Journal of Applied Biomechanics, 2008, 24, 262-270
© 2008 Human Kinetics, Inc.
Continuous Relative Phase Variability During an Exhaustive
Run in Runners With a History of Iliotibial Band Syndrome
Ross H. Miller,1 Stacey A. Meardon,2 Timothy R. Derrick,2 and Jason C. Gillette2
1
University of Massachusetts and 2Iowa State University
Previous research has proposed that a lack of vari-
ability in lower extremity coupling during running is
associated with pathology. The purpose of the study
was to evaluate lower extremity coupling variability
in runners with and without a history of iliotibial
band syndrome (ITBS) during an exhaustive run.
Sixteen runners ran to voluntary exhaustion on a
motorized treadmill while a motion capture system
recorded reflective marker locations. Eight runners
had a history of ITBS. At the start and end of the
run, continuous relative phase (CRP) angles and
CRP variability between strides were calculated for
key lower extremity kinematic couplings. The ITBS
runners demonstrated less CRP variability than con-
trols in several couplings between segments that have
been associated with knee pain and ITBS symptoms,
including tibia rotation–rearfoot motion and rearfoot
motion–thigh ad/abduction, but more variability in
knee flexion/extension–foot ad/abduction. The ITBS
runners also demonstrated low variability at heel
strike in coupling between rearfoot motion–tibia
rotation. The results suggest that runners prone to
ITBS use abnormal segmental coordination patterns,
particular in couplings involving thigh ad/abduction
and tibia internal/external rotation. Implications for
variability in injury etiology are suggested.
Keywords: running, variability, injury, dynamical
systems, iliotibial band syndrome
Miller is with the Department of Kinesiology, University of Massachu-
setts, Amherst, MA, and Meardon, Derrick, and Gillette are with the
Department of Kinesiology, Iowa State University, Ames, IA.
262
Knowledge of system variability is essential for
understanding the intrinsic behavior of a dynamical
system. Between-stride variability during locomotion
can play both beneficial and detrimental roles depend-
ing on the parameter under investigation. Increased
variability in stride length (Nakamura et al., 1996) and
stride time (Hausdorff et al., 2001) has been associated
with increased risk of falling. Studies of segmental
coordination patterns have suggested a beneficial role
for variability in allowing flexibility to transition to more
stable coordination patterns in response to perturbations
(Van Emmerik & Van Wegen, 2000; Van Emmerik et al.,
2004; Pollard et al., 2005).
While the relationship between variability and
pathology in runners is poorly understood, variability
in segmental coordination patterns may be an important
etiological factor in running injuries (DeLeo et al., 2004).
In the kinematic chain, motion of one segment must
influence the motions of adjacent segments. Rather than
focusing on the mechanics of single joints and segments,
some researchers have considered the coupling, or coor-
dination and relative timing, between segments (Hamill
et al., 1992; Nigg et al., 1993; McClay & Manal, 1997;
Hintermann & Nigg, 1998). In the dynamical systems
approach to lower extremity coordination, continuous
relative phase (CRP) provides a continuous measure of
segmental coordination throughout the entire stride cycle,
and allows an assessment of the flexibility in coordination
(Hamill et al., 1999). Previous research using CRP and
other coordination measures in runners has suggested
that a lack of variability in lower extremity coordination
is indicative of a pathological state, particularly in run-
ners with patellofemoral pain (PFP). Hamill et al. (1999)
found that runners with PFP were less variable across
the entire stride cycle than healthy runners in couplings
involving tibial rotation. Heiderscheit et al. (2002) found

that runners with PFP displayed less variability in thigh
rotation–tibia rotation coupling near heel strike compared
with healthy controls. A certain degree of lower extremity
coordination variability thus appears to be an important
component of pain-free running.
The results of Hamill et al. (1999) and Heiderscheit
et al. (2002) suggest that runners with pathology use a
narrow range of coordination patterns. During the cycli-
cal motions of running, repetitive coordination patterns
could lead to repetitive local tissue stress, from which
an injury may develop. However, interpretations of
segmental coordination assessed by CRP must be made
carefully. CRP is typically derived from sinusoidal oscil-
lators (Kelso, 1995) that provide intuitive results. With
the possible exception of hip flexion/extension, the time
series of most running kinematics are predominantly
nonsinusoidal. Interpretations of CRP for running kine-
matics are therefore limited to relationships between the
signals’ phase planes and should not be used to describe
Downloaded by Purdue Univ on 09/17/16, Volume 24, Article Number 3
relationships between the original kinematic time series
(Peters et al., 2003).
