Journal of Electromyography and Kinesiology 23 (2013) 696–703

Contents lists available at SciVerse ScienceDirect

Journal of Electromyography and Kinesiology

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

Knee flexor strength and bicep femoris electromyographical activity is lower

in previously strained hamstrings
David A. Opar a,⇑, Morgan D. Williams b, Ryan G. Timmins a, Nuala M. Dear a, Anthony J. Shield a
a
School of Exercise and Nutrition Science and Institute of Health and Biomedical Innovation, Queensland University of Technology, Brisbane, Australia
b
Faculty of Health, Sport and Science, University of Glamorgan, Pontypridd, Wales, United Kingdom

a r t i c l e i n f o a b s t r a c t

Article history: The aim of this study was to determine if athletes with a history of hamstring strain injury display lower
Received 24 April 2012 levels of surface EMG (sEMG) activity and median power frequency in the previously injured hamstring
Received in revised form 22 October 2012 muscle during maximal voluntary contractions. Recreational athletes were recruited, 13 with a history of
Accepted 6 November 2012
unilateral hamstring strain injury and 15 without prior injury. All athletes undertook isokinetic
dynamometry testing of the knee flexors and sEMG assessment of the biceps femoris long head (BF)
and medial hamstrings (MHs) during concentric and eccentric contractions at ±180 and ±60° s 1. The
Keywords:
Muscle
knee flexors on the previously injured limb were weaker at all contraction speeds compared to the unin-
Strain injury jured limb (+180° s 1 p = 0.0036; +60° s 1 p = 0.0013; 60° s 1 p = 0.0007; 180° s 1 p = 0.0007) whilst
Surface electromyography sEMG activity was only lower in the BF during eccentric contractions ( 60° s 1 p = 0.0025; 180° s 1
Maladaptation p = 0.0003). There were no between limb differences in MH sEMG activity or median power frequency
Neuromuscular from either BF or MH in the injured group. The uninjured group showed no between limb differences
in any of the tested variables. Secondary analysis comparing the between limb difference in the injured
and the uninjured groups, confirmed that previously injured hamstrings were mostly weaker (+180° s 1
p = 0.2208; +60° s 1 p = 0.0379; 60° 1 p = 0.0312; 180° s 1 p = 0.0110) and that deficits in sEMG were
confined to the BF during eccentric contractions ( 60° s 1 p = 0.0542; 180° s 1 p = 0.0473). Previously
injured hamstrings were weaker and BF sEMG activity was lower than the contralateral uninjured
hamstring. This has implications for hamstring strain injury prevention and rehabilitation which should
consider altered neural function following hamstring strain injury.
Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction retrospective nature of these findings cannot be taken to suggest

that hamstring injury has resulted in these deficits, it is agreed that
Hamstring strain injuries, characterised by acute pain in the hamstring strain injury does lead to maladaptation (Opar et al.,
posterior thigh and disruption of hamstring muscle fibres, are the 2012). Importantly, prospective studies in both sprinters and soc-
primary injury sustained in a number of sports (Orchard and cer players have identified eccentric knee flexor strength deficits
Seward, 2010; Woods et al., 2004; Drezner et al., 2005) and as elevating hamstring strain injury risk (Croisier et al., 2008;
re-injury rates are also high (Orchard and Seward, 2010). The high Sugiura et al., 2008). These findings suggest the importance of
rate of injury and re-injury, combined with the fact that a previous eccentric strength for the prevention of hamstring strain injury
hamstring strain is the most significant risk factor for future injury and that eccentric weakness should be corrected following injury
(Arnason et al., 2004), suggests that our understanding of the neu- to reduce the risk of a recurrence. However a clear understanding
romuscular maladaptations that occur following hamstring strain of the mechanisms underpinning the decline in eccentric strength
requires further attention. following hamstring strain injury is required in order to develop
Previous hamstring strain injury has been associated with more appropriate exercise interventions. Whilst evidence does
between-limb differences in eccentric strength that is typically exist of persistent atrophy of biceps femoris long head (BF) up to
greater than concentric strength deficits (Croisier et al., 2002; Lee 23 months following grade I and II hamstring strain injuries (Silder
et al., 2009). Furthermore these deficits in eccentric strength are et al., 2008) this muscular maladaptation does not explain why the
still present despite athletes returning to full training and decline in hamstring strength appears to be greater in eccentric
competition (Croisier et al., 2002; Lee et al., 2009). Whilst the actions (Croisier et al., 2002; Lee et al., 2009).
Surprisingly the impact of strain injuries on the neural function
⇑ Corresponding author. Tel.: +61 7 31385865; fax: +61 7 3138 3980. of the involved musculature has been largely overlooked. Ham-
E-mail address: d.opar@qut.edu.au (D.A. Opar). string strain injury has been reported to result in acute (Verrall

1050-6411/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jelekin.2012.11.004
 D.A. Opar et al. / Journal of Electromyography and Kinesiology 23 (2013) 696–703 697

