Wattananon et al.

BMC Musculoskeletal Disorders (2017) 18:455

DOI 10.1186/s12891-017-1820-x

RESEARCH ARTICLE Open Access

Kinematic characterization of clinically

observed aberrant movement patterns in
patients with non-specific low back pain: a
cross-sectional study
Peemongkon Wattananon1* , David Ebaugh2, Scott A. Biely3, Susan S. Smith4,5, Gregory E. Hicks5 and
Sheri P. Silfies2

Abstract
Background: Clinical observation of aberrant movement patterns during active forward bending is one criterion
used to identify patients with non-specific low back pain suspected to have movement coordination impairment.
The purpose of this study was to describe and quantify kinematic patterns of the pelvis and trunk using a dynamics
systems approach, and determine agreement between clinical observation and kinematic classification.
Method: Ninety-eight subjects performed repeated forward bending with clinical observation and kinematic data
simultaneously collected. Kinematic data were plotted using angle-angle, coupling-angle, and phase-plane
diagrams. Accuracy statistics in conjunction with receiver operating characteristic curves were used to determine
agreement between clinical observation and kinematic patterns.
Results: Kinematic patterns were consistent with clinical observation and definitions of typical and aberrant
movement patterns with moderate agreement (kappa = 0.46–0.50; PABAK = 0.49–0.73). Early pelvic motion
dominance in lumbopelvic coupling-angle diagram ≥59° within the first 38% of the movement represent observed
altered lumbopelvic rhythm. Frequent disruptions in lumbar spine velocity represented by phase-plane diagrams
with local minimum occurrences ≥6 and sudden decoupling in lumbopelvic coupling-angle diagrams with sum of
local minimum and maximum occurrences ≥15 represent observed judder.
Conclusion: These findings further define observations of movement coordination between the pelvis and lumbar
spine for the presence of altered lumbopelvic rhythm and judder. Movement quality of the lumbar spine segment
is key to identifying judder. This information will help clinicians better understand and identify aberrant movement
patterns in patients with non-specific low back pain.
Keywords: Low back pain, Multi-segment kinematics, Clinical observation, Aberrant movement patterns

Background pain that is not attributable to a recognizable or known

Low back pain is one of the most common health prob- specific pathology is referred to as a non-specific low back
lems in the United States [1]. More importantly, high pain (NSLBP). Non-specific low back pain is accountable
prevalence and recurrence rates increase the number of for approximately 85% of all low back pain [4].
medical visits, hospitalization, and utilization of health Current clinical research suggests patients with NSLBP
care services, including physical therapy [1–3]. Low back demonstrate different clinical characteristics that may re-
sult from different underlying contributory mechanisms
[5–9]. Impairment in inter-segmental movement coordin-
* Correspondence: [email protected]
1
Motor Control and Neural Plasticity Laboratory, Faculty of Physical Therapy, ation (e.g., coordination between the lumbar spine, pelvis
Mahidol University, 999 Phuttamonthon 4 Road, Salaya, Nakhon Pathom and hip) has been proposed as one cause of NSLBP
73170, Thailand [8, 10]. Movement coordination impairment (MCI) is
Full list of author information is available at the end of the article

© The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Wattananon et al. BMC Musculoskeletal Disorders (2017) 18:455 Page 2 of 12

defined as poorly coordinated or controlled spine and pel- clinicians with the ability to enhance their knowledge and
vis position and movement during functional tasks that understanding of inter-segment coordination and move-
places repeated abnormal stresses on musculoskeletal ment control associated with different aberrant movement
tissues eventually contributing to tissue injury and patterns observed during forward bending. This could
pain [11]. Clinicians have assumed that MCI is associ- lead to better identification and treatment of MCI,
ated with impaired neuromuscular control that can be and provide a significant step toward quantification of
identified by clinical observation of aberrant move- aberrant movement.
ment patterns [9, 11–17].
Evidence supports that aberrant movement patterns Methods
observed during an active forward bending task is one Subjects
identifier of patients with MCI and that these patients Ninety-eight subjects with both clinical observation and
benefit from exercises focused on trunk muscles and de- kinematic data recorded simultaneously during a series of
signed to improve coordination and control (e.g., core forward bending tasks were used in this secondary data
stabilization or lumbar stabilization) [7, 18]. Recent work analysis [19]. Subjects were between 18 and 65 years of
has demonstrated that clinical observation of aberrant age and took part in a study conducted within a university
movement patterns during standing forward bend has and private physical therapy clinic. This study was ap-
fair to almost perfect (kappa = 0.35–0.89) inter-rater reli- proved by the university institutional review board, and all
ability when motion was observed simultaneously by subjects provided written informed consent prior to par-
two experienced clinicians [19]. Findings from this study ticipation. Thirty-five subjects had no history of LBP, 29
also revealed that aberrant movement patterns were sig- subjects were experiencing a current episode of LBP that
nificantly associated with NSLBP providing construct val- started within the past 7 weeks, and 34 had a history of
idity for the association of aberrant movement patterns LBP but were currently pain free (Table 1). Exclusion cri-
with current symptoms [19]. Furthermore, a greater fre- teria for all subjects consisted of: 1) clinical signs of sys-
quency of aberrant patterns can be seen in patients with temic disease; 2) definitive neurologic signs including
current NSLBP compared to healthy controls when per- weakness or numbness in the lower extremity; 3) previous
forming multiple repetitions of forward bend [19]. spinal operation; 4) diagnosed osteoporosis, severe spinal
Although evidence supports the use of clinical obser- stenosis, and/or inflammatory joint disease; 5) pregnancy;
vation for identifying patients with MCI, investigators 6) any lower extremity condition that would potentially
have not systematically captured, described, or quanti- alter trunk movement in standing; 7) vestibular dysfunc-
fied typical and aberrant movement patterns using con- tion; 8) extreme psychosocial involvement; or 9) active
tinuous kinematic data of multiple body segments treatment of another medical condition that would pre-
(femur, pelvis, lumbar spine, and thoracic spine) during a clude participation in any aspect of the study.
forward bending motion. As a result, clinicians have lim-
ited information about which segments and movement Procedures and kinematic instrumentation
characteristics (range, velocity, and/or timing) significantly Subjects performed 6 repetitions of an active forward bend
contribute to the observed aberrant movement patterns. task. Two experienced physical therapists observed the for-
Kinematic data have been widely used for investigating ward bend task while kinematic data was simultaneously
the amount of trunk and pelvic motion during forward collected. These therapists had at least 5 years of experience
bending, with limited investigation into the movement in spinal rehabilitation and completed a 2-h training session
patterns and underlying neuromuscular control [20–22]. that standardized the definitions of aberrant patterns prior
Kinematics, in conjunction with a dynamic systems ap- to data collection. For each subject, the therapists, who
proach, can be used to better understand movement pat- were blinded to the group assignment, independently rated
terns [22–25]. By plotting continuous angle changes the movement pattern as typical or aberrant. Table 2 pro-
between different body segments, or continuous angle vides operational definitions of typical and aberrant move-
changes against segmental instantaneous angular vel- ment patterns used by these clinicians to assess movement
ocity, kinematic data can be used to represent patterns during standing forward bending [7, 12, 15, 17, 19, 27–29].
of movement (inter-segment coordination, and movement An electromagnetic tracking system (3Space Fastrak,
control) during functional motions [26]. The purposes of Polhemus Inc., Colchester, VT) was utilized to capture
this study were to 1) describe and quantify temporal and position and orientation of thoracic and lumbar spine, pel-
spatial 3-dimensional multi-segmental kinematics of the vis, and thigh segments at 30 Hz during forward bend and
pelvis and trunk using a dynamics systems approach, and return to standing. Kinematic sensors were mounted to
2) determine agreement between clinical observation and orthoplast and attached to the subject at the following
kinematic classification of movement patterns. Detailed body landmarks (Fig. 1): 1) right femur (15 cm. superior
kinematic descriptions of these patterns should provide to the right femoral lateral epicondyle), 2) pelvis (over the
Wattananon et al. BMC Musculoskeletal Disorders (2017) 18:455 Page 3 of 12

