Human Movement Science 54 (2017) 2433

Contents lists available at ScienceDirect

Human Movement Science

journal homepage: www.elsevier.com/locate/humov

Full Length Article

Applying an active lumbopelvic control strategy during lumbar

extension exercises: Eect on muscle recruitment patterns of the
MARK
lumbopelvic region
Jessica Van Oosterwijcka,1, Eline De Riddera,1, Andry Vleeminga,b,

Guy Vanderstraetena, Stijn Schouppea, Lieven Danneelsb,
a
Department of Rehabilitation Sciences and Physiotherapy, Faculty of Medicine and Health Sciences, Ghent University, De Pintelaan 185, 1B3, B-
9000 Ghent, Belgium
b
Department of Anatomy, Center for Excellence in the Neurosciences, Medical Faculty, University of New England, Biddeford, ME, USA

AR TI CLE I NF O AB S T R A CT

Keywords: Objective: Examine whether implementing an active lumbopelvic control strategy during high
Electromyography load prone lumbar extension exercises aects posterior extensor chain recruitment and
Exercise therapy lumbopelvic kinematics.
Lumbosacral region Methods: Thirteen healthy adults acquired an optimal active lumbopelvic control strategy during
Paraspinal muscles
guided/home-based training sessions. During the experimental session electromyography was
Rehabilitation
Training
used to evaluate the activity of the posterior extensor chain muscles during high load trunk/
bilateral leg extension exercises with/without application of the strategy. Video-analysis was
used to evaluate thoracic/lumbar/hip angles.
Results: Implementing the active lumbopelvic control strategy decreased the lordotic angle
during trunk (p = 0.045; 3.2) and leg extension exercises (p = 0.019; 10). The hip angle
was solely aected during trunk extension (p < 0.001; +9.2). The posterior extensor chain (i.e.
mean of the relative activity of all muscles (%MVIC) was recruited to a higher extent (p = 0.026;
+9%) during trunk extension exercises performed with active lumbopelvic control. Applying the
strategy during leg extension exercises lead to less activity of longissimus thoracic (p = 0.015;
10.2%) and latissimus dorsi (p = 0.010; 4.4%), and increased gluteus maximus activity
(p 0.001; +16.8%).
Conclusions: When healthy people are taught/instructed to apply an active lumbopelvic control
strategy, this will decrease the degree of lumbar (hyper)lordosis and this inuences the
recruitment patterns of trunk and hip extensors. Hence, the possible impact on predetermined
training goals should be taken into account by trainers.

1. Introduction

Lumbar extension exercises are widely used in training regimens to enhance endurance, strength and functionality of the posterior
extensor chain as this will enhance performance levels (Steele, Bruce-Low, & Smith, 2015; Verna et al., 2002). This chain consists of
the extensor muscles located in the thoracic, lumbar and pelvic region of the posterior side of the body. More specically, the

Corresponding author.
E-mail address: Lieven.Danneels@Ugent.be (L. Danneels).
1
The rst 2 authors contributed equally to this manuscript.

http://dx.doi.org/10.1016/j.humov.2017.03.002
Received 10 August 2016; Received in revised form 14 March 2017; Accepted 15 March 2017
0167-9457/ 2017 Elsevier B.V. All rights reserved.
J. Van Oosterwijck et al. Human Movement Science 54 (2017) 2433

