Content uploaded by Pablo B. Costa
All content in this area was uploaded by Pablo B. Costa on Oct 21, 2016
Content may be subject to copyright.
Journal of Exercise Physiologyonline
Volume 16 Number 3
Tommy Boone, PhD, MBA
Todd Astorino, PhD
Julien Baker, PhD
Steve Brock, PhD
Lance Dalleck, PhD
Eric Goulet, PhD
Robert Gotshall, PhD
Alexander Hutchison, PhD
M. Knight-Maloney, PhD
Len Kravitz, PhD
James Laskin, PhD
Yit Aun Lim, PhD
Lonnie Lowery, PhD
Derek Marks, PhD
Cristine Mermier, PhD
Robert Robergs, PhD
Chantal Vella, PhD
Dale Wagner, PhD
Frank Wyatt, PhD
Ben Zhou, PhD
Official Research Journal
of the American Society of
Official Research Journal of
the American Society of
Electromyographic Activity of Lower Body Muscles
during the Deadlift and Stiff-Legged Deadlift
Ewertton Souza Bezerra1, Roberto Simão2, Steven J Fleck3, Gabriel
Paz2, Marianna Maia2, Pablo B. Costa4, Alberto Carlos Amadio5,
Humberto Miranda2, Julio Cerca Serrão5
1Federal University of Amazonas, AM, Brazil; 2Federal University of
Rio de Janeiro, RJ, Brazil; 3Colorado College, CO, USA, 4California
State University, CA, USA, 5São Paulo University, SP, Brazil
Bezerra ES, Simão, R, Fleck SJ, Paz G, Maia M, Costa PB,
Amadio AC, Miranda H, Serrão JC. Electromyographic Activity of
Lower Body Muscles during the Deadlift and Stiff-Legged Deadlift.
JEPonline 2013;16(3):30-39. The purpose of this study was to
analyze eletromyographic (EMG) signal of biceps femoris (BF), vastus
lateralis (VL), lumbar multifidus (LM), anterior tibialis (AT), and medial
gastrocnemius (MG) during the deadlift (DL) and stiff-legged deadlift
(SLDL). Fourteen men (26.71 ± 4.99 yrs; body mass 88.42 ± 12.39
kg; 177.71 ± 8.86 cm) voluntarily participated in this study. The data
were obtained on three non-consecutive days separated by 48 hrs. In
the first day, anthropometric measures and the repetition maximum
testing (1 RM) for both exercises were applied in a counter-balanced
cross-over design. On the second day, the 1 RM was re-tested. On
the third day, both exercises were performed at 70% of 1 RM and the
EMG data were collected. Parameters related to the RMS during the
movement, temporal activation patterns, and relative times of
activation were analyzed for each muscle. The maximum activation
level for VL during the DL (128.3 ± 33.9% of the EMG peak average)
was significantly different (P = 0.027) from the SLDL (101.1 ± 14% of
the EMG peak average). These findings should be useful when
emphasizing different muscle groups in a resistance training program
Key Words: Resistance Training, Electromyography, Muscle Strength
The deadlift (DL) and variations are usually prescribed by strength and conditioning professionals to
strengthen the legs, hips, back, and torso musculature (6). The traditional DL begins with the knees
flexed in a squat type position. The elbows are extended and an alternating handgrip is used to grip
the bar, which is positioned over the metatarsal region of the lifter’s feet. During the concentric
exercise movement, the bar is raised from the floor to a mid-thigh position by extending the hip and
knee joints (8). In the stiff-legged deadlift (SLDL), the concentric phase begins with the knees almost
completely straight and the bar is moved from the floor to a mid-thigh position mainly by hip extension
keeping the knees slightly bent throughout the exercise movement.
Studies have compared the DL and the Sumo style deadlift using 3D kinematic analysis (10) and 2D
kinematic analysis (18). Kinematic analysis has also been used to compare DL technique of skilled
and unskilled lifters (4). Although several studies have examined the DL, only a few researchers have
investigated muscle activation during this exercise (5-8). The DL and Sumo technique have been
compared using electromyography (EMG) analysis. The data indicate that the vastus lateralis, vastus
medialis, gastrocnemius (medial head), and tibialis anterior showed greater muscle activation during
the Sumo style compared to the DL (9).