Table 1 Demographic data for control, current episode of LBP, and history of LBP subjects
N %Female Age ± SD (years) NPRS ± SD (score0–10) ODI ± SD (score0–100)
Control 35 57.1 40.9 ± 9.3 N/A N/A
Current episode of LBP 29 48.3 43.6 ± 12.3 4 ± 2.6 28 ± 14.1
History of LBP 34 50.0 46.7 ± 9.5 N/A N/A
Total 98 52.0 43.7 ± 10.5 N/A N/A
Group comparison p value N/A 0.75a
0.07b
N/A N/A
LBP Low back pain, NPRS Numeric pain rating scale, ODI Oswestry disability index, SD Standard deviation
a
Group comparison using a chi-square test
b
Group comparison using a one-way analysis of variance (ANOVA)

spinous process of S2), 3) lumbar spine (over the spinous Kinematic data reduction was completed using custom
process of L1), and 4) thoracic spine (over the spinous LabVIEW programs (National Instruments, Austin, TX.).
process of T3). Based upon the recommendations of the Data were converted to segmental angular rotations
International Society of Biomechanics (ISB), the following using Euler angles following a Cardan sequence of x
anatomical landmarks were digitized to create a local ref- (flexion/extension), y (lateral bend to the right/left, and z
erence frame for each body segment: 1) thorax (sternal (rotation to the right/left) (Fig. 1). Segmental rotations in-
notch, xyphoid process, T8, and C7); 2) pelvis (right ASIS, cluded: 1) total trunk motion (FT; thoracic spine motion
left ASIS, right PSIS, and left PSIS); 3) lumbar spine with respect to the femur); 2) pelvic motion (FP; pelvic mo-
(xyphoid process, T10, L3-L5); and 4) femur (medial epi- tion with respect to the femur); 3) lumbar motion (PL; lum-
condyle, lateral epicondyle, and femoral head) [30]. bar spine motion with respect to the pelvis); 4) thoracic
Preliminary work conducted in our lab established the motion (LT; thoracic spine motion with respect to the lum-
intra- and inter-session coefficient of multiple correl- bar spine); and 5) thoracolumbar motion (PT; combined
ation (CMC) for measuring movement patterns with the lumbar-thoracic spine motion with respect to the pelvis).
electromagnetic tracking system. The CMC was fair to Based on total trunk motion, a LabVIEW program was
excellent (intra-session CMC = 0.76–0.95 and inter- used to determine the start and stop points for each
session CMC = 0.51–0.95) across segments demonstrat- repetition of the forward bend motion. Kinematic data
ing consistency of the movement patterns in standing were then filtered with a dual pass Butterworth filter
forward bend for time-series range of motion and angu- (2nd order low pass frequency at 5 Hz) and time-
lar velocity. The lower CMC values were associated with normalized to 51 data points (0–50) to represent 100%
movements in the frontal plane. of the forward bend motion. Typical and aberrant move-
ment patterns were represented by the following kine-
Data reduction matic diagrams: 1) angle-angle, 2) coupling-angle, and 3)
Mutual agreement on clinical observation by the two ex- phase-plane diagrams [23, 31–33].
perienced clinicians was used for stratification of individ-
ual kinematic patterns derived from the forward bending Kinematic representation and interpretation of movement
task (98 subjects × 6 repetitions = 588 movement patterns) patterns
into typical or aberrant patterns of movement. This strati- Inter-segmental coordination
fication was performed independent of the subject’s low Coordination of movement between the lumbar spine
back pain status. and pelvis is clinically referred to as lumbopelvic rhythm

Table 2 Operational definitions of clinically observed typical and aberrant movement patterns during a standing forward bend and
return motion
Movement pattern type Operational definition
Typical During the forward bend phase, hip and lumbar spine motion occur simultaneously with lumbar spine motion
predominating in the first 1/3rd and hip motion predominating in the last 1/3rd of the movement. During the
return to upright phase, hip and lumbar spine motion occur simultaneously with hip motion predominating in
the first 1/3rd and lumbar spine motion predominating in the last 1/3rd of the movement. Movement should
be smooth (gradual increase and decrease in velocity) and remain in the sagittal plane.
Altered lumbopelvic rhythm (aLPR) During the forward bend phase, hip motion is greater than lumbar spine motion during the first 1/3rd and/or
lumbar motion greater than hip motion during the last 1/3rd of the movement, or during return to an upright
position, lumbar spine motion is greater than hip motion during the first 1/3rd and/or hip motion greater than
lumbar spine motion during the last 1/3rd of the movement.
Judder (JUD) Observation of a sudden deceleration and acceleration, or quick out of sagittal plane movements during trunk
forward bending or return.
Wattananon et al. BMC Musculoskeletal Disorders (2017) 18:455 Page 4 of 12