posterior extensor chain consists of the Latissimus Dorsi (LD), the Thoracic Erector Spinae (TES) which is formed by the Longissimus
thoracis pars Thoracic (LT) and Iliocostalis lumborum pars Thoracis (IT), the Lumbar Erector Spinae (LES) which is formed by the
Longissimus thoracis pars Lumborum (LL) and Iliocostalis lumborum pars Lumborum (IL), the Lumbar Multidus (LM), the Gluteus
Maximus (GM) and the Biceps Femoris (BF), which are functionally coupled via the thoracolumbar fascia (Vleeming, Pool-
Goudzwaard, Stoeckart, van Wingerden, & Snijders, 1995). Despite the fact that only some of these muscles directly attach on to the
lumbar vertebrae, a contraction of one of the muscles will inuence the lumbar region, even if this is not the muscle its primary
function. Although many dierent modalities of lumbar extension exercises exist (Mayer et al., 1999; Mayer, Mooney, & Dagenais,
2008; Mayer, Verna, Manini, Mooney, & Graves, 2002; Plamondon, Serresse, Boyd, Ladouceur, & Desjardins, 2002) trainers often
implement modalities performed from prone position into training programs (Mayer, Udermann, Graves, & Ploutz-Snyder, 2003;
Verna et al., 2002). By extending either the trunk or the legs, lumbar extension and activation of the muscles generating this
movement is induced. These exercises are high load as they activate the muscles from the posterior extensor chain at high degrees,
namely between 40 and 70% of their maximal voluntary isometric contraction (MVIC) (Clark, Manini, Mayer, Ploutz-
Snyder, & Graves, 2002; Dickx et al., 2010; Mayer et al., 1999, 2002; Plamondon et al., 2002). Because these trunk and leg
extension exercise modalities from prone position do not require expensive training devices they can be easily performed in the gym,
at home, and during eld training.
However, it has been shown that lumbar extension exercises from prone position can cause high spinal compressive loads (up to
6000 N) due to excessive anterior pelvic tilt and hyperlordosis of the lumbar spine, which could diminish the positive training eects
(Callaghan, Gunning, & McGill, 1998; Granata, Lee, & Franklin, 2005; McGill, 2002). It has been advocated that these disadvantages
can be limited by applying active lumbopelvic control techniques during exercise (Oh, Cynn, Won, Kwon, & Yi, 2007). In normal
trunk function co-contraction of the deep lumbopelvic muscles, i.e. LM, transversus abdominis (TA), and the pelvic oor muscles,
precedes the activation of prime movers during movements which jeopardize the trunk stability, providing mechanical stability for
spinal loads exceeding 1500 N (Demoulin, Distree, Tomasella, Crielaard, & Vanderthommen, 2007; Panjabi, 2003). As contraction of
this lumbopelvic muscle corset contributes signicantly to lumbar segmental control, the use of these deep muscles should be
optimized using training and implemented during lumbar extension exercises (Jull & Richardson, 1994). The lumbopelvic control
training focuses on teaching subjects to actively co-contract the deep lumbopelvic muscles and to maintain this co-contraction while
performing activities such as prone lumbar extension exercises (Cameron & Monroe, 2011; Oh et al., 2007). The continuous tonic low
level activation of the deep lumbopelvic muscles will form a cylinder around the lumbar spine providing functional control during
these activities and exercises (Cholewicki & VanVliet, 2002; Hodges, 1999, 2003).
Currently, little is known on how active involvement of this lumbopelvic muscle corset during lumbar extension exercises
inuences the recruitment of the posterior extensor chain in healthy people. Several studies have demonstrated that contracting the
lumbopelvic muscle corset alters the muscle recruitment patterns when performing low load exercises or daily activities (Oh et al.,
2007; Stevens et al., 2007; Watanabe, Eguchi, Kobara, & Ishida, 2007). For instance, it has been demonstrated that an abdominal
drawing-in maneuver, used to facilitate activation of the TA, during prone unilateral leg extension reduces the LES activity but
increases the activity of the hip extensors (Oh et al., 2007). On the other hand, it has been shown that contraction of the lumbopelvic
muscle corset during active sitting enhances the activity of the LES and LM (Watanabe et al., 2007). As these studies have examined
the recruitment patterns during dierent activities or positions, are limited to low load activities or exercises, and have not examined
the activity of the thoracic extensors, it is dicult to compare ndings and to conclude how recruitment of the posterior extensor
chain is exactly inuenced when an active lumbopelvic control strategy is implemented to high load lumbar extension exercises. In
the same context, it is relevant to note that most studies investigating muscle recruitment patterns during lumbar extension exercises
have overlooked the contribution of the hip extensors. A trunk extension consists of a combined extension movement of the thoracic
and lumbar spine as well as anterior rotation of the pelvis and hips (Graves et al., 1994; Pollock et al., 1989), whereas a leg extension
is composed of an extension of the lumbar spine and a rotation of the hips and pelvis (Oh et al., 2007). From this biomechanical
perspective it is clear that lumbar extension exercises do not only require activation of the trunk extensor muscles but also of the hip
extensor muscles (Arokoski et al., 1999; Kankaanpaa et al., 1998; Plamondon et al., 2002; Sparto & Parnianpour, 1998). Furthermore,
it has been shown that the active lumbopelvic control strategy is able to reduce the degree of lumbar lordosis during sitting
(Watanabe et al., 2007) and unilateral leg extension exercise (Oh et al., 2007), but no studies have examined whether this is also the
case during high load lumbar extension exercises from prone position such as trunk and bilateral leg extension exercises.
To examine whether the implementation of an active lumbopelvic control strategy during high load prone lumbar extension
exercises aects the lumbopelvic recruitment patterns and kinematics, the activation levels of the posterior extensor chain during
prone trunk and leg extension exercises were studied with and without the implementation of an active lumbopelvic control strategy.
Since the LES and (the deep bers of) the LM contribute to lumbar spine control (Cholewicki & VanVliet, 2002; MacDonald,
Moseley, & Hodges, 2006; Wilke, Wolf, Claes, Arand, & Wiesend, 1995), we hypothesized an increased recruitment of these muscles
and a reduced degree of lumbar lordosis when the lumbopelvic muscle corset is contracted during prone lumbar extension exercises.

2. Materials and methods

2.1. Subjects

A convenience sample consisting of 13 healthy subjects (9 females, 4 males) of 22.6 2.1 years participated in this study.
Subjects their mean height and weight were 172 7.3 cm and 61.3 9.5 kg respectively. Subjects were excluded from
participation if they; 1) reported previous back surgery or established spinal deformities, 2) had consulted a physician regarding

25
J. Van Oosterwijck et al. Human Movement Science 54 (2017) 2433

Fig. 1. Dynamic and functional positions and movements during which the active lumbopelvic control strategy is performed during training.

low back pain (LBP) in the past year, 3) currently experienced LBP, neck or hip pain, 4) had a history of severe neurologic,
respiratory, cardiovascular, or orthopedic disorders, 5) were elite athletes. The study was approved by the local Ethics Committee and
all subjects provided written consent. Subjects attended three training sessions and three test sessions.

2.2. Training sessions

Subjects were taught to acquire an optimal active control strategy of the lumbopelvic region by participating in three individually
guided training sessions (40 min 1/week). The subjects were instructed in accordance with the principles often used in training
lumbar control (Jull & Richardson, 1994; OSullivan, 2000; Richardson & Jull, 1995). In summary, the rst training session was used
to inform subjects about the basic anatomy and function of the muscles which form the lumbopelvic muscle corset, to learn to control
the lumbar spine in neutral position in dierent postures (sitting, standing, and 4-point kneeling), and to learn to contract the
lumbopelvic muscle corset without substitution strategies of supercial muscles and with focus on the continuation of normal
breathing. Once subjects were able to contract the lumbopelvic muscle corset whilst maintaining a normal breathing pattern and
without substitution of supercial muscles, the duration of the contractions and the repetitions were increased so that subjects were
able to sustain each contraction for 10 s and repeat this 10 times. When subjects were able to perform the latter exercise, the active
lumbopelvic control strategy was integrated within more dynamic and functional activities. During the second session controlled leg
movements were added in supine and 4-point kneeling position. During the last training session, subjects were instructed to sustain
the co-contraction during more complicated exercises as presented in Fig. 1. All exercises were performed in 3 sets of 15 repetitions
and the researchers gave tactile and verbal feedback about the performance of the co-contraction. In addition subjects received a
leaet describing the exercises and were asked to perform them daily at home.
At the end of the training program, the ability to optimally contract the lumbopelvic muscle corset was evaluated by two
independent researchers as this was a prerequisite to participate in the test sessions. This implied that subjects were able to perform a
10 s during isolated contraction of the lumbopelvic muscle corset in 4-point kneeling stance without substitution or compensatory
strategies whilst maintaining a normal breathing pattern, and repeat this 10 times. The contraction of the LM and the TA was assessed
through palpation lateral from the spinous process at L3L5, and 2 cm inferior and medial of the anterior superior iliac spine
respectively (Gill, Teyhen, & Lee, 2007; Hides, Richardson, & Jull, 1998). After training all subjects were able to perform an optimal
contraction of the lumbopelvic muscle corset and thus participated to the test sessions.