Currently, the relationship between lower extremity
coordination variability and iliotibial band syndrome
(ITBS), a common overuse running injury, is unknown.
ITBS is an overuse injury of the lateral knee believed to
develop when the iliotibial band (ITB) becomes inflamed
due to impingement (Fredericson & Wolf, 2005) or lateral
compression (Fairclough et al., 2006) between the ITB
and the lateral femoral epicondyle. Previous research
has found kinematic discriminators between healthy
runners and runners prone to ITBS. Runners with a his-
tory of ITBS have demonstrated greater foot inversion,
foot adduction, and knee flexion at heel strike (Miller et
al., 2007). Runners who prospectively developed ITBS
demonstrated greater hip adduction and knee internal
rotation before displaying symptoms (Noehren et al.,
2007). In addition, runners with a history of ITBS have
shown an increased likelihood of impingement with the
lateral femoral epicondyle near the end of an exhaustive
run and more strain in the ITB while running compared
with healthy controls (Miller et al., 2007). If variability
in the spatial and temporal coordination of these kinemat-
ics is related to pathology, then CRP variability in these
couplings should discriminate between the groups.
Knowledge of lower extremity coordination vari-
ability in runners prone to ITBS, as well as the relation-
ship between variability and fatigue, may be useful in
treating and preventing ITBS from a clinical perspective.
Dynamical systems theory predicts high variability during
transitions (Schoner & Kelso, 1988), e.g., from swing
to stance. Heiderscheit (2000) suggests that treatments
of disordered movement may be most effective when
variability is high. Based on these suggestions, it seems
relevant to identify phases (stance, swing), instances (heel
strike), or conditions (rested, fatigued) under which the
locomotor system is or is not highly variable, and how
this variability relates to the etiology of running inju-
ries. However, these topics have not been investigated
extensively.
Continuous Relative Phase Variability   263
The purpose of the study was to investigate the role
of lower extremity coordination variability in runners
with retrospective cases of ITBS during an exhaustive
run. It was hypothesized that the ITBS runners would
exhibit less variable coordination than control runners in
couplings between kinematics previously associated with
ITBS. Further, it was hypothesized that the ITBS runners
would decrease their lower extremity coordination vari-
ability between the beginning and end of the run. A long-
term goal of this research is to determine implications for
variability in the etiology of ITBS. Suggestions for these
implications are drawn from the present results.
Methods
Subjects
Sixteen recreational runners were recruited. Runners
completed a self-report on their history of injuries and
means of diagnoses. Eight runners who had been diag-
nosed by a physician or a physical therapist with ITBS
comprised the ITBS group, mean (SD): age = 27.5 (9.0)
years, height = 170.1 (6.9) cm, mass = 68.7 (15.9) kg.
All runners in the ITBS group had experienced recurrent
symptoms in one (two of eight) or both (six of eight) legs,
with recurrence defined as losing at least a week of train-
ing time due to symptoms, but were symptom free for at
least 4 months before data collection. Eight runners with
no history of injuries were age matched within 4 years
of an ITBS runner and comprised the control group, age
= 26.4 (7.7) years, height = 172.8 (8.9) cm, mass = 71.3
(14.4) kg. No runner had a history of major lower extrem-
ity injury other than ITBS. The protocol was approved
by the university’s institutional review board. Subjects
gave informed written consent.
Experimental Setup
An eight-camera 120-Hz Peak Motus motion capture
system (Vicon Peak, Centennial, CO) captured three-
dimensional optical data while subjects ran on an adjust-
able speed Quinton treadmill (Stairmaster, Kirkland,
WA).
Data Collection
Retroreflective markers were attached to both legs on the
toe, heel, dorsifoot, medial and lateral malleoli, anterior
calf, medial and lateral femoral epicondyles, anterior
thigh, greater trochanter, anterior-superior iliac spines,
and L5/S1. Subjects performed a standing calibration
trial with their feet shoulder width apart while the cam-
eras recorded marker positions for 1 s. Subjects ran on
the treadmill at a level grade, with instructions to select
a constant pace that would exhaust them within 20 min.
Self-selected paces were allowed to better capture the
subjects’ normal training kinematics. Subjects warmed
up at a slower pace for 2 min before beginning the run.