Movement control
Movement control of a body segment was captured and
described using phase-plane and plane angle-angle dia-
grams (Fig. 4). Movement control is characterized by
smoothness of the segment’s velocity. Disruptions in
control can be identified by the number of local mini-
mum (LMin) and maximum (LMax) occurrences [23].
These occurrences represent sudden deceleration and
acceleration during movement tasks that are clinically
referred to as judder (JUD). Quick out of sagittal plane
movement or off axis motion is another sign of impaired
control that is another focus of clinical definition of jud-
der. This presentation of poor movement control can also
be captured and described using plane angle-angle and
phase-plane diagrams. Additionally, changes in movement
control of one segment might cause changes in the rela-
tive coordination between segments. Therefore, phase-
plane, plane angle-angle, and coupling-angle diagrams can
be used to detect the segment responsible for coordin-
ation changes identified in a coupling-angle diagram.

Statistical analysis
For the first purpose of this study, temporal and spatial
3-dimensional kinematics of the pelvis and trunk seg-
ments (lumbar, thoracic) associated with both typical
Fig. 1 Location of kinematic sensors on the femur, pelvis, lumbar
spine, and thoracic spine. Cardan sequence was x (flexion positive), y and aberrant forward bend movement patterns were de-
(right side bend positive), and z (right rotation positive) scribed using means and standard deviations of derived
kinematic variables (Table 3). The kinematic data were
also graphed and additional descriptors were developed.
Individual kinematic variables were tested for normal-
(LPR). This characteristic of movement can be captured ity and homogeneity of variance assumptions. Independ-
and described using a segment angle-angle diagram ent t-test (1-tailed) was used to test for differences
(Fig. 2). The shape or trajectory of the diagram provides between typical and aberrant movement patterns when
information regarding qualitative coordination between those assumptions were met, while Mann-Whitney U
two segments. A diagonal straight line indicates that the test was used if those assumptions were violated. We
two segments are moving at a constant ratio. Horizontal intended to initially remove kinematic variables that did
or vertical lines indicate that one segment is moving, not differentiate between typical and aberrant movement
whereas the other segment is not [33]. A limiting factor patterns, but we did not wish to exclude any potentially
of using angle-angle diagrams to represent LPR arises useful kinematic variables. Therefore, we decided to use
when subjects move through different amounts of mo- a liberal approach, in which individual kinematic vari-
tion. Vector coding (Fig. 2) can be used to address this ables with p-value less than 0.10 (p < 0.10) were retained
limitation by standardizing a segment’s contribution by as potential key variables of segment and movement
calculating a vector (coupling-angle) between two adja- characteristics that would then be used to determine
cent points relative to the right horizontal [31, 34]. A agreement between clinical observation and kinematic
coupling-angle diagram (Fig. 3) also represents coordin- classification. Additionally, the mean and standard devi-
ation between segments, and quantifies the shape or tra- ation for each kinematic data point from typical move-
jectory of movement coordination relative to the percent ment patterns were used to generate a mean typical
of movement. A coupling angle of 45° indicates 1:1 mo- movement pattern along with standard deviation bands
tion between segments, greater than 45° indicates distal that represented typical movement variability. Aberrant
segment (pelvis) dominance; while less than 45° indicates movement patterns were then plotted against these typical
proximal segment (lumbar spine) dominance [33]. These patterns to further describe differences in movement qual-
diagrams were also used to determine when, during the ity. We found that the derived angle and velocity changes
motion (% of movement), one segment dominated the at the start and end of motion often caused errors in
motion relative to another segment. coupling-angle and phase-plane diagrams secondary to
Wattananon et al. BMC Musculoskeletal Disorders (2017) 18:455 Page 5 of 12

Fig. 2 Example of a lumbopelvic (FP-PL) segment angle-angle diagram from one subject. One segment’s angular displacement (x axis) in sagittal
plane versus another segment’s angular displacement (y axis) in sagittal plane during forward bend. Altered lumbopelvic rhythm was defined as
slope greater than 45° in the first 1/3rd of the movement. To quantify coordination changes observed in a segment angle-angle diagram, a
coupling angle (insert), which is the angle between the vector formed between two adjacent data points relative to the right horizontal, coupling
angle (θ) = atan [(Yi + 1–Yi)/(Xi + 1–Xi)], was used to standardize segment coordination across trials and subjects

Fig. 3 Example of a lumbopelvic (FP-PL) coupling-angle diagram that plots percentage total angular displacement during forward bending (x axis) versus
coupling angles (y axis). Relative timing of a shift from lumbar domination to pelvic domination within the movement pattern is defined by % total angular
displacement when the coupling angles are greater than 46°. Local minimum (LMin) or maximum (LMax) occurrences (insert) representing coordination
changes in the coupling-angle diagrams was identified by the greatest (local maximum) or least (local minimum) values (X)
Wattananon et al. BMC Musculoskeletal Disorders (2017) 18:455 Page 6 of 12

Fig. 4 a Example of lumbar spine phase-plane diagram representing percentage of total angular displacement (x axis) versus instantaneous angular
velocity (y axis) during forward bending. Local minimum and maximum occurrence of the phase-plane diagram (+, insert) represent disruptions in
angular velocity (sudden deceleration and acceleration) that are associated with judder. This pattern can also be characterized by quick out of plane
deviation in the pelvis or lumbar spine plane angle-angle diagram (b) over a short period of time (as indicated by +, in both diagrams). These out of
plane deviations are consistent with sudden decreases and increases in angular velocity in the pelvis or lumbar spine phase plane diagram