26
J. Van Oosterwijck et al. Human Movement Science 54 (2017) 2433

2.3. Test sessions

The rst test session consisted of the indirect determination of the 1-RM. Prone trunk and leg extension exercises were performed
with and without the implementation of the active lumbopelvic control strategy at 60% of 1-RM during the second and third test
session.

2.3.1. Indirect determination of the 1-RM

At least three days before the rst exercise session, the exercise load, expressed as a percentage of the one repetition maximum (%
of 1-RM), was estimated using the Holten-diagram which describes the relation between the performed number of repetitions and the
exercise intensity (Danneels, Cools, et al., 2001; Dickx et al., 2010). Therefore all subjects were asked to execute the maximal amount
of repetitions of dynamic prone trunk/leg extension with the weight of their upper/lower body as the exercise weight (which is
estimated as 70% and 30% of the total body weight respectively). The total body weight was determined on-site on a calibrated body
scale. The number of repetitions each subject was able to perform during both types of exercises using this method was registered.
The exercise intensity was individually adjusted at 60%RM, calculated using the following formula; exercise weight (kg) = (upper or
lower body weight (kg) exercise load (i.e. 60% of 1-RM))/exercise load determined on the testing day (% of 1-RM derived from
Holten-diagram) (Dickx et al., 2010).

2.3.2. Lumbar extension exercises

During the remaining two test sessions 4 dierent lumbar extension exercise modalities were performed i.e. dynamic trunk
extension, dynamic bilateral leg extension, dynamic trunk extension with the instruction to implement the active lumbopelvic control
strategy, and dynamic bilateral leg extension with the instruction to apply the active lumbopelvic control strategy. During each
exercise session 2 modalities were performed. The order of the exercises modalities was randomized by lottery. To prevent muscular
fatigue, only two dierent exercise modalities were performed during one test session and an interval of 30 min was obligated
between these two exercise modalities.
All extension exercises were performed in prone position on a variable angle chair with the trunk or legs positioned at 45 of
exion and the superior border of the anterior iliac placed on the edge of the variable angle chair (De Ridder, Van Oosterwijck,
Vleeming, Vanderstraeten, & Danneels, 2013). For the trunk extension exercises the legs were strapped to the table at the ankles, and
subjects their hands were placed on the opposite shoulder. To perform the leg extension exercises the upper body was strapped to the
table at the level of the angulus inferior of the scapulae, and subjects their hands were placed under their forehead. The exercise
positions are presented in Fig. 2.
One repetition consisted of lifting the trunk/legs to the horizontal in 2 s and returning to starting position in 2 s. A metronome
(60 beats/min) was used to ensure appropriate timing of the contractions. To reach the horizontal position, tactile feedback was given
by a rope between two vertical stands. Subjects performed one set of 10 dynamic repetitions. Before starting the exercises which
included the active lumbopelvic control strategy following instructions were given; Keep your lumbar region in neutral position and
maintain a lumbopelvic muscle co-contraction during the whole exercise.

2.4. Measures

2.4.1. Electromyography (EMG)

The EMG signals of 8 muscles, were bilaterally measured using a 16 channel telemetric surface EMG system (TeleMyo 2400 G2

Fig. 2. Position of the prone lumbar extension exercises.

27
J. Van Oosterwijck et al. Human Movement Science 54 (2017) 2433

Telemetry System, Noraxon, USA).

The skin was shaved and rubbed with alcohol to reduce the impedance and improve skin contact. Noraxon surface dual electrodes
with a xed inter-electrode distance of 2 cm and an electric surface contact of 1 cm diameter were bilaterally attached, parallel to the
muscle ber orientation, in analogy to Coorevits, Danneels, Cambier, Ramon, and Vanderstraeten (2008), Danneels, Cagnie, et al.
(2001) and Danneels, Vanderstraeten, et al. (2001) over the GM, BF, LM, LT, LL, IT, and IL. A single reference electrode was placed on
the angulus inferior of the right scapula.
Before starting the exercises, EMG reference data were obtained by performing 3 MVICs of 4 s against manual resistance for each
muscle (group) in prone position (Danneels, Vanderstraeten, et al., 2001; De Ridder et al., 2013). Each MVIC was followed by 30 s of
rest.
Raw signals were bandpass-ltered between 10 and 500 Hz, amplied (common mode rejection ratio > 100 dB, overall gain
1000, noise < 1 V RMS), and analogue-to-digital (16-bit) converted at a sampling rate of 1500 Hz. The signal processing consisted
of full wave rectication and smoothing, using a RMS algorithm with a 100 ms time constant. Muscle activity was measured during
the whole exercise. The mean activity level for each muscle was calculated over 5 separate repetitions (repetitions 26) and used for
further analysis. Because the muscle activity during the exercises was measured on two separate days, which could result in
dierences in skin resistance, the EMG signals of the muscles were normalized against their MVICs.

2.4.2. Kinematics
To measure the kyphotic angle reective markers were placed on the spinous process of the 7th cervical vertebra, the most
prominent point of the thoracic kyphosis also described as the thoracic apex (approximately T7), the vertebral level where the spine
transitions from lordosis to kyphosis also known as the thoracolumbar junction or inection point (corresponds with spinous process
of the 1st lumbar vertebra. To determine the lordotic angle reective markers were placed on the infection point, the middle of the
deepest point of the lumbar lordosis curve also described as the lumbar apex (approximately L3), and the posterior superior iliac
spine. The hip extension angle was measured via reective markers on the anterior superior iliac spine, the greater trochanter, and
the lateral malleolus. The locations of the marker placements were determined by one examiner via visual inspection and palpation
(Dolphens et al., 2012).
Video analysis was limited to 2D and recordings were made throughout the entire exercise using a digital video recorder (Sony
Handycam DCR-HC 37 capturing 50 frames per second at 1000 Hz) placed in a standardized position perpendicular to the plane in
which the movement took place and real-time synchronized with the EMG recordings within the EMG software. The 3 angles
mentioned above were calculated using the video frames in which the subject reached the horizontal position during the exercise via
the freeware motion-analysis software Kinovea (version 0.8.15, www.kinovea.org) which allows angle calculation with a margin of
error of 1 degree. The mean angle was calculated from the measured angles during repetitions 2, 4 and 6, with a margin of error of 1
degree. The analysis was performed by a trained researcher who was unaware whether the exercise was performed with or without
the instruction to actively control the lumbopelvic region.