Marker positions were recorded every 2 min for 10 s until
263
 264   Miller et al.
the subject could no longer maintain the pace. Subjects
wore black Spandex shirt and shorts and their own run-
ning shoes.
Data Reduction
Marker positions were smoothed by a fourth-order recur-
sive Butterworth filter with a 15-Hz cutoff selected by
residual analysis (Winter, 2004). A MATLAB program
(MathWorks, Natick, MA) modeled the lower extremity
as a rigid-segment linkage (pelvis, thigh, calf, foot), with
three degrees of rotational freedom at each joint. Three-
dimensional joint and segment angles were calculated by
Cardan transformation matrices of three successive rota-
tions (flexion/extension, adduction/abduction, internal/
external rotation; Robertson et al., 2004). Angular veloci-
ties were calculated by the central difference method.
Timings of heel strike and toe-off were determined using
the zero-jerk algorithm (Hreljac & Marshall, 2000).
Phase plots (Figure 1) were constructed for thigh
Downloaded by Purdue Univ on 09/17/16, Volume 24, Article Number 3
adduction/abduction (thigh AD/AB), tibia internal/
external rotation (tibia IR/ER), foot adduction/abduction
(foot AD/AB), foot inversion/eversion (foot IN/EV), knee
flexion/extension (knee FL/EX), and knee adduction/
abduction (knee AD/AB). Phase plots consisted of the
angle on the horizontal axis and its respective angular
velocity on the vertical axis. Phase plots were normalized
by the following equations:
Figure 1 — A normalized phase angle plot for knee flexion
data demonstrating the calculation of the phase angle. θ and
ω are the normalized angular position and angular velocity,
respectively. φ is the phase angle.
264
(1)
(2)
where θ is the angle, ω is the angular velocity, and the
subscript i indicates the data point within the stride. Nor-
malization was performed to account for frequency and
amplitude differences in the time series before computing
phase angles and to allow comparisons with previous
research on CRP in pathological running (Hamill et al.,
1999). Equations 1 and 2 normalize the phase plot such
that the origin corresponds to the midpoint of the angu-
lar range of motion and zero angular velocity. Equation
2 accounts for frequency differences in the signals by
ensuring that CRP angles between signals of different
frequencies provide intuitive results (Peters et al., 2003)
and avoids losing information regarding zero velocity
(Hamill et al., 2000). The phase angle φ was defined as
the angle between the right horizontal axis and a given
data point (θi,norm, ωi,norm):
(3)
CRP angles were calculated for five couplings of interest:
thigh AD/AB–tibia IR/ER, thigh AD/AB–foot IN/EV,
tibia IR/ER–foot IN/EV, knee FL/EX–foot AD/AB, and
knee AD/AB–foot IN/EV. Couplings were selected based
on their likelihood of imposing strain on the ITB, the
importance of the kinematics in previous ITBS research
(Miller et al., 2007; Noehren et al., 2007), and previous
research on kinematic coupling in runners (Bates et
al., 1978; Hamill et al., 1992, 1999; McClay & Manal,
1997). A CRP angle was defined as the proximal segment
φ minus the distal segment φ. CRP variability (VCRP)
was calculated as the between-stride standard deviation
in CRP for a single subject at each time step. VCRP was
calculated between the first seven strides from the first,
middle, and last 10-s data blocks. VCRP was averaged
across the full stride, stance, or swing to obtain single-
value representations of variability during these periods,
and then averaged across subjects. VCRP at heel strike
(0% of stance) was also assessed.
Statistical Analysis
VCRP were compared by a 2 × 2 factorial ANCOVA with
group (ITBS or control) and time (start of the run or end
of the run) as factors, with time as a repeated measure
and running speed as a covariate. Post hoc comparisons
were made by linear contrasts. Significant differences
were reported at p < .05. Trends toward group differences
were reported at p < .10.
 Continuous Relative Phase Variability   265

Results
Runners reached voluntary exhaustion after an average
of 16 (SD, 5) min, with no difference between groups.
Average selected speeds were 9.5 (2.2) min/mile for the
Downloaded by Purdue Univ on 09/17/16, Volume 24, Article Number 3
ITBS group and 7.8 (1.0) min/mile for the control group.
Figure 2 shows exemplar CRP and VCRP data for both
groups. There were no significant main effects for time
and no significant interactions between group and time
on VCRP.