significant variability associated with values fluctuating of forwarding bending movement) the pelvis dominated
around zero angular motion or velocity. Therefore, we the motion relative to the lumbar spine.
used data between 5% and 95% of total trunk motion for For judder (smoothness of movement), local minimum
further analysis. (LMin) and local maximum (LMax) occurrences in the
The second purpose of this study was to determine phase-plane diagrams for each segment (FT, FP, PL, LT,
the level of agreement between clinical observation and and PT) and coupling-angle diagram for FP-PL were used
kinematic classification derived from kinematic diagrams to quantify movement control and inter-segment coordin-
that represent segments and movement characteristics ation, respectively. LMin and LMax in the phase-plane
contributing to the clinically observed aberrant move- and coupling-angle diagrams correspond to the two oper-
ment patterns. An accuracy statistics approach in con- ational definitions of judder (sudden deceleration and ac-
junction with receiver operating characteristic curves celeration, and quick off axis or out of plane movement).
(ROC) was used for this analysis [35–38]. For altered However, clinical observation data did not indicate what
lumbopelvic rhythm (aLPR, inter-segment coordination), type of judder had been identified. Therefore, after key
we did not know at which point in the movement pat- segment and characteristics were identified, we further
tern, clinicians perceived onset of pelvic domination classified judder into quick out of plane movement based
(pelvic-dominated angle). Therefore, we varied the lum- on corresponding plane angle-angle diagrams and calcu-
bopelvic coupling angle from 45°-90°, then derived a lated prevalence of this type of judder.
variable, called “timing” for each 1° pelvic-dominated Contingency tables and receiver operating characteris-
angle increase, to determine when during the motion (% tic curves (ROC) were created using the total number of
typical and each aberrant movement pattern based on
Table 3 Kinematic variables used to describe aberrant clinical observation (reference standard) and the kine-
movement pattern matic variables derived from quantification of kinematic
Aberrant movement pattern Variable diagrams. The pelvic-dominated angle and segment that
aLPR Slope of AA (Mean CA), Timing generated the optimal area under the ROC curve (AUC)
were then identified. The ROC of identified pelvic-
JUD LMin, LMax
dominated angle or segment and its kinematic variable was
aLPR Altered lumbopelvic rhythm, JUD Judder, AA Angle-angle diagram, CA
Coupling angle diagram, Timing When in the movement aLPR occurred, LMin
used to determine the cut-off point that maximized agree-
Local minimum occurrence, LMax Local maximum occurrence ment on kinematic variables. Kappa values were used to
Wattananon et al. BMC Musculoskeletal Disorders (2017) 18:455 Page 7 of 12

assess the agreement between clinical observation and were significantly greater for JUD than the typical patterns
kinematic classification. Sensitivity, specificity, and positive (Tables 5 and 6). These kinematic variables were retained
and negative likelihood ratios (LR) were calculated for the to determine the agreement between clinical observation
cut point. Statistical analysis was performed using custom and kinematic classification of judder.
LabVIEW (National Instruments, Austin, TX) and SPSS
(IBM SPSS Statistics for Windows, Version 21.0. Armonk,
Clinical and kinematic agreement
NY) software.
Accuracy statistics and ROC analysis revealed that lum-
bar spine segment kinematics were the key for separat-
Results
ing typical from aberrant movement patterns. Table 7
Movement pattern classification: Clinical observation
shows the kinematic variables used for classification,
Based upon clinical observation, two experienced clini-
AUC, kappa, and accuracy statistics at ROC cut-off point
cians mutually agreed on 195 out of 588 movement trials
for each aberrant pattern. In addition, we found that 7
(33%). One hundred and eight forward bend movement
out of 30 judder patterns (23%) were further classified as
trials (18%) were classified as a typical, 57 trials (10%) were
quick out of plane movement based on lumbar spine
classified as demonstrating altered inter-segment coordin-
plane angle-angle diagram. Figures 5 and 6 demonstrate
ation (LPR), and 30 trials (5%) were classified as demon-
examples of typical and aberrant patterns along with the
strating poor movement control (JUD). In most trials, the
kinematic variables used to identify the aberrant motion
thoracic segment’s movement pattern did not assist in the
observed by the clinicians.
identification of any aberrant pattern. Therefore, this seg-
ment was not included in statistical analyses related to our
second purpose. Discussion
Our kinematic analysis of forward bend movement pat-
Kinematic classification terns provides information that enhances the under-
Inter-segment coordination standing of pelvis and trunk movement characteristics
The overall slope of the lumbopelvic (FP-PL) angle-angle associated with typical and aberrant movement patterns.
diagram for those patterns with aLPR was significantly Additionally, our findings identified the key segments
steeper than that of the typical movement pattern (pelvis and lumbar spine) and movement characteristics
(Table 4). When broken down into specific ranges of the of these segments that best described aberrant move-
forward bend motion, the aLPR angle-angle slopes in the ment patterns observed during clinical examination of
first and second third of the motion were significantly forward bending. Kinematic classifications representing
steeper than typical; however, in the last third of the mo- these segments (Table 7) demonstrated high specificity
tion, aLPR slopes were significantly less steep than typical. and positive likelihood ratio, which indicates that our
Analysis of the coupling-angle diagram revealed that kinematic variables have the ability to detect clinically
the time (% of movement) when the pelvic contribution observed aberrant movement patterns from the kine-
was greater than the lumbar contribution occurred sig- matic data. Detailed kinematic descriptions of typical
nificantly earlier in the aLPR patterns (21.9 ± 14.1) when and aberrant movement patterns are discussed in detail
compared with the typical pattern (30.6 ± 16.5). This within the following paragraphs.
variable was retained for analysis of agreement between
clinical observation and kinematic classification of aLPR. Table 5 Mean and standard deviation local minimum (LMin),
local maximum (LMax), and sum of local minimum and maximum
(LSum) occurrences of clinically observed typical and judder (JUD)
Movement control patterns using phase-plane diagram for each segment
The numbers of LMin, and LMax occurrences, as well as
Segment Group LMin LMax LSum
the sum of local minimum and maximum (LSum) occur-
FT Typical 2.7 ± 1.4* 3.6 ± 1.4* 6.3 ± 2.7*
rences on the phase-plane and coupling-angle diagrams JUD 5.1 ± 2.4 5.9 ± 2.2 11.0 ± 4.5
FP Typical 2.1 ± 1.4* 4.1 ± 1.3* 7.3 ± 2.6*
Table 4 Mean and standard deviation of slope of typical and JUD 5.5 ± 3.0 6.1 ± 2.7 11.6 ± 5.7
altered lumbopelvic rhythm (aLPR) angle-angle diagram (mean PL Typical 3.0 ± 1.3* 3.7 ± 1.2* 6.7 ± 2.5*
coupling angle) of lumbopelvic segments (FP-PL) for the first, JUD 5.6 ± 2.5 6.1 ± 2.2 11.7 ± 4.6
second, and last 1/3rd of motion, and overall motion
PT Typical 1.3 ± 1.0* 1.9 ± 0.9* 3.2 ± 1.9*
Segment Group Slope JUD 2.6 ± 1.5 3.1 ± 1.2 5.7 ± 2.6
First 1/3 Second 1/3 Last 1/3 Overall FT Total trunk (Thoracic spine (T3) with respect to right femur), FP Pelvic
FP-PL Typical 32.0 ± 16.6* 51.3 ± 11.0* 69.17 ± 18.1* 51.99 ± 7.0* segment (Pelvis (S2) with respect to right femur), PL Lumbar segment (Lumbar
spine (L1) with respect to pelvis (S2)), PT Thoracolumbar segment (Thoracic
aLPR 44.7 ± 17.5 63.0 ± 10.4 58.91 ± 15.7 60.26 ± 7.4
spine (T3) with respect to pelvis (S2))
* = statistical significance (p < 0.10) * = statistical significance (p < 0.10)
Wattananon et al. BMC Musculoskeletal Disorders (2017) 18:455 Page 8 of 12