2.5. Statistical analysis

Statistical analysis was performed using SPSS 19.0 (IBM corporation, USA). Due to the symmetry of the exercises and the lack of a
side eect (left vs right), the EMG values of the muscles of both sides were averaged. The Kolmogorov-Smirnov was used to test the
normality of all variables and descriptive statistics were computed for the anthropometric characteristics and muscle activity. The
recruitment pattern of the posterior extensor chain (%MVIC) (dependent variable) during the lumbar extension exercises was
analyzed using a linear mixed model analysis, with the EMG values as the dependent variables and the following 3 factors: instruction
to apply the active lumbopelvic control strategy (exercise with vs without the instruction to actively control the lumbopelvic region),
muscle (LD, IT, IL, LT, LL, LM, GM, BF) and extension modality (trunk extension vs leg extension). Secondarily, because the inuence of
active lumbopelvic control strategy depends on the moving body part (instruction to apply active lumbopelvic control strategy -
body part), trunk and leg extension exercise were analyzed separately. Therefore two new mixed models were conducted. When
required, post hoc comparisons were made and adjusted using a Bonferroni-correction. Paired sample T-tests were used to determine
the alterations in kinematics between the lumbar extension exercise with and without the instruction to apply the active lumbopelvic
control strategy. Statistical signicance for all tests was set at p 0.05.

3. Results

3.1. Kinematics

The instruction to apply an active lumbopelvic control strategy during dynamic bilateral leg extension and dynamic trunk
extension resulted in signicant less lumbar lordosis (resp. 10, t(3) = 4.7, p = 0.019, 95%CI [0.95, 6.23]; 3.2, t(5) = 2.65,
p = 0.045, 95%CI [4.32, 21.15]) compared to the same exercises during which this strategy was not used. The thoracic angle showed
no signicant dierences between the lumbar extension exercises performed with or without the active lumbopelvic control strategy
(p between 0.059 and 0.082).
The hip extension angle during actively controlled trunk extension was signicantly larger compared to the non-controlled
modality (+9.2, t(4) = 15.78, p < 0.001, 95%CI [7.58, 10.81]). No signicant dierences in hip angle could be found between
the leg extension exercises with and without active lumbopelvic control (p = 0.061).

28
J. Van Oosterwijck et al. Human Movement Science 54 (2017) 2433

Table 1
Mean kyphotic, lordotic and hip angle and standard deviation (SD) during prone lumbar extension exercise modalities.

Angles Trunk extension Leg extension

Without active lumbopelvic control With active lumbopelvic control Without active lumbopelvic control With active lumbopelvic control
strategy strategy strategy strategy
Mean (SD) Mean (SD) Mean (SD) Mean (SD)

Kyphotic 145.0 (3.6) 135.0 (2.7) 154.8 (5.5) 147.3 (2.9)

Lordotic 37.2 (6.5)* 34.0 (5.9) 39.3 (4.0)* 29.3 (1.5)
Hip 145.6 (2.1)* 154.8 (2.5) 152.0 (3.2) 155.3 (3.6)

* Signicant dierence at p 0.05 between the exercise modality without and with active lumbopelvic control strategy.

Mean angles and standard deviations are presented in Table 1.

3.2. General eect of an active lumbopelvic control strategy during lumbar extension exercises

No signicant 3 way-interactions (instruction to apply an active lumbopelvic control strategy body part muscle) was found
(F = 1.997, p = 0.053). The trunk and leg extension exercises were analyzed separately (instruction to apply an active lumbopelvic
control strategy muscle). A signicant 2-way interaction for the leg extension exercises was shown (F = 3.73, p = 0.001).
Regarding the trunk extension exercises no 2-way interaction eect was found, but a signicant eect of the main factors was
observed (instruction to apply an active lumbopelvic control strategy F = 27.17, p = 0.026; muscle F = 39.14, p < 0.001).

3.3. The eect of an active lumbopelvic control strategy during trunk extension

When the instruction was given to apply an active lumbopelvic control strategy during dynamic trunk extension exercise this
resulted in a signicant (F = 27.17, p = 0.026, 95%CI [5.62, 12.43]) higher recruitment level (+9%) of the posterior extensor chain
(i.e. mean of the relative activity of all muscles (%MVIC)).

3.4. The eect an active lumbopelvic control strategy during leg extension

The mean relative activity of the trunk and hip extensor muscles did not signicantly dier between leg extension exercises
performed with active lumbopelvic control (44.1%MVIC) and those without active lumbopelvic control (45.9%MVIC). However, a
signicant interaction eect between leg extension with or without active lumbopelvic control and muscle was noticeable, suggesting
that the inuence of an active lumbopelvic control strategy during leg extension depends on the analyzed muscle.
The LD and LT showed signicant lower levels of mean muscle activity during controlled leg extensions (respectively 4.4%
MVIC, F = 7.03, p = 0.010, 95%CI [7.75, 1.09]; 10.2%MVIC, F = 6.24, p = 0.015, 95%CI [18.39, 2.04]). The
recruitment of the IT, LL, IL, LM and BF did not signicantly (p > 0.05) change when the active control strategy was added,
whereas the GM was clearly activated at a higher degree (+16.8%MVIC, F = 14.57, p < 0.001, 95%CI [8.02, 25.65]) when
active lumbopelvic control was performed during the exercise.
The inuence of the active lumbopelvic control strategy on the mean muscle activity during leg extension is presented in Table 2.

Table 2
Mean muscle activity (%MVC) and standard deviation (SD) during prone lumbar extension exercise modalities.

Muscles Trunk extension Leg extension

Without active lumbopelvic control With active lumbopelvic control Without active lumbopelvic control With active lumbopelvic control
strategy strategy strategy strategy
Mean (SD) Mean (SD) Mean (SD) Mean (SD)

LD 20.9 (8.8)* 30.5 (14.4) 12.0 (8.6)* 7.6 (3.3)

LT 52.8 (20.7) 59.8 (23.6) 61.5 (19.6)* 51.3 (11.8)
IT 51.1 (20.2) 54.9 (17.6) 48.6 (22.7) 51.0 (14.7)
LL 57.7 (22.6) 65.4 (18.4) 62.2 (22.8) 53.6 (16.8)
IL 58.4 (21.0) 67.5 (19.2) 56.8 (19.6) 51.2 (20.6)
LM 57.2 (20.4) 63.5 (21.0) 64.2 (27.5) 57.1 (18.1)
GM 23.8 (10.1)* 32.4 (21.6) 22.0 (7.7)* 38.8 (24.1)
BF 44.2 (20.3)* 64.4 (20.8) 39.9 (22.5)* 42.3 (18.1)

LD = Latissimus Dorsi, LT = Longissimus thoracis pars Thoracic, IT = Iliocostalis lumborum pars Thoracis, LL = Longissimus thoracis pars Lumborum,
IL = Iliocostalis lumborum pars Lumborum, LM = Lumbar Multidus, GM = Gluteus Maximus, BF = Biceps Femoris.
* Signicant dierence at p 0.05 between the exercise modality without and with active lumbopelvic control strategy.