et al., 2001) and chronic pain (Croisier et al., 2002; Jönhagen et al., graphical activity from the MH and BF. After preparation of the skin
1994). This muscular pain has the potential to alter central nervous via shaving, light abrasion and sterilisation, electrodes were placed
function at both the spinal and supraspinal level (Mense, 2003), on the posterior thigh half way between the ischial tuberosity and
and might therefore be expected to result in a restriction of electr- tibial epicondyles with electrodes oriented parallel to the line be-
omyographical activity and the median power frequency of this tween these two land marks, as per SENIAM guidelines (Hermens
activity during contraction. Furthermore this restriction may be et al., 2000). The reference electrode was placed on the ipsilateral
specifically confined to the muscle and contraction mode responsi- head of the fibula. Muscle bellies were identified via palpation dur-
ble for the noxious stimulus. Therefore the purpose of this study ing forceful isometric knee flexion and correct placement was con-
was to assess concentric and eccentric hamstring torque, surface firmed by observing sEMG activity during active internal and
EMG (sEMG) activity and the median power frequency of the sEMG external rotation of the flexed knee to assess cross talk between
signal of recreational athletes with and without a history of unilat- MH and BF.
eral hamstring strain injury. It was hypothesised that the previ-
ously injured hamstrings would display strength, sEMG activity 2.4. Isokinetic dynamometry
and median power frequency deficits during fast and slow eccen-
tric contractions, but not concentric contractions, compared to Assessment of concentric and eccentric knee flexor strength
the contralateral limb. Furthermore, we hypothesised that lower was performed on a Biodex Systems 3 Dynamometer (Biodex Med-
levels of sEMG activity and median power frequency would be con- ical Systems, Shirley, NY). Participants were seated on a custom
fined specifically to the previously injured hamstring muscle (i.e. pad, placed on top of the original seat, which contained two holes
BF or medial hamstrings (MHs)). It was also hypothesised that at the level of the posterior mid thigh to minimise movement arte-
the control group would display no differences in any of the afore- fact from sEMG electrodes on the dynamometer seat. The hips
mentioned variables between dominant and non-dominant limbs. were flexed at 85° from neutral with the lateral epicondyle of the
As a confirmatory secondary analysis, it was also hypothesised that femur carefully aligned with fulcrum of the dynamometer. The
the between limb differences in eccentric hamstring torque, sEMG tested leg was attached to the lever of the dynamometer via a Vel-
and median power frequency would be greater in previously in- cro strap and padded restraints were fastened across the trunk,
jured athletes compared to the control group. hips and mid thigh of the tested leg to isolate movement to the
knee joint. The range of motion was set at 5–90° of knee flexion
(0° = full knee extension) and correction for limb weight was
2. Materials and methods
performed.
Three sets of four submaximal contractions of the knee exten-
2.1. Participants
sors and flexors were performed at +240° s 1 as a warm-up to pre-
pare the participant for maximal effort in the following sets.
Twenty-eight recreationally active males participated in the
Concentric testing for both legs consisted of three sets of three con-
study, with most competing in Australian football, rugby, soccer
secutive maximum voluntary contractions (MVC) of the knee
or sprinting. Thirteen athletes (26.2 ± 5.8 years; 1.80 ± 0.04 m;
extensors and flexors at velocities of +60° s 1 and +180° s 1 with
83.0 ± 14.8 kg) had at least one unilateral hamstring strain injury
30 s rest between sets. Athletes were motivated verbally by the
(INJ) within the last 18 months and all had suffered a grade II injury
investigators to encourage maximal effort throughout the range
previously. Another 15 athletes (26.7 ± 5.8 years; 1.8 ± 0.05 m;
of motion. Eccentric testing ( 60° s 1 and 180° s 1) was identical
83.5 ± 7.9 kg) had no history of hamstring strain injury (UI). All par-
except that only eccentric contraction of the knee flexors was per-
ticipants were free of any other injury to the lower limbs and were
formed by the participant (whereby the knee joint was extended
fully active in their chosen sport at the time of testing. All testing
despite active contraction of the knee flexors) and at the comple-
procedures were approved by the University Human Research Eth-
tion of each contraction the investigators returned the lever to
ics Committee. Participants gave informed written consent prior to
the starting position. The leg and velocity testing orders were ran-
testing after having all procedures explained to them.
domised but concentric contractions were always performed be-
fore eccentric contractions. All participants were required to
2.2. Injury questionnaire attend at least one familiarisation session to ensure consistency
of MVCs and one testing session with P7 days between sessions.
Following recruitment, participants completed an injury ques-
tionnaire with their chosen practitioner (i.e. physiotherapist) 2.5. Data analysis
who had previously diagnosed and treated all the athletes ham-
string strain injuries. As per previous investigations (Sole et al., Dynamometer torque and lever position data were transferred
2011), the notes taken from clinical examination were used to de- to computer at 1 kHz and stored for later analysis. Average peak
tail the date of injury and return to pre-injured levels of training torque was defined as the mean maximal torque of the six highest
and competition, severity (grades I, II or III) (Blankenbaker and torque contractions at each velocity. Surface EMG was sampled
Tuite, 2010), location (dominant or non-dominant limb; BF or simultaneously with dynamometer data at 1 kHz through a 16-
MH head; proximal or distal) and rehabilitation details of all previ- bit PowerLab26T AD recording unit (ADInstruments, New South
ous hamstring strain injuries. Limb dominance was determined as Wales, Australia) (amplification = 1000 between 10 Hz and 1 kHz;
the preferred kicking limb. Athletes were considered to be success- common mode rejection ratio = 110 dB) and stored for later analy-
fully rehabilitated when they returned to pre-injured levels of sis where it was fourth order Butterworth filtered between 20 and
training and were available for competition (Fuller et al., 2006). 500 Hz (24 dB roll off) using MATLAB (MathWorks, Natick, Massa-
Athletes who were unable to obtain data on all prior hamstring chusetts) and then full wave rectified using the root-mean-square
strains from their practitioner were excluded from the study. method across a 100 ms window. At each velocity, sEMG data were
averaged across a knee joint ROM between 15° and 35° as this is
2.3. EMG recording where deficits in sEMG have been noted previously (Sole et al.,
2011). Data at all velocities was then normalised to the maximal
Bipolar pre-gelled Ag/AgCl sEMG electrodes (10 mm diameter, averaged sEMG amplitude recorded during MVCs at + 180° s 1
25 mm inter-electrode distance) were used to record electromyo- (Aagaard et al., 2000; Seger and Thorstensson, 1994; Westing
698 D.A. Opar et al. / Journal of Electromyography and Kinesiology 23 (2013) 696–703

