absorb some of the impact forces. At the end of stance, during push off, the rear foot
inverts and the knee extends. Caught between the subtalar joint and the knee is the tibia,
which, given the shape of the distal and proximal articulations, must rotate internally with
~~------------------
104
rear foot eversion and knee flexion and then rotate externally with rear foot inversion and
knee extension. If rear foot eversion is prolonged so that peak eversion and peak knee
flexion do not occur at the same time, then the knee will start extending while the foot is
still everted. This will rotate the ends of the tibia in opposite directions, inducing torsion
stress within the bone and soft tissue structures of the knee. Several authors have
suggested this mal-coupling could be a major contributor to running injuries at the knee
(Dierks & Davis, 2007; Stergiou & Bates, 1997; Stergiou et aI., 1999; Tiberio, 1987).
Figures 36 and 37 shows the rear foot eversion curves plotted with the knee
flexion curves for two sample feet. These curves are averages from all the trials used in
the single subject analysis for these feet and results from both marker conditions are
shown. The solid lines show results from the shoe markers trials while the dashed lines
show results from the heel windows markers trials. When examining the data from the
shoe markers condition, maximum rear foot eversion and maximum knee flexion appear
to be fairly synchronous and one might conclude there are no coupling related issues with
these parameters. However, when the same parameters are examined using data from the
heel windows markers a noticeable shift in the coupling pattern is apparent, with maximal
eversion occurring later in the stance phase. This means these subjects will have a time
period when their knee is extending while their foot is still everting, a scenario which, as
described above, might potentially contribute to injury of the soft tissue structures at the
knee.
Figures 36 and 37 are only two examples of how the shoe markers might mask
potentially injurious biomechanics. Overall five of the seventeen feet analyzed in the
105
single subject analysis demonstrated similar patterns, with the timing of peak eversion
excursion and knee flexion being noticeably different when measured with the heel
windows markers instead of the shoe markers. This many individual feet showing
differences suggests this observation is not a fluke due to the unique kinematics of one or
two individuals. However, these results were not found in every individual foot
examined, suggesting this mal-coupling occurs on an individual basis. This example both
reinforces how potentially injurious biomechanical patterns could be masked when
assessed using both the shoe markers and a group analysis method and provides more
evidence for the importance of using the heel windows method for a truer measure of rear
foot motion.
-- -- -- - - ------
R06R - Eversion Excursion and Knee Flexion
60 25
----.
-+ 20 ~'-"
::: 50 (t> ~
----..8 >-t15 CIl
o ;.< o' e;'-"~ 10 =::l '"rj~--~ 40 ----. 0~ 5 + 0'-" .....tTl
::: 30
- -.- 0 -<
0 . (l)
.... . >-t
;.< ..
-5 CIl.
_.
~ 20 . 0
-
.
.
=::l~'7' .. -10 tTl>~'-" ..~ ::: ..
-15 -< =::l~ .~ 10 .. (t> -> _ c•• _ >-t(JQ.. ..
-20 CIl (t>::: ...............
_.
o ----.E 0 =::l-3;.<
-25 ----.I:il I
'-"0 10 20 30 40 50 60 70 80 90 100
Percent Stance
Figure 36. Sample rear foot eversion and knee flexion curves for both the shoe markers
and heel windows marker conditions. The solid lines are from the shoe markers
condition while the dashed lines are from the heel windows marker condition. Knee
flexion is on the left vertical axis while rear foot eversion is on the right vertical axis.
106
R05R - Eversion Excursion and Knee Flexion
.--., 60
+
'-'
.--., ,§ 50
~~
(1)-
bb~ 40~
o
35
.1 §'
.", 30 "
•• I (D
.- I ~ ~
•
••••• 25 g' ~
20 .--., 'Tj~ + 0
••• 15 '-' S
10
5
o
-5
-10
-15 '7'
'-'
o 10 20 30 40 50 60 70 80 90 100
Percent Stance
Figure 37. A second example of rear foot eversion and knee flexion curves for both the
shoe markers and heel windows marker conditions. The solid lines are from the shoe
markers condition while the dashed lines are form the heel windows marker condition.
Knee flexion is shown on the left vertical axis while rear foot eversion is shown on the
right vertical axis.
Differences in Maximal Instantaneous Vertical Loading Rates
An interesting finding in the single subject analysis was that while a substantial
number of feet demonstrated significant differences between marker conditions for the
kinematic variables, only two feet demonstrated a significant difference for the kinetic
variable. One potential reason could be that, due to the number of trials required for the
single subject analysis, fewer feet were analyzed. If more feet had been analyzed,
perhaps more would have demonstrated significant differences between conditions.
However, it is interesting to note that the percent changes from the group design showed,
107
on average, this particulate parameter had the smallest percent change between the two
marker conditions. The two feet with significant differences in the single subject analysis
were the exceptions to this trend, suggesting that even if more feet had been added, they
most likely would not have shown significant differences between conditions.
One potential reason for the lack of differences between conditions involves a
reconsideration of the role of the impact forces experienced during running. In general,
the vertical ground reaction force experienced by heel striking runners has two peaks, an
impact peak and an active peak. The impact peak is often around 1.5 times body weight
while the active peak is around 2.5 to 3 times body weight, though these magnitudes vary
with running speed (Cavanagh & Lafortune, 1980). Traditionally, it has been thought
that excessively large ground reaction forces, especially during the impact peak, playa
role in the development of running related injuries (Cavanagh & Lafortune, 1980; Hreljac
et aI., 2000; Messier & Pittala, 1988; Milner, Ferber et aI., 2006). However, Nigg (2001)
has proposed new way of looking at ground reaction forces and their role in running
InJurIes.
According to Nigg's model, impact forces in running can be thought of as input
signals to the body which contain both amplitude and frequency components. This input
signal induces vibrations in both the soft and bony tissues of the lower limb. The
vibrations within the soft tissue are not only uncomfortable, but also inefficient from a
metabolic standpoint since they cost energy and require larger oxygen consumption from
the individual. Therefore, based on the input signal from one foot strike, the central
nervous system will tune the leg muscles to reduce muscle vibrations on the next foot
108
strike. In essence, the central nervous system is anticipating a certain level of impact
force and will tune the leg muscles appropriately. Subtle kinematic differences from
stride to stride are a result of this tuning.
There is a growing body of evidence supporting this theory, though it should be
noted that the bulk of this research was carried out in the same laboratory. Studies
examining muscle activity during standing have shown that muscle activity changes in
response to vibrations applied to the lower limb (Wakeling & Nigg, 2001; Wakeling,
Nigg, & Rozitis, 2002). Studies using a pendulums to apply impact forces to the plantar
surface of a foot have found similar results (Wakeling, Von Tschamer, Nigg, & Stergiou,
2001). Studies using actual running have also indicated that muscle activity changes
from foot strike to foot strike (Boyer & Nigg, 2004; Wakeling, Pascual, & Nigg, 2002).
Lastly, computer modeling simulations have also shown support for this model,
suggesting that muscle tuning could result from attempts to keep either soft tissue
vibrations or the amount force applied to the body consistent from foot strike to foot
strike (Nigg & Liu, 1999; Zadpoor & Nikooyan, 2010).
Though the concept of muscle tuning is fairly new, it could potentially explain the
lack of statistically significant differences in loading rates between the two marker
conditions. While this study calculated loading rates not impact forces, the loading rates
could easily be influenced by the magnitude of the impact force. Therefore, ifmuscle
tuning is happening, either the vibrations applied to the body or the forces applied to the
body are held constant. Under this model one would not expect to see differences in
loading rates between the two marker conditions.