Figure 2 — CRP angles during the complete stride cycle for representative (a–e) ITBS subject and (f–j) control subject. Data are
from the beginning of the subjects’ runs. 0% and 100% on the x-axis represent consecutive heel strikes. Vertical lines indicate toe-
off. Dashed lines are CRP angles. Solid lines are plus or minus one between-stride standard deviation. Note the relatively lower
variability in the ITBS subject in thigh AD/AB–tibia IR/ER, thigh AD/AB–foot IN/EV, and tibia IR/ER–foot IN/EV.
265
 266   Miller et al.

Figure 3 shows VCRP in couplings over the complete
stride cycle. Compared with the control group, runners
with a history of ITBS were more variable in knee FL/
EX–foot AD/AB at the start of the run (18.6° for ITBS
vs. 15.3° for control, p = .02), less variable in thigh AD/
AB–foot IN/EV at the end of the run (30.5° for ITBS vs.
33.1° for control, p = .03), and tended to be less variable
in thigh AD/AB–tibia IR/ER at the end of the run (31.4°
for ITBS vs. 33.5° for control, p = .09). VCRP in other
couplings were not significantly different.
Figure 4 shows VCRP in couplings during swing.
Compared with the control group, runners with a history
of ITBS were more variable in knee FL/EX–foot AD/
AB at the start of the run (18.8° for ITBS vs. 15.4° for
control, p = .04), tended to be more variable in knee FL/
EX–foot AD/AB at the end of the run (18.0° for ITBS vs.
15.5° for control, p = .06), and tended to be less variable
in thigh AD/AB–tibia IR/ER at both the start (31.4° for
ITBS vs. 34.3° for control, p = .07) and end (32.5° for
Downloaded by Purdue Univ on 09/17/16, Volume 24, Article Number 3
= .02) and end (19.0° for ITBS vs. 14.5° for control, p
= .003) of the run. VCRP in other couplings were not
significantly different.
Table 1 summarizes the group differences in VCRP
at heel strike. Among the five couplings analyzed, runners
with a history of ITBS were less variable than controls in
four couplings at the start of the run and four couplings
at the end of the run. Most of the differences were not
significant due to large intersubject variance; however,
the ITBS group was significantly less variable in tibia IR/
ER–foot IN/EV at the start of the run (13.3° for ITBS vs.
24.2° for control, p = .004) and showed a trend to be less
variable in thigh AD/AB–tibia IR/ER at the end of the run
(19.3° for ITBS vs. 24.0° for control, p = .09).
Discussion
In this study, lower extremity coordination variability
assessed by continuous relative phase (CRP) in runners

ITBS vs. 35.0° for control, p = .07) of the run and in thigh
with a history of iliotibial band syndrome (ITBS) dif-
AD/AB–foot IN/EV at the end of the run (32.2° for ITBS
fered from age-matched control runners during a run to
vs. 34.4° for control, p = .09). VCRP in other couplings
voluntary exhaustion, but only in particular segmental
were not significantly different.
couplings and only at certain times during the run. Over
Figure 5 shows VCRP in couplings during stance.
the full stride cycle and during swing, the ITBS group
Compared with the control group, runners with a history
was less variable in couplings between thigh AD/AB–foot
of ITBS were more variable in knee FL/EX–foot AD/AB
IN/EV and thigh AD/AB–tibia IR/ER, although the only
at both the start (18.6° for ITBS vs. 15.3° for control, p

Figure 3 — CRP variability (VCRP) over the complete stride cycle. Data are for the start, middle, and end of the run. Circles are
the ITBS group. Squares are the control group. * indicates a significant difference in group means (p < .05). † indicates a trend (p
< .10).

266
Downloaded by Purdue Univ on 09/17/16, Volume 24, Article Number 3

Figure 4 — CRP variability (VCRP) during swing. Data are for the start, middle, and end of the run. Circles are the ITBS group.
Squares are the control group. † indicates a trend toward a difference in group means (p < .10).

Figure 5 — CRP variability (VCRP) during stance. Data are for the start, middle, and end of the run. Circles are the ITBS group.
Squares are the control group. * indicates a significant difference in group means (p < .05).