et al., 1991). For this process the data was separated in tertiles Table 1
throughout the ROM (15–35°, 35–60°, 60–80°) and the tertile Hamstring strain injury information for most recent injury for athletes recruited to
the injured group.
exhibiting the highest amplitude of sEMG was used for normalisa-
tion. Median power frequency was determined from the non- Subject Time Rehabilitation Location Total
rectified sEMG signal via Fast Fourier transform with Hann since HSI duration HSIs
(months) (weeks) sustained
window function applied (Aagaard et al., 2000) across the entire
ROM using LabCart 7.3 (ADInstruments, New South Wales, 1 2 4 Dominant, proximal BF 1
2 3 4 Non-dominant, proximal BF 3
Australia) with 1 Hz frequency resolution. This resulted in 1.08 and 3 8 4 Non-dominant, distal BF 1
0.36 s time epochs for analysis of contractions at ±60 and 180° s 1 4 7 2 Non-dominant, proximal BF 2
respectively. Median power frequency was analysed over a larger 5 3 4 Dominant, proximal BF 4
ROM (15–80°) than sEMG activity to allow for a valid estimation of 6 5 2 Non-dominant, distal BF 2
7 18 4 Non-dominant, distal BF 1
frequency. Median power frequency was defined as the frequency
8 4 4 Non-dominant, proximal BF 2
at which 50% of total power was reached for each time epoch. 9 2 5 Non-dominant, proximal BF 2
10 5 3 Non-dominant, proximal BF 4
2.6. Statistical analysis 11 2 2 Dominant, proximal BF 2
12 3 6 Non-dominant, distal BF 4
13 7 3 Non-dominant, proximal BF 3
Data were analysed using JMP version 10.0 Pro Statistical Dis-
covery Software (SAS Inc). In the primary analysis, comparisons HSI, hamstring strain injury; BF, biceps femoris. All prior injuries were confined to
the same leg and muscle as most recent injury however location on muscle
were made between the injured and uninjured limbs in the INJ
(proximal or distal) differed in some instances.
group and between dominant and non-dominant limbs in the UI
group. Dependent variables were compared using one tailed paired
t tests for both groups to allow an equal likelihood for finding sig-
nificant differences between limbs (Lee et al., 2009). For the pri-
mary analysis data are presented as means and standard
deviation. Bonferroni corrections were performed to account for
four comparisons made for each dependent variable across the
velocities used, with significance set at p < 0.0125. In the confirma-
tory secondary analysis independent t tests for unequal variance
were used to compare the between limb differences of the depen-
dent variables in the INJ (uninjured limb minus injured limb) and
UI groups (dominant limb minus non-dominant limb) as assump-
tions for equal variance between groups was not met. For the sec-
ondary analysis significance was set at p < 0.05 and data are
presented as mean differences and 95% confidence intervals. To as-
sess the magnitudes of the differences for the primary and second-
ary analyses Cohen’s d was calculated to report effect size (ES).

3. Results

3.1. Participants

There was no significant difference between the UI and INJ

groups with respect to age, height or body mass. The details of
injury histories of all athletes from the INJ group can be found in
Table 1. All athletes from the INJ group reported largely standard
rehabilitation progression (i.e. (Heiderscheit et al., 2010)) guided
by their physiotherapist.

3.2. Average peak torque

There were significant differences in average peak torque be-

tween limbs in the INJ group, with the previously injured limb
weaker at all contraction modes and velocities (Fig. 1a and Table 2).
No differences in average peak torque were noted between limbs
in the UI group (Fig. 1b and Table 2). Between limb differences in
torque were significantly greater in the INJ group compared to
the UI group at all contraction modes and velocities, except for Fig. 1. Knee flexor average peak torque at four different isokinetic velocities from
concentric contractions at 180° s 1 (Table 5). the: (A) injured athletes and (B) uninjured athletes. Negative movement velocities
are indicative of eccentric contractions and positive velocities indicate concentric
contractions. Error bars display standard deviation. ⁄p < 0.0125 injured vs. unin-
3.3. sEMG activity jured limbs.

BF electromyographical activity was significantly lower in the

previously injured limb compared to the contralateral uninjured between limbs in the INJ group for MH electromyographical activ-
limb in the INJ group during eccentric contractions but not concen- ity at any contraction mode or velocity (Fig. 3a and Table 3). In the
tric contractions (Fig. 2a and Table 3). There were no differences UI group there were no differences in activation between limbs for
 D.A. Opar et al. / Journal of Electromyography and Kinesiology 23 (2013) 696–703 699

Table 2
Knee flexor torque of athletes with and without a history of unilateral hamstring strain injury during concentric and eccentric contraction.
1
Movement velocity (° s ) Injured limb Uninjured limb p ES
Injured group
+180 109.29 (±13.14) 118.64 (±12.47) 0.0036* 0.78
+60 132.00 (±21.28) 146.01 (±15.49) 0.0013* 0.70
60 166.76 (±30.19) 185.02 (±25.22) 0.0007* 0.57
180 163.82 (±30.43) 184.37 (±22.33) 0.0007* 0.74
Dominant limb Non-dominant limb p ES
Uninjured group
+180 127.13 (±22.12) 122.73 (±21.24) 0.0608 0.20
+60 154.93 (±24.27) 151.59 (±25.10) 0.1558 0.14
60 199.71 (±31.46) 198.68 (±33.30) 0.4341 0.03
180 194.84 (±25.97) 194.60 (±28.84) 0.4828 0.01

Negative movement velocities are indicative of eccentric contractions and positive velocities indicate concentric contractions. Data are
presented as mean (± standard deviation).
*
Significance was set at p < 0.0125. Cohen’s d was used to calculate effect size.

