Hamstrings Hypertrophy Is Specific to the

Training Exercise: Nordic Hamstring versus

Lengthened State Eccentric Training
SUMIAKI MAEO1,2, THOMAS G. BALSHAW2,3, DARREN Z. NIN2,4,5, EMMET J. MC DERMOTT2,6,
THOMAS OSBORNE2,7, NAOMI B. COOPER2, GARRY J. MASSEY2,8, PUI W. KONG4,
MATTHEW T. G. PAIN2, and JONATHAN P. FOLLAND2,3,9
1
Faculty of Sport and Health Science, Ritsumeikan University, Kusatsu, JAPAN; 2School of Sport, Exercise and Health Sciences,
Loughborough University, Loughborough, UNITED KINGDOM, 3Versus Arthritis Centre for Sport, Exercise and Osteoarthritis
Research, Loughborough University, Loughborough, UNITED KINGDOM; 4National Institute of Education, Nanyang
Technological University, SINGAPORE; 5Department of Surgery, Massachusetts General Hospital, Harvard Medical School,
Boston, MA; 6Department of Physical Education and Sport Sciences, University of Limerick, Limerick, IRELAND; 7Leeds
Institute of Rheumatic and Musculoskeletal Medicine (LIRMM), University of Leeds, Leeds, UNITED KINGDOM; 8School of

BASIC SCIENCES
Sport and Health Sciences, University of Exeter, Devon, UNITED KINGDOM; and 9National Institute for Health and Care
Research (NIHR) Leicester Biomedical Research Centre, Leicester, UNITED KINGDOM

ABSTRACT
MAEO, S., T. G. BALSHAW, D. Z. NIN, E. J. MC DERMOTT, T. OSBORNE, N. B. COOPER, G. J. MASSEY, P. W. KONG, M. T. G.
PAIN, and J. P. FOLLAND. Hamstrings Hypertrophy Is Specific to the Training Exercise: Nordic Hamstring versus Lengthened State Eccen-
tric Training. Med. Sci. Sports Exerc., Vol. 56, No. 10, pp. 1893-1905, 2024. Introduction: The hamstring muscles play a crucial role in
sprint running but are also highly susceptible to strain injuries, particularly within the biceps femoris long head (BFlh). This study compared
the adaptations in muscle size and strength of the knee flexors, as well as BFlh muscle and aponeurosis size, after two eccentrically focused
knee flexion training regimes: Nordic hamstring training (NHT) vs lengthened state eccentric training (LSET, isoinertial weight stack resis-
tance in an accentuated hip-flexed position) vs habitual activity (no training controls: CON). Methods: Forty-two healthy young males com-
pleted 34 sessions of NHT or LSET over 12 wk or served as CON (n = 14/group). Magnetic resonance imaging-measured muscle volume of
seven individual knee flexors and BFlh aponeurosis area, and maximum knee flexion torque during eccentric, concentric, and isometric con-
tractions were assessed pre- and post-training. Results: LSET induced greater increases in hamstrings (+18% vs +11%) and BFlh (+19% vs
+5%) muscle volumes and BFlh aponeurosis area (+9% vs +3%) than NHT (all P ≤ 0.001), with no changes after CON. There were distinctly
different patterns of hypertrophy between the two training regimes, largely due to the functional role of the muscles; LSET was more effective
for increasing the size of knee flexors that also extend the hip (2.2-fold vs NHT), whereas NHT increased the size of knee flexors that do not
extend the hip (1.9-fold vs LSET; both P ≤ 0.001). Changes in maximum eccentric torque differed only between LSET and CON (+17% vs
+4%; P = 0.009), with NHT (+11%) inbetween. Conclusions: These results suggest that LSET is superior to NHT in inducing overall ham-
strings and BFlh hypertrophy, potentially contributing to better sprint performance improvements and protection against hamstring strain in-
juries than NHT. Key Words: MUSCLE VOLUME, APONEUROSIS SIZE, ECCENTRIC STRENGTH

T
Address for correspondence: Jonathan P. Folland, Ph.D., School of Sport, Ex- he hamstrings are the primary knee flexors and play
ercise and Health Sciences, Loughborough University, Leicestershire, LE11 a major role in horizontal force production during
3TU, United Kingdom; E-mail: [email protected].
Submitted for publication March 2024. sprinting (1). Indeed, sprint performance is associated
Accepted for publication May 2024. with hamstrings muscle size (2,3) and can be improved by re-
Supplemental digital content is available for this article. Direct URL citations sistance training of the knee flexors (4). However, hamstring
appear in the printed text and are provided in the HTML and PDF versions strain injuries (HSI) are highly prevalent in many sports such
of this article on the journal’s Web site (www.acsm-msse.org). as American football (5), rugby (6), and track and field (7),
0195-9131/24/5610-1893/0 and account for 12–22% of all injuries in football/soccer (8,9).
MEDICINE & SCIENCE IN SPORTS & EXERCISE® HSI typically occur during high-speed running, specifically
Copyright © 2024 The Author(s). Published by Wolters Kluwer Health, Inc. on during the late swing phase of sprinting (i.e., when the hip is
behalf of the American College of Sports Medicine. This is an open-access article
distributed under the terms of the Creative Commons Attribution-Non Commercial- flexed and the knee is extended) (10,11). The late swing phase
No Derivatives License 4.0 (CCBY-NC-ND), where it is permissible to down- involves peak force production by the hamstrings (12), while
load and share the work provided it is properly cited. The work cannot be changed contracting eccentrically at a relatively long length (peak length
in any way or used commercially without permission from the journal. for the gait cycle) (13). Thus, establishing an effective training
DOI: 10.1249/MSS.0000000000003490 modality for increasing hamstrings muscle size as well as knee

1893
 flexor strength, particularly eccentric strength, will benefit many which was approved by the Loughborough University Ethics
athletes and coaches for both performance improvement and Review Sub-Committee (R17-P054) and Nanyang Technolog-
injury prevention purposes (1,14,15). ical University Institutional Review Board (IRB-2017-07-030).
Nordic hamstring training (NHT), an eccentric training mo- Participants were first assigned to either CON or training in a
dality for the knee flexors, has been widely demonstrated to re- 1:2 ratio depending on schedule availability (i.e., whether they
duce the risk of new and recurrent HSI (16–18). This may be at could visit the laboratory two to three times a week for 12 wk),
least partly explained by an increase in hamstring muscle size and then training participants were randomly assigned to LSET
and eccentric knee flexor strength induced by NHT (14). How- or NHT after the preintervention measurements. A total of six
ever, the hip joint remains relatively extended (i.e., ~ anatom- participants withdrew from the study because of personal rea-
ical position) throughout NHT. Furthermore, weaker partici- sons unrelated to study participation; 42 participants completed
pants may lack the strength to control the lowering movement the study.
beyond the initial phase of the contraction during NHT (there-
after presumably falling with relatively low neuromuscular ac- Overview
tivation at more extended angles). Therefore, the length of the
Participants visited the laboratory for a familiarization ses-
biarticular hamstring muscles during the active phase of NHT
sion involving voluntary maximum isometric, concentric, and
appears shorter than during the late swing phase of running
eccentric contractions. Height, body mass, and physical activity
(19). In addition, growing evidence suggests that training at
levels with the International Physical Activity Questionnaire
BASIC SCIENCES

long muscle length promotes muscle hypertrophy (15,20,21).