The SLDL has been compared to the leg curl (LC) and back squat (BS) using EMG techniques (19).
The results indicate that greater muscle activation of the biceps femoris (long head) and the
semitendinosus muscles takes place in each of the exercises during the concentric phase compared
to the eccentric phase. There were differences among the three exercises, with the LC and the SLDL
demonstrating greater biceps femoris and semitendinosus muscle activation vs. the BS exercise (19).
However, there is a lack of evidence to compare muscle activation between the DL and SLDL. Such
information may help coaches and strength and conditioning practitioners to optimize the resistance
training prescription, and also specify the performance of a target muscle during the execution of an
exercise. Thus, the purpose of this study was to analyze the EMG signal of biceps femoris (BF),
vastus lateralis (VL), lumbar multifidus (LM), anterior tibialis (AT), and medial gastrocnemius (MG)
during the DL and SLDL.
Fourteen men (26.71 ± 4.99 yrs; 88.42 ± 12.39 kg; 177.71 ± 8.86 cm; biacromial diameter 42.44 ±
2.46 cm; and bi-trochanteric diameter 44.54 ± 5.44 cm) voluntarily participated in the study. All
subjects had at least 2 yrs of recreational resistance training experience, no current injury to the lower
extremities, and experience in both the DL and the SLDL resistance exercises. Following an
explanation of the experimental procedures, the subjects read and signed an informed consent form.
This study was approved by the research ethics committee.
To investigate muscle activation of selected lower body muscles during the DL and SLDL, data were
collected on three nonconsecutive test days. Forty-eight hours was chosen as the time period
between the 3 tests sessions as this is the minimum rest period needed to recovery between one
repetition maximum (1 RM) attempt (17). On the first test day, anthropometric measurements and the
1 RM for both exercises were determined in a counter-balanced cross-over design. On the second
test session, the 1 RM was re-tested. On third test session, both exercises were performed with 70%
of 1 RM and EMG data were measured for the BF, VL, LM, AT, and MG. Three repetitions using 70%
of the 1 RM was used as the percentage of 1 RM during collection of EMG data because it is often
used when performing resistance training (8,19). On third test session, 20 min of rest were provided
between the exercises (which were performed in a crossover design manner).
1 RM Test
The mass of all weight plates and bar (Buick®, São Paulo, SP, Brazil) used for measuring 1 RM were
determined with a precision scale. The data were assessed on two non-consecutive days, separated
by 48 hrs in a counter-balanced cross-over design. To minimize possible errors in the 1 RM tests, the
following strategies were adopted: (a) all subjects received standard instructions before testing on the
general routine of data assessment and the exercise technique of each exercise; (b) the exercise
technique of subjects during all testing sessions was monitored and corrected as needed; and (c) all
subjects were given verbal encouragement during the tests. Each subject’s 1 RM was determined
with a maximum of five 1 RM attempts for each exercise and 3 to 5 min rest intervals between
attempts. After the 1 RM for either the DL or SLDL was determined, a 10 min rest period was
provided before the first 1 RM for the second exercise was performed. Standard exercise techniques
were followed for both exercises. No pause was allowed between the eccentric and concentric
phases of a repetition. In addition, for a repetition to be successful, a complete range of motion as is
normally defined for the exercise had to be completed.
Maximum 1 RM tests were determined on 2 d separated by a 48-hr interval in order to determine test-
retest reliability. The subjects were not allowed to perform any exercise other than normal daily
activity during the period between the testing sessions. Excellent day-to-day reliability for each 1 RM
exercise was shown by this protocol. The 1 RM testing on the two occasions showed intraclass
correlation coefficients of r = 0.96 and r = 0.94 (P<0.05) for the DL and SLDL, respectively.
Additionally, the t tests revealed no significant difference between the 1 RM tests for either of the
Characterization of the Movements Analyzed
The DL can be characterized with the barbell initially on the floor. The subject starts the exercise
movement with ~90° knee angle with the thigh parallel to the floor. The bar is grasped with an
alternating handgrip. The hips are flexed with the torso close to 45° from vertical with the scapulae
partially abducted. The hands are placed on the bar at approximately biacromial breadth apart.