3.4. Median power frequency

One participant from the INJ group was a clear outlier (median
power frequency was more than three standard deviations above
the mean for eccentric contractions) and was removed from anal-
ysis. There were no differences in median power frequency at any
velocity between legs in the INJ group for BF or MH (Table 4). A
similar lack of differences was noted at all velocities for the UI
group for BF or MH median power frequency (Table 4). The be-
tween limb differences in median power frequency did not differ
between the INJ and UI groups at any contraction mode or velocity
(Table 5).

4. Discussion

It is accepted that a prior hamstring strain injury results in mal-

adaptation of the previously injured tissue (Opar et al., 2012).
Whilst a number of muscular maladaptations have been reported
previously (Brockett et al., 2004; Croisier et al., 2002; Lee et al.,
2009; Silder et al., 2008; Silder et al., 2010; Worrell et al., 1991),
the impact of a prior hamstring strain injury on neural function
has been scarcely examined (Sole et al., 2011). The current study
used between limb comparisons of normalised sEMG activity and
median power frequency to determine differences in neural ham-
string function between injured and uninjured limbs. This method
eliminates a number of confounding factors by ensuring that mus-
cle lengths and electrode locations are identical between trials
within and between limbs and has been used extensively to assess
relative muscle activation in maximal concentric and eccentric
contraction (Aagaard et al., 2000; Seger and Thorstensson, 1994;
Westing et al., 1991).
From the INJ group in the current study, the novel findings were
that the previously injured limb, when compared to the contralat-
eral uninjured limb displayed (1) a lower level of sEMG activity
Fig. 2. Biceps femoris long head normalised surface electromyography (sEMG) at specifically in the previously injured muscle (BF) during slow and
four different isokinetic velocities from the: (A) injured athletes and (B) uninjured fast eccentric contractions (Fig. 2a and Table 3); and (2) there
athletes. Negative movement velocities are indicative of eccentric contractions and was no difference in the median power frequency in either the pre-
positive velocities indicate concentric contractions. Error bars display standard
viously injured BF or uninjured MH (Table 4). Furthermore, lower
deviation.⁄p < 0.0125 injured vs. uninjured limbs.
levels of strength were observed across all contraction modes
and velocities in the injured limb compared to the uninjured limb
BF (Fig. 2b and Table 3) or MH (Fig. 3b and Table 3) at any in the INJ group (Fig. 1a). In contrast the UI group showed no dif-
contraction mode or velocity. Between limb differences in electr- ferences between dominant and non-dominant limbs in any of
omyographical activity were greater in the INJ group compared the tested variables indicating there is no influence of limb
to the UI group only for BF at 180° s 1 (Table 5). All other dominance (Figs. 1b, 2b, 3b; Tables 2–4). These findings were
between limb differences in electromyographical activity were mostly supported by confirmatory analysis which indicated that
similar between INJ and UI groups, although a trend existed at the between limb differences in knee flexor torque at all contrac-
60° s 1 (Table 5). tion modes and velocities, except for the fastest concentric
700 D.A. Opar et al. / Journal of Electromyography and Kinesiology 23 (2013) 696–703

Table 3
Normalised electromyographical activity of the biceps femoris long head and medial hamstrings of athletes with and without a history of unilateral hamstring strain injury during
concentric and eccentric contraction.
1
Movement velocity (°s ) Biceps femoris Medial hamstrings
Injured limb Uninjured limb p ES Injured limb Uninjured limb p ES
Injured group
a a
+180 0.96 (±0.06) 0.99 (±0.02) 0.0894 0.95 (±0.07) 0.98 (±0.06) 0.0622
+60 0.89 (±0.20) 0.93 (±0.12) 0.2255 0.18 0.91 (±0.23) 0.96 (±0.13) 0.2412 0.09
60 0.58 (±0.17) 0.71 (±0.17) 0.0025* 0.47 0.58 (±0.21) 0.64 (±0.12) 0.1296 0.06
180 0.53 (±0.20) 0.66 (±0.18) 0.0003* 0.58 0.52 (±0.22) 0.61 (±0.15) 0.0770 0.26
Biceps femoris Medial hamstrings
Dominant limb Non-dominant limb p ES Dominant limb Non-dominant limb p ES
Uninjured group
a a
+180 0.97 (±0.06) 0.99 (±0.02) 0.1602 0.94 (±0.11) 0.94 (±0.12) 0.4444
+60 0.95 (±0.16) 0.97 (±0.18) 0.2703 0.12 0.93 (±0.26) 0.97 (±0.23) 0.2890 0.16
60 0.70 (±0.21) 0.69 (±0.17) 0.4275 0.05 0.64 (±0.25) 0.67 (±0.16) 0.3077 0.14
180 0.60 (±0.26) 0.61 (±0.14) 0.4052 0.05 0.56 (±0.23) 0.59 (±0.15) 0.2538 0.15

Negative movement velocities are indicative of eccentric contractions and positive velocities indicate concentric contractions. Data are presented as mean (± standard
deviation).
*
Significance was set at p < 0.0125. Cohen’s d was used to calculate effect size (ES).
a
ES for electromyographical activity could not be calculated given the use of this data in the normalisation process.