(IPAQ, short format [31]) were also measured in this session.
Importantly, hamstrings muscle hypertrophy was found to be
Thereafter, two duplicate neuromuscular measurement sessions
>50% greater after knee flexion (leg curl) training performed
were conducted both pre (sessions 4–5 d apart before the first
at long lengths (hip flexed) compared with short lengths (hip
training session) and post (2–3 d after the last training session
extended) (15). Furthermore, rehabilitation emphasizing eccen-
and 4–5 d apart) 12 wk of training or control intervention. This
tric knee flexion training at long lengths (accentuated hip-flexed
approach of duplicate measurement sessions at each time point
position), named lengthened state eccentric training (LSET)
is thought to reduce measurement error and may be particularly
(22), resulted in a significantly lower HSI recurrence rate com-
useful in the context of training adaptations (e.g., Heritage Fam-
pared with noncompliant athletes (23). Considering these find-
ily Study [32]) and between-group comparisons across two time
ings, LSET may produce greater increases in hamstrings muscle
points (29,33). All measurements were of the dominant leg, and
size and strength than NHT, with implications for injury pre-
the neuromuscular measurement sessions involved recordings
vention. However, no study has compared the functional and
of knee flexion torque and surface electromyography (EMG)
morphological adaptations of LSET versus NHT.
of hamstring muscles during voluntary maximum isometric,
The biceps femoris long head (BFlh) has the highest suscep-
concentric, and eccentric contractions. Axial T1-weighted mag-
tibility to HSI (5,16,18). Thus, morphological adaptations of the
netic resonance imaging (MRI) scans of the thigh were also
BFlh to training interventions are of particular interest, includ-
conducted pre- (5 d before the first training session) and
ing size of the muscle, and its aponeurosis, which is integral to
post-training (2–3 d after the final training session), always pre-
force transmission. Furthermore, a small BFlh proximal aponeu-
ceding the first neuromuscular measurement sessions. Partici-
rosis has been suggested as a risk factor for HSI by concentrating
pants in the training groups completed 12 wk (34 sessions) of
mechanical strain on the surrounding muscle tissue (24–27).
systematic, progressive (load and volume of repetitions in-
In the vastus lateralis muscle, aponeurosis size appears respon-
creased) knee flexor training of both legs. All participants were
sive to resistance training (28–30). To the authors’ knowledge,
instructed to maintain their habitual physical activity and diet
however, no study has investigated whether BFlh aponeurosis
throughout the study, other than the supervised training inter-
size changes after resistance training.
ventions. Participants were instructed to eat and drink normally
The main purpose of this study was to compare changes in
and avoid strenuous exercise and alcohol intake for 36 h, and
muscle size and strength of the knee flexors, as well as BFlh
caffeine consumption for 6 h, before all measurement sessions.
muscle and aponeurosis size, after 12 wk of LSET versus
Measurement sessions were conducted at a consistent time of
NHT or habitual activity (control, CON). We hypothesized
day for each participant between 9:00 and 20:00 for both
that LSET would induce greater increases in hamstrings muscle
MRI and neuromuscular sessions.
size, BFlh aponeurosis size, as well as eccentric knee flexor
strength than NHT.
Resistance Training Interventions
METHODS The training program consisted of 34 supervised sessions
over 12 wk (three times per week apart from weeks 1 and 12
Participants
(two times per week); Supplemental Table 1, Number of ec-
Forty-eight healthy young males with no history of lower centric sets × repetitions completed in each session of both
extremity injury or systematic exercise training of any kind training programs, Supplemental Digital Content, http://links.
in the last 18 months provided written informed consent and lww.com/MSS/D32) with each session separated by ≥36 h.
completed preintervention measurements within this study, The two training regimes were inherently different as NHT

1894 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

is a bilateral, primarily body weight, exercise involving purely weight between each repetition. Specifically, knee joint angle
eccentric contractions, whereas for LSET, we employed a con- at the start/end of each repetition (i.e., most extended position)
ventional isoinertial weight stack machine (modified for greater was increased weekly by 5–7° in weeks 1–5 from 37° to 32°,
hip flexion) and concentric contractions/lifts to then be able to 26°, 19°, and finally to 14° (0° = full extension) in week 5 on-
lower/return the load eccentrically (see Supplemental Fig. 1, ward, based on the goniometer measurements.
Illustrations and training load of the lengthened state eccentric Nordic hamstring training. Participants knelt on a pad-
training and Nordic hamstring training, Supplemental Digital ded 30-cm high box, with the lower leg horizontal and both
Content, http://links.lww.com/MSS/D32). To achieve a high ankles protruding over the rear end of the box, while the thighs
level of eccentric loading with LSET, the concentric load was and torso were initially vertical (Supplemental Fig. 2B, Sup-
lifted with two legs and lowered eccentrically with one leg. plemental Digital Content, http://links.lww.com/MSS/D32).
Nonetheless, the number of eccentric sets and repetitions were Each ankle was restrained by an inextensible strap, placed
standardized across both training regimes. The number of ec- 4 cm above the medial malleolus, and in series with an S-beam
centric sets with each leg increased from 2 to 4, and the num- strain gauges (Force Logic, Swallowfield, UK). From this ini-
ber of eccentric repetitions per set from 6 to 10, throughout the tial position, participants slowly leaned/lowered themselves
training program. All training sessions began with a standard- forward from the knees by eccentrically contracting their ham-
ized cycling warm-up (5 min, 70 rpm, 150 W; Ergomedic 874 E, strings. They were instructed to take ~4 s to perform this con-
Monark Exercise AB, Sweden). This was followed by 2× of trolled lowering eccentric contraction, keeping the hips and

BASIC SCIENCES
~15 s of static stretches of the hamstrings of each leg in a torso straight and their arms close to their chest for as long
standing position (with the involved knee extended, contralat- as possible, before being unable to further control the descent
eral knee flexed, and hip flexed to lean the upper body forward and falling onto a crash matt placed on the ground in front of
toward the extended leg). Consistent verbal encouragement them. The analogue force from both strain gauges was sampled
was provided for both groups throughout the training. at a frequency of 2000 Hz using an A/D converter (CED Power
Lengthened state eccentric training. Participants 1401 mk II, CED, UK) and a personal computer (Spike 2, CED,
were positioned on a modified seated leg curl machine (Seated UK), and displayed on a screen placed on the ground in front
Leg Curl SL40, LifeFitness, USA), specifically with a modi- of the participant to provide real-time visual feedback of force
fied back rest so that the hip was maintained in a flexed posi- during the NHT. When participants could control their descent
tion (120°, 0° = anatomical position) to ensure the hamstrings to within 15° of horizontal , progression involved additional
were trained in a lengthened state (Supplemental. Fig. 2A, load, added by use of a weighted vest starting at 1 kg and
Supplemental Digital Content, http://links.lww.com/MSS/ progressing up to 21 kg in one participant.
D32). The knee joint center was aligned with the axis of rota-
tion of the machine’s lever arm, with each participant’s seating
Pre- and Postintervention Measurements
position and lever arm length noted and replicated throughout
the study. Adjustable straps were tightly secured across the Dynamometry and EMG. While seated on an isokinetic
hips, chest, and knee to prevent extraneous movement and to dynamometer (Contrex MJ, CMV AG, Dubendorf, Switzerland)
maintain the hip angle. To facilitate high eccentric knee flexor with a hip flexion angle of 60° (0° = full extension), strapping
loading, participants first flexed the knee by pulling the lever was secured across the participant’s waist, shoulders, and dis-
arm down and back (i.e., concentric knee flexor contraction) tal thigh just above the patella of the involved lower extremity
using both legs until the knee joint angle was ~90°, then using to minimize extraneous bodily movement during contractions.
only one leg performed a slow and controlled knee extension A high-density foam shin pad was secured behind the shank of
~4 s (i.e., eccentric knee flexor contraction) returning the lever the dominant leg ~2 cm above the medial malleolus. The shank
arm to its original position. Participants alternated the leg that was then strapped to the dynamometer crank arm at ~15% of
performed eccentric sets until the required number of sets was shank length (i.e., the distance between the lateral malleolus
completed on both legs with a rest period of 5 min between and knee joint space) above the medial malleolus. The knee
legs and 2 min between sets. The starting load/weight of each joint space was aligned with the dynamometer axis of rotation
participant was based on pretest strength measurements then during a submaximal knee flexor contraction, whereas the knee
iteratively adjusted/increased (typically by 2 kg) when partic- joint was positioned at a midrange angle. Analogue torque, crank
ipants could perform all the prescribed repetitions of the final angle, and crank angular velocity signal outputs from the dy-
(if two or three sets) or penultimate (if four sets) set, and all namometer were recorded using an A/D converter (Micro 1401,
training loads were recorded in training logs/sheets. To reduce CED, UK) and associated computer software (Spike 2, CED,
the chance of hamstring injury from contracting the muscle at UK) during isometric, concentric, and eccentric knee flexion
unaccustomed long lengths in the early weeks of the training, contractions. Torque, crank position, and crank velocity data
knee joint range of motion, and thus the lengthened state of the were smoothed at 15 Hz for analysis purposes.
hamstrings muscle, was progressively increased during the Surface EMG from the lateral and medial hamstrings was
first 5 wk of training. This was done by manipulating the length recorded using a wireless EMG system (Trigno, Delsys, USA)
of the cable between the training machine lever arm and the during all maximum knee flexion contractions. Skin preparation
weight stack and ensuring that participants fully lowered the (shaving, abrading, and swabbing with 70% ethanol) preceded