During the concentric phase of the movement, the bar passes the shins while the hips and knees
extend. The trunk is raised to an upright standing position while the scapulae are adducted. The
concentric phase is complete once the upright position is achieved. The eccentric phase is performed
by returning the bar to the floor with all joint movements performed in reverse order. In the SLDL start
position the barbell is on the floor, the feet are spread to approximately bitrochanteric width, knees
are slightly flexed, shoulders are in a neutral position, scapula adducted, and hands are holding the
bar with an alternating grip at approximately biacromial breadth. During the concentric phase, the bar
passes the shins, while the hips extend, raising the trunk to an upright standing position while
extending the shoulders and adducting the scapulae. The concentric movement is completed once
the upright position is achieved. The eccentric phase is performed by returning the bar to the start
position with all joint movements performed in the reverse manner compared to the concentric phase.
Electromyographic and Kinematic Data
To examine muscle activity, surface electromyography signals were collected from the muscles to be
analyzed (EMG 1000, Lynx Inc. São Paulo, São Paulo, Brazil). Pre-amplified active electrodes, with a
20 times gain, band pass up to 4 KHz set on a polyurethane structure with two silver plates positioned
10 mm apart were used for all analyses of the muscles examined. Before the application of the
electrodes, the skin was shaved, abraded, and cleansed with alcohol. The electrodes were then
placed between the motor point and the distal tendon in each muscle studied in the direction of the
muscle’s fibers (14).
For the assessment of the kinematic data, spherical plastic markers (2.5 cm in diameter) covered with
reflective tape were positioned over the following bony landmarks: lateral malleolus of the right ankle,
proximal upper edge of the lateral tibial plateau of the right knee, greater trochanter of the right femur,
and lateral acromion process of the right shoulder. In addition, a piece of reflective tape (1 cm2) was
positioned on the third metatarsal head of the right foot. Data were collected at 30 HZ using a video
camera (SONY®, São Paulo, São Paulo, Brazil) during the performance of each exercise. The images
were analyzed using a Vicon Motion Analysis System (Vicon Corporation, Los Angeles, CA, USA).
The beginning and ending position of the eccentric and concentric phase for both exercises was
determined with the hip flexed to its greatest extent during the exercise movement. Hip angle was
defined as 0° in the fully flexed hip position.
The EMG of the BF, VL, LM, AT, and MG muscles were analyzed during the DL and the SLDL. These
muscles were chosen because they are superficial and biomechanically involved in the exercise
movements (1,8,19). Although the LM is regarded as a deep muscle in the lumbar region, it is slightly
more superficial and can be located by palpation of the spinous process of the 5th vertebra . All
EMG signals were recorded at a sample frequency of 1000 Hz. For each muscle, temporal activation
patterns, muscle activation level, and contraction time were analyzed. Temporal activation patterns
were obtained using a linear wrap trace from the results of the EMG signal of each muscle after
normalization. Muscle activation level was obtained using the Root Mean Square (RMS) measure.
The RMS represents the greatest value obtained during the movement . For the relative time of
activation, a time interval was determined for each muscle in which muscular activity was maintained
at a level over 50% of the peak EMG signal during the movement cycle including both eccentric and
concentric phases. The relative time of activation was expressed as a percentage representing how
long the EMG was above at least 50% of the peak EMG during the temporal activation patterns
(movement cycle) of each exercise.
The original signal of each muscle was smoothed using a butterworth filter (second order butterworth
low pass filter with a frequency of 500 Hz). After filtering, normalization of the EMG signal was
performed using the peak average for each muscle in three repetitions of the DL or SLDL. Briefly, the
maximum EMG value for each muscle was determined for each movement cycle, an average was
calculated, and then a peak muscle activation value for each subject was calculated. This value was
used as a reference value for 100% muscle activation. Thus, the entire signal was normalized using
this value that allowed for comparison among the different muscle groups, exercises, and subjects.
After normalization, the starting and ending points for each of the three repetitions were determined
and then, subsequently, the average EMG calculated. The muscle activation intensity value
represented by the muscular intensity estimation (RMS) was obtained from the original signal. For
normalization, RMS, and the relative time of activation of each muscle, an EMGONIO1 routine was
used (MATLAB 6.0 software; MathworksInc, Natick, Massachusetts, USA). The ORIGIN 6.0 software
(Microcal Software Inc, Massachusetts, USA) was used for graphic representations.