This study is, to our knowledge, the first to identify lower levels
of sEMG activity specifically in the previously injured BF muscle
compared to a contralateral uninjured BF. Recent evidence examin-
ing a similar phenomenon did not find a muscle specific, between
limb differences in sEMG activity following a hamstring strain in-
jury (Sole et al., 2011). The discrepancies between the findings
from the current study and the previous study by Sole et al.
(2011) may be attributed to the inclusion of athletes with bilateral
injury histories in their study, which may have contributed to the
lack of difference in sEMG activity between the injured leg and the
contralateral control limb (Sole et al., 2011). However our finding
that, when comparing BF sEMG across the two groups, only during
eccentric contractions at 180° s 1 was the between limb differ-
ence significantly greater in the INJ compared to the UI group,
somewhat confirms a previous similar finding by Sole et al.
(2011). Whilst there was no significant between limb difference
in BF sEMG during eccentric contractions at 60° s 1 when com-
paring the two groups in the current study, the large ES
(d = 0.74) indicates that a significant difference may have been de-
tected with an increased sample size.
Reductions in muscle activation during eccentric contractions is
due to reduced motor unit recruitment and/or firing rates (Webber
and Kriellaars, 1997) which impact upon maximal torque genera-
tion capabilities. Following hamstring strain injury it has been sug-
gested that the purpose of reduced hamstring activation would be
to protect the damaged tissue from high force contraction (Opar
et al., 2012). Hamstring strain injuries themselves are character-
ised by acute pain in the posterior thigh (Verrall et al., 2001) with
reports of chronic pain not uncommon (Croisier et al., 2002; Jönhagen
et al., 1994) and this has the potential to result in long-term re-
organisation of the nervous system at the spinal and supraspinal
levels (Mense, 2003). The current study confirms that, even in ath-
letes who have been successfully rehabilitated and have returned
to competition, sEMG activity of the BF remains suppressed. This
would indicate that, for the current cohort, contemporary rehabil-
itation practices were unsuccessful at addressing deficits in the
Fig. 3. Medial hamstring normalised surface electromyography (sEMG) at four activation of BF. This is of concern from the perspective of HSI
different isokinetic velocities from the: (A) injured athletes and (B) uninjured
recurrence given submaximal stimulation of in situ animal muscle
athletes. Negative movement velocities are indicative of eccentric contractions and
positive velocities indicate concentric contractions. Error bars display standard reduces the amount of stress that muscle can withstand before the
deviation. occurrence of stretch induced failure (Garrett et al., 1987). This
may indicate that the previously injured BF is unable to withstand
the same amount of stress before failure compared to an uninjured
contractions, and BF sEMG during fast eccentric contraction was muscle, thus increasing the likelihood of re-injury. The observation
greater in INJ group compared to the UI group (Table 5). of no between limb differences in median power frequency in the
 D.A. Opar et al. / Journal of Electromyography and Kinesiology 23 (2013) 696–703 701

Table 4
Median power frequency of the biceps femoris long head and medial hamstrings of athletes with and without a history of unilateral hamstring strain injury during concentric and
eccentric contraction.
1
Movement velocity (° s ) Biceps femoris Medial hamstrings
Injured limb Uninjured limb p ES Injured limb Uninjured limb p ES
Injured group
+180 61.70 (±5.82) 64.70 (±9.00) 0.1005 0.40 67.75 (±6.25) 71.15 (±8.34) 0.1680 0.47
+60 60.30 (±6.64) 62.11 (±7.80) 0.2220 0.25 58.70 (±7.48) 62.78 (±9.57) 0.1655 0.48
60 64.78 (±7.83) 66.92 (±9.35) 0.2530 0.24 62.85 (±9.63) 66.03 (±15.53) 0.2950 0.25
180 63.04 (±6.38) 68.03 (±13.73) 0.1030 0.50 64.68 (±9.42) 70.43 (±18.49) 0.2140 0.41
Biceps femoris Medial hamstrings
Dominant limb Non-dominant limb p ES Dominant limb Non-dominant limb p ES
Uninjured group
+180 63.57 (±11.35) 62.82 (±7.41) 0.3580 0.08 74.84 (±13.24) 72.04 (±7.71) 0.2460 0.26
+60 62.71 (±7.60) 62.84 (±7.51) 0.4670 0.02 69.44 (±10.44) 66.28 (±6.28) 0.1025 0.37
60 63.25 (±9.37) 63.38 (±6.89) 0.4620 0.02 70.24 (±15.52) 66.42 (±13.50) 0.2075 0.26
180 64.22 (±12.62) 66.05 (±8.26) 0.2400 0.17 70.21 (±18.21) 71.05 (±13.62) 0.4275 0.05

Negative movement velocities are indicative of eccentric contractions and positive velocities indicate concentric contractions. Data are presented as mean (± standard
deviation). Significance was set at p < 0.0125. Cohen’s d was used to calculate effect size (ES).