109
CHAPTER VIII
CONCLUSIONS, LIMITATIONS, FUTURE DIRECTIONS
Conclusions
The purpose of this study was to assess the effects of two marker systems on
kinematic and kinetic measurements commonly made during a running gait analysis. A
secondary purpose was to recommend a marker placement method for a planned
longitudinal study on overuse injuries in runners. As the data analysis for the study
evolved it became apparent that the originally planned analysis may not have been
suitable. As such, a tertiary purpose of this study was to explore any differences when
the data was analyzed using a single subject analysis instead of a traditional group
analysis. From the results of this study the following conclusions are drawn.
1) It is recommended that the heel windows marker placement method be used
when conducting a running gait analysis, especially when rear foot motion is one of the
parameters of interest. The heel windows method may yield insights into potential injury
related mechanisms such as changes in movement variability or joint and segment
coupling in the lower extremity that are not evident when markers are placed directly on
the shoes. It is suggested that using the heel windows method and placing markers
110
directly on the skin provides a more accurate measurement of the rear foot motion during
running than does placing markers on the exterior of the shoe.
2) For a longitudinal study on overuse running injuries, it is recommended that each
subject obtain a baseline measurement and then a single subject analysis methodology be
used to compare each subject to their own baseline. The results of this study suggest
group designs mask individual changes in kinematics which could be important
biomechanical markers in regards to overuse running injuries. A similar result will occur
if subjects are compared to some normative average value for a given parameter. This
average will not pick up small changes in an individual which may be precursors to or
results of, an injury. Therefore this study highlights the importance of a single subject
design for a longitudinal evaluation of injury risk in runners.
This has important implications for study design and planning. Studies intent on
using single subject design should consider how the number of subjects affects the
statistical power of the results and plan accordingly. Most likely, this suggests more
subjects will be needed than would be needed for a traditional group design and more
trials per subject will have to be collected to ensure sufficient statistical power in the
analysis. Additionally, researchers should consider the expected nature of their data for
adherence to the requirements of autocorrelation and normalcy required for performing
statistical analysis on single subject design data.
111
Limitations
Perhaps the most visible limitation to this study is the extent to which the running
the subjects performed in this study actually reflects their true running kinematics. The
subjects were running indoors in a laboratory, not outdoors over ground where they
normally would run. Several subjects indicated the pace they ran was similar to their
daily run pace, while others thought it was a bit slower. Additionally, data was collected
over the course of a 10 meter straight in the laboratory and even though subjects were
instructed to run at an even pace they may have inadvertently accelerated through the
data collection zone.
Though subjects could run continuous laps, the track around which they ran was
fairly small with tight turns. Though it did not appear this way from watching the trials,
there is the possibility the subjects were not completely upright and using their normal
running gait when passing through the data collection zone. If they were still slightly at
an angle this would artificially enhance their leg varus with respect to the floor. It has
been suggested that increased tibial varus leads to excessive compensatory pronation
(James et aI., 1978). If this happened more during the heel windows trials it could
potentially suggest kinematic differences which were simply a result of the running
course. However, visual inspection of the trials suggests this did not occur and the results
found in this study match those presented in the literature, suggesting this should not be
considered a significant limitation.
112
Another, potentially more serious limitation to this study was the fact that the
markers in the heel windows condition had to be taped to the heels of the individual, the
effects of which are not known. Though all subjects indicated the shoe felt normal even
with the marker bases taped to their calcaneous, the effect to which this extra material
subconsciously modified their kinematics is unknown. There is the possibility that the
presence of the tape against the skin provided the subject with additional sensory
feedback during the heel windows marker trials which allowed them to subtly adjust their
gait from stride to stride. This phenomenon has previously been observed in other
situations where enhanced cutaneous feedback has been used to modify kinematics such
as placing tape over the vastus medialis muscle in individuals with patellofemoral pain
syndrome (Christou, 2004; MacGregor, Gerlach, Mellor, & Hodges, 2005; Tiberio,
1987)).
A final limitation to this study involves the model used to represent the motion of
the foot. This model focused on rear foot motion, and thus three markers were placed on
the rear foot, a method which, in reality, describes motion of the calcaneous relative to
the tibia. Motion between the calcaneous, talus, and navicular, and motion between the
mid-tarsal joints play an important role in the motion of the entire foot. By improving the
model to include a multi segmented foot one might be more able to accurately model the
true motion of the entire foot.
113
Future Directions
This study found significant differences in kinematics between the shoe markers
and heel windows markers conditions and it was suggested that using the heel windows
markers has the potential to mask injury related biomechanical markers. While this study
did show some preliminary work comparing the timing of peak knee flexion and rear foot
eversion which suggested this might be true, an in depth investigation of this possibility
was not the main focus of this study. Future studies could reinforce the conclusions of
this study that heel windows markers should be used for running motion capture, when
possible, by exploring the effects of the two marker systems on coordination patterns
between the lower limbs. This would be especially important considering the limited
nature of the coordination assessments in the current study.
Another direction which should be explored in future studies is the applicability
of the heel windows concept to the front of the shoe in addition to the heel counter. This
method has been used for assessing foot motion while walking (Wolf et aI., 2008).
However, to the best ofthis author's knowledge this has never been done for a running
gait analysis using standard running shoes. Exploring this possibility might allow the use
of a multi-segmented foot model and, by extension, more accurate quantification of the
motion of the entire foot, not just the calcaneous.
APPENDIX A
SUBJECT INFORMED CONSENT FORM
114
115
CONSENT FORM
You are invited to participate in a research study conducted by Drs. Li-Shan Chou, Louis
Osternig, Stan James, and graduate student James Becker. We hope to develop a protocol
for a thorough clinical and biomechanical assessment ofrunners, and using this protocol,
track the runners over time to see if there are any changes in these parameters prior to,
during, or post injury. At this point we are testing, refining, and trying to validate the
protocol.
If you decide to participate, you will be tested in the Motion Analysis Laboratory at the
University of Oregon.
TESTING PROCEDURES: The assessments in the Motion Analysis Lab will include
both clinical and biomechanical evaluations. The clinical evaluation will include
measures of your body alignment, joint range of motion, and muscle strength. The
alignment and range of motion assessments will be made by a trained clinician while the
strength measures will be tested. For the running gait analysis reflective markers will be
placed on your at selected bony landmarks and muscle surfaces to record the motion of
each individual body segment. You will run laps around the laboratory space and your
body movement (indicated by motion of reflective makers) during running will be
recorded by our optoelectronic cameras for further analysis. With your approval we may
also record your running with traditional video cameras and/or take photographs of the
marker set up placed on your body. We will record your running under several different
conditions. In the first condition the markers for your feet will be placed directly on the
outside of your shoe. For the second condition we will drill holes in your running shoe
and the markers will be directly attached to your heel through the shoe. You will be
asked to wear a pair ofpaper physical therapy shorts and sleeveless shirt (tank top) during
testing. The testing session will require a maximum of 3 hours of your time.
COMPENSATION: You will be compensated $75 for participating in this study as
reimbursement for a new pair of running shoes. You should understand that your old
shoes will no longer be usable after your participation in the study.