NORDIC HAMSTRING VS LENGTHENED STATE TRAINING Medicine & Science in Sports & Exercise® 1895
 sensor fixation on the skin with the use of adhesive interfaces. Maximum concentric and eccentric knee flexion
Single differential Trigno Standard EMG sensors, constituting contractions. Isovelocity concentric and eccentric strength
a bipolar configuration, were situated over the lateral and me- measurements involved passive torque assessment, followed
dial hamstrings at 50% of thigh length above the popliteal by warm-up and MVC. Passive torque was assessed during
fossa. EMG signals were amplified at source (x300, 20 to four passive knee flexion–extension repetitions (middle two
450 Hz bandwidth) before further amplification (x909, overall used for analysis) through 0–95° crank arm range of motion
effective gain) and sampled at 2000 Hz. Correction for the in- at an isovelocity of 50°⋅s−1, while the participant was instructed
herent 48-ms delay of the analogue signal from the Trigno EMG to remain completely relaxed. Thereafter, participants per-
system was performed before analysis to time align EMG data formed two warm-up sets (at ~50% and 80% of maximum ef-
with torque, angle, and angular velocity signals, with all vari- fort) of two concentric–eccentric repetition cycles at 50°⋅s−1
ables synchronously recorded using the same A/D converter throughout their full range of movement (0–95°) with 30 s be-
and computer software. tween sets. Participants were then instructed to “pull as hard as
Maximum isometric knee flexion contractions. Par- you can throughout the range of movement” during both the
ticipants performed two isometric maximum voluntary con- concentric and eccentric phases, and completed two maxi-
tractions (MVC) of the dominant limb for 3–5 s at each crank mum effort sets of two concentric–eccentric repetition cycles
angle of 10°, 95°, 38°, and 66° (in the order listed, 0° = full ex- at 50°⋅s−1 with 45 s rest between sets. Real-time torque bio-
tension), following a series of incremental isometric knee flex- feedback was provided throughout the contractions with the
BASIC SCIENCES

ion contractions (~3–5 s per contraction; 3 × 50%, 3 × 75%, highest concentric and eccentric torque achieved so far indi-
and 1 × 90% of perceived maximum effort) at the initial crank cated and intense verbal encouragement provided during all
angle (10°). During MVC, participants were instructed to “pull maximum efforts.
as hard as possible” until they were provided with the signal Maximum concentric and eccentric knee flexion torque was
to cease the contraction, with intense verbal encouragement defined as the instantaneous highest torque registered within
provided during all maximum isometric efforts. Real-time bio- the isovelocity period (i.e., within 10% of the target velocity
feedback, displayed on a computer screen in front of the par- of 50°⋅s−1), corrected for angle-specific passive limb torque.
ticipant, was provided to indicate the highest isometric torque Root mean square EMG amplitude of both hamstrings sensors
achieved at each angle and motivate participants to improve was measured during 200-ms epochs (100 ms either side) at
their performance relative to the previous maximum effort. both concentric and eccentric flexion maximum torque, and
Isometric torque data were gravity corrected by subtracting then averaged across both sensors. Concentric and eccentric
baseline (passive) torque. Within an individual test session, maximum torque values were taken as a mean of the two test
isometric maximum voluntary torque was defined at each sessions at pre or post if there was less than a 10% difference
crank position as the highest torque achieved during the between measurements from each session; otherwise, a weighted
two maximum efforts. Hamstrings EMG amplitude was the mean was derived. Concentric and eccentric hamstring EMG
average of the root mean square of both hamstrings sensors amplitude was taken as the mean of the two test sessions at
measured during a 500-ms epoch at isometric knee flexion each time point.
maximum voluntary torque (250 ms either side) at each crank Magnetic resonance imaging. A 3-T MRI scanner (GE
position. Healthcare Discovery MR750w 3.0-T MRI scanner) was used
Isometric maximum knee flexion torque at each crank angle to scan the dominant leg in the supine position with the hip
was taken as a mean of the two test sessions at pre or post if and knee in the extended/anatomical position. Using body array
there was less than a 10% difference; otherwise, a weighted and spine coils, T1-weighted axial images were acquired from
mean was derived (weighting, in favor of the higher score, in- the anterior superior iliac spine to below the insertion of the
creased as the percentage difference between the two scores popliteus (POP) on the tibia, in three overlapping blocks. Fish
increased). Actual knee joint angles during the maximal con- oil capsules were placed on the lateral aspect of the partici-
tractions were derived from video camera recordings as de- pant’s thigh to allow blocks to be aligned during analysis.
tailed in the Supplemental Digital Content (Supplemental The following imaging parameters were used: imaging ma-
Digital Content, http://links.lww.com/MSS/D32). Knee joint trix = 512 × 512, field of view = 260 mm × 260 mm, spatial
angles at each crank position (10°, 38°, 66°, and 95°) were resolution = 0.508 mm × 0.508 mm, slice thickness = 5 mm,
collapsed across all four test sessions and corresponded to interslice gap = 0 mm, repetition time = 600 ms, echo time =
knee joint angles of 35 ± 5°, 55 ± 6°, 77 ± 8° and 98 ± 8°, re- 7.648 ms. MRI data were anonymized before analysis (i.e., in-
spectively. Quadratic functions were fitted to the relationship vestigators were made blinded to the conditions/groups). Pre-
between the measured torque–angle relationship for each par- and postimages were analyzed side by side to allow for consis-
ticipant at each time point (pre and post) and used to derive tent analysis (in terms of inclusion/exclusion of noncontractile
knee flexion torque at 10° intervals between 35° and 95° knee tissues such as aponeurosis, blood vessels and nerves) within
joint angles for each participant. Maximum isometric torque each participant, with a convolution filter (sharpen 5 × 5) ap-
was taken as the highest value produced at any of the 10° in- plied to sharpen the images.
tervals in this range. Hamstring EMG amplitude was taken Muscle volume. Anatomical cross-sectional areas of the
as the mean of the two test sessions at each time point. biceps femoris short head (BFsh), BFlh, semitendinosus (ST),