267
 268   Miller et al.
Table 1  Group Differences in CRP Variability at Heel Strike
Start of run
Coupling ITBS Control
Thigh AD/AB–tibia IR/ER 19.3 (8.7) 19.5 (11.4)
Thigh AD/AB–foot IN/EV 22.5 (7.0) 24.4 (16.2)
Tibia IR/ER–foot IN/EV 13.3 (8.2)* 24.2 (13.2)
Knee FL/EX–foot AD/AB 13.8 (6.7) 16.2 (5.7)
Knee AD/AB–foot IN/EV 22.9 (9.2) 19.3 (12.9)
Note. Values are means for all subjects in the group. * indicates significantly less variable than control (p < .05). † indicates a trend to be less variable
than control (p < .10). Negative differences (Diff) indicate that the ITBS group’s average variability was smaller than the control group’s average
variability. Units are in degrees.
statistically significant difference in these couplings was
found for thigh AD/AB–foot IN/EV over the complete
stride cycle at the end of the run (see Figures 3 and 4). The
ITBS group was more variable in knee FL/EX–foot AD/
Downloaded by Purdue Univ on 09/17/16, Volume 24, Article Number 3
AB, particularly during stance (Figure 5). There were no
significant main effects for time in either group, indicat-
ing that variability did not change during the course of
the run. However, when comparing variability between
the two groups at the start and end of the run, the ITBS
group was less variable over the complete stride cycle in
thigh AD/AB–foot IN/EV at the end of the run, but not
at the start of the run (Figure 3b), and showed a trend to
be less variable in thigh AD/AB–tibia IR/ER at the end
of the run, but not at the start of the run (Figure 3a). We
hypothesize that these differences may be compensatory
mechanisms used to avoid painful coordination patterns,
or patterns that were painful in the past.
The findings of less variability in selected couplings
for the retrospective ITBS group during the full stride
cycle and during swing are similar to the results of pre-
vious studies on VCRP in injured runners (Hamill et al.,
1999). The finding of less variability at heel strike in the
injured group is also consistent with Hamill et al. (1999)
and Heiderscheit et al. (2002), although their subjects
were runners with PFP who were currently symptomatic.
At heel strike, variability in tibia IR/ER–foot IN/EV
was particularly low for runners with a history of ITBS.
Although we cannot infer a direct relationship between
the temporal synchrony of tibia IR/ER and foot IN/
EV from this result, the lack of variability indicates an
abnormal phasic relationship between these segments at
heel strike. Tibia IR/ER and foot IN/EV are mechanically
linked due to the anatomy of the ankle joint. Excessive
eversion is suspected to lead to excessive knee internal
rotation by internally rotating the tibia, potentially
increasing soft tissue strain at the knee (Bates et al.,
1978; McClay & Manal, 1997). In a prospective study,
runners who developed ITBS demonstrated greater peak
knee internal rotation than runners who did not become
injured (Noehren et al., 2007). Greater peak knee internal
rotation, in conjunction with a lack of variability in tibia
IR–foot IN/EV coupling, could accumulate trauma by
stressing the ITB through a repetitive phasic relationship
between these segments.
268
End of run
Diff ITBS Control Diff
−0.2° 19.3 (10.3)† 24.0 (10.8) −4.7°
−1.9° 23.0 (11.9) 25.1 (13.1) −2.1°
−10.9° 20.2 (12.2) 25.0 (12.0) −4.8°
−2.4° 16.9 (10.4) 14.1 (8.5) +2.8°
+3.6° 24.5 (16.6) 29.3 (10.0) −4.8°
Although couplings between other kinematics not
considered in this study may also be significant dis-
criminators between groups, couplings must be chosen
by a rationale that relates to the functional mechanics of
the system. For this study, couplings between kinemat-
ics that have been associated with ITBS (knee flexion,
thigh adduction, rearfoot motion, tibia internal rotation;
Miller et al., 2007; Noehren et al., 2007) were assessed.
Along with lower tibia IR/ER–foot IN/EV variability
at heel strike compared with controls, runners with a
history of ITBS demonstrated lower thigh AD/AB–foot
IN/EV variability during the full stride cycle at the end
of the run (Figure 3), greater knee AD/AB–foot IN/EV
variability at heel-strike at the start of the run, but lower
knee AD/AB–foot IN/EV variability at heel-strike at
the end of the run (Table 1). Thigh AD/AB–foot IN/EV
and knee AD/AB–foot IN/EV are notable because both
of motions occur in the frontal plane, and could act to
increase lateral compression of the ITB as proposed by
Fairclough et al. (2006).