29
J. Van Oosterwijck et al. Human Movement Science 54 (2017) 2433

4. Discussion

This study is the rst to demonstrate that an active control strategy of the lumbopelvic region during high load dynamic prone
lumbar extension exercises aects the degree of lumbar lordosis and alters the recruitment patterns of the posterior extensor chain.
The nding that this active control strategy is able to limit the increases in lumbar (hyper)lordosis which normally take place during
the performance of lumbar extension exercises is an important element to consider in clinical practice when creating exercise
programs, as high load lumbar extension exercises in which (hyper)lordosis occurs have been shown to be associated with increased
spinal loads and thus risks for spinal injuries (Callaghan et al., 1998).
The EMG data of this study indicate that the posterior extensor chain is more active when trunk extension is performed while
actively controlling the lumbopelvic region. The increase in total muscle work during a controlled trunk extension can be explained
by a more detailed analysis of the moving body parts during the exercise. Earlier studies consider a trunk extension as a compound
movement of the thoracic and lumbar spine, combined with a rotation of the pelvis and extension of the hips (Graves et al., 1994;
Pollock et al., 1989). In the present study it was demonstrated that the use of an active lumbopelvic control strategy eectively
prevents excessive lumbar lordosis during trunk extension, resulting in an augmentation of hip extension. Co-contracting the deep
muscles of the lumbopelvic corset provides a higher segmental control and reduces lumbar displacement, which explains the decrease
in lumbar extension (Barr, Griggs, & Cadby, 2005; Brown, Vera-Garcia, & McGill, 2006; Vera-Garcia, Elvira, Brown, & McGill, 2007).
By integrating an active control strategy there is a decrease in lumbar extension. Therefore the trunk extension needs to be performed
with more hip extension in order to reach the horizontal end-position. As a result of the changed biomechanics, the lever arm to lift
the trunk will be greater in the actively controlled exercise condition, and thus more eort of the muscles is necessary. In line with the
altered lever arm, the muscles conjoining the thorax with the pelvis, have a more stable basis to lift the trunk when the exercise is
performed whilst actively controlling the lumbopelvic region. In this condition the muscles will be able to pull the trunk as a solid
mass upwards which could explain their increased recruitment.
Since the LES and, (the deep) LM in specic, are assumed to highly contribute to lumbar spinal control (Cholewicki & VanVliet,
2002; MacDonald et al., 2006; Wilke et al., 1995), an increased recruitment of these muscles was expected when the lumbopelvic
muscle corset was contracted during the prone lumbar extension exercises. While our study results could not specically verify this
hypothesis, a previous study supports this assumption by showing a higher lumbar muscle contribution (i.e. 25% increase in the EMG
activity of the lumbar extensors) when maintaining a neutral lumbar position during a dynamic trunk extension exercise from prone
position (Mayer et al., 2002). The present study did not specically show an increase of the LES or LM activity, but showed an
increase of the relative activity of the posterior extensor chain during a dynamic trunk extension exercise performed from prone
position while applying an active lumbopelvic control strategy. This dierence between our ndings and the results of the study of
Mayer et al. (2002) could be caused by three dierent elements; 1) the exercise intensity, 2) the type of control strategy, and 3) the
muscle evaluation technique. In the study of Mayer et al. (2002) the weight of the trunk was used as the exercise load, which is
probably lower than the intensity of 60% 1-RM used in our study. As a result of the high exercise load in our study, large levels of
lumbar muscle activity were already generated in the non-controlled condition. In addition, not only did the activity of the lumbar
extensor muscles increase in our study but also the thoracic and hip extensor muscles were recruited to a higher extent, whereas
Mayer et al. (2002) found no inuence of the instruction to maintain the lumbar position on the activity of the hip extensors.
Subsequently, one could assume that the overall increased recruitment of the posterior extensor chain in our study could have
compensated for the expected higher lumbar muscle activity when actively controlling the lumbopelvic region. The combination of
both the high exercise intensity and the overall increase in recruitment of the posterior extensor chain could explain the rather
identical net contribution of the dierent extensor muscles during both exercise conditions seen in the present study. The
dissimilarities with our results could also be explained by dierences in the strategies which were used. Whereas in the current study
subjects were learned to co-contract the deep stabilizing muscles of the lumbopelvic region during dierent training sessions and
instructed to integrate this strategy during the trunk extension exercise, the study of Mayer et al. (2002) only instructed to maintain
the lumbar lordosis without focusing on the deep muscles. To date, both strategies, without consensus about which strategy is the
most benecial, are used in clinical rehabilitation and exercise programs. With regard to the evaluation technique, both studies used
surface EMG to capture the signals of LES, but while the study of Mayer et al. (2002) examined the global LES activity the current
study made a dierence between IL and LL.
This is the rst study to investigate the eect of an active lumbopelvic control strategy during bilateral leg extension exercises. It
was shown that although the use of this strategy did not inuence the overall amount of posterior extensor chain, excessive lumbar
lordosis was prevented and the distribution of the activity between the muscles was altered.
When the dynamic high load leg extension exercise was performed with and without the use of an active lumbopelvic control
strategy, it was observed that the implementation of the control technique resulted in a decrease of the relative activity levels of the
LD and LT, whereas the GM activity increased. The present observation is in part conrmed by the study of Oh et al. (2007) in which
a decreased anterior pelvic tilt, higher GM activity and a decrement in LES activity was established when performing prone unilateral
leg extension during which the lumbar region was actively controlled. Although in the present study the lower lumbar extensor
activity levels also decreased, dierences were not statistically signicant. To complete the leg extension exercise the thoracic
extensors are activated, likely causing an initial increase in the degree of lumbar lordosis. However, when the lumbopelvic region is
controlled the action of these muscles is partly inhibited, resulting in a reduced lumbar angle. In contrast to the trunk extension
exercise, no compensation in the hip extension angle occurs in order to reach the horizontal. This can be explained by the fact that
even in the non-actively controlled leg extension exercise a large hip extension movement occurs. These changes in kinematics are
supported by the alterations in the activity levels of the relevant muscles. As a consequence of the limited lumbar extension, the