Table 5
Comparison of between limb differences in knee flexor torque and normalised electromyographical activity and median power frequency of the biceps femoris long head and
medial hamstrings in athletes with and without a history of hamstring strain injury, during concentric and eccentric contraction.
1
Movement velocity (° s ) Knee flexor torque
Injured group Uninjured group p ES
+180 9.34 (3.03–15.66) 4.40 ( 1.33 to 10.13) 0.2208 0.48
+60 14.01 (5.98–22.02) 3.34 ( 3.48 to 10.16) 0.0379* 0.83
60 18.26 (8.68–27.84) 1.03 ( 12.10 to 14.17) 0.0312* 0.85
180 20.55 (9.72–31.37) 0.24 ( 11.56 to 12.04) 0.0110* 1.03
Biceps femoris Medial hamstrings
Injured group Uninjured group p ES Injured group Uninjured group p ES
Normalised electromyographical activity
a a
+180 0.03 ( 0.01 to 0.07) 0.01 ( 0.05 to 0.02) 0.0919 0.03 ( 0.01 to 0.06) 0.00 ( 0.08 to 0.07) 0.4070
+60 0.04 ( 0.07 to 0.15) 0.03 ( 0.11 to 0.06) 0.3271 0.41 0.05 ( 0.10 to 0.21) 0.04 ( 0.17 to0.10) 0.3661 0.36
60 0.13 (0.05–0.22) 0.01 ( 0.09 to 0.11) 0.0542 0.74 0.07 ( 0.06 to 0.20) 0.03 ( 0.15 to 0.09) 0.2395 0.46
180 0.13 (0.07–0.19) 0.02 ( 0.15 to 0.12) 0.0473* 0.82 0.09 ( 0.04 to 0.21) 0.03 ( 0.13 to 0.07) 0.1210 0.61
Median power frequency
+180 3.00 ( 1.86 to 7.85) 0.74 ( 3.55 to 5.04) 0.4570 0.29 3.40 ( 4.04 to 10.84) 2.80 ( 5.71 to 11.30) 0.9078 0.04
+60 1.81 ( 3.21 to 6.84) 0.12 ( 3.24 to 3.00) 0.4835 0.37 4.08 ( 4.75 to 12.90) 3.16 ( 1.94 to 8.26) 0.8462 0.08
60 2.15 ( 4.72 to 9.01) 0.13 ( 3.02 to 2.75) 0.5122 0.27 3.18 ( 9.44 to 15.80) 3.82 ( 5.93 to 13.57) 0.9315 0.03
180 4.99 ( 3.18 to 13.15) 1.83 ( 7.26 to 3.59) 0.1442 0.60 5.76 ( 9.63 to 21.14) 0.84 ( 10.52 to 8.85) 0.4377 0.31

Negative movement velocities are indicative of eccentric contractions and positive velocities indicate concentric contractions. Data are presented as mean differences (95%
confidence intervals).
*
Significance was set at p < 0.05. Cohen’s d was used to calculate effect size (ES).
a
ES for electromyographical activity could not be calculated given the use of this data in the normalisation process.

INJ group suggests that prior hamstring strain injury may not im- forces at relatively long muscle lengths in running (Thelen et al.,
pact upon average muscle fibre conduction velocity (Linnamo 2005). Late stage rehabilitation involving more forceful eccentric
et al., 2000). It should also be acknowledged that a number of other contractions at long muscle lengths might be expected to over-
factors also influence the median power frequency of the electr- come these maladaptations (Lynn and Morgan, 1994), however,
omyographical signal and further investigation examining these suppression of hamstring activation, as reported in the current
factors is warranted. study, would reduce the stimulus the previously injured muscle
It has been proposed previously that the suppression of ham- is exposed to, thus potentially compromising the adaptive re-
string muscle activation following hamstring strain injury has the sponse to rehabilitation. The present study suggests that chronic
potential to limit adaptation during the rehabilitation process lowering of hamstring activation following strain injury could sab-
(Opar et al., 2012). This model suggests early to middle stage reha- otage the rehabilitation process. Still, the full impact of prior ham-
bilitation for hamstring strain injury typically involves avoidance string strain injury on neurological control of the involved muscle/
of excessive stretching of the involved tissue and submaximal s and impact on adaptation requires further attention.
exercise performed through limited range of motion in an attempt The current study found strength at all velocities and contrac-
to prevent proliferation of scar tissue (Heiderscheit et al., 2010). tion modes was lower in the previously injured limb compared
Such an approach might be expected to result in a reduction of to the uninjured limb. Previous work has found eccentric but not
in-series sarcomeres (Williams and Goldspink, 1978) and induce concentric declines in strength (Lee et al., 2009) or greater eccen-
atrophy (Silder et al., 2008) potentially reducing the optimal length tric deficits (22–24%) compared to concentric deficits (10–11%) fol-
of the hamstrings (Brockett et al., 2004) which would be unfavour- lowing hamstring strain injury (Croisier et al., 2002). As muscle
able given the need for the hamstrings to generate high eccentric shortening velocity is known to influence maximal tension
702 D.A. Opar et al. / Journal of Electromyography and Kinesiology 23 (2013) 696–703