The data were descriptively analyzed in which the mean and standard deviation for each dependent
variable were calculated. Data normality was checked using the Shapiro-Wilk test. The t test for
paired data was used to determine significant differences in maximum RMS and relative activation
time of muscles between the DL and SLDL exercises. The alpha level was set at P<0.05 for all
analyses. All statistical analyses were performed using SPSS 20.0 software for Windows (SPSS Inc.,
Chicago, IL, USA).
Comparisons of RMS revealed significant differences (P<0.05) between the DL and the SLDL for the
VL and MG muscles. However, no differences were found for the BF, LM, and AT muscles (Table 1).
Relative time of activation between the DL and the SLDL showed significant differences for the VL
only (P<0.05). No significant differences were found for the others muscles (BF, LM, AT, and MG)
Table 1. Mean ±SD of EMG Variables Analyses.
RMS: Percentage normalization of the EMG peak average; EMG/TIME: relative time activation as
percentage of movement cycle above 50% of RMS; BF: biceps femoris; VL: vastuslateralis; LM:
lumbar multifidus; MG: medial gastrocnemius; AT: anterior tibialis; DL: Deadlift; SLDL: Stiff-legged
deadlift. *Significant differences between the DL and the SLDL (P<0.05).
The temporal activation between the DL and SLDL for the BF, LM, MG, and AT muscles
demonstrated similar patterns. However, the VL muscle showed a different activation pattern. During
the DL, the VL showed higher activation at the beginning of the ascent and ending of the descent
phases. In contrast, during the SLDL, the VL showed its highest activation at ~60° of ascent phase
Figure 1. Mean ±SD of the RMS for the Analyzed Muscles. BF: biceps femoris; VL: vastus
lateralis; LM: lumbar multifidus; MG: medial gastrocnemius; AT: anterior tibialis; DL: Deadlift; SLDL:
Stiff-legged deadlift. *Significant differences between the DL and the SLDL (P<0.05).
The key findings of this study were the differences in RMS for the VL and MG muscles between the
DL and SLDL exercises. In addition, the VL demonstrated a higher relative activation time (i.e., time
above 50% RMS) than the other muscles during the DL. The VL muscle had a peak of activity during
the first 20° of the ascent phase due to its role in knee extension and indirectly hip extension (in that
the movement is a closed kinetic chain exercise). This finding is consistent with results from the
parallel squat (2), mini-squat (6), leg press, squat, and deadlift (8,11); all demonstrated an increased
level of activity in the VL muscle in the beginning of the concentric phase.
The changes in the VL muscle EMG potential during the DL may be associated with its role during
knee joint extension (ascent phase) and flexion (descent phase) of the DL movement and the
concomitant decrease of motor units used at the same resistance during the descent phase of the
movement. The percentage of activity in the data presented in this study is greater than reported by
Gullett and colleagues (12), who analyzed the differences in EMG activity of lower limb muscles as a
function of bar position (front or back to the trunk) in the squat exercise. In the Gullet et al. study (12),
the VL muscle activation was 60% of maximal voluntary isometric contraction (MVIC). However, it is
noteworthy that the forms of normalization of EMG signals were different, since the aforementioned
study used MVIC.
Proposed Mechanisms to Explain the EMG Activity
In the descent phase of the DL (between -40° to 0°), the RMS signal for the VL also increased but not
to the extent as it did in the ascent phase. The gradual increase of this muscle’s activity is due to the
changing need to exert more strength by the time the knee flexion becomes more acute during the
descent phase of the DL movement. Similar findings were noticed by Escamilla et al. (8,9) and Gullett
et al. (12) during the squat and the conventional style DL. The VL RMS during the SLDL showed a
constant activity during the descent phase of the movement (between -80° and 0°). However, it
shows an activity peak between 40° and 60° in the ascent phase of the SLDL movement, possibly
due to the factors of needing to increase the muscle activity at this knee angle of the movement and
to counteract the co-contraction of the long head of the BF that has an increased muscle activity
during the same movement phase. The increased VL RMS may also be essential for knee joint
stabilization, but more research is needed to confirm this point.