Table 6 Mean and standard deviation of local minimum (LMin), Altered inter-segmental coordination
local maximum (LMax), and sum of local minimum and maximum aLPR is characterized by either shared movement be-
(LSum) occurrences of clinically observed typical and judder (JUD) tween the pelvis and lumbar spine, or pelvic dominated
using the lumbopelvic (FP-PL) coupling-angle diagram motion during the first 1/3rd of the forward bend mo-
Segment Group LMin LMax LSum tion. Motion continues with increased pelvis domination
FP-PL Typical 3.6 ± 1.4* 4.3 ± 1.3* 8.0 ± 2.7* in the second 1/3rd of the motion. In the last 1/3rd of the
JUD 5.5 ± 3.3 6.3 ± 3.3 11.8 ± 6.7 motion the pattern is dominated by lumbar spine mo-
* = statistical significance (p < 0.10) tion (Table 4 and Fig. 5c). Overall this pattern is the re-
verse of a typical forward bend pattern.
Coupling-angle diagrams revealed sharp increases in
Typical forward bend movement pattern the coupling angle in the first 1/3rd of forward bend in-
Inter-segment coordination or lumbopelvic rhythm dur- dicating shared motion between the pelvis and lumbar
ing a forward bend motion is described as smooth and spine that occurs much earlier than the typical pattern.
continuous motion with the first third being dominated In the second 1/3rd of forward bend, the coupling angle
by lumbar spine motion, the second third shared motion increased indicating pelvis domination, which also ap-
between the lumbar and pelvic segments, and the last peared earlier than the typical pattern. In the last 1/3rd
third being dominated by pelvic motion (Table 4). Typ- of forward bend, the coupling angle decreased indicating
ical lumbopelvic coordination, represented in an angle- a reversed pattern (Fig. 5d).
angle diagram, is characterized by a smooth concave line Data suggested that patterns of lumbopelvic coupling
with gradual changes in segmental dominance (Fig. 5a). angles are key for identifying aLPR (Table 7 and Fig. 5d).
When plotted in a coupling-angle diagram, typical lum- Clinicians seemed to perceive pelvic domination when
bopelvic rhythm is represented by a diagonal line with a coupling angle (pelvic-dominated angle) reached 59°
positive slope from lower left corner to upper right cor- (pelvic-lumbar ratio = 1.66:1). At a coupling angle of 59°,
ner (Fig. 5b). This kinematic description is consistent timing (relative to the % of motion during forward bend-
with the clinical definition of typical lumbopelvic rhythm ing task) that maximizes the agreement between clinical
described by Calliet and Farfan [27, 29]. observation and kinematic classification derived from
The control of each segmental movement was demon- coupling-angle diagram demonstrated transition from
strated by a smooth gradual increase in velocity to mid- lumbar spine domination to pelvis domination in the
point of forward bend, and then a smooth gradual second 1/3rd of the movement (38%). This slight differ-
decrease in velocity to the end of the forward bend mo- ence between the clinical definition (shift within first
tion. Overall the phase-plane diagram was bell shaped 33% of motion) and kinematic cut point (38% of motion)
(Fig. 6a), with a minimal number of local minimum and is likely related to time normalization and averaging data
maximum occurrences (Table 5). No quick out of sagit- across subjects. However, the lumbopelvic coupling-
tal plane movements were noticed during the forward angle diagram represents what the clinicians observed as
bend motion. Inter-segmental movement coordination altered lumbopelvic rhythm with focus on the amount of
was also smooth and continuous with a minimal number pelvis contribution (pelvis domination) in the early
of local minimum and maximum occurrences (Table 6 phase (first 38%) of the movement.
and Fig. 6b). These kinematic descriptions are also con- A limited number of studies exist that describe trunk
sistent with clinical definition of typical (normal) for- and pelvic angular motion during a standing trunk for-
ward bending described by Paris [17]. ward bend task [20, 21, 39, 40]. In these studies, the

Table 7 Agreement (95% CI) between clinical observation and kinematic lumbopelvic segment movement characteristics and
accuracy statistics of the kinematic variables for predicting the observed movement pattern
Type Variable Diagram AUC %Agreement Kappa PABAK χ2 p value Sensitivity Specificity +LR -LR
aLPR CA_FP_PL@59 CA 0.73 74.55 0.47 0.49 38.33 <0.001 0.60 0.86 4.27 0.46
and Timing < 38% (0.32–0.60) (0.33–0.62) (0.52–0.67) (0.79–0.91) (2.49–7.62) (0.36–0.61)
JUD LMin_PL ≥ 6 PP 0.85 85.93 0.50 0.73 40.55 <0.001 0.43 0.98 22.75 0.58
(0.30–0.59) (0.58–0.84) (0.30–0.49) (0.94–0.99) (5.44–143.44) (0.51–0.74)
LSum_CA ≥ 15 CA 0.66 85.19 0.46 0.71 36.75 <0.001 0.37 0.99 38.5 0.64
(0.26–0.51) (0.55–0.82) (0.25–0.40) (0.96–0.1.00) (5.69–796.05) (0.60–0.79)
Typical (N = 108); Altered lumbopelvic rhythm (aLPR; N = 57); Judder (JUD; N = 30)
AUC Area under the receiver operating characteristic curve, PABAK Prevalence-adjusted bias-adjusted kappa, +LR Positive likelihood ratio, −LR Negative likelihood
ratio, CA Coupling-angle diagram, PP Phase-plane diagram, CA_FP_PL@59 and Timing < 38% Coupling angle of pelvis and lumbar spine at 59 degrees cut-off and
38% of movement in coupling angle reached 59 degrees for FP-PL, LMin_PL Number of local minimum occurrences (lumbar spine), LSum_CA Total number of local
minimum and maximum occurrences (lumbopelvic coupling angle)
Wattananon et al. BMC Musculoskeletal Disorders (2017) 18:455 Page 9 of 12