generating capacity (Fenn and Marsh, 1935) the different concen- work, notwithstanding the difficulty in recruiting athletes for the
tric velocities used in previous work may explain the inconsistent INJ group.
findings for this contraction mode. In line with this, the percentage In conclusion, this study is the first to report that athletes with a
difference in strength between previously injured and uninjured history of unilateral hamstring strain injury display reductions in
limbs tested at a comparable velocities (+60° s 1) is similar in the the sEMG activity of a previously injured BF during eccentric con-
current study (10.9%) and previous work (11%) (Croisier et al., tractions and no difference in the median power frequency of
2002). The much larger decline in eccentric strength reported else- either hamstring head during concentric or eccentric contractions.
where (Croisier et al., 2002) is less likely to be due to differences in Furthermore strength was suppressed during both contraction
eccentric testing velocities as eccentric strength is largely unaf- modes in the injured limb compared to the uninjured limb. Previ-
fected by lengthening velocity (Edman et al., 1978). It may be, ous hamstring strain injury may result in between limb alterations
however, explained by differences in rehabilitation practices of in neuromuscular function and rehabilitation practices need to
the respective cohorts given the greater appreciation for eccentric consider the recovery of strength and activation during eccentric
conditioning in hamstring strain injury prevention in recent times contractions as markers of successful rehabilitation as this may as-
(Petersen et al., 2011). Perhaps not surprisingly, more recent sist in reducing the incidence of hamstring strain injury recurrence.
studies have reported smaller eccentric strength differences in
the order of 13% (Lee et al., 2009), which is comparable to the Acknowledgment
10.9–12.5% differences reported in the current study.
Uniformly lower concentric and eccentric strength, as observed The authors would like to thank Associate Professor Timothy
in the current study, would be expected if strength was deter- Carroll from the University of Queensland for his assistance in
mined solely from muscle cross sectional area and volume, given the preparation of this manuscript.
the noted atrophy of BF following hamstring strain injury (Silder
et al., 2008). Interestingly, sEMG activity was lower only during References
eccentric contractions, despite lower strength across contraction
modes and velocities. This suggests that reductions in BF activity Aagaard P, Simonsen EB, Andersen JL, Magnusson SP, Halkjaer-Kristensen J, Dyhre-
contribute to prolonged eccentric, but not concentric, weakness Poulsen P. Neural inhibition during maximal eccentric and concentric
quadriceps contraction: effects of resistance training. J Appl Physiol
following hamstring strain injury. It might therefore be expected 2000;89(6):2249–57.
that the decline in eccentric strength following hamstring strain Arnason A, Sigurdsson SB, Gudmundsson A, Holme I, Engebretsen L, Bahr R. Risk
injury would be of a greater relative magnitude than concentric factors for injuries in football. Am J Sports Med 2004;32(Suppl. 1):5S–16S.
Blankenbaker DG, Tuite MJ. Temporal changes of muscle injury. Semin
strength, but this is not supported by the current data. It may be
Musculoskelet Radiol 2010;14(2):176–93.
that other muscles which contribute to knee flexion, that were Brockett CL, Morgan DL, Proske U. Predicting hamstring strain injury in elite
not examined in the current study, such as the short head of biceps athletes. Med Sci Sports Exerc 2004;36(3):379–87.
femoris, gastrocnemius and sartorius, increase their involvement Croisier JL, Forthomme B, Namurois MH, Vanderthommen M, Crielaard JM.
Hamstring muscle strain recurrence and strength performance disorders. Am
during maximal eccentric contraction in a previously injured leg J Sports Med 2002;30(2):199–203.
to help overcome the limitation in sEMG activity of BF. Indeed, Croisier JL, Ganteaume S, Binet J, Genty M, Ferret JM. Strength imbalances and
compensatory hypertrophy of the short head of biceps femoris prevention of hamstring injury in professional soccer players: a prospective
study. Am J Sports Med 2008;36(8):1469–75.
has been reported previously (Silder et al., 2008), suggesting ham- Drezner J, Ulager J, Sennett MD. Hamstring muscle injuries in track and field
string strain injury may lead to increased use of uninjured muscu- athletes: a 3-year study at the Penn Relay Carnival. Clin J Sport Med
lature, however further examination of this area is warranted. 2005;15(5):386 [abstract].
Edman KA, Elzinga G, Noble MI. Enhancement of mechanical performance by
There are some limitations in the present study’s methodology. stretch during tetanic contractions of vertebrate skeletal muscle fibres. J Physiol
The retrospective nature of the study does not allow for the deter- 1978;281:139–55.
mination of whether the reduction in sEMG activity of BF is the Fenn WO, Marsh BS. Muscular force at different speeds of shortening. J Physiol
1935;85(3):277–97.
cause of or the result of injury. Prospective studies are required Fuller CW, Ekstrand J, Junge A, Andersen TE, Bahr R, Dvorak J, et al. Consensus
to determine if low levels of BF activity elevates the risk of sustain- statement on injury definitions and data collection procedures in studies of
ing a future hamstring strain injury. It should be noted, however, football (soccer) injuries. Br J Sports Med 2006;40(3):193–201.
Garrett Jr WE, Safran MR, Seaber AV, Glisson RR, Ribbeck BM. Biomechanical
that whilst prospective studies have determined that a between
comparison of stimulated and nonstimulated skeletal muscle pulled to failure.
limb eccentric strength difference of approximately 4.5% is associ- Am J Sports Med 1987;15(5):448–54.
ated with future hamstring strain injury (Sugiura et al., 2008), Heiderscheit BC, Sherry MA, Silder A, Chumanov ES, Thelen DG. Hamstring strain
post-injury eccentric weakness is reported to be between 13% injuries: recommendations for diagnosis, rehabilitation, and injury prevention. J
Orthop Sports Phys Ther 2010;40(2):67–81.
and 24% (Croisier et al., 2002; Lee et al., 2009), suggesting ham- Hermens HJ, Freriks B, Disselhorst-Klug C, Rau G. Development of recommendations
string injury enhances eccentric knee flexor weakness, most prob- for SEMG sensors and sensor placement procedures. J Electromyogr Kinesiol
ably via neuromuscular maladaptation. Also using the maximal 2000;10(5):361–74.
Jönhagen S, Németh G, Eriksson E. Hamstring injuries in sprinters. The role of
activation data from the fastest concentric movement velocity concentric and eccentric hamstring muscle strength and flexibility. Am J Sports
(+180° s 1) to normalise the sEMG data as per previous investiga- Med 1994;22(2):262–6.
tions (Aagaard et al., 2000) has the potential to mask any between- Lee MJ, Reid SL, Elliott BC, Lloyd DG. Running biomechanics and lower limb strength
associated with prior hamstring injury. Med Sci Sports Exerc
limb differences in sEMG activity at this velocity, however given 2009;41(10):1942–51.
the important nature of eccentric strength in hamstring strain in- Linnamo V, Bottas R, Komi PV. Force and EMG power spectrum during and after
jury aetiology, sEMG activity during eccentric contraction was of eccentric and concentric fatigue. J Electromyogr Kinesiol 2000;10(5):293–300.
Lynn R, Morgan D. Decline running produces more sarcomeres in rat vastus
most interest. Finally, the power of the current study may have intermedius muscle fibers than does incline running. J Appl Physiol
been too small to detect between limb differences in variables 1994;77(3):1439–44.
not determined to be significantly different in current study. We Mense S. The pathogenesis of muscle pain. Curr Pain Headache Rep
2003;7(6):419–25.
have reported ES for all comparisons (Tables 2–4) to further illus-
Opar DA, Williams M, Shield AJ. Hamstring strain injuries: factors that lead to injury
trate the strength of the between limb differences. The ES data sug- and reinjury. Sports Med 2012;42(3):209–26.
gests that, in particular, the study may have been underpowered to Orchard JW, Seward H. Injury report 2009: Australian football league. Sport Health
detect differences in the electromyographical activity of the MH 2010;28:10–9.
Petersen J, Thorborg K, Nielsen MB, Budtz-Jørgensen E, Hölmich P. Preventive effect
and the median power frequency between injured and uninjured of eccentric training on acute hamstring injuries in men’s soccer: a cluster-
limbs. A larger sample size should be a consideration for future randomized controlled trial. Am J Sports Med 2011;39(11):2296–303.
 D.A. Opar et al. / Journal of Electromyography and Kinesiology 23 (2013) 696–703 703