The MG activity during the DL and SLDL exercises increased slightly during the ascent phase. The
increased MG activity during the ascent phase could be explained by an increased plantar flexion
moment during this phase as shown by Escamilla et al. (2000) for the conventional style DL. Da Silva
et al. (5) found that the gastrocnemius muscle is more activated during leg press exercise with low
foot placement and 45° (near 80% of MVIC). This may have happened due to the increased plantar
flexion movement in the two exercises. However, the RMS showed significant differences with the
SLDL that may be attributed to the initial imbalance of the body caused by not bending the knee,
which required a greater involvement of stabilizing muscles with MG.
The BF behavior during the DL showed a muscle activity peak in the beginning of the ascent phase
between 20° and 40° of hip extension followed by a decrease in the last degrees (-20°) of movement
during the descent phase (hip flexion). However, this factor may not be attributed to a reduction of
muscle activation in the descent phase, since as previously shown, at the same force output eccentric
actions compared to concentric actions involve a smaller activation of motor units and, therefore, a
decrease in the EMG amplitude. The RMS average observed for BF during the initial ascent phase of
the DL is due to the role it muscle plays during hip joint extension. These findings were similar to
Escamilla and colleagues (9) who observed that for DL the movement during the execution of the
technique variations (e.g. sumo and conventional) and during movements such as the squat showed
higher muscle activity of BF during hip joint extension (19). During the SLDL, the BF showed the
same behavior as during the DL, with a peak of muscle activity in the ascent phase between 20° and
40° during hip extension. Similar results were observed for DL, it was showed also a decreasing in BF
activity in the descent phase (-80° to -20° of the movement) while the hip joint flexion was performed.
The behavior of the BF during the DL and the SLDL in ascent and descent phases were similar and
may be attributed to its role during both flexion (descent phase) and extension (ascent phase) of the
hip joint during the exercise movement.
Synergist Muscle Behavior during the Resistance Exercises
The findings of this study disagree with the findings of the Yamishita (20) study, who suggested that
agonist-antagonist concurrent activation may generate an inhibition in the activation of the muscles.
This hypothesis indicated that the single-joint exercise such as the SLDL may provide an increase in
hamstring muscle activity. Wright et al. (19) also observed that squat promoted less muscle activation
of BF and semitendinosus when compared to LC and SLDL. These studies are also in contrary to the
results found by Luttgens and Wells (16), who did not observed any significant difference in the
hamstring activity during hip and knee extension.
The LM activation during the DL and the SLDL may be characterized as a normal pattern of muscle
activity, since little variation was showed during the ascent and descent phases of the movements.
Similar results were observed by Escamilla et al. (9) for the sumo and conventional style deadlift.
Despite kinematic differences caused by an increased forward flexion of the trunk during the
conventional DL, variations regarding muscle activity of the LM did not occur, indicating that the
action of the LM does not change significantly during the DL. Hamlyn et al. (13) found that the erector
spine muscles in the lumbosacral region showed an increase of 34.5% in muscle activity while
performing a squat compared to the DL.
Although in the current study, there were no significant differences between the two movements. It is
noteworthy that the muscle activity was high (i.e., above 80% of MVIC) (3), which may indicate a high
stabilizing effect. The AT muscle activity remained constant at a low intensity during the DL and
SLDL. Relative time of activation showed high standard deviation values indicating a high variation
among subjects for both exercises, which indicates individual technique may be a factor that affects
muscle activity during the DL and SLDL.
Limitations and Implications
One of the limitations of the current study is that only one set of each exercise was performed. A
traditional resistance training session is composed by multiple sets and exercises. Also, the relative
time of activation may be influenced by the subjects’ resistance training experience. However, in the
current study all subjects had at least 2 yrs of resistance training experience with the DL and SLDL
exercises. This suggests that other factors may be responsible for the relative time of activation such
as the number of sets, loads, and velocity. The BF and AT showed a similar mean relative time of
activation for both the DL and SLDL. The VL and MG had higher mean relative time of activation
during the DL while the LM had a higher mean during the SLDL. The lack of agreement in the
scientific literature related to muscle relative time of activation in dynamic movements (including the
DL and SLDL) is a limitation, which make the comparison between the exercises difficult. Thus, in
future studies other variables should be evaluated such as the influence of number of sets, exercise
velocity, load intensity and muscle groups in the muscle activation and relative time activation during
back squat exercises.