1896 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

semimembranosus (SM), sartorius (SAR), and gracilis (GRA) poor image quality: all MRI variable data from three CON par-
muscles were outlined every third slice, and that of the POP ticipants; SAR/GRA/POP volume data from one NHT partic-
was outlined every slice, from the most distal to proximal im- ipant; POP volume data from one LSET participant; aponeu-
ages using image analysis software (Horos software, version rosis data from one NHT participant (detailed in each table
1.1.7). The volume of each muscle was calculated using cubic and figure).
spline interpolation of the anatomical cross-sectional areas
along the limb (100 points, Origin 2021, OriginLab Corpora-
tion). The volumes of the hamstrings (HAMS) and overall
RESULTS
knee flexors (KF) were calculated by summing the volumes Group characteristics at baseline. Height (LSET
of the four hamstrings and seven knee flexors muscles, respec- 1.78 ± 0.06; NHT 1.76 ± 0.08; CON 1.78 ± 0.07 m), body
tively. We also calculated the volume of the knee flexors that mass (LSET 77 ± 11; NHT 76 ± 13; CON 73 ± 6 kg), age
extend the hip (KF and HE; sum of BFlh, ST, SM) and the (LSET 25 ± 4; NHT 27 ± 3; CON 24 ± 3 yr) and habitual
knee flexors that do not extend the hip (KF not HE; sum of physical activity (IPAQ: LSET 1580 ± 479; NHT
BFsh, GRA, SAR, POP). 1198 ± 392; CON 1342 ± 458 MET·min·wk−1) did not differ
BFlh aponeurosis morphology. The contact interface between groups at baseline (ANOVA, 0.073 ≤ P ≤ 0.752).
distance between the BFlh muscle and the proximal aponeuro- Similarly, there were no baseline between-group differences
sis was outlined in each image in which the aponeurosis was in knee flexor muscle volume (individual muscles and mus-
cle groups; 0.065 ≤ P ≤ 0.976), BFlh aponeurosis morphol-

BASIC SCIENCES
identifiable (24). The contact interface distance in each slice
included both the internal and external aponeurosis, and the ogy (0.375 ≤ P ≤ 0.834), maximum knee flexion torque
highest contact interface distance across slices was considered (across contraction types; 0.314 ≤ P ≤ 0.433), or hamstring
maximum width. BFlh aponeurosis area was calculated as the EMG (across contraction types; 0.256 ≤ P ≤ 0.912)
product of the contact interface distance multiplied by the slice (Tables 1 and 2).
thickness (24). In addition, aponeurosis length was calculated Training quantification for LSET and NHT. The ec-
by multiplying the number of images in which the aponeurosis centric phase load in the LSET group had increased by
was identifiable by slice thickness. week 4 (+26% vs week 1; P < 0.001) and increased further
by week 12 (+41% vs week 1; P = 0.017 vs week 4) (Supple-
mental Fig. 2A′, Supplemental Digital Content, http://links.
Data and Statistical Analysis
lww.com/MSS/D32). Maximum eccentric force during the
All statistical analyses were performed using SPSS software Nordic hamstring exercise in the NHT group increased by
(v22, IBM Corporation, USA). Data normality was assessed week 4 (+25% vs week 1; P = 0.047), with a subtle nonsignif-
using the Shapiro–Wilk test for each variable on pretest values icant further increase by week 12 (+37% vs week 1) (Supple-
pooled across all groups. Three variables (SAR muscle volume, mental Fig. 2B′, Supplemental Digital Content, http://links.
isometric EMG, and concentric EMG) were found to be non- lww.com/MSS/D32).
normally distributed, and these data were log10 transformed Muscle size. Following LSET, with the exception of the
for further analysis. One-way analysis of variance (ANOVA) POP (P = 0.066), within-group increases occurred (paired
was conducted on all pretest variables to assess whether t-test, all P < 0.001) in the volume of all four individual constit-
baseline differences existed between groups. To examine uent hamstrings muscles (BFsh +6%; BFlh +19%; ST +27%;
training load progression within each training group, one- SM +14%), SAR (+8%), GRA (+24%), overall hamstrings
way ANOVA was conducted on the eccentric phase load at (+18%), KF and HE (+20%), KF not HE (+11%), and over-
week 1, 4, 8, and 12 followed by least significant differences all knee flexors (+17%; Table 1). After NHT, with the ex-
(LSD) tests corrected for multiple comparisons. Within- ception of the SM (P = 0.423) and POP (P = 0.130), there
group pre- to postintervention changes for absolute data were were pre- to postincreases in all individual constituent mus-
evaluated using paired t-tests. Comparisons of between-group cles of the hamstrings (BFsh +22%; P < 0.001, BFlh +5%;
adaptations to the intervention were assessed with repeated P < 0.021, ST +20%; P < 0.001), SAR (+18%; P < 0.001),
measures analysis of covariance (ANCOVA; group (LSET GRA (+30%; P < 0.001), overall hamstrings (+11%;
vs NHT vs CON) × time (pre vs post)), with corresponding P < 0.001), KF and HE (+9%; P < 0.001), KF not HE
pretraining values used as covariates. When group × time in- (+22%; P < 0.001), and overall knee flexors (+14%;
teraction effects displayed P < 0.05, post hoc tests were con- P < 0.001). After CON, there were no within-group changes
ducted. Specifically, absolute change values were calculated in the volume of any muscle or muscle group (paired t-test,
for the variables that had significant interaction effects, and 0.173 ≤ P ≤ 0.955).
were compared among groups by one-way ANCOVA All the muscle volume measurements (ANCOVA (all)
followed by LSD tests corrected for multiple comparisons. P < 0.001), except for POP (P = 0.437; Table 1), showed sig-
Data are presented as mean ± SD in the text/tables and nificant group × time effects. LSET resulted in greater abso-
mean ± SE within the figures. Of the 42 participants who lute muscle volume increases in the BFlh and SM compared
completed all the measurements (n = 14/group), the following with NHT (LSD (all) P < 0.001) and CON ((all) P < 0.001),
MRI variables were excluded from the analysis because of but the changes in these muscles did not differ between NHT