RISKS AND DISCOMFORTS: We expect that there will be no more risk for you during
these tests than there normally is for you when outside of the laboratory. However,
running in the laboratory is different than running outside. You will be asked to speed up
then slow down over a 20 meter distance. Running laps in the laboratory will require
negotiating tight comers. We will do our best to arrange the lab equipment and furniture
to minimize any discomforts and provide as much room as possible. If you are not
comfortable you may stop the trials at any time. You may feel fatigue during or after the
testing. Our staff member will check with you frequently and provide any required
116
assistance. You will be given frequent breaks as requested. Drilling the holes in your
running shoes will require the removal of the inner lining so there is the possibility of
rubbing or discomfort on your feet. We will do our best to reduce these effects, and
should they still be present you may request additional modifications or stop the trials at
any time. There is also the possibility ofdiscomfort involved in removing adhesive tape
(used for marker placement) from skin at the end of the experiment.
ADDITIONAL INFORMATION: Any information that is obtained in connection with this
study and that can be identified with you will remain confidential and will not be shared
without your permission. Subject identities will be kept confidential by coding the data as to
study, subject pseudonyms, and collection date. The code list will be kept separate and
secure from the actual data files.
Your participation is voluntary. Your decision whether or not to participate will not affect
your relationship with the Department of Human Physiology or University ofOregon. You
do not waive any liability rights for personal injury by signing this form. In spite of all
precautions, you might develop medical complications from participating in this study. If
such complications arise, the researchers will assist you in obtaining appropriate medical
treatment. In addition, ifyou are physically injured because of the project, you and your
insurance company will have to pay your doctor bills. Ifyou are a University ofOregon
student or employee and are covered by a University ofOregon medical plan, that plan
might have terms that apply to your injury. If you have any questions about your rights as a
research subject, you can contact the Office for Protection of Human Subjects, 5219
University of Oregon, Eugene, OR 97403, (541) 346-2510. This office oversees the review
ofthe research to protect your rights and is not involved with this study.
If you decide to participate, you are free to withdraw your consent and discontinue
participation at any time without penalty.
If you have any questions, please feel free to contact Dr. Li-Shan Chou, (541) 346-3391,
Department of Human Physiology, l12C Esslinger Hall, University of Oregon, Eugene
OR, 97403-1240. You will be given a copy of this form to keep. Your signature indicates
that you have read and understand the information provided above, that you willingly
agree to participate, that you may withdraw your consent at any time and discontinue
participation without penalty, that you will receive a copy of this form, and that you are
not waiving any legal claims, rights or remedies.
Name:
------------------
Signature: _
Date:
--------------
APPENDIXB
CLINICAL EVALUAnON FORM
117
118
Running Study Subject Questionnaire and Clinical Evaluation Form
Subject Code: _ Date: _
Age: _
Year in college, if applicable: _
Number of Years Running: _
Approximate Mileage Run per Week: _
Over the course of your running career, have you sustained any running related
injuries? Y N
IfYes then please describe the nature of the injury, diagnosis by a physician,
extent or duration of the injury, and treatment protocols you underwent to
relieve symptoms:
Other Comments ofHistory Information:
General Lower Body Alignment and Mobility Assessment
119
Angle of Gait Subtalar Joint
Motion
Toe In L R Inversion L: R:
Straight L R Eversion L: R:
Toe Out L R ForefootAlignment
Tibial Torsion Neutral L R
Leg Varus to L: R: Varus L R
Floor
Extremity Length L: R: Valgus L R
Standing Arch Position 1st Ray
Type
High L R Plantar Flexed L R
Medium L R Dorsiflexed L R
AnIde Dorsi Neutral L R
Flexion
Knee Extended L: R: Motion 1st Ray
Knee Flexed L: R: Normal L R
Ankle Plantar Mod. Restricted L R
Flexion
Knee Extended L: R: Restricted L R
Knee Flexed L: R: 1st MPJ Joint
Prone Hip Restricted L R
Rotation
Internal L: R: Dosiflexion L: R:
External L: R: Toe Position
Foot Motion Straight L R
Loose L R Heel Varus @ STN L: R:
Tight L R
Normal L R
120
Measurement of Arch Height
(from Williams, McClay, Hammill, and Buchanan (2001). Lower Extremity Kinetic and Kinematic
Differences in High and Low Arched Runners. J. Applied Biomechanics. Vol. 17, pp. 153-161)
1. Height of Dorsum of foot @ 50% foot length: L
2. Truncated Foot Length L R
_____R
(measured from most posterior point of calcaneous to medial joint space of fIrst metatarsal phalangeal
joint).
3. Arch Height Ratio: L
(measurement 1 divided by measurement 2)
____R
Measurements of Strength and Range of Motion
Hip Rotation Strength (manual testing since cannot do this test on the Biodex)
Int.: L R
Ext.: L R
General Flexibility and Range of Motion
Gastroc (w/STJ in IT Band (Ober's)
neutral)
Left Left
Right Right
Hamstrings Inversion
Left Left
Right Right
I Quadriceps Eversion
Left Left
Right Right
APPENDIXC
METHODOLGY FOR ESTABLISHING THE ANATOMIC AND TRACKING
MARKER COORDINATE SYSTEMS
121
122
C.l.l - Markers used to establish the anatomic coordinate system and tracking marker
coordinate system for the foot.
MM: Medial Malleolus
LM: Lateral Malleolus
AJC: Ankle Joint Center
TM: Toe Marker
HT: Heel Top
HB: Heel Bottom
H: Heel
FcoM
Foot Center of
mass
On the most prominent part of the medial malleolus.
On the most prominent part of the lateral malleolus.
The midpoint between MM and LM.
On the shoe approximately over the space between the first
and second metatarsals.
The upper marker on the vertical bisection of the heel
counter.
The lower marker on the vertical bisection of the heel
counter.
Virtual marker located at the midpoint between HT and HB,
collinear with TM
Virtual marker located at the foot center of mass.
123
C.l.2 - Definition of the XYZ anatomic coordinate system for the foot.
The origin of the foot segment is located at the foot center of mass (FCOM) marker. This
point is defined by the heel marker (H) and the toe marker (TM).
COMFoot = H+ 0.44*(TM - H) (2)
The X axis for the foot segment pointed anteriorly and the markers were placed so that it
was parallel with the floor in a normal standing position. It was the normalized unit
vector generated from the vector running from the virtual heel marker (H) to the toe
marker (TM) and defined as:
x=
fM-H
Iffi-HI (3)
The Z axis for the foot segment pointed to the right. It was the normalized vector
generated from the cross product of the vector running from the virtual heel marker (H)
to the toe marker (TM) and the vector running from the virtual heel marker (H) to the
virtual ankle joint center marker (AJC). This vector was defined as:
Ii = (fM - H)x (AJC - H)
I(ffi - H)x (AJC - H)I (4)
The Y axis for the foot segment pointed superiorly and was defined as the cross product
of the vectors defining the z and x axes. It was defined as:
j = kxl. (5)
124
C.1.3 - Definition of the xyz coordinate systems for the tracking markers of the foot.
The y axis of the tracking marker coordinate system was the normalized vector generated
from the vector running from the heel bottom marker (HB) to the heel top marker (HT).
It was defined as:
-> HT-HB
y= IHT-HB\ (6)
The x axis of the tracking marker coordinate system was the normalized vector generated
from the cross product ofthe vector running from the heel bottom marker (HB) to the
heel top marker (HT) and the vector running from the heel bottom marker (HB) to the
lateral heel marker (HL). It was defined as:
->
x=
cm: - HB) x (HT - HB)
ICm: - HB) x (HT - HB)I . (7)
The z axis of the tracking marker coordinate system was the cross product of the vectors
defining the x and y axes of the tracking marker coordinate system. It was defined as:
(8)
125
C.2.I - Markers used to establish the anatomic and tracking marker coordinate systems
of the shank
MM: Medial Malleolus On the most prominent part of the medial malleolus.