The EMG data indicate that the DL is more effective for activating the VL muscle than the SLDL.
However, the MG muscle showed higher muscle activation during the SLDL than the DL. These
findings should be useful when emphasizing different muscle groups in a resistance training program.
Dr. Humberto Miranda is grateful to the Research and Development Foundation of Rio de Janeiro
Address for correspondence: Miranda H, PhD, School of Physical Education and Sports, Federal
University of Rio de Janeiro, Rio de Janeiro, RJ, Brazil, ZIP CODE: 21941-590, Phone:+55 21 2287-
9329, Email: email@example.com.
1. Bojadsen T, Silva E, Rodrigues A, et al. Comparative study of multifidi in lumbar and thoracic
spine. J Eletromyogr Kinesiol. 2000;10:143-149.
2. Boyden G, Kingman J, Dyson R. A comparison of quadriceps eletromyographic activity with
the position of the foot during the parallel squat. J Strength Cond Res. 2000;14:379-382.
3. Brennecke A, Guimarães Leone R, et al. Neuromuscular activity during bench press exercise
performed with and without the preexhaustion method. J Strength Cond Res. 2009;23:1933-
4. Brown E, Abani K. Kinematics and kinetics of the deadlift in adolescent power lifters. Med Sci
Sports Exerc. 1985;17:554-566.
5. Da silva E, Brentano M, Cadore E, et al. Analysis of muscle activation during different leg
press exercises at submaximum effort levels. J Strength and Cond Res. 2008;22:1059-1065.
6. Earl J, Schmitz R, Arnold B. Activation of de VMO and VL during dynamic mini-squat exercises
with and without isometric hip adduction. J Eletromyogr Kinesiol. 2011;11:59-67.
7. Ebben W, Leigh D. The role of the back squat as hamstring training stimulus. Nat Strength
and Cond Association. 2000;22:15-17.
8. Escamilla R, Fleisig G, Zheng N, et al. Biomechanics of the knee during closed kinetic chain
and open kinetic chain exercises. Med Sci Sports Exerc. 1998;30:556-569.
9. Escamilla R, Francisco A, Kayes A, et al. An electromyographic analysis of sumo and
conventional style deadlifts. Med Sci Sports Exerc. 2002;34:682-688.
10. Escamilla R, Francisco A, Fleisig GA. Three dimensional biomechanical analyses of sumo
and conventional styles deadlifts. Med Sci Sports Exerc. 2000;32:265-1275.
11. Escamilla R, Fleisig G, Zheng N, et al. Effects of technique variations on knee biomechanics
during the squat and leg press. Med Sci Sports Exerc. 2001;33:1552-1566.
12. Gullett J, Tillman M, Gutierrez G, et al. A biomechanical comparison of back and front squats
in healthy trained individuals. J Strength and Cond Res. 2008;23:284-292.
13. Hamlyn N, Behm D, Young W. Trunk muscle activation during dynamic weight-training
exercises and isometric instability activities. J Strength and Cond Res. 2007;21:1108-1112.
14. Hermens H, Freriks B, Diseelhorst-Klug D, et al. Development of recommendations for SEMG
sensors and sensor placement procedures. J Eletromyogr Kinesiol. 2000;10:361-374.
15. Hoehn L, Niven I. Averages on the move. Math Mag. 1985;58:151-156.
16. Luttgens K, Wells K. Kinesiology: Scientific Basis of Human Motion. (7th Edition).
Philadelphia: Saunders College Publishing, 1982
17. Matuszak M, Fry A, Weiss L, et al. Effect of rest interval length on repeated 1 repetition
maximum back squats. J Strength and Cond Res. 2003;17:634-637.
18. Mcguigan M, Wilson B. Biomechanical analysis of the deadlift. J Strength and Cond Res.
19. Wright G, Delong T, Gehlsen G. Electromyographic activity of the hamstrings during
performance of the leg curl, stiff-leg deadlift and back squat movements. J Strength and
Cond Res. 1999;13:168-174.
20. Yamishita N. EMG activities in mono- and bi-articular thigh muscles in combined hip and knee
extension. Eur J Appl Physiol. 1988;58:274–277.
The opinions expressed in JEPonline are those of the authors and are not attributable to JEPonline,
the editorial staff or the ASEP organization.