Seger JY, Thorstensson A. Muscle strength and myoelectric activity in prepubertal Ryan G. Timmins completed his B.App.Sci(HM)(Hons)
and adult males and females. Eur J Appl Physiol 1994;69(1):81–7. at the Queensland University of Technology in Bris-
Silder A, Heiderscheit BC, Thelen DG, Enright T, Tuite MJ. MR observations of long- bane, Australia. He completed his honours thesis under
term musculotendon remodeling following a hamstring strain injury. Skeletal the supervision of Dr Anthony Shield, focussing on
Radiol 2008;37(12):1101–9. neuromuscular alterations in the hamstring muscles
Silder A, Reeder SB, Thelen DG. The influence of prior hamstring injury on following repeat sprint running. He is currently work-
lengthening muscle tissue mechanics. J Biomech 2010;43(12):2254–60. ing as an Associate Lecturer at the Queensland Uni-
Sole G, Milosavljevic S, Nicholson HD, Sullivan SJ. Selective strength loss and versity of Technology in the School of Exercise and
decreased muscle activity in hamstring injury. J Orthop Sports Phys Ther Nutrition Sciences.
2011;41(5):354–63.
Sugiura Y, Saito T, Sakuraba K, Sakuma K, Suzuki E. Strength deficits identified with
concentric action of the hip extensors and eccentric action of the hamstrings
predispose to hamstring injury in elite sprinters. J Orthop Sports Phys Ther
2008;38(8):457–64.
Thelen DG, Chumanov ES, Hoerth DM, et al. Hamstring muscle kinematics during
treadmill sprinting. Med Sci Sports Exerc 2005;37(1):108–14.
Verrall GM, Slavotinek JP, Barnes PG, Fon GT, Spriggins AJ. Clinical risk factors for
hamstring muscle strain injury: a prospective study with correlation of injury Nuala M. Dear completed her BAppSc(HM)(Hons) at
by magnetic resonance imaging. Br J Sports Med 2001;35(6):435–9. Queensland University of Technology and is now
Webber S, Kriellaars D. Neuromuscular factors contributing to in vivo eccentric undertaking her Masters of Physiotherapy Studies in
moment generation. J Appl Physiol 1997;83(1):40–5. the School of Rehabilitation Sciences at the University
Westing SH, Cresswell AG, Thorstensson A. Muscle activation during maximal of Queensland in Brisbane, Australia. She has a keen
voluntary eccentric and concentric knee extension. Eur J Appl Physiol
interest in the application of sports science research in
1991;62(2):104–8.
the injury rehabilitation of athletes and is a sessional
Williams P, Goldspink G. Changes in sarcomere length and physiological properties
academic in the School of Exercise and Nutrition at
in immobilized muscle. J Anat 1978;127(3):459–68.
Woods C, Hawkins RD, Maltby S, Hulse M, Thomas A, Hodson A. The Football Queensland University of Technoology.
Association Medical Research Programme: an audit of injuries in professional
football: analysis of hamstring injuries. Br J Sports Med 2004;38(1):36–41.
Worrell TW, Perrin DH, Gansneder BM, Gieck JH. Comparison of isokinetic strength
and flexibility measures between hamstring injured and noninjured athletes. J
Orthop Sports Phys Ther 1991;13(3):118–25.

David A. Opar completed his B.App.Sci(HM)(Hons) at Anthony J. Shield current research interests are
RMIT University in Melbourne, Australia and is now a focussed on hamstring strain injury. Dr Shield and his
PhD student at the School of Exercise and Nutrition team are currently investigating the effects of prior
Sciences at Queensland University of Technology. His hamstring injury on hamstring muscle activation and
thesis, supervised by Dr Anthony Shield and Dr Morgan architecture. He completed his PhD at Southern Cross
Williams, is focused on alterations to neuromuscular University, Australia and is currently a Senior Lecturer
hamstring function in previously injured athletes. in the School of Exercise and Nutrition Sciences at
Previously he has worked as an Associate Lecturer at Queensland University of Technology.
RMIT University in the Division of Exercise Sciences.

Morgan D. Williams completed a PhD at ACU in Mel-

bourne, Australia. He is currently a Senior Lecturer in
Sports Science at the University of Glamorgan, Wales,
UK.