NORDIC HAMSTRING VS LENGTHENED STATE TRAINING Medicine & Science in Sports & Exercise® 1897
 TABLE 1. Muscle volume of constituent knee flexor muscles, anatomical and functional muscle groups, and BFlh aponeurosis morphology pre and post LSET (n = 14), NHT (n = 14), and control
(CON, n = 11) interventions.
LSET NHT CON ANCOVA Interaction
Pre Post Pre Post Pre Post P
Volume of individual muscles (cm3)
BFsh 108 ± 22 115 ± 20*** 103 ± 31 126 ± 33*** 106 ± 32 105 ± 33 <0.001
BFlh 189 ± 32 224 ± 34*** 200 ± 50 211 ± 51* 184 ± 34 185 ± 37 <0.001
ST 202 ± 44 257 ± 54*** 227 ± 61 273 ± 70*** 206 ± 40 201 ± 46 <0.001
SM 237 ± 38 270 ± 35*** 248 ± 72 252 ± 77 236 ± 36 239 ± 35 <0.001
SAR 153 ± 44 166 ± 44*** 140 ± 28 166 ± 35*** 146 ± 39 144 ± 39 <0.001
GRA 94 ± 34 117 ± 41*** 98 ± 32 127 ± 37*** 103 ± 38 102 ± 39 <0.001
POP 21.7 ± 3.8 22.1 ± 4.0 18.0 ± 3.4 18.4 ± 3.4 21.9 ± 5.3 22.0 ± 5.4 0.437
3
Volume of muscle groups (cm )
Hamstrings 736 ± 95 867 ± 99*** 778 ± 191 861 ± 207*** 732 ± 113 731 ± 122 <0.001
KF and HE 628 ± 82 752 ± 88*** 675 ± 166 735 ± 178*** 626 ± 87 625 ± 97 <0.001
KF not HE 377 ± 91 420 ± 100*** 355 ± 81 432 ± 90*** 377 ± 97 374 ± 101 <0.001
Overall KF 1008 ± 168 1176 ± 179*** 994 ± 161 1129 ± 182*** 1003 ± 179 999 ± 192 <0.001
Aponeurosis
Area (cm2) 35.3 ± 7.8 38.5 ± 8.2*** 37.6 ± 9.4 38.8 ± 9.5* 33.0 ± 5.9 33.5 ± 6.3* <0.001
Maximum width (cm) 3.42 ± 0.76 3.68 ± 0.72*** 3.72 ± 0.95 3.74 ± 0.95 3.29 ± 0.77 3.34 ± 0.80 0.016
Length (cm) 19.2 ± 3.6 19.1 ± 3.6 18.4 ± 4.2 18.4 ± 4.2 18.5 ± 2.7 18.5 ± 2.7 0.300
Data are means ± SD. Within-group effects of time were determined from paired t-tests and are denoted by *P < 0.05 or ***P < 0.001. ANCOVA interaction effects of time (pre vs post) × group
(LSET vs NHT vs CON) are reported. Post hoc comparisons of between-group changes are shown in Figures 1, 2, and 4. Hamstrings, the sum of BFsh, BFlh, ST, and SM. KF and HE, the sum of
BASIC SCIENCES

BFlh, ST, and SM. KF not HE, the sum of BFsh, SAR, GRA, and POP. Overall KF, the sum of all seven individual knee flexors. Participant numbers are as stated previously other than the following:
POP, KF not HE, and overall KF in LSET (n = 13); SAR, GRA, POP, KF not HE, overall KF, and all aponeurosis variables in NHT (n = 13).

and CON (LSD 0.053 ≤ P ≤ 0.949; Fig. 1A). In contrast, NHT P = 0.030; Table 1). The absolute increases in BFlh aponeuro-
produced greater increases in absolute volume of the BFsh and sis area were greater for LSET than NHT (LSD P = 0.001) and
SAR (Fig. 1B) than LSET (LSD 0.001 ≤ P < 0.010) or CON CON (P < 0.001; Fig. 4A) but did not differ between NHT and
([both] P < 0.001), and these muscles also had greater increases CON (P = 0.292). Within-group increases in BFlh aponeuro-
after LSET than CON (0.010 ≤ P ≤ 0.027). LSET and NHT pro- sis maximum width only occurred after LSET (+8%; paired
duced similar muscle volume increases in ST (P = 0.072) and t-test, P < 0.001), not NHT (P = 0.788) or CON (P = 0.446;
GRA (P = 0.113), and both training groups increased by more Table 1). Absolute increases in BFlh aponeurosis maximum
than CON (LSET, LSD (all) P < 0.001; NHT LSD (all) width were greater for LSET than NHT (LSD P = 0.031) or
P < 0.001). Overall hamstring volume change was different CON (P = 0.038; Fig. 4B) but did not differ between NHT
between all three groups (LSET > NHT > CON; both and CON (P = 0.876). BFlh aponeurosis length did not increase
P < 0.001; Fig. 2). KF and HE as well as KF not HE volume within any group (paired t-test, 0.336 ≤ P ≤ 0.337) and showed
changes also showed differences between all three groups no group × time effect (ANCOVA P = 0.300; Table 1).
but with opposite patterns LSET > NHT > CON for KF and Maximum eccentric, isometric and concentric knee
HE (both P < 0.001), but NHT > LSET > CON for KF not flexion strength. Maximum eccentric torque increased from
HE (both P ≤ 0.001). Overall knee flexor volume increases pre to post within the LSET (+17%; paired t-test, P = 0.002)
were greater for both LSET and NHT ((all) P < 0.001) com- and NHT groups (+11%; P = 0.048), but not for CON (+4%;
pared with CON but did not differ between the two training P = 0.397; Table 2). The absolute increase in maximum eccentric
groups (P = 0.095). Percentage change values (based on pre torque following LSET was greater than CON (LSD, P = 0.009;
to post mean changes [34]) for each muscle and muscle groups Fig. 5A) but did not differ between LSET and NHT (P = 0.237)
are summarized in Figure 3. or NHT and CON (P = 0.104). Within-group increases in max-
BFlh aponeurosis. BFlh aponeurosis area showed within- imum isometric torque occurred after LSET (+27%; paired
group increases from pre to post after LSET (+9%; paired t-test, P < 0.001), NHT (+25%; P < 0.001), and CON (+14%;
t-test, P < 0.001), NHT (+3%; P = 0.026), and CON (+2%; P < 0.001). The absolute increases in maximum isometric

TABLE 2. Maximum knee flexion torque during eccentric, isometric, and concentric contractions pre and post LSET (n = 14), NHT (n = 14), and control (CON, n = 14) interventions.
LSET NHT CON ANCOVA Interaction
Pre Post Pre Post Pre Post P
Knee flexion torque (N·m)
Eccentric (50°/s) 145 ± 23 169 ± 24** 143 ± 45 159 ± 39* 128 ± 33 133 ± 27 0.013
Isometric (0°/s) 124 ± 26 158 ± 23*** 132 ± 48 166 ± 39*** 111 ± 32 127 ± 27*** <0.001
Concentric (50°/s) 120 ± 20 142 ± 19** 122 ± 39 139 ± 30* 108 ± 29 118 ± 27* 0.063
Hamstring EMG (mV)
Eccentric (50°/s) 0.099 ± 0.048 0.135 ± 0.063*** 0.086 ± 0.043 0.106 ± 0.068 0.087 ± 0.040 0.092 ± 0.050 0.125
Isometric (0°/s) 0.122 ± 0.073 0.177 ± 0.077*** 0.099 ± 0.059 0.155 ± 0.088** 0.086 ± 0.032 0.104 ± 0.038* 0.046
Concentric (50°/s) 0.118 ± 0.056 0.148 ± 0.058** 0.111 ± 0.054 0.137 ± 0.076* 0.111 ± 0.050 0.119 ± 0.053 0.278
Data are means ± SD. Within-group effects of time were determined from paired t-tests and are denoted by *P < 0.05, **P < 0.01, or ***P < 0.001. ANCOVA interactions for time (pre vs
post) × group (LSET vs NHT vs CON) are reported. Post hoc comparisons of between-group changes are shown in Figure 5.