This study assessed five different kinematic cou-
plings, but found significant differences only in some
couplings. One possible explanation is that these cou-
plings play different roles in the loading of the iliotibial
tract. Functionally, the ITB stabilizes the lateral hip and
resists hip adduction. At the end of the run, runners with
a history of ITBS were less variable in thigh AD/AB–foot
IN/EV (Figure 3b) and tended to be less variable in thigh
AD/AB–tibia IR/ER (Figure 3a). Recent studies have
suggested that abnormal hip muscle function (Fairclough
et al., 2007) and weak hip abductors (Fredericson et al.,
2000) are etiological factors in ITBS. Low variability
in couplings involving thigh AD/AB could indicate
dysfunctional hip abductors. Further research on hip
muscle mechanics in healthy and pathological running
is needed.
We hypothesized that runners with a history of ITBS
would decrease VCRP between the start and end of the
run. Statistical analysis indicated that neither group sig-
nificantly changed VCRP in any coupling between the
start and end of the run. There were also no interactions
between group and time, indicating that time did not
have a differential effect on VCRP between the groups.
These findings suggest that fatigue did not have a large,

consistent effect on variability in the five couplings
analyzed. Previously, the runners with a history of ITBS
demonstrated kinematic adjustments between the start
and end of the run (increased knee flexion at heel strike,
increased maximum foot inversion, increased maximum
foot adduction) that may increase strain in the ITB (Miller
et al., 2007). These combined findings suggest that the
body can adjust the kinematics of individual joints and
segments (possibly to attenuate impacts or reduce injury
risks) without altering lower extremity coordination
patterns. At this stage, the dynamical systems approach
to running injury etiology is still relatively new, and the
kinematic adjustments in individual segments during
the run are probably more clinically meaningful than
the lack of fatigue-induced changes in VCRP. More
studies are needed to improve our understanding of the
lower extremity as a dynamical system in healthy and
pathological running.
An important question arises in regards to the pres-
Downloaded by Purdue Univ on 09/17/16, Volume 24, Article Number 3
ent results: what are the physiological consequences of
variability on running coordination, tissue trauma, and
pathological states? Hamill et al. (1999) proposed that
a lack of variability indicates a narrow range of relative
segment actions that allow for pain-free running in run-
ners with PFP. Low segmental coordination variability
in couplings involving thigh AD/AB could limit the
range of coordination patterns the runners with a history
of ITBS used to successfully complete the movement,
and could repeatedly stress the ITB through hip adduc-
tion as proposed by Noehren et al. (2007). In a previous
study on the same group of runners, the runners with a
history of ITBS demonstrated more strain than controls
in a musculoskeletal model of the ITB (Miller et al.,
2007). Unpublished results indicate similar findings
in a prospective study of runners who developed ITBS
(Hamill et al., 2007). Therefore, we propose that the ITBS
group’s low variability in thigh AD/AB couplings limits
the range of coordination patterns they use to success-
fully complete the movement task. Consequently, these
runners performed repetitive segmental motion patterns
that continually stressed the ITB, leading to pain and
injury over time.
Runners with a history of ITBS were more variable in
knee FL/EX–foot AD/AB during both stance and swing.
Excessive variability in knee FL/EX–foot AD/AB may
indicate a lack of stability in this coupling, both during
weight-bearing (stance) and during recovery from and
preparation for weight-bearing and impact (swing). In
this context, the term stability must be used cautiously
because the variability and stability of a dynamical system
are not necessarily synonymous (Li et al., 2005). Messier
et al. (1995) found that the knee extensors of runners with
ITBS were weaker than healthy runners. In conjunction
with the thigh AD/AB coupling results, these findings
suggest three conclusions:
1. Muscular strength may play a role in VCRP.
2. A lack of coordinative flexibility (low variability)
and a lack of consistent coupling control (high
Continuous Relative Phase Variability   269
variability) should both be considered when
investigating running injuries with a dynamical
systems approach.
3. The swing phase should not be neglected when
investigating running injuries.
The first conclusion is made cautiously because strength
was not assessed for these particular runners. The second
conclusion highlights the importance of careful data
interpretation. Even if low variability is deemed “bad,”
high variability is not necessarily “good.” Within any
system, there is likely a midlevel of variability at which
it performs optimally.