LM: Lateral Malleolus On the most prominent part of the lateral malleolus.
AJC: Ankle Joint Center The midpoint between MM and LM.
LK: Lateral Knee On the most prominent part of the lateral femoral
epicondyle.
MK: Medial Knee On the most prominent part of the medial femoral
epicondyle.
KJC: Knee Joint Center The midpoint between MK and LK.
S: Shank On the inferior 1/3 of the medial portion of the shank
collinear with MK and MM.
SCOM
Shank Center of Virtual marker located at the shank center of mass
Mass
I~d'·MK.~.LK
f ,:1,
"/
ly ,I
...._--_._ .._-------
126
C.2.2 - Definition of the XYZ anatomic coordinate system for the shank.
The origin of the shank segment is located at the shank center of mass (SCOM) marker.
This point is defined by the virtual ankle joint center marker (AlC) and the virtual knee
joint center marker (KlC).
CO MShank = KjC + 0.42*(KjC - AjC) (9)
The Y axis for the shank segment pointed superiorly. It was the normalized unit vector
generated from the vector running from the virtual ankle joint center (AlC) to the knee
joint center (KlC) and defined as:
KjC-AjCJ = ---:------,.-
IKjC -AjCI (10)
The X axis for the shank segment pointed anteriorly. It was the normalized vector
generated from the cross product of the vector running from the virtual ankle joint center
(AlC) to the knee joint center (KlC) and the vector running from the virtual ankle joint
center (AlC) to the medial malleolus marker (MM). This vector was defined as:
f=
(KjC - AjC) x (KiM - AjC)
I(KjC - AlC) x (MM - AjC)1 (11)
The Z axis for the shank segment pointed to the right and was defined as the cross
product of the vectors defining the X and Yaxes. It was defined as:
k = LX] (12)
127
C.2.3 - Definition of the xyz coordinate systems for the tracking markers of the shank.
The y axis of the tracking marker coordinate system was the normalized vector generated
from the vector running from the lateral malleolus marker (LM) to the lateral knee
marker (LK). It was defined as:
~ IK-rM
y = IIK- rMl (13)
The x axis of the tracking marker coordinate system was the normalized vector generated
from the cross product of the vector running from the lateral malleolus marker (LM) to
the shank marker (8) and the vector running from the lateral malleolus marker (LM) to
the lateral knee marker (LK). It was defined as:
x=
(s - rM) x (LK - LM)
I(s - rM) x (LK - rM)1 (14)
The z axis of the tracking marker coordinate system for the shank was the cross product
of the vectors defining the x and y axes. It was defined as:
(15)
128
C.3.1 - Markers used to establish the anatomic and tracking marker coordinate systems
of the thigh
T: Thigh Marker
HJC: Hip Joint Center
LK: Lateral Knee
MK: Medial Knee
KJC: Knee Joint Center
TeoM
Thigh Center of
Mass
Thigh marker placed collinear with LK and the greater
trochanter.
The hip joint center defined based on anthropometric
measurements of the ASIS as explained in Vaughan, Davis,
& O'Connor (1999).
On the most prominent part of the lateral femoral
epicondyle.
On the most prominent part of the medial femoral
epicondyle.
The midpoint between MK and LK.
Virtual marker located at the thigh center of mass
y
z
\x
\
\
} MK
LK.'~•
.' KJC)
\
, ..1.:,'
,"
129
C.3.2 - Definition of the XYZ anatomic coordinate system for the thigh.
The origin of the thigh segment is located at the thigh center of mass (TCOM) marker.
This point is defined by the virtual hip joint center marker (HJC) and the virtual knee
joint center marker (KJC).
COMThigh= HjC + O.39*(KfC - HjC) (16)
The Y axis for the thigh segment pointed superiorly. It was the normalized unit vector
generated from the vector running from the virtual knee joint center (KJC) to the virtual
hip joint center (HJC) and defined as:
HjC-KfCJ = -=""""':'"""'"--==
IHjC - KTCI (17)
The X axis for the thigh segment pointed anteriorly. It was the normalized vector
generated from the cross product of the vector running from the virtual knee joint center
(KJC) to the thigh marker (T) and the vector running from the virtual knee joint center
(KJC) to the virtual hip joint center marker (HJC). This vector was defined as:
f=
(T - RIC) x (T - RIC)
I(T - RIC) x (T - RIC) I (18)
The Z axis for the shank segment pointed to the right and was defined as the cross
product of the vectors defining the X and Yaxes. It was defined as:
k = LX] (19)
130
C.2.3 - Definition of the xyz coordinate systems for the tracking markers ofthe thigh.
The y axis of the tracking marker coordinate system was the normalized vector generated
from the vector running from the lateral knee marker (LK) to the thigh marker (T). It was
defined as:
(20)
The x axis of the tracking marker coordinate system was the normalized vector generated
from the cross product of the vector running from the lateral knee marker (LK) to the
thigh marker (T) and the vector running from the lateral knee marker (LK) to the virtual
hip joint center marker (HJC). It was defined as:
....
x=
(T-rK)x (T-rK)
I(T - rK) x (T - 0<)1 (21)
The z axis of the tracking marker coordinate system for the thigh was the cross product of
the vectors defining the x and y axes. It was defmed as:
(22)
131
Placed on the left anterior superior iliac spine.
Placed on the right anterior superior iliac spine.
Place midway between the two posterior superior iliac
spines.SacralSC:
CA.1 - Markers used to establish the anatomic and tracking marker coordinate systems
of the pelvis
LASIS: Left ASIS
RASIS: Right ASIS
PCOM
Pelvis Center of
Mass
Calculated as the point half way along the vector
connecting the PSIS marker with the midpoint of the
vector connecting the two ASIS markers.
\
y
If""" \
"
I ~\
" .1
I
132
C.3.2 - Definition of the XYZ anatomic coordinate system and the xyz tracking marker
coordinate system for the pelvis.
The anatomic and tracking marker coordinate systems are the same for the pelvis
segment since only three markers are being placed on this segment.
The origin of the pelvis segment is located at the pelvic center of mass marker (PCOM).
This point is defined by the midpoint along a vector connecting the PSIS marker with the
midpoint of the vector between the right and left ASIS markers
(23)
The Z and z axes for the pelvis segment pointed to the right. It was the normalized unit
vector generated from the vector running from the left ASIS (LASIS) to the right ASIS
(RASIS) and is defined as:
------+ ------>k = RASIS - LASIS
IRASIS - IASiSI (24)
The Y and y axes for the pelvic segment pointed superiorly. It was the normalized vector
generated from the cross product of the vector connecting the two ASIS markers and the
vector running from the left ASIS marker to the PSIS marker. This vector was defined
as:
CRASIS - IASiS) x (PSIS - IASiS)r = ---:....---:------:"------,----.,..-
ICRASIS - IASiS) x (PSIS - IASiS) I (25)
The X and x axes for the pelvic segment pointed anteriorly and was defined as the cross
product of the vectors defining the Z and Y axes. It was defined as:
" "kl = ] X (26)
133
BIBLIOGRAPHY
Areblad, M., Nigg, B. M., Ekstrand, J., Olsson, K. 0., & Ekstron, H. (1990). Three
Dimensional Measurement of Rearfoot Motion During Running. Journal of
Biomechanics, 23(9), 933 - 940.