1898 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

 BASIC SCIENCES
FIGURE 1—Absolute changes (pre to post) in the volume of seven constituent knee flexor muscles following LSET (n = 14), NHT (n = 14), and control
(CON, n = 11) interventions. Symbols indicate between-group differences in the magnitude of pre to post changes where post hoc tests displayed LSD
P < 0.05: *different from CON, †different from LSET, §different from NHT. Data are means ± SE. Participant numbers are as stated previously other than
the following: POP in LSET (n = 13); SAR, GRA, and POP in NHT (n = 13).

torque for LSET (LSD P = 0.002) and NHT (P = 0.001) were (+9%; P = 0.027), but no group × time effect was observed
both greater than CON but did not differ between LSET and (ANCOVA P = 0.063).
NHT (P = 0.697). Maximum concentric torque showed Isometric knee flexion torque–angle relationships.
within-group increases following LSET (+18%; paired t-test, After LSET (+18% to +27%; paired t-test, (all) P ≤ 0.001;
P = 0.001), NHT (+13%; P = 0.042), and CON interventions Fig. 6A) and NHT (+25% to +29%; 0.001 ≤ P ≤ 0.004;

FIGURE 2—Absolute changes (pre to post) in the volume of anatomical and functional muscle groups following LSET (n = 14), NHT (n = 14), and control
(CON, n = 11) interventions. Symbols indicate between-group differences in the magnitude of pre to post changes where post hoc tests displayed LSD
P < 0.05: *different from CON, §different from NHT, †different from LSET. Data are means ± SE. Overall KF, the sum of all seven individual knee flexors.
HAMS, the sum of the four hamstring muscles. KF and HE, the sum of BFlh, ST, and SM. KF not HE, the sum of BFsh, SAR, GRA, and POP. Participant
numbers are as stated previously other than the following: overall KF and KF not HE in LSET and NHT (n = 13).

NORDIC HAMSTRING VS LENGTHENED STATE TRAINING Medicine & Science in Sports & Exercise® 1899
 FIGURE 3—Summary of the percentage changes in muscle volume of the individual knee flexor muscles, and anatomical and functional muscle groups
based on pre to post mean changes for each muscle or muscle group after LSET, NHT and control (CON) interventions. Symbols indicate between-
group differences in the magnitude of pre to post changes where post hoc tests displayed LSD P < 0.05: *different from CON, §different from NHT, †dif-
BASIC SCIENCES

ferent from LSET. KF, the sum of all seven individual knee flexors. HAMS, the sum of BFsh, BFlh, ST, and SM. KF and HE, the sum of BFlh, ST, and SM.
KF not HE, the sum of BFsh, SAR, GRA, and POP.

Fig. 6B), there were within-group increases in isometric torque torque for LSET compared with CON occurred from 35°
at all knee joint angles between 35° and 95°. After CON, there to 75° (LSD 0.008 ≤ P ≤ 0.047), but not at 85° and 95°
were pre- to postincreases in isometric torque between 35° and (0.082 ≤ P ≤ 0.088).
55° (+11% to +18%; paired t-test, 0.001 ≤ P ≤ 0.012), but Surface EMG. After LSET, there were within-group in-
not between 65° and 95° (0.061 ≤ P ≤ 0.636; Fig. 6C). Sig- creases in eccentric (+37%; paired t-test, P < 0.001), isometric
nificant group × time effects were observed for isometric (+45%; P < 0.001), and concentric (+25%; P = 0.002) hamstring
torque at all knee joint angles between 35° and 95° EMG (Table 2). After NHT, there were pre- to postincreases in
(0.001 ≤ P ≤ 0.020). Absolute increases in isometric torque isometric (+56%; P = 0.004) and concentric (+23%; paired
were greater for NHT compared with CON for all knee joint t-test, P = 0.027), but not eccentric (P = 0.081), hamstring
angles (i.e., 35° to 95°; LSD 0.001 ≤ P ≤ 0.017; Fig. 6D) EMG. After CON, there were within-group increases in isomet-
and were also greater for NHT compared with LSET for 55° ric (+21%; P = 0.044), but not eccentric or concentric (paired
and 75° (LSD 0.035 ≤ P ≤ 0.040), but not at other angles t-test, 0.475 ≤ P ≤ 0.651), hamstring EMG. No group × time ef-
(0.062 ≤ P ≤ 0.766). Greater increases in absolute isometric fects were detected for eccentric or concentric hamstring EMG

FIGURE 4—Absolute changes (pre to post) in BFlh aponeurosis area (A) and maximum width (B) following lengthened state training (LSET, n = 14), NHT
(n = 13), and control (CON, n = 11) interventions. Symbols indicate differences in the magnitude of pre to post changes where post hoc tests displayed LSD
P < 0.05: *different from CON, §different from NHT. Data are means ± SE.

1900 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

 BASIC SCIENCES
FIGURE 5—Absolute changes (pre to post) in maximum knee flexion torque and hamstring EMG during eccentric, concentric, and isometric contractions
following lengthened state training (LSET, n = 14), NHT (n = 14), and control (CON, n = 14) interventions. Symbols indicate differences in the magnitude of
pre to post changes where post hoc tests displayed LSD P < 0.05: *different from CON. Data are means ± SE.

(ANCOVA 0.125 ≤ P ≤ 0.278). A significant group × time for increasing KF not HE size (approximately twofold vs
effect was observed for isometric hamstring EMG (ANCOVA LSET). The different pattern and magnitude of responses after
P = 0.046), but post hoc comparisons of absolute change LSET supported the first part of our hypothesis and suggests
data did not reveal any between-group differences (LSD that LSET is superior to NHT in inducing greater hypertro-
0.065 ≤ P < 1.00; Fig. 5B). phy of the hamstrings as well as the size of the BFlh muscle
and aponeurosis, potentially contributing to better sprint
DISCUSSION performance and protection against HSI, which frequently
occur within this muscle. However, contrary to the second
The main findings of this study were that LSET induced part of our hypothesis, there were no differences in knee
greater increases in the volume of the hamstrings and BFlh flexor eccentric strength gains between the two training re-
muscle as well as BFlh aponeurosis size than NHT. In addi- gimes, perhaps because of similar increases in overall KF
tion, there was a distinctly different pattern of hypertrophy be- muscle volume.
tween the training regimes, with larger increases in the BFlh Hypertrophic adaptations. After 12 wk of the inter-
and SM after LSET (more than threefold vs NHT), but greater vention, both LSET and NHT significantly increased the vol-
increases in BFsh and SAR after NHT (more than twofold vs ume of all the knee flexor muscles, except for SM after NHT
LSET). These hypertrophic differences between exercises ap- and the smallest muscle (POP) in both groups, whereas the
peared to be largely due to the functional role of the muscles; control group remained very consistent across all seven muscles
LSET was more effective for increasing KF and HE size (Table 1). However, there was no significant difference in overall
(more than twofold vs NHT) and NHT was more effective knee flexor volume changes between LSET and NHT (Fig. 2).

NORDIC HAMSTRING VS LENGTHENED STATE TRAINING Medicine & Science in Sports & Exercise® 1901
BASIC SCIENCES

FIGURE 6—Knee flexion maximum isometric torque–angle relationships pre and post (A) LSET (n = 14), (B) NHT (n = 14), and (C) control (CON, n = 14)
interventions. D, Absolute changes (pre to post) in maximum knee flexion torque at knee joint angles from 35° to 95° (0° = full extension). A–C, Symbols
denote significant within-group increases in torque from pre to post at the angle marked determined by paired t-tests as follows: *P < 0.05, **P < 0.01,
or ***P < 0.001. Data are means ± SD. D, Symbols indicate differences in the magnitude of pre to post changes where post hoc tests displayed LSD
P < 0.05: *different from CON, †different from LSET. Data are means ± SE.