It is notable that in present study, as well as previous
studies, pathological runners have failed to show less
variability than the healthy group in couplings during
the stance phase (Hamill et al., 1999; Heiderscheit et al.,
2002). These findings seem counterintuitive to the theory
that low variability is related to repetitive tissue stress.
If pathological runners use a narrow range of coordina-
tion patterns that repeatedly stress the same tissues, they
should be less variable during stance, when the actual pain
and heavy tissue stress occurs for injuries such as ITBS
and PFP. It is possible during recovery from their injuries
(recall that the runners in the ITBS group were previously
injured but symptom free during data collection), the
ITBS group learned to select new coordination patterns
that allowed pain-free running, leading to an increase in
stance-phase variability. Low variability during swing
in the runners with a history of ITBS may be an artifact
of their earlier injured states that has remained due to
the lack of weight bearing and lack of ITB pain during
swing. However, runners currently experiencing PFP
showed similar trends in variability (Hamill et al., 1999),
and three of the eight runners in the ITBS group have
experienced recurrent symptoms since data collection in
late 2005/early 2006 (personal interviews). In any case,
because variability in the runners with a history of ITBS
may have been different before and after the injury, the
present results cannot definitively confirm a causal or
compensational role for variability in ITBS. Prospective
studies observing variability before, during, and after
recovery from injury would contribute significantly to
our understanding of the interaction between variability,
pain, and running injuries.
Although the statistical analysis accounted for run-
ning speed as a covariate, it could be argued that the
study was limited by allowing subjects to run at self-
selected paces and not a standardized pace. Although the
group running times were statistically similar, the ITBS
group tended to run slower. Since ITBS is an overuse
injury likely developed during training, the decision to
allow self-selected paces was made to better capture the
runners’ normal training kinematics and coordination
patterns.
In summary, during an exhaustive run, runners with
a history of ITBS demonstrated less CRP variability than
controls in thigh AD/AB–foot IN/EV, but were more
variable in knee FL/EX–foot AD/AB. The same runners
269
 270   Miller et al.
previously demonstrated more strain in a model of the
iliotibial band. The results indicate that a history of ITBS
is associated with altered variability in coupled segmental
motions related to the injury, and suggest that variability
in injury-prone runners may be indicative of abnormal
segmental coordination patterns and the potential for
repetitive tissue stress.
Acknowledgments
The authors would like to thank Joseph Hamill and Graham
Caldwell for their constructive comments on the manuscript.
References
Bates, B.T., James, S.L., & Osternig, L.R. (1978). Foot function
during the support phase of running. Running, 1978(fall),
24–31.
DeLeo, A.T., Dierks, T.A., Ferber, R., & Davis, I.S. (2004).
Lower extremity joint coupling during running: a cur-
Downloaded by Purdue Univ on 09/17/16, Volume 24, Article Number 3
rent update. Clinical Biomechanics (Bristol, Avon), 19,
983–991.
Fairclough, J., Hayashi, K., Toumi, H., Lyons, K., Bydder, G.,
Philips, N., et al. (2006). The functional anatomy of the
iliotibial band during flexion and extension of the knee:
implications for understanding iliotibial band syndrome.
Journal of Anatomy, 208, 309–316.
Fairclough, J., Hayashi, K., Toumi, H., Lyons, K., Bydder, G.,
Philips, N., et al. (2007). Is iliotibial band syndrome really
a friction syndrome? Journal of Science and Medicine in
Sport, 10, 74–76.
Fredericson, M., Cookingham, C.L., Chaudhari, A.M., Dowdell,
B.C., Oestreicher, N., & Sahrmann, S.A. (2000). Hip
abductor weakness in distance runners with iliotibial
band syndrome. Clinical Journal of Sport Medicine, 10,
169–175.
Fredericson, M., & Wolf, C. (2005). Iliotibial band syndrome
in runners: innovations in treatment. Sports Medicine
(Auckland, N.Z.), 35, 451–459.
Hamill, J., Bates, B.T., & Holt, K.G. (1992). Timing of lower
extremity joint actions during treadmill running. Medicine
and Science in Sports and Exercise, 24, 807–813.
Hamill, J., Van Emmerik, R.E.A., Heiderscheit, B.C., & Li, L.
(1999). A dynamical systems approach to lower extremity
running injuries. Clinical Biomechanics (Bristol, Avon),
14, 297–308.