30
J. Van Oosterwijck et al. Human Movement Science 54 (2017) 2433

thoracic muscles, particularly the LD and LT, contribute less during dynamic leg extension with active control of the lumbopelvic
region. Despite our expectations, there was no signicant eect on the lumbar extensor muscle activity. As stated above this could be
due to the high exercise intensity in our study, which will automatically induce high activity levels of the lumbar muscles. Although,
there was no signicant increase in hip extension, actively controlling the lumbopelvic region enhanced the activation of the GM, but
not the BF. A possible explanation is that a decrease in lumbar angle is associated with a posterior pelvic tilt, which is mainly caused
by the contraction of the GM (Kendall, Kendall, & Wadswoth, 1971), while the BF acts as prime mover and actually lifts the legs
during the leg extension exercise.
Our results concerning the relative activity of thoracic, lumbar and hip extensors between the exercise conditions, can support the
assumption that no trunk muscle is superior in controlling the spine during movements (Cholewicki & VanVliet, 2002; Kavcic,
Grenier, & McGill, 2004). The higher contribution of the large torque producing muscles during high load prone extension exercises
indicates that control of the lumbar spine is not only achieved by the deep lumbopelvic muscles but also by the large torque
producing muscles, which probably induce a more global stability. This co-operation between the deep lumbopelvic muscles and the
large torque producing muscles will assist in the control of spinal buckling and intervertebral motion via compression. In this respect,
the activity of the GM and LD will contribute to spinal control through the thoracolumbar fascia (McGill, 2002; Vleeming et al.,
1995). Moreover, the higher contraction of the GM in the actively controlled exercise conditions could help the deep muscles to
control the lordotic angle, by generating a posterior pelvic tilt in high load conditions.
Although it was not our main scope to compare both exercises, the degree of lumbar lordosis during a non-actively controlled
trunk extension exercise varied from 31 to 43.5 and from 35 to 43 during the non-actively controlled leg extension exercise. During
the actively controlled trunk extension lordotic angles between the 28 and 40 were demonstrated, whereas only small variations in
lumbar lordosis were demonstrated during the leg extension exercise (2831). This emphasizes that the eect of contracting the
lumbopelvic muscle corset on the lordotic angle is higher during the leg extension exercise than during the trunk extension exercise.
The large dierences in the kinematics during a prone leg extension versus a prone trunk extension may be related to the freedom of
movement of the pelvis. Whereas the pelvis rests on the table during prone trunk extension exercises, the pelvis is unsupported during
the leg extension exercises and hence a larger movement range is possible in the latter condition. This is armed by the lordotic and
hip extension angles demonstrated during leg extension.
The present study has several limitations that need to be taken into account. The study was executed on a small population of
young healthy individuals. As altered muscle recruitment patterns occur in case of LBP and this condition is related to poor muscle
function, future research in these patients would be appropriate. After a short training program, the ability to contract the deep
lumbopelvic muscles independently of more supercial muscles was judged via observation and palpation since this techniques is
commonly used in clinical practice, although the use of ultrasound may have provided a more accurate indication. Similar, the
thoracolumbar and lumbopelvic angles were assessed using a clinical video set-up implying 2D analysis which is less accurate than
3D. As surface EMG was used to measure muscle activity, crosstalk from surrounding muscles cannot be excluded. In this light, the
utilization of ne wire EMG or mfMRI to evaluate deep muscle work would be more appropriate and allows to identify possible
dierences among the LES and LM. Another consideration is that the type of training which was used in the study can be combined
with more extensive sensorimotor control training when indicated. The eventual goal would be to obtain an automatic activation of
the deep lumbopelvic muscles during activities and exercises. The current study design did not allow to exclude the possibility that an
automatic activation of the deep muscles took place during the exercise condition in which no instructions regarding the application
of the lumbopelvic control strategy were given. However, even if this would be the case, the dierential results between the exercise
condition with and without instruction to actively apply the control strategy show that providing these instructions does make a
dierence. Although it has been previously shown that when high-load lumbar extension exercises are performed with less lumbar
lordosis this reduces the spinal loads (Callaghan et al., 1998), the current study only examined how the implementation of an active
lumbopelvic control strategy during lumbar extension exercise performance inuences the degree of lumbar lordosis and thus
conclusions regarding the eects on spinal loading cannot be drawn from this study.

5. Conclusions

The present study showed that the instruction to apply an acquired active lumbopelvic control strategy, existing of co-contraction
of the deep lumbopelvic muscles, is able to reduce lumbar (hyper)lordosis during high load dynamic trunk extension exercises and
bilateral leg extension exercises performed by healthy people. The observation that an active lumbopelvic control strategy, as used in
this study can diminish the degree of lumbar lordosis during prone extension exercises, is highly valuable for clinical practice. When
integrating these exercises in training programs, it is advisable to learn and instruct subjects to use the active control strategy of the
lumbopelvic region in order to reduce the degree of lumbar lordosis and prevent the associated risk of spinal injury. However, the
implementation of this active lumbopelvic control strategy aects the muscle recruitment patterns of the posterior extensor chain
during these exercises. In particular, when the active lumbopelvic control technique is implemented during bilateral leg extensions
exercises from prone position, the recruitment of the large torque producing muscles is changed in order to complete the lumbar
extension exercises. Moreover, based on our ndings it can be assumed that both the deep muscles and large torque producing
muscles are working harmoniously in order to control and move the lumbar spine during high load prone extension exercises.

Funding statement

Jessica Van Oosterwijck is a postdoctoral research fellow funded by the Special Research Fund of Ghent University. Stijn

31
J. Van Oosterwijck et al. Human Movement Science 54 (2017) 2433

Schouppe is nancially supported by an Interdisciplinary Research grant (BOF14/IOP/067) from the Special Research Fund of Ghent
University.

Conict of interest

The authors declare that there are no conicts of interest to report.