Fig. 5 Example of typical lumbopelvic rhythm in a lumbopelvic (FP-PL) angle-angle diagram (a) and a lumbopelvic (FP-PL) coupling-angle diagram (b),
and an example of altered lumbopelvic rhythm in a lumbopelvic (FP-PL) angle-angle diagram (c), and a lumbopelvic (FP-PL) coupling-angle diagram (d).
Solid line represents individual pattern, dotted line represents mean typical pattern, and dashed lines represent ±1 standard deviation of the typical
pattern. “X”s are placed to divide total movement into first, second and last 1/3rd of the movement in the angle-angle diagram (a and c). Typical
angle-angle diagram demonstrates the sequence of lumbar spine domination, shared, then pelvis domination; whereas altered lumbopelvic rhythm
demonstrates a reversed sequence as pelvis domination in first 2/3rd, then lumbar spine domination (a and c). In coupling-angle diagram (b and d),
altered lumbopelvic rhythm pattern (d) is identified if it falls in altered area (coupling angle greater than 59° before completion of 38% of the forward bend)

researchers investigated ratios of pelvis to lumbar seg- continuous lumbar spine and pelvis movement coordin-
ment motion at discrete points in the movement, and ation was a study that used lumbopelvic segment angle-
reported means and standard deviations. Although they angle diagrams to represent typical lumbopelvic rhythm
report differences between healthy and low back pain [41]. The finding from this study was similar to our find-
groups, this approach does not provide continuous in- ing in which typical lumbopelvic rhythm demonstrated
formation about inter-segment coordination and control. that the overall diagram was a concave line. The lumbar
Therefore, this existing body of work cannot fully de- spine had a greater contribution in the early stage of the
scribe altered lumbopelvic rhythm and pinpoint when motion followed by shared motion between the lumbar
transition from lumbar domination to pelvic domination and pelvic segments, and the pelvis had a greater contri-
occurred. The only reported approach that focused on bution in the last stage of the motion. Our approach of
Wattananon et al. BMC Musculoskeletal Disorders (2017) 18:455 Page 10 of 12

Fig. 6 Example of typical movement pattern in a lumbar phase-plane diagram (a) with local minimum occurrences (X) and a lumbopelvic (FP-PL)
coupling angle (b) with local minimum and maximum occurrences (+), and example of judder in a lumbar phase-plane diagram (c) with local
minimum occurrences (X) and a lumbopelvic (FP-PL) coupling angle diagram (d) with local minimum and maximum occurrences (+). Solid line
represents individual pattern, dotted line represents mean typical pattern, and dashed lines represent ±1 standard deviation of the typical pattern

using lumbopelvic coupling-angle diagrams provides trunk forward bend (Fig. 6a and c). This kinematic de-
typical timing (relative to % of movement) for when the scription was consistent with clinical observation of JUD
transition from lumbar to pelvic domination occurs. (a sudden deceleration and acceleration). Additionally,
This spatial and temporal information can be used to quick out of plane movement is best defined as when
further explain altered lumbopelvic rhythm. the pattern momentarily deviates away from sagittal
plane in lumbar spine plane angle-angle diagram (Fig. 4).
Altered movement control Quick out of plane movements were not frequently
Qualitative assessment of segment control during for- demonstrated in our dataset. We found that the occur-
ward bend suggests that the frequency of disruptions, or rence of quick out of plane movement was consistent
sudden decreases and increases in angular velocity, in with disruption in angular velocity in the lumbar spine
the JUD group were significantly greater than the typical phase-plane diagram. This suggests that clinical observa-
group (Table 5). Coupling-angle diagrams also revealed tion of judder based on lumbar spine velocity may be
that the JUD group had a greater number of sudden de- sufficient and observation of quick out of plane motion
coupling instances in inter-segmental coordination than might not be necessary for determination of judder. To
the typical group (Table 6). date, no researcher has investigated the primary segment
The data indicate that the kinematic patterns in and movement control characteristics that represent
phase-plane diagram considered as JUD are best defined JUD during standing trunk forward bend.
as the number of local minimum occurrences equal to The data also indicate that JUD can be quantified as
or greater than 6 in the lumbar spine segment. Clini- number of local minimum and maximum occurrences
cians appear to focus on lumbar spine angular velocity equal to or greater than 15 in coupling-angle diagram
or smoothness of the movement during the standing (Fig. 6b and d). Although the kinematic description
Wattananon et al. BMC Musculoskeletal Disorders (2017) 18:455 Page 11 of 12