Bandholm, T., Boysen, L., Haugaard, S., Kreutzfeldt-Zebis, M., & Beneke, 1. (2008).
Medial Longitudinal Arch Deformation During Quiet Standing and Gait in
Subjects with Medial Tibial Stress Syndrome. Journal ofFoot and Ankle Surgery,
47(2), 89 - 95.
Bates, B. (1996). Single Subject methodology: an alternative approach. Medicine and
Science in Sports and Exercise, 28(5), 631 - 638.
Bates, B., Dufek, J. S., & Davis, H. (1992). The effect of trial size on statistical power.
Medicine and Science in Sports and Exercise, 24(9), 1059 - 1068.
Bates, B., James, C. R., & Dufek, J. S. (2000). Singe Subject Analysis. In N. Stergiou
(Ed.), Innovative Analyses ofHuman Movement. Champaign, Ill: Human Kinetics.
Bates, B., Ostemig, L., Sawhill, J., & James, S. (1983). An assessment of subject
variability, subject-shoe interaction, and the evaluation of running shoes using
ground reaction force data. Journal ofBIomechanics, 16(3), 181 - 191.
Bates, B., Ostemig, L. R, Mason, B. R, & James, S. L. (1979). Functional variability of
the lower extremity during the support phase of running. Medicine and Science in
Sports and Exercise, 11(4),328 - 331.
Bennell, K., Crossley, K., Jayarajan, J., Walton, E., Warden, S., Kiss, Z., et al. (2004).
Ground Reaction Forces and Bone Parameters with Tibial Stress Fracture.
Medicine and Science in Sports and Exercise, 36(3),397 - 404.
Bennett, 1., Reinking, M., Pluemer, B., Pentel, A., Seaton, M., & Killian, C. (2001).
Factors Contributing to the Development of Medial Tibial Stress Syndrome in
High School Runners. Journal ofOrthopaedic & Sports Physical Therapy, 31(9),
504 - 510.
Boyer, K., & Nigg, B. (2004). Muscle activity in the leg is tuned in response to impact
force characteristics. Journal ofBiomechanics, 37, 1583 - 1588.
Butler, R., Davis, I., & Hamill, J. (2006). Interaction of Arch Type and Footwear on
Running Mechanics. American Journal ofSports Medicine, 34(12), 1998 - 2005.
Butler, R, Hamill, J., & Davis, I. (2007). Effect of footwear on high and low arched
runners' mechanics during a prolonged run. Gait and Posture, 26, 219 - 225.
Butler, R., Hillstrom, H., Song, 1., Richards, c., & Davis, I. S. (2008). Arch height index
measurement system - Establishment of reliability and normative values. Journal
ofthe American Podiatric Medical Association, 98(2), 102-106.
134
Caster, B. L., & Bates, B. T. (1995). The assessment of mechanical and neuromuscular
response strategies during landing. Medicine and Science in Sports and Exercise,
27(5), 736 - 744.
Cavanagh, P. R, & Lafortune, M. (1980). Ground reaction forces in distance running.
Journal ofBiomechanics, 13, 397 - 406.
Chang, R, Van Emmerik, R, & Hamill, 1. (2008). Quantifying rearfoot-forefoot
coordination in human walking. Journal ofBiomechanics, 41,3101 - 3105.
Christina, K., White, S., & Gilchrist, L. (2001). Effect of localized muscle fatigue on
vertical ground reaction forces and ankle joint motion in running. Human
Movement Sciences, 20(3), 257 - 276.
Christou, E. (2004). Patella taping increases vastus medialis oblique activity in the
presence ofpatellofemoral pain. Journal ofElectromyography and Kinesiology,
14, 495 - 504.
Cornwall, M. W., & McPoil, T. G. (1995). Comparison of2 dimensional and 3
dimensional rearfoot motion during walking. Clinical Biomechanics, 10(1),36-
40.
Crossley, K., Bennell, K. L., Wrigley, T., & Oakes, B. W. (1999). Ground reaction
forces, bone characteristics and tibial stress fractures in male runners. Medicine
and Science in Sports and Exercise, 31, 1088 - 1093.
De Wit, B., De Clercq, D., & Aerts, P. (2000). Biomechanical analysis of the stance
phase during barefoot and shod running. Journal ofBIomechanics, 33, 269 - 278.
De Wit, B., De Clercq, D., & Lenoir, M. (1995). The effect ofvarying midsole hardness
on impact forces and foot motion during foot contact in running. Journal of
Applied Biomechanics, 11(4),395 - 406.
DeLeo, A., Dierks, T., Ferber, R, & Davis, 1. (2004). Lower Extremity Joint Coupling
During Running: A Current Update. Clinical Biomechanics, 19,983 - 991.
Derrick, T., Dereu, D., & McLean, S. (2002). Impacts and kinematic adjustments during
an exhaustive run. Medicine and Science in Sports and Exercise, 34(6),998 -
1002.
Devita, P., & Skelly, W. (1990). Intrasubject variability oflower extremity joint moments
offorce during the stance phase of running. Human Movement Science, 9(2),99 -
115.
Dierks, T., & Davis, 1. (2007). Discrete and continuous joint coupling relationships in
uninjured recreational runners. Clinical Biomechanics, 22,581 - 591.
Dierks, T., Manal, K., Hamill, J., & Davis, 1. (2008). Proximal and Distal Influences on
Hip and Knee Kinematics in Runners with Patellofemoral Pain During a
Prolonged Run. Journal ofOrthopaedic & Sports Physical Therapy, 38(8), 448 -
456.
135
Donoghue, O. A., Harrison, A. J., Laxton, P., & Jones, R. K. (2008). Lower limb
kinematics of subjects with chronic achilles tendon injury during running. Print,
16(1),23-38.
Dufek, J. S., & Bates, B. (1991). Dynamic performance assessment of selected sport
shoes on impact forces. Medicine and Science in Sports and Exercise, 23(9), 1062
- 1067.
Dufek, J. S., Bates, B., & Davis, H. (1995). The effect of trial size and variability on
statistical power. Medicine and Science in Sports and Exercise, 27(2),288 - 295.
Dufek, 1. S., Bates, B. T., Stergiou, N., & James, C. R. (1995). Interactive effects
between group and single subject response patterns. Human Movement Science,
14,301 - 323.
Duysens, J., & Van de Crommert, H. (1998). Neural control of locomotion; Part 1: The
central pattern generator from cats to humans. Gait and Posture, 7, 131 - 141.
Edington, C., Frederick, E. C., & Cavanagh, P. (1990). Rear Foot Motion in Distance
Running. In P. R. Cavanagh (Ed.), Biomechanics ofDistance Running.
Champaign, IL: Human Kinetics.
Eslami, M., Begon, M., Farahpour, N., & Allard, P. (2007). Forefoot-rearfoot coupling
patterns and tibial internal rotation during stance phase of barefoot versus shod
running. Clinical Biomechanics, 22, 74 - 80.
Ferber, R., McClay Davis, 1., and Williams, D. (2005). Effect of foot orthotics on rearfoot
and tibia joint coupling patterns and variability. Journal ofBiomechanics, 38,477
- 483.
Grau, S., Maiwald, C., Krauss, 1., Axmann, D., Janssen, P., & Horstmann, T. (2008).
What are causes and treatment strategies for patellar-tendinopathy in female
runners? J Biomech, 41(9),2042-2046.