The fact that overall knee flexor hypertrophy was similar for vs +6.5%) and SM (2.3-fold, +8.2% vs +3.6%) compared with
LSET and NHT may suggest that the hypertrophic stimulus a more modest difference in the ST (1.2-fold, +23.6% vs 19.3%).
was comparable between the two types of training despite their Kellis and Blazevich (19) suggest that the contribution of the
many differences, including different muscle lengths and pos- BFlh and SM to knee flexion torque production is much higher
tures, bilateral versus unilateral eccentrics, concentric contrac- than the other two constituents when the hamstrings are in a
tions with LSET (even if at a low load), and body weight vs lengthened position (i.e., in a hip-flexed and knee-extended po-
weight stack resistance. Despite the similar overall knee flexor sition; see Figs. 4 and 5 of Ref. [19]). Therefore, the current
hypertrophy, there were many differences between the train- study together with these previous studies (15,19) indicates
ing regimes for smaller muscle groups and individual muscles that LSET is the better choice than NHT when aiming to elicit
as discussed below in detail. hamstrings hypertrophy and especially of the constituent BFlh
LSET not only resulted in greater hypertrophy of the ham- and SM muscles.
strings compared with NHT (1.7-fold) but also produced a dif- Interestingly, NHT resulted in no/small hypertrophy of the
ferent pattern of hypertrophy between muscles, larger in- BFlh and SM (similar to CON and <LSET) but clear hypertro-
creases in BFlh (3.5-fold) and SM (9.7-fold), similar increases phy of the ST and BFsh (>CON, and similar or >LSET, re-
in the ST (1.3-fold), but smaller increases in BFsh (3.8-fold spectively). The lack of BFlh hypertrophy after NHT may be
greater after NHT; Fig. 3). Thus, this study found pronounced surprising based on acute EMG studies that indicate a high
evidence for training-specific adaptations in the amount and level of BFlh activation during this exercise (35,36). However,
pattern of hypertrophy with different knee flexion exercises. EMG studies may be misleading because of difficulties in accu-
In accordance with Maeo et al. (15,20), this suggests that exer- rately locating electrodes over individual muscles and cross-talk
cise selection can markedly affect the morphological changes (37). Using functional MRI, Bourne et al. (38) found BFlh and
with resistance training even when exercises involve the same SM activation during NHT to be significantly lower than the
joint action. The greater hamstrings hypertrophy after LSET BFsh and especially ST, which broadly mimics the pattern of hy-
than that after NHT and the pattern of the individual ham- pertrophy seen after NHT in the current study. Moreover, Bourne
strings muscle changes are similar to the findings of Maeo et al. et al. (39) observed similar hypertrophic effects of NHT to the
(15) who also found greater hamstrings muscle hypertrophy current study after 10 wk (20 sessions), with no changes in BFlh
after 12 wk of knee flexion training at long lengths (hip flexed, and SM compared with CON, but BFsh and ST showing marked
seated) versus short lengths (hip extended, prone), with the hypertrophy. The current study with nearly double the training
most pronounced differences for the BFlh (2.2-fold, +14.4% sessions (34 sessions in 12 wk) reinforces the finding that NHT

1902 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

produces no/negligible hypertrophy of the SM and BFlh but of the BFlh (muscle and aponeurosis size was unchanged)
substantial hypertrophy of the BFsh and ST (20–23%). but perhaps because increases in size and strength of the
In fact, NHT produced greater hypertrophy of the BFsh and other muscles (BFsh particularly, but also SAR, ST) reduce
SAR compared with LSET, with no between-group difference the demands placed on the BFlh. Finally, it is notable that
in GRA (Fig. 3) and POP showing no hypertrophic response to BFlh fascicle length has been shown to be associated, pro-
either type of training, perhaps because of either reduced accu- spectively, with HSI (longer BFlh fascicles, lower HSI
racy in assessing the volume of this small muscle, or its primary risks) (42), and NHT is reported to increase BFlh fascicle
role as a knee joint stabilizer rather than a knee flexor (40,41). length (14). Although the capability of LSET to increase BFlh
Collectively, the nonhip extending knee flexors (KF not HE; fascicle length is unknown, it is likely possible because in-
BFsh, SAR, GRA, POP) were more responsive to NHT (ap- creased muscle volume, which occurred in BFlh after LSET,
proximately twofold LSET), whereas the hip extending knee can result from both longitudinal and radial growth of muscles
flexors (KF and HE; BFlh, SM, ST) were more responsive (43). Thus, further research is needed to investigate whether
to LSET (more than twofold NHT). As discussed previously, BFlh aponeurosis size as well as fascicle length and their change
during LSET, the long length of the KF and HE muscles (i.e., after LSET and/or other training interventions are related to
biarticular hamstrings), but not the KF muscles that are not HE future HSI.
muscles, is the likely explanation for their differing hypertro- Functional adaptations. Maximum eccentric knee flex-
phic response to this type of training. Considering NHT, al- ion torque increased in LSET (+17%) and NHT (+11%) but

BASIC SCIENCES
though no convincing data are available, the lack of high ex- not in CON (+4%) (Table 2). Although only LSET increased
ternal resistance to hip extension during this exercise (i.e., hip eccentric strength compared with CON, there was no signifi-
extension torque is restrained by gravity acting on the trunk cant difference between LSET and NHT. Eccentric knee flexion
and antagonist co-activation) may limit the contribution/ strength is considered a key factor in HSI prevention (14,42,44),
activation of the hip extending knee flexors and place greater and the current results suggest that LSET and NHT may have
reliance on the nonhip extending knee flexors. This point is similar efficacy for improving eccentric strength. However, as
partly supported by the finding that peak forces during NHT with the BFlh aponeurosis size, longitudinal investigation of
coincided with low BFlh and SM muscle activities (40), sug- training-induced increases in eccentric knee flexion strength
gesting other muscles may be more heavily involved in this on HSI needs to be examined in future studies.
exercise, and this agrees with our finding of no hypertrophy Maximum isometric and concentric torque increased in all
of BFlh and SM after NHT. The SAR, being a biarticular groups including CON (Table 2), suggesting some learning ef-
hip flexor, would also have likely been at longer lengths dur- fects despite the familiarization session and two duplicate
ing NHT (hip extended) than LSET (hip flexed), which may measurement sessions at each time point in the current study.
also explain its greater hypertrophic response to NHT. Maeo This learning effect may be because the knee flexor muscle
et al. (15) also found greater SAR hypertrophy when trained group gets relatively low habitual use in daily life, particularly
by knee flexion exercise at long (hip extended, prone) vs for performing maximum contractions at long lengths where
short (hip flexed, seated) lengths, collectively indicating that the largest isometric strength improvements occurred. Our pre-
muscle lengths during exercise influence training-induced vious study (33) using the same approach (one familiarization
muscle hypertrophy. and two duplicate measurement sessions at each time point),
Aponeurosis adaptations. BFlh aponeurosis size, assessed but measurements of the knee extensors in the middle of the
as contact interface area and maximum width, had larger in- range of motion, did not find such learning effects in a control
creases after LSET (+8–9%) compared with NHT (+1–3%) group. Nevertheless, the greater gains in maximum isometric
and CON (+1–2%), with no significant difference found be- strength of both LSET and NHT compared with CON (Fig. 5A)
tween NHT and CON (Fig. 4). Although previous studies may be at least partly attributable to similar increases in over-
have found vastus lateralis aponeurosis size to increase with all knee flexor volume for LSET and NHT (Fig. 2), although
training (28,29), this is the first study to document training- this did not translate into between-group differences in con-
induced increases in BFlh aponeurosis size after LSET but centric strength. Changes in hamstring EMG during the max-
not NHT. As mentioned earlier, a small BFlh aponeurosis size imum contractions appeared to have a similar pattern to those
has been suggested as a risk factor for HSI by concentrating of maximum knee flexion torque (Fig. 5B), but none of these
mechanical strain on the surrounding muscle tissue (24–27). changes were significantly different between groups. This may
Given that NHT, which induced no/negligible increase in be partly because EMG measurements were from only two
BFlh aponeurosis size in this study, has been shown to be ef- hamstring muscles, whereas knee flexion torque is produced
fective in reducing the risk of new and recurrent HSI (16–18), by up to nine individual muscles. Indeed, training-induced
it is possible that LSET may be more effective than NHT in changes in EMG often align with those of torque when EMG
preventing strain injuries. However, it is also possible that a is taken from most of the muscles producing the intended torque
small BFlh aponeurosis size is unrelated to future injury occur- (33,45–48). Thus, future studies should consider more careful
rence, as currently there is no prospective study confirming familiarization (multiple sessions) and a greater range of EMG
this relationship. Another possibility is that the benefits of NHT measurement sites when assessing the knee flexor muscles in
reducing injury risk in the BFlh are not due to adaptations training studies.