using coupling-angle diagram was not matched with represent both typical and aberrant (altered lumbopelvic
clinical observation of JUD, it seemed clinicians’ classifi- rhythm, or judder) movements during standing forward
cation of this aberrant movement pattern is made bend. Coordination of the movement between the pelvis
through particular attention to changes in inter-segment and lumbar spine can be assessed for presence of altered
coordination between the pelvis and lumbar spine. Po- lumbopelvic rhythm and judder. The lumbar spine seg-
tentially, this kinematic description of judder could be ment appears to be the key segment to observe judder.
used to refine clinical observation of judder. These detailed kinematic descriptions should provide cli-
Collectively, the findings from our study provide de- nicians with direction for identifying aberrant movement
tailed descriptions of temporal and spatial 3-dimensional patterns. Collectively, these data can be used to help im-
multi-segmental kinematics of the pelvis and trunk seg- prove understanding of typical and aberrant movement
ments for typical and aberrant movements during stand- patterns, train clinicians in their clinical observation of
ing forward bend motion. This information can be used typical and aberrant movement patterns and to test the
to enhance knowledge and understanding of inter- efficacy of interventions to change inter-segmental co-
segment coordination and movement control, and may ordination and control.
help refine operational definitions that clinicians use to
Abbreviations
identify aberrant movement patterns. This information AA: Angle-angle diagram; aLPR: Altered lumbopelvic rhythm; ASIS: Anterior
may also be useful for future studies that investigate superior iliac spine; AUC: Area under the receiver operating characteristic
typical and aberrant movement patterns or are designed curve; CA: Coupling angle diagram; CMC: Coefficient of multiple correlation;
FP: Pelvic motion with respect to the femur; FT: Thoracic spine motion with
to determine the ability of exercise and motor control respect to the femur; ISB: International Society of Biomechanics; JUD: Judder;
based therapeutic interventions to alter these patterns. LBP: Low back pain; LMax: Local maximum; LMin: Local minimum;
The findings of this study should be considered in LPR: Lumbopelvic rhythm; LR: Likelihood ratio; LSum: Sum of local minimum
and maximum; LT: Thoracic spine motion with respect to the lumbar spine;
light of the following limitations. Our data interpretation MCI: Movement coordination impairment; NSLBP: Non-specific low back
was based on the observations of two experienced ortho- pain; PABAK: Prevalence-adjusted bias-adjusted kappa; PL: Lumbar spine
pedic physical therapists which limits generalizability. motion with respect to the pelvis; PP: Phase-plane diagram; PSIS: Posterior
superior iliac spine; PT: Combined lumbar-thoracic spine motion with respect
Data interpretation may also be influenced by their clin- to the pelvis; ROC: Receiver operating characteristic curve
ically imposed thresholds of aberrance. These thresholds
directly affect the prevalence of typical and aberrant rat- Acknowledgements
We would like to thank the Rehabilitation Sciences Research Laboratories of
ings. Our approach to analysis was from an accuracy sta- Drexel University and Physiotherapy Associates Clinic for providing spaces for
tistics perspective using maximum agreement to develop data collection. We would also like to thank all subjects who participated in
thresholds. It is possible that these thresholds or criteria this study.
are not the same as those used by other clinicians. We
Funding
also had a relatively low percentage of mutual agreement This study was funded in part by the Orthopaedic Section of the American
between two experienced clinicians when we included Physical Therapy Association.
only those repetitions where both raters indicated a typ-
Availability of data and materials
ical pattern or only one type of aberrant pattern on the The datasets used and/or analyzed during this study would be available from
same repetition. This was done to ensure that movement corresponding author upon reasonable request.
patterns we analyzed were clear representations of a typ-
Authors’ contributions
ical or aberrant pattern. Our prior work focusing on SPS, SAB, SSS, and DE have contributed to the conception, research design,
clinical agreement (clinician’s come to the same overall and manuscript preparation and edition. PW have substantially contributed
decision about typical or aberrant pattern for the sub- to data collection, data analysis, and drafting and revising the manuscript.
GEH has significantly contributed to data analysis, as well as editing and
ject) demonstrated moderate to almost perfect agree- revising the manuscript. All authors read and approved the final manuscript.
ment (kappa = 0.46–0.83) [18]. But this does suggest that
multiple repetitions are likely necessary for clinical Ethics approval and consent to participate
The study was approved by the Drexel University Institutional Review Board
agreement on movement patterns. We also acknowledge (project number: 1042185). All subjects provided written informed consent
the limitations associated with the use of the same data prior to participation.
set to develop and test accuracy of the kinematic vari-
Consent for publication
ables. However, this works serves as a starting point for All subjects provided written informed consent. The consent stated that all
quantification of aberrant movement patterns and we data collected would be used for publication.
recognize that further work and analysis is warranted.
Competing interests
The authors declare that they have no competing interests.
Conclusion
Angle-angle, coupling-angle, and phase-plane diagrams
Publisher’s Note
can be used to qualitatively and quantitatively describe Springer Nature remains neutral with regard to jurisdictional claims in published
3-dimensional multi-segmental kinematic patterns that maps and institutional affiliations.
Wattananon et al. BMC Musculoskeletal Disorders (2017) 18:455 Page 12 of 12