Acknowledgements

The authors would like to thank Marieke De Simpelaere, Janis Demeurisse and Bram Van Wittenberge for their assistance during
data collection.

References

Arokoski, J. P., Kankaanpaa, M., Valta, T., Juvonen, I., Partanen, J., Taimela, S., ... Airaksinen, O. (1999). Back and hip extensor muscle function during therapeutic
exercises. Archives of Physical Medicine and Rehabilitation, 80(7), 842850.
Barr, K. P., Griggs, M., & Cadby, T. (2005). Lumbar stabilization: Core concepts and current literature, Part 1. American Journal of Physical Medicine and Rehabilitation,
84(6), 473480.
Brown, S. H., Vera-Garcia, F. J., & McGill, S. M. (2006). Eects of abdominal muscle coactivation on the externally preloaded trunk: Variations in motor control and its
eect on spine stability. Spine, 31(13), E387E393. http://dx.doi.org/10.1097/01.brs.0000220221.57213.25.
Callaghan, J. P., Gunning, J. L., & McGill, S. M. (1998). The relationship between lumbar spine load and muscle activity during extensor exercises. Physical Therapy,
78(1), 818.
Cameron, M. H., & Monroe, L. G. (2011). Physical rehabilitation for the physical therapist assistant.
Cholewicki, J., & VanVliet, J. J. T. (2002). Relative contribution of trunk muscles to the stability of the lumbar spine during isometric exertions. Clinical Biomechanics
(Bristol, Avon), 17(2), 99105.
Clark, B. C., Manini, T. M., Mayer, J. M., Ploutz-Snyder, L. L., & Graves, J. E. (2002). Electromyographic activity of the lumbar and hip extensors during dynamic trunk
extension exercise. Archives of Physical Medicine and Rehabilitation, 83(11), 15471552. http://dx.doi.org/10.1053/apmr.2002.34828.
Coorevits, P., Danneels, L., Cambier, D., Ramon, H., & Vanderstraeten, G. (2008). Assessment of the validity of the Biering-Sorensen test for measuring back muscle
fatigue based on EMG median frequency characteristics of back and hip muscles. Journal of Electromyography and Kinesiology, 18(6), 9971005. http://dx.doi.org/
10.1016/j.jelekin.2007.10.012.
Danneels, L. A., Cagnie, B. J., Cools, A. M., Vanderstraeten, G. G., Cambier, D. C., Witvrouw, E. E., & De Cuyper, H. J. (2001). Intra-operator and inter-operator
reliability of surface electromyography in the clinical evaluation of back muscles. Manual Therapy, 6(3), 145153. http://dx.doi.org/10.1054/math.2001.0396.
Danneels, L. A., Cools, A. M., Vanderstraeten, G. G., Cambier, D. C., Witvrouw, E. E., Bourgois, J., & de Cuyper, H. J. (2001). The eects of three dierent training
modalities on the cross-sectional area of the paravertebral muscles. Scandinavian Journal of Medicine and Science in Sports, 11(6), 335341.
Danneels, L. A., Vanderstraeten, G. G., Cambier, D. C., Witvrouw, E. E., Stevens, V. K., & De Cuyper, H. J. (2001). A functional subdivision of hip, abdominal, and back
muscles during asymmetric lifting. Spine, 26(6), E114E121.
De Ridder, E. M., Van Oosterwijck, J. O., Vleeming, A., Vanderstraeten, G. G., & Danneels, L. A. (2013). Posterior muscle chain activity during various extension
exercises: An observational study. BMC Musculoskeletal Disorders, 14, 204. http://dx.doi.org/10.1186/1471-2474-14-204.
Demoulin, C., Distree, V., Tomasella, M., Crielaard, J. M., & Vanderthommen, M. (2007). Lumbar functional instability: A critical appraisal of the literature. Annales de
Radaptation et de Mdecine Physique: Revue Scientique de la Socit Franaise de Rducation Fonctionnelle de Radaptation et de Mdecine Physique, 50(8),
http://dx.doi.org/10.1016/j.annrmp.2007.05.007 677-684, 669-676.
Dickx, N., Dhooge, R., Cagnie, B., Deschepper, E., Verstraete, K., & Danneels, L. (2010). Magnetic resonance imaging and electromyography to measure lumbar back
muscle activity. Spine, 35(17), E836E842. http://dx.doi.org/10.1097/BRS.0b013e3181d79f02.
Dolphens, M., Cagnie, B., Vleeming, A., Vanderstraeten, G., Coorevits, P., & Danneels, L. (2012). A clinical postural model of sagittal alignment in young adolescents
before age at peak height velocity. European Spine Journal, 21(11), 21882197. http://dx.doi.org/10.1007/s00586-012-2416-7.
Gill, N. W., Teyhen, D. S., & Lee, I. E. (2007). Improved contraction of the transversus abdominis immediately following spinal manipulation: A case study using real-
time ultrasound imaging. Manual Therapy, 12(3), 280285. http://dx.doi.org/10.1016/j.math.2006.06.014.
Granata, K. P., Lee, P. E., & Franklin, T. C. (2005). Co-contraction recruitment and spinal load during isometric trunk exion and extension. Clinical Biomechanics,
20(10), 10291037. http://dx.doi.org/10.1016/j.clinbiomech.2005.07.006.
Graves, J. E., Webb, D. C., Pollock, M. L., Matkozich, J., Leggett, S. H., Carpenter, D. M., ... Cirulli, J. (1994). Pelvic stabilization during resistance training: Its eect on
the development of lumbar extension strength. Archives of Physical Medicine and Rehabilitation, 75(2), 210215.
Hides, J. A., Richardson, C. A., & Jull, G. A. (1998). Use of real-time ultrasound imaging for feedback in rehabilitation. Manual Therapy, 3(3), 125131. http://dx.doi.
org/10.1016/S1356-689X(98)80002-7.
Hodges, P. W. (1999). Is there a role for transversus abdominis in lumbo-pelvic stability? Manual Therapy, 4(2), 7486. http://dx.doi.org/10.1054/math.1999.0169.
Hodges, P. W. (2003). Core stability exercise in chronic low back pain. Orthopedic Clinics of North America, 34(2), 245254. http://dx.doi.org/10.1016/S0030-
5898(03)00003-8.
Jull, G. A., & Richardson, C. (1994). Rehabilitation of active stabilization of the lumbar spine. In L. T. Twomey, & J. R. Taylor (Eds.), Physical therapy of the low back
(pp. 251283). (2nd ed.). New York: Churchill Livingstone.
Kankaanpaa, M., Laaksonen, D., Taimela, S., Kokko, S. M., Airaksinen, O., & Hanninen, O. (1998). Age, sex, and body mass index as determinants of back and hip
extensor fatigue in the isometric Sorensen back endurance test. Archives of Physical Medicine and Rehabilitation, 79(9), 10691075. http://dx.doi.org/10.1016/
S0003-9993(98)90173-3.
Kavcic, N., Grenier, S., & McGill, S. M. (2004). Determining the stabilizing role of individual torso muscles during rehabilitation exercises. Spine, 29(11), 12541265.
Kendall, H. O., Kendall, F. P., & Wadswoth, G. E. (1971). Muscle testing and function(2nd ed.). Baltimore: Williams and Wilkins.
MacDonald, D. A., Moseley, G. L., & Hodges, P. W. (2006). The lumbar multidus: Does the evidence support clinical beliefs? Manual Therapy, 11(4), 254263. http://
dx.doi.org/10.1016/j.math.2006.02.004.
Mayer, J. M., Graves, J. E., Robertson, V. L., Pierra, E. A., Verna, J. L., & Ploutz-Snyder, L. L. (1999). Electromyographic activity of the lumbar extensor muscles: Eect
of angle and hand position during Roman chair exercise. Archives of Physical Medicine and Rehabilitation, 80(7), 751755.
Mayer, J., Mooney, V., & Dagenais, S. (2008). Evidence-informed management of chronic low back pain with lumbar extensor strengthening exercises. Spine Journal,
8(1), 96113. http://dx.doi.org/10.1016/j.spinee.2007.09.008.
Mayer, J. M., Udermann, B. E., Graves, J. E., & Ploutz-Snyder, L. L. (2003). Eect of roman chair exercise training on the development of lumbar extension strength.
Journal of Strength and Conditioning Research, 17(2), 356361.
Mayer, J. M., Verna, J. L., Manini, T. M., Mooney, V., & Graves, J. E. (2002). Electromyographic activity of the trunk extensor muscles: Eect of varying hip position
and lumbar posture during Roman chair exercise. Archives of Physical Medicine and Rehabilitation, 83(11), 15431546.
McGill, S. (2002). Low back disorders: Evidence-based prevention and rehabilitation(3rd ed.). .
OSullivan, P. B. (2000). Lumbar segmental instability: Clinical presentation and specic stabilizing exercise management. Manual Therapy, 5(1), 212. http://dx.doi.