Hamill, J., Haddad, J.M., & McDermott, W.M. (2000). Issues in
quantifying variability from a dynamical systems perspec-
tive. Journal of Applied Biomechanics, 16, 407–418.
Hamill, J., Miller, R.H., Noehren, B.W., & Davis, I.S. (2007).
A strain model of the iliotibial band. 25th International
Symposium on Biomechanics in Sports, Ouro Preto, Brazil,
August 23-27, 2007.
Hausdorff, J.M., Rios, D.A., & Edelberg, H.K. (2001). Gait
variability and fall risk in community-living older adults:
a 1-year prospective study. Archives of Physical Medicine
and Rehabilitation, 82, 1050–1056.
Heiderscheit, B.C. (2000). Movement variability as a clinical
measure for locomotion. Journal of Applied Biomechan-
ics, 16, 419–427.
Heiderscheit, B.C., Hamill, J., & Van Emmerik, R.E.A. (2002).
Variability of stride characteristics and joint coordination
270
among individuals with unilateral patellofemoral pain.
Journal of Applied Biomechanics, 18, 110–121.
Hintermann, B., & Nigg, B.M. (1998). Pronation in runners.
Implications for injuries. Sports Medicine (Auckland,
N.Z.), 26, 169–176.
Hreljac, A., & Marshall, R.N. (2000). Algorithms to determine
event timing during normal walking using kinematic data.
Journal of Biomechanics, 33, 783–786.
Kelso, J.A.S. (1995). Dynamic Patterns: The Self-Organization
of Brian and Behavior. Cambridge, MA: MIT Press.
Li, L., Haddad, J.M., & Hamill, J. (2005). Stability and variabil-
ity may respond differently to changes in walking speed.
Human Movement Science, 24, 257–267.
McClay, I.S., & Manal, K. (1997). Coupling parameters in
runners with normal and excessive pronation. Journal of
Applied Biomechanics, 13, 109–124.
Messier, S.P., Edwards, D.G., Martin, D.F., Lowery, R.B.,
Cannon, D.W., James, M.K., et al. (1995). Etiology of
iliotibial band friction syndrome in distance runners. Medi-
cine and Science in Sports and Exercise, 27, 951–960.
Miller, R.H., Lowry, J.L., Meardon, S.A., & Gillette, J.C.
(2007). Lower extremity mechanics of iliotibial band
syndrome during an exhaustive run. Gait & Posture, 26,
407–413.
Nakamura, T., Meguro, K., & Sasaki, H. (1996). Relationship
between falls and stride length variability in senile demen-
tia of the Alzheimer type. Gerontology, 42, 108–113.
Nigg, B.M., Cole, G.K., & Nachbauer, W. (1993). Effects of
arch height of the foot on angular motion of the lower
extremities in running. Journal of Biomechanics, 26,
909–916.
Noehren, B.W., Davis, I.S., & Hamill, J. (2007). Prospective
study of the biomechanical factors associated with iliotibial
band syndrome. Clinical Biomechanics (Bristol, Avon),
22, 951–956.
Peters, B.T., Haddad, J.M., Heiderscheit, B.C., Van Emmerik,
R.E.A., & Hamill, J. (2003). Limitations in the use and
interpretation of continuous relative phase. Journal of
Biomechanics, 36, 271–274.
Pollard, C.D., Heiderscheit, B.C., Van Emmerik, R.E.A., &
Hamill, J. (2005). Gender differences in lower extremity
coupling variability during an unanticipated cutting maneu-
ver. Journal of Applied Biomechanics, 21, 143–152.
Robertson, D.G.E., Caldwell, G.E., Hamill, J., Kamen, G., &
Whittlesey, S.N. (2004). Research Methods in Biomechan-
ics. Champaign, Illinois: Human Kinetics.
Schoner, G., & Kelso, J.A.S. (1988). Dynamic pattern gen-
eration in behavioral and neural systems. Science, 239,
1513–1520.
Van Emmerik, R.E.A., & Van Wegen, E.E.H. (2000). On
variability and stability in human movement. Journal of
Applied Biomechanics, 16, 394–406.
Van Emmerik, R.E.A., Rosenstein, M.T., McDermott, W.J.,
& Hamill, J. (2004). A nonlinear dynamics approach to
human movement. Journal of Applied Biomechanics, 20,
396–420.
Winter, D.A. (2004). Biomechanics and Motor Control of
Human Movement. Hoboken, New Jersey: Wiley.