Grood, E., & Suntay, W. (1983). A joint coordinate system for the clinical description of
three dimensional motions: applications to the knee. Journal ofBiomechanical
Engineering, 105, 136 - 144.
Hamill, 1., Bates, B., & Holt, K. (1992). Timing of lower extremity joint actions during
treadmill running. Medicine and Science in Sports and Exercise, 24(7),807 - 813.
Hamill, J., van Emmerik, R., Heiderscheit, B., & Li, L. (1999). A dynamical systems
approach to lower extremity running injuries. Clinical Biomechanics, 14, 297 -
308.
Heiderscheit, B., Hamill, 1., & Van Emmerik, R. (2002). Variability of Stride
Characteristics and Joint Coordination Among Individuals with Unilateral
Patellofemoral Pain. Journal ofApplied Biomechanics, 18, 110 - 121.
Hintermann, B., & Nigg, B. (1998). Pronation in Runners: Implications for Injuries.
Sports Medicine, 26(3), 169 - 176.
136
Holden, J. P., & Cavanagh, P. R (1991). The free moment of ground reaction in distance
running and its changes with pronation. J Biomech, 24(10), 887-897.
Hreljac, A (2005). Etiology, prevention, and early intervention of overuse injuries in
runners: a biomechanical perspective. Physical medicine and Rehabilitation
Clinics ofNorth America, 16,651 - 667.
Hreljac, A, Marshall, R N., & Hume, P. A (2000). Evaluation oflower extremity
overuse injury potential in runners. Medicine and Science in Sports and Exercise,
32(9), 1635 - 1641.
Ireland, M., Willson, 1., Ballantyne, B., & McClay Davis, 1. (2003). Hip Muscle Strength
in Females With and Without Patellofemoral Pain. Journal ojOrthopaedic &
Sports Physical Therapy, 33(11),671 - 676.
Jacobs, S., & Berson, B. (1986). Injuries to Runners: A study of entrants to a 10,000
meter race. American Journal ofSports Medicine, 14(2), 151 - 155.
James, S., Bates, B., & Osternig, L. (1978). Injuries to Runners. American Journal oj
Sports Medicine, 6(2),40 - 50.
Jones, B., & Knapik, J. (1999). Physical Training and Exercise Related Injuries:
Surveillance, Research, and Injury Prevention in Military Populations. Sports
Medicine, 27(2), 111 - 125.
Kadaba, M., Ramakrishnan, H., & Wootten, M. (1990). Measurement of Lower
Extremity Kinematics During Level Walking. Journal ojOrthopaedic Research,
8,383 - 392.
Kadaba, M., Ramakrishnan, H., Wootten, M., Gainer, 1., Gorton, G., & Cochran, G.
(1989). Repeatability of Kinematic, Kinetic, and Electromyographic Data in
Normal Adult Gait. Journal ojOrthopaedic Research, 7(6),849 - 860.
Kaufman, K., Brodine, S., Shaffer, R, Johnson, C., & Cullison, T. (1999). The Effect of
Foot Structure and Range ofMotion on Musculoskeletal Overuse Injuries.
American Journal ojSports Medicine, 27(5),585 - 593.
Kibler, W., Goldber, C., & Chandler, T. (1991). Functional biomechanical deficits in
running athletes with plantar fasciitis. American Journal ofSports Medicine,
19(6), 66 - 71.
Korpelainen, R, Orava, S., Karpakka, J., Siira, P., & Hulkko, A (2001). Risk Factors for
Recurrent Stress Fractures in Athletes. American Journal ofSports Medicine,
29(3),304 - 310.
Lun, V., Meeuwisse, W. H., Stergiou, P., & Stefanyshyn, D. (2003). Relation between
running injury and static lower limb alignment in recreational runners. British
Journal ojSports Medicine, 38(9),576 - 580.
Lysholm, J., & Wiklander, J. (1987). Injuries in Runners. American Journal ofSports
Medicine, 15(2), 168 - 170.
137
MacGregor, K., Gerlach, S., Mellor, R., & Hodges, P. (2005). Cutaneous stimulation
from patella tape causes a differential increase in vasti muscle activity in people
withpatellofemoral pain. Journal ofOrthopaedic Research, 23(2),351 - 358.
MacKay-Lyons, M. (2002). Central Pattern Generation of Locomotion: A Review of the
Evidence. Physical Therapy, 82(1),69 - 83.
MacLean, C., McClay Davis, 1., & Hamill, J. (2006). Influence of a custom foot orthotic
intervention on lower extremity dynamics in healthy runners. Clinical
Biomechanics, 21,623 - 630.
Marti, B., Vader, J. P., Minder, C., & Abelin, T. (1988). On the etiology of running
injuries The 1984 Bern Grand-Prix Study. American Journal ofSports Medicine,
16(3), 285 - 293.
McClay, 1. (1995). The Use of Gait Analysis to Enhance the Understanding of Running
Injuries. In R. L. Craik, and Oatis, CA (Ed.), Gait Analysis: Theory and
Application. Champaign, 11: Human Kinetics.
McClay, 1., & Manal, K. (1997). Coupling Parameters in Runners with Normal and
Excessive Pronation. Journal ofApplied Biomechanics, 13, 109 - 124.
McClay, 1., & Manal, K. (1998a). A Comparison of three dimensional lower extremity
kinematics during running between excessive pronators and normals. Clinical
Biomechanics, 13(3), 195 - 203.
McClay, 1., & Manal, K. (1998b). The Influence ofFoot Abduction on Difference
between Two Dimensional and Three Dimensional Rearfoot Motion. Foot and
Ankle International, 19(1), 26 - 31.
McClay, 1., & Manal, K. (1999). Three dimensional kinetic analysis of running:
significance of secondary planes of motion. Medicine and Science in Sports and
Exercise, 31(11), 1629.
Messier, S. P., & Pittala, K. A. (1988). Etiologic factors associated with selected running
injuries. Med Sci Sports Exerc, 20(5),501-505.
Milani, T. L., & Hennig, E. M. (2000). Measurements of Rearfoot Motion during
Running. Sportverl Sportschad, 14, 115 - 120.
Miller, R., Meardon, S., Derrick, T., & Gillette, J. (2008). Continuous Relative Phase
Variability During and Exhaustive Run in Runners with a History of Iliotibial
Band Syndrome. Journal ofApplied Biomechanics, 24, 262 - 270.
Milner, C., Davis, 1., & Hamill, J. (2006). Free moment as a predictor of tibial stress
fracture in distance runners. Journal ofBiomechanics, 39,2819 - 2825.
Milner, c., Ferber, R., Pollard, c., Hamill, J., & Davis, 1. (2006). Biomechanical Factors
Associated with Tibial Stress Fracture in Female Runners. Medicine and Science
in Sports and Exercise, 38(2),323 - 328.
138
Molloy, J. M., Christie, D. D., Teyhen, D. S., Yeykal, N. S., Tragord, B. S., Neal, M. S.,
Nelson, E.S., et aL (2009). Effect ofRunning Shoe Type on the Distribution and
Magnitude ofPlantar Pressures in Individuals with Low-or High Arched Feet.
Journal ofthe American Podiatric Medical Association, 99(4),330 - 338.
Morio, C., Lake, M., Guegen, N., Rao, G., & Baly, L. (2009). The influence of footwear
on foot motion during walking and running. Journal ofBiomechanics, 42,2081 -
2088.