32
J. Van Oosterwijck et al. Human Movement Science 54 (2017) 2433

org/10.1054/math.1999.0213.
Oh, J. S., Cynn, H. S., Won, J. H., Kwon, O. Y., & Yi, C. H. (2007). Eects of performing an abdominal drawing-in maneuver during prone hip extension exercises on hip
and back extensor muscle activity and amount of anterior pelvic tilt. Journal of Orthopaedic and Sports Physical Therapy, 37(6), 320324. http://dx.doi.org/10.
2519/jospt.2007.2435.
Panjabi, M. M. (2003). Clinical spinal instability and low back pain. Journal of Electromyography and Kinesiology, 13(4), 371379.
Plamondon, A., Serresse, O., Boyd, K., Ladouceur, D., & Desjardins, P. (2002). Estimated moments at L5/S1 level and muscular activation of back extensors for six
prone back extension exercises in healthy individuals. Scandinavian Journal of Medicine and Science in Sports, 12(2), 8189. http://dx.doi.org/10.1034/j.1600-
0838.2002.120204.x.
Pollock, M. L., Leggett, S. H., Graves, J. E., Jones, A., Fulton, M., & Cirulli, J. (1989). Eect of resistance training on lumbar extension strength. American Journal of
Sports Medicine, 17(5), 624629.
Richardson, C. A., & Jull, G. A. (1995). Muscle control-pain control. What exercises would you prescribe? Manual Therapy, 1(1), 210. http://dx.doi.org/10.1054/
math.1995.0243.
Sparto, P. J., & Parnianpour, M. (1998). Estimation of trunk muscle forces and spinal loads during fatiguing repetitive trunk exertions. Spine, 23(23), 25632573.
http://dx.doi.org/10.1097/00007632-199812010-00011.
Steele, J., Bruce-Low, S., & Smith, D. (2015). A review of the specicity of exercises designed for conditioning the lumbar extensors. British Journal of Sports Medicine,
49(5), 291297. http://dx.doi.org/10.1136/bjsports-2013-092197.
Stevens, V. K., Coorevits, P. L., Bouche, K. G., Mahieu, N. N., Vanderstraeten, G. G., & Danneels, L. A. (2007). The inuence of specic training on trunk muscle
recruitment patterns in healthy subjects during stabilization exercises. Manual Therapy, 12(3), 271279. http://dx.doi.org/10.1016/j.math.2006.07.009.
Vera-Garcia, F. J., Elvira, J. L., Brown, S. H., & McGill, S. M. (2007). Eects of abdominal stabilization maneuvers on the control of spine motion and stability against
sudden trunk perturbations. Journal of Electromyography and Kinesiology, 17(5), 556567. http://dx.doi.org/10.1016/j.jelekin.2006.07.004.
Verna, J. L., Mayer, J. M., Mooney, V., Pierra, E. A., Robertson, V. L., & Graves, J. E. (2002). Back extension endurance and strength: The eect of variable-angle roman
chair exercise training. Spine, 27(16), 17721777.
Vleeming, A., Pool-Goudzwaard, A. L., Stoeckart, R., van Wingerden, J. P., & Snijders, C. J. (1995). The posterior layer of the thoracolumbar fascia. Its function in load
transfer from spine to legs. Spine, 20(7), 753758.
Watanabe, S., Eguchi, A., Kobara, K., & Ishida, H. (2007). Inuence of trunk muscle co-contraction on spinal curvature during sitting for desk work. Electromyography
and Clinical Neurophysiology, 47(6), 273278.
Wilke, H. J., Wolf, S., Claes, L. E., Arand, M., & Wiesend, A. (1995). Stability increase of the lumbar spine with dierent muscle groups A biomechanical in-vitro
study. Spine, 20(2), 192198. http://dx.doi.org/10.1097/00007632-199501150-00011.

33