Author details 21. McClure PW, Esola M, Schreier R, Siegler S. Kinematic analysis of lumbar and
1
Motor Control and Neural Plasticity Laboratory, Faculty of Physical Therapy, hip motion while rising from a forward, flexed position in patients with and
Mahidol University, 999 Phuttamonthon 4 Road, Salaya, Nakhon Pathom without a history of low back pain. Spine. 1997;22(5):552–8.
73170, Thailand. 2Physical Therapy & Rehabilitation Sciences Department, 22. Wong TKT, Lee RYW. Effects of low back pain on the relationship between the
Drexel University, 1601 Cherry Street, Philadelphia, PA 19102, USA. 3Physical movements of the lumbar spine and hip. Hum Movement Sci. 2004;23(1):21–34.
Therapy Program, Neumann University, One Neumann Drive, Aston, PA 1901, 23. Kurz MJ, Stergiou N. Applied dynamic system theory for the analysis of
USA. 4College of Nursing and Health Professions, Drexel University, 245 N movement. In: Innovative analyses of human movement: analytical tools for
15th St, Philadelphia, PA 19102, USA. 5Department of Physical Therapy, human movement research. Champaign: human kinetics; 2004. p. 93–119.
University of Delaware, 540 S. College Ave, Suite 210E, Newark, DE 19713, 24. Silfies SP, Mehta R, Smith SS, Karduna AR. Differences in feedforward trunk
USA. muscle activity in subgroups of patients with mechanical low back pain.
Arch Phys Med Rehab. 2009;90(7):1159–69.
Received: 26 April 2017 Accepted: 8 November 2017 25. Stergiou N, Jensen JL, Bates BT, Scholten SD, Tzetzis G. A Dynamical systems
investigation of lower extremity coordination during running over
obstacles. Clin Biomech. 2001;16(3):213–21.
26. Spinelli BA, Wattananon P, Silfies S, Talaty M, Ebaugh D. Using kinematics
References and a dynamical systems approach to enhance understanding of clinically
1. Pengel LHM, Herbert RD, Maher CG, Refshauge KM. Acute low back pain: observed aberrant movement patterns. Manual Ther. 2015;20(1):221–6.
systematic review of its prognosis. BMJ. 2003;327:323. 27. Cailliet R. Low back pain syndrome. Philadelphia: F.A. Davis Co; 1988.
2. Andersson GBJ. Epidemiological features of chronic low-back pain. Lancet. 28. Cook C, Brismee J-M, Sizer PS. Factors associated with physiotherapists'
1999;354:581–5. confidence during assessment of clinical cervical and lumbar spine
3. Walker BF. The prevalence of low back pain: a systematic review of the instability. Physiother Res Int. 2005;10(2):59–71.
literature from 1966 to 1998. J Spinal Disord Tech. 2000;13(3):205–17. 29. Farfan HF. Muscular mechanism of the lumbar spine and the position of
4. Carey TS, Garrett JM, Jackman AM. Beyond the good prognosis: power and efficiency. Orthop Clin North Am. 1975;6(1):135–44.
examination of an inception cohort of patients with chronic low back pain. 30. Wu G, Siegler S, Allard P, Kirtley C, Leardini A, Rosenbaum D, et al. ISB
Spine. 2000;25(1):115–20. recommendation on definitions of joint coordinate system of various joints
5. Dankaerts W, O'Sullivan PB, Burnett AF, Straker LM. The use of a for the reporting of human joint motion—part I: ankle, hip, and spine. J
mechanism-based classification system to evaluate and direct management Biomech. 2002;35(4):543–8.
of a patient with non-specific chronic low back pain and motor control 31. Hamill J, Haddad JM, McDermott WJ. Issues in quantifying variability from a
impairment–a case report. Manual Ther. 2007;12(2):181–91. dynamical systems perspective. J Appl Biomech. 2000;16(4):407–18.
32. Miller RH, Chang R, Baird JL, Van Emmerik REA, Hamill J. Variability in
6. Flynn TW, Fritz JM, Whitman JM, Wainner RS, Magel JS, Rendeiro DG, et al. A
kinematic coupling assessed by vector coding and continuous relative
clinical prediction rule for classifying patients with low back pain who
phase. J Biomech. 2010;43(13):2554–60.
demonstrate short-term improvement with spinal manipulation. Spine.
33. Winstein CJ, Garfinkel A. Qualitative dynamics of disordered human
2002;27(24):2835–43.
locomotion: a preliminary investigation. J Motor Behav. 1989;21(4):373–91.
7. Hicks GE, Fritz JM, Delitto A, McGill SM. Preliminary development of a
34. Chang R, Van Emmerik R, Hamill J. Quantifying rearfoot-forefoot
clinical prediction rule for determining which patients with low back pain
coordination in human walking. J Biomech. 2008;41(14):3101–5.
will respond to a stabilization exercise program. Arch Phys Med Rehab.
35. Deyo RA, Centor RM. Assessing the responsiveness of functional scales to
2005;86(9):1753–62.
clinical change: an analogy to diagnostic test performance. J Chron Dis.
8. O'Sullivan PB. Diagnosis and classification of chronic low back pain
1986;39(11):897–906.
disorders. Maladaptive movement and motor control impairments as
36. Portney LG, Watkins MP. Foundations of clinical research: applications to
underlying mechanism. Manual Ther. 2005;10(4):242–55.
practice. 3rd ed. Upper Saddle River, NJ: Pearson & Prentice Hall; 2009.
9. Van Dillen LR, Sahrmann SA, Norton BJ, Caldwell CA, McDonnell MK, Bloom
37. Sackett DLA. Primer on the precision and accuracy of the clinical
NJ. Movement system impairment-based categories for low back pain:
examination. JAMA. 1992;267(19):2638–44.
stage 1 validation. J Orthop Sport Phys. 2003;33(3):126–42.
38. Sim J, Wright CC. The kappa statistic in reliability studies: use, interpretation,
10. O'Sullivan PB. Masterclass. Lumbar segmental 'instability': clinical presentation
and sample size requirements. Phys Ther. 2005;85(3):257–68.
and specific stabilizing exercise management. Manual Ther. 2000;5(1):2–12.
39. Mayer TG, Tencer AF, Kristoferson S, Mooney V. Use of noninvasive
11. Sahrmann SA. Diagnosis and treatment of movement impairment techniques for quantification of spinal range-of-motion in normal subjects
syndromes. Saint Louis, MO: Mosby Publishers; 2001. and chronic low-back dysfunction patients. Spine. 1984;9(6):588–95.
12. Delitto A, Erhard RE, Bowling RWA. Treatment-based classification approach 40. Paquet N, Malouin F, Richards CL. Hip-spine movement interaction and
to low back syndrome: identifying and staging patients for conservative muscle activation patterns during sagittal trunk movements in low back
treatment. Phys Ther. 1995;75(6):470–85. pain patients. Spine. 1994;19(5):596–603.
13. van Dijk M, Smorenburg N, Visser B, Heerkens YF, Nijhuis-van der Sanden 41. Lee RY, Wong TK. Relationship between the movements of the lumbar
MWG. How clinicians analyze movement quality in patients with non- spine and hip. Hum Movement Sci. 2002;21(4):481–94.
specific low back pain: a cross-sectional survey study with Dutch allied
health care professionals. BMC Musculoskel Dis. 2017;18(1):288.
14. Delitto A, George SZ, Van Dillen LR, Whitman JM, Sowa G, Shekelle P, et al.
Low back pain. J Orthop Sport Phys. 2012;42(4):A1–A57.
15. Farfan HF, Gracovetsky S. The nature of instability. Spine. 1984;9(7):714–9.
16. Kirkaldy-Willis WH, Farfan HF. Instability of the lumbar spine. Clin Orthop Submit your next manuscript to BioMed Central
Relat R. 1982;165:110–23. and we will help you at every step:
17. Paris SV. Physical signs of instability. Spine. 1985;10(3):277–9.
18. Rabin A, Shashua A, Pizem K, Dickstein R, Dar G. A clinical prediction rule to • We accept pre-submission inquiries
identify patients with low back pain who are likely to experience short-term • Our selector tool helps you to find the most relevant journal
success following lumbar stabilization exercises: a randomized controlled • We provide round the clock customer support
validation study. J Orthrop Sport Phys. 2014;44(1):6–B13.
19. Biely SA, Silfies SP, Smith SS, Hicks GE. Clinical observation of standing trunk • Convenient online submission
movements: what do the aberrant movement patterns tell us? J Orthop • Thorough peer review
Sport Phys. 2014;44(4):262–72. • Inclusion in PubMed and all major indexing services
20. Esola MA, McClure PW, Fitzgerald GK, Siegler S. Analysis of lumbar spine and
• Maximum visibility for your research
hip motion during forward bending in subjects with and without a history
of low back pain. Spine. 1996;21(1):71–8.
Submit your manuscript at
www.biomedcentral.com/submit