Mundermann, A, Nigg, B. M., Humble, R., & Stefanyshyn, D. (2003). Foot orthotics
affect lower extremity kinematics and kinetics during running. Clinical
Biomechanics, 18, 254 - 262.
Nawoczenski, D., Saltzman, C., & Cook, T. (1998). The Effect of Foot Structure on the
Three Dimensional Kinematic Coupling Behavior of the Leg and Rear Foot.
Physical Therapy, 78(4),404 - 416.
Nigg, B. (1986). Experimental Techniques used in running shoe research. In B. Nigg
(Ed.), Biomechanics ofRunning Shoes. Champaign, II: Human Kinetics.
Nigg, B., & Liu, W. (1999). The effect of muscle stiffness and damping on simulated
impact force peaks during running. Journal ofBiomechanics, 32,849 - 856.
Nigg, B., (1997). Impact forces in running. Current Opinion in Orthopaedics, 8(6),43 -
47.
Nigg, B. M. (2001). The Role of Impact Forces and Foot Pronation: A New Paradigm.
Clinical Journal ofSports Medicine, 11(1), 2 - 9.
Noehren, B., Davis, I., & Hamill, J. (2007). Prospective Study of the biomechanical
factors associated with iliotibial band syndrome. Clinical Biomechanics, 22, 951 -
956.
Novacheck, T. (1988). The Biomechanics ofRunning. Gait and Posture, 7, 77 - 95.
O'Connor, K., & Hamill, J. (2004). The role of selected extrinsic foot muscles during
running. Clinical Biomechanics, 19, 71 - 77.
Paavola, M., Kannus, P., Jarvinen, T., Khan, K., Jozsa, L., & Jarvinen, M. (2002).
Current Concepts Review: Achilles Tendinopathy. Journal ofBone and Joint
Surgery, 84-A(11), 2062 - 2076.
Pohl, M., Hamill, J., & Davis, I. (2009). Biomechanical and Anatomic Factors Associates
with a History of Plantar Fasciitis in Female Runners. Clinical Journal ofSports
Medicine, 19(5),372 - 376.
Pohl, M., Messenger, N., & Buckley, J. (2006). Changes in foot and lower limb coupling
due to systematic variations in step width. Clinical Biomechanics, 21, 175 - 183.
Pohl, M., Messenger, N., & Buckley, J. (2007). Forefoot, rearfoot and shank coupling:
Effect of variations in speed and mode of gait. Gait and Posture, 25,295 - 302.
139
Pohl, M., Mullineaux, D., Milner, c., Hamill, J., & Davis, 1. (2008). Biomechanical
Predictors of retrospective tibial stress fracture in runners. Journal of
Biomechanics, 41, 160 - 1165.
Powell, K., Kohl, H., Caspersen, C., & Blair, S. (1986). An Epidemiological Perspective
on the Causes of Running Injuries. Physician and Sports Medicine, 14(6), 100 -
114.
Powers, C. (2003). The Influence of Altered Lower Extremity Kinematics on
Patellofemoral Joint Dysfunction: A Theoretical Perspective. Journal oj
Orthopaedic & Sports Physical Therapy, 33(11),639 - 646.
Rauh, M., Koepsell, T., & Rivara, F. (2006). Epidemiology of musculoskeletal injuries
among high school cross country runners. American Journal ojEpidemiology,
163, 151 - 159.
Reber, L., Perry, 1., & Pink, M. (1993). Muscular Control of the Ankle in Running.
American Journal ojSports Medicine, 21(6),805 - 810.
Reinking, M. (2006). Exercise Related Leg Pain in Female Collegiate Athletes: The
Influence of Intrinsic and Extrinsic Factors. American Journal ojSports Medicine,
34(9), 1500 - 1507.
Reinking, M., & Hayes, A. (2006). Intrinsic factors associated with exercise-related leg
pain in collegiate cross-country runners. Clin J Sport Med, 16(1), 10-14.
Reinschmidt, C., & Nigg, B. (1995). Influence of heel height on ankle joint moments in
running. Medicine and Science in Sports and Exercise, 27(3),410 - 416.
Reinschmidt, C., Stacoff, A., & Stussi, E. (1992). Heel Movement within a court shoe.
Medicine and Science in Sports and Exercise, 24(12), 1390 - 1395.
Reinschmidt, C., van den Bogert, A. J., Murphy, N., Lundberg, A., & Nigg, B. M. (1997).
Tibiocalcaneal motion during running, measured with external and bone markers.
Clinical Biomechanics, 12(1), 8 - 16.
Richards, S., Taylor, R., Ramasamy, R., & Richards, R. (1999). Single Subject Research:
Applications in Educational and Clinical Settings. San Diego, CA: Singular
Publishing Group.
Ryan, M. B., MacLean, C. L., & Taunton, 1. E. (2006). A review of anthropometric,
biomechanical, neuromuscular, and training related factors associated with
injuries in runners. International Journal ojSports Medicine, 7(2), 120 -137.
Scott, S., & Winter, D. (1990). Internal Forces at Chronic Running Injury Sites. Medicine
and Science in Sports and Exercise, 22(3),357 - 369.
Snyder, K., Earl, J., O'Connor, K., & Ebersole, K. (2009). Resistance training is
accompanied y increases in hip strength and changes in lower extremity
biomechaniC$ during running. Clinical Biomechanics, 24,26 - 34.
140
Soutas-Little, R W., Beavis, G. C., Verstraete, M. C., & Markus, T. L. (1987). Analysis
of foot motion during running using ajoint co-ordinate system. Medicine and
Science in Sports and Exercise, 19(3), 285 - 293.
Souza, R, & Powers, C. (2009). Predictors of Hip Internal Rotation During Running.
American Journal o/Sports Medicine, 37(3),579 - 587.
Stackhouse, C., McClay Davis, 1., & Hamill, J. (2004). Orthotic intervention in forefoor
and rearfoot strike running patterns. Clinical Biomechanics, 19, 64 - 70.
Stacoff, A., & Luethi, S. (1986). Special Aspects of shoe contruction and foot anatomy.
In B. Nigg (Ed.), Biomechanics o/Running Shoes. Champaign, II: Human
Kinetics.
Stacoff, A., Nigg, B., Reinchmidt, C., van den Bogert, A. J., & Lundberg, A. (2000).
Tibiocalcaneal kinematics of barefoot versus shod running. Journal of
Biomechanics, 33, 1387 - 1395.
Stacoff, A., Reinchmidt, C., & Stussi, E. (1992). The movement of the heel within a
running shoe. Medicine and Science in Sports and Exercise, 24(6),695 - 701.
Stacoff, A., Reinschmidt, C., Nigg, B. M., van den Bogert, A. J., Lundberg, A., Denoth,
J., et al. (2001). Effects of shoe sole construction on skeletal motion during
running. Medicine and Science in Sports and Exercise, 33(2),311 - 319.
Stefanyshyn, D., Stergiou, P., Lun, V., Meeuwisse, W., & Worobets, J. (2006). Knee
Angular Impulse as a Predictor of Patellofemoral Pain in Runners. American
Journal o/Sports Medicine, 34(11), 1844 - 1851.
Stergiou, N., & Bates, B. (1997). The relationship between subtalar and knee joint
function as a possible mechanism for running injuries. Gait and Posture, 6, 177 -
185.
Stergiou, N., Bates, B., & James, S. (1999). Asynchrony between subtalar and knee joint
function during running. Medicine and Science in Sports and Exercise, 31(11),
1645 - 1655.
Stergiou, N., Harboume, R, & Cavanaugh, J. (2006). Optimal Movement Variability: A
New Theoretical Perspective for Neurologic Physical Therapy. Journal of
Neurologic Physical Therapy, 30(3), 120 - 129.
Taunton, J. E., Ryan, M.B., Clement, D.B., McKenzie, D.C., Lloyd-Smith, D.R, and
Zumbo, B.D. (2002). A retrospective case control analysis of 2002 running
injuries. British Journal ofSports Medicine, 36, 95 - 101.
Taunton, J. E., Ryan, M.B., Clement, D.B., McKenzie, D.C., Lloyd-Smith, D.R, and
Zumbo, B.D. (2003). A prospective study of running injuries: the Vancouver Sun
Run "In Training" Clinics. British Journal ofSports Medicine, 37,239 - 244.
141
Tiberio, D. (1987). The Effect of Excessive Subtalar Joint Pronation on Patellofemoral
Mechanics: A Theoretical Model. Journal ofOrthopaedic & Sports Physical
Therapy, 9(4), 160 - 165.
Tweed, J. L., Campbell, J. A., & Avil, S. 1. (2008). Biomechanical risk factors in the
development of medial tibial stress syndrome in distance runners. JAm Podiatr
Med Assoc, 98(6),436-444.
Vagenas, G., & Hoshizake, B. (1992). A multivariable analysis oflower extremity
kinematic asymmetry in running. Journal ofApplied Biomechanics, 8(1), 11 - 29.
Van de Crommert, H., Mulder, T., & Duysens, J. (1998). Neural control oflocomotion;
Part 2: sensory control of the central pattern generator and its relation to treadmill
training. Gait and Posture, 7,251 - 263.
Van Gheluwe, B., Tielemans, R., & Roosen, P. (1995). The Influence of Heel Counter
Rigidity on Rearfoot Motion During Running. Journal ofApplied Biomechanics,
11,47-67.
Van Ghent, R. N., Siem, D., Middelkoop, v. M., van Os, A. G., Bierma-Zeinstra, S. M.
A, & Koes, B. W. (2007). Incidence and determinants of lower extremity running
injuries in lond distance runners: a systematic review. British Journal ofSports
Medicine, 41,469 - 480.
Van Mechelen, W. (1992). Running Injuries: A review of the epidemiological literature.
Sports Medicine, 14(5), 320 - 335.
van Middelkoop, M., Kolkman, J., van Ochten, J., Bierma-Zeinstra, S., & Koes, B.
(2006). Course and Predicting Factors of Lower Extremity Injuries After Running
a Marathon. Clinical Journal ofSports Medicine, 17(1), 25 - 30.
Vaughan, C., Davis, B., & O'Connor, J. (1999). Dynamics ofHuman Gait (2 ed.).
Howard Place, Western Cape, South Africa: Kiboho Publishers.
Verdejo, R., & Mills, N. 1. (2004). Heel-shoe interactions and the durability of EVA
foam running-shoe midsoles. Journal ofBiomechanics, 37(9), 1379 - 1386.
Wakeling, J., & Nigg, B. (2001). Modification of soft tissue vibrations in the leg by
muscular activity. Journal ofApplied Physiology, 90(2),412 - 420.
Wakeling, 1., Nigg, B., & Rozitis, A (2002). Muscle activity damps the soft tissue
resonance that occurs in response to pulsed and continuous vibrations. Journal of
Applied Physiology, 92(3), 1093 - 1103.
Wakeling, J., Pascual, S. A, & Nigg, B. (2002). Altering muscle activity in the lower
extremities by running with different shoes. Medicine and Science in Sports and
Exercise, 34(9), 1529 - 1532.
Wakeling, J., Von Tschamer, V., Nigg, B., & Stergiou, P. (2001). Muscle activity in the
leg is tuned in response to ground reaction forces. Journal ofApplied Physiology,
91(3), 1307 - 1317.
142
Walter, S., Hart, L., McIntosh, l, & Sutton, l (1989). The Ontario Cohort Study on
Running Injuries. Archives ofInternal Medicine, 149(11),2561 - 2564.
Warren, B. L. (1984). Anatomical factors associated with predicting plantar fasciitis in
long-distance runners. Med Sci Sports Exerc, 16(1),60-63.
Warren, B. L. (1990). Plantar fasciitis in runners. Treatment and prevention. Sports Med,
10(5),338-345.
Wen, D. Y., Puffer, J. C., & Schmalzried, T. P. (1997). Lower Extremity alignment and
risk of overuse injuries in runners. Medicine and Science in Sports and Exercise,
29(10), 1291 - 1298.
Wen, D. Y., Puffer, l C., & Schmalzried, T. P. (1998). Injuries in Runners: a prospective
study of alignment. Clinical Journal ofSports Medicine, 8(3), 187 - 194.
Willems, T., Witvrouw, E., De Cock, A., & De Clercq, D. (2007). Gait Related Risk
Factors for Exercise-Related Lower Leg Pain During Shod Running. Medicine
and Science in Sports and Exercise, 39(2),300 - 339.
Williams, D., McClay Davis, L, & Baitch, S. (2003). Effect of Inverted Orthoses on
Lower Extremity Mechanics in Runners. Medicine and Science in Sports and
Exercise, 35(12),2060 - 2068.
Williams, D., McClay Davis, L, Scholz, J., Hamill, J., & Buchanan, T. (2004). High
arched runners exhibit increased leg stiffness compared to low arched runners.
Gait and Posture, 19,263 - 269.
Williams, D., & McClay, L (2000). Measurements Used to Characterize the Foot and the
Medial Longitudinal Arch: Reliability and Validity. Physical Therapy, 80(9),864
- 871.
Williams, D., McClay, L, & Hamill, J. (2001). Arch Structure and injury patterns in
runners. Clinical Biomechanics, 16, 341 - 347.
Williams, D., McClay, L, Hamill, J., & Buchanan, T. (2001). Lower Extremity Kinematic
and Kinetic Differenced in Runners with High and Low Arches. Journal of
Applied Biomechanics, 17, 153 - 163.
Winter, D. (2005). Biomechanics and motor control ofhuman movement. Hoboken, NJ:
Wiley & Sons, Inc.
Wolf, S., Simon, l, Patikas, D., Schuster, W., Armbrust, P., & Doderlein, L. (2008). Foot
motion in children shoes: a comparison of barefoot walking with shod walking in
traditional and flexible shoes. Gait and Posture, 27,51 - 59.
Wu, G. (2002). ISB recommendation on definitions of joint coordinate system of various
joints for reporting of human joint motion: - part I: ankle, hip, and spine. Journal
ofBiomechanics, 25,543 - 548.
143
Zadpoor, A., & Nikooyan, A. (2010). Modeling muscle activity to study the effects of
footwear on the impact forces and vibrations of the human body during running.
Journal ofBiomechanics, 43, 186 - 193.
Zifchock, R, Davis, 1., & Hamill, J. (2006). Kinetic Asymmetry in female runners with
and without retrospective tibial stress fractures. Journal ofBiomechanics, 31(15),
2792 - 2797.
Zifchock, R., Davis, 1., Higginson, 1., McCaw, S., & Royer, T. (2008). Side to side
difference in overuse running injury susceptibility: A retrospective study. Human
Movement Sciences, 27, 888 - 902.
Zifchock, R., Davis, 1., Higginson, 1., McCaw, S., & Royer, T. (2009). Side to side
difference in overuse running injury susceptibility: A retrospective study. Human
Movement Sciences, 27, 888 - 902.