NORDIC HAMSTRING VS LENGTHENED STATE TRAINING Medicine & Science in Sports & Exercise® 1903
 Isometric knee flexion torque increased at a wide range of concentrically lifting the load with two legs to eccentrically
joint angles after both LSET (+18–22%) and NHT (+25–29%), lower/return the load with one leg. Thus, LSET involved a signif-
whereas CON also increased torque at extended knee joint angles icant volume of concentric work, albeit at a relatively low load.
(+11–18%), again suggesting some learning effects (Fig. 6A–C). Nonetheless, despite these numerous differences between the
The isometric strength changes across the range of knee joint two training regimes, they produced similar overall knee flexor
angles were overall greater for both NHT (all angles) and LSET hypertrophy but very different patterns of hypertrophy within
(the five most extended angles out of the seven) than CON, and the individual muscles. Furthermore, adding more work/training
also greater for NHT than LSET at intermediate angles 55–75° volume to NHT seems unlikely to significantly affect SM and
(Fig. 5D). The reason for the differences between NHT and BFlh hypertrophy or BFlh aponeurosis size as discussed previ-
LSET is unclear but may be partly attributable to the fact that ously (Figs. 1 and 4). Finally, the distinct patterns of hypertrophy
NHT involves contracting at relatively short muscle lengths within the knee flexors after LSET vs NHT (e.g., hamstrings
than LSET. It should be recognized that the measurements in vs SAR) seem likely to be specific to the nature of the exercise
this study did not extend beyond the angle of peak knee flexion performed rather than the loading magnitude or volume per se.
torque. This was because of the difficulties in measuring knee
flexion torque at long muscle lengths due to the discrepancy CONCLUSIONS
in crank angle and actual knee joint angle; during MVC at ex- In summary, the main findings of this study were that LSET
tended angles, the discrepancy was >25°, likely resulting from induced greater increases in hamstring muscle size including
BASIC SCIENCES

the compliance and misalignment of the segments to the crank. larger increases in BFlh muscle volume (Fig. 3) and BFlh apo-
Manipulating hip joint angle (e.g., accentuated hip-flexed posi- neurosis size (Fig. 4). Moreover, the training regimes induced
tion, similar to LSET) during knee flexion torque measurements distinctly different patterns of hypertrophy that appeared to be
could help overcome this issue. This should be taken into ac- largely due to the functional role of the muscles; LSET was
count in future studies to better understand the effects of train- more effective for increasing KF and HE muscle size (2.2-fold
ing interventions including LSET and/or NHT on strength im- vs NHT) and NHT for increasing KF not HE size (1.9-fold vs
provements across wide joint angles/muscle lengths. LSET). These results suggest that LSET is superior to NHT for
Limitations. This study compared the effects of two ec- inducing hypertrophy of the hamstrings and BFlh muscle, po-
centrically focused knee flexor training regimes that are inher- tentially contributing to better sprint performance improve-
ently different exercises: e.g., loading mechanism (weight stack, ments and providing a stronger protective effect against HSI,
LSET vs mainly body weight, NHT), joint positions (hip flexed, which often occur in the BFlh muscle.
LSET vs hip extended, NHT), bilateral (NHT) vs unilateral
The results of the study are presented clearly, honestly, and without
(LSET) eccentrics, and concentric component (LSET only). fabrication, falsification, or inappropriate data manipulation. The results
Therefore, this study did not isolate a single experimental var- of the present study do not constitute endorsement by the American
iable; rather, it compared two quite distinct training regimes. College of Sports Medicine. The authors would like to thank all the vol-
unteers for their time and efforts in completing the study. This study
Given the differential training effects we have observed, fur- was supported by the National Institute of Education Academic Re-
ther studies should strive to isolate the specific variables ac- search Fund (RI 4/15 KPW), Singapore. The authors declare no con-
counting for these differences. Moreover, LSET was designed flicts of interest.
T.G.B., D.Z.N., P.W.K., M.T.G.P., and J.P.F. conceived and de-
to provide a practical (i.e., widely accessible) resistance train- signed the study. All authors contributed to data collection or analysis,
ing regime for high eccentric loading of the knee flexors at as well as interpretation of the results. S.M. drafted the manuscript, and
long lengths with a minor modification (adjusted backrest hip all authors read and approved the final version of the manuscript, includ-
ing the order of presentation of the authors. This study was approved by
angle) to a widely used type of knee flexion weight stack ma- the Loughborough University Ethics Review Sub-Committee (Ethics ap-
chine rather than sophisticated inaccessible equipment (e.g., a proval number R17-P054) and Nanyang Technological University Institu-
motorized isokinetic dynamometer) previously used for LSET tional Review Board (Reference number IRB-2017-07-030). Written in-
formed consent was obtained from each participant. All data are avail-
(20,21). However, to achieve high eccentric loading without able in the main text. Additional data related to this study will be made
motorized apparatus or manual assistance, LSET involved available from the corresponding author upon reasonable request.

REFERENCES
1. Morin J-B, Gimenez P, Edouard P, et al. Sprint acceleration mechan- 5. Elliott MCCW, Zarins B, Powell JW, Kenyon CD. Hamstring muscle
ics: the major role of hamstrings in horizontal force production. Front strains in professional football players: a 10-year review. Am J Sports
Physiol. 2015;6:404. Med. 2011;39(4):843–50.
2. Sugisaki N, Kobayashi K, Tsuchie H, Kanehisa H. Associations 6. Brooks JH, Fuller CW, Kemp SP, Reddin DB. Incidence, risk, and
between individual lower-limb muscle volumes and 100-m sprint time prevention of hamstring muscle injuries in professional rugby union.
in male sprinters. Int J Sports Physiol Perform. 2018;13(2):214–9. Am J Sports Med. 2006;34(8):1297–306.
3. Miller R, Balshaw TG, Massey GJ, et al. The muscle morphology of 7. Malliaropoulos N, Papacostas E, Kiritsi O, Papalada A, Gougoulias
elite sprint running. Med Sci Sports Exerc. 2021;53(4):804–15. N, Maffulli N. Posterior thigh muscle injuries in elite track and field
4. Ishoi L, Holmich P, Aagaard P, Thorborg K, Bandholm T, Serner A. athletes. Am J Sports Med. 2010;38(9):1813–9.
Effects of the Nordic hamstring exercise on sprint capacity in male 8. Ekstrand J, Hägglund M, Waldén M. Epidemiology of muscle inju-
football players: a randomized controlled trial. J Sports Sci. 2018; ries in professional football (soccer). Am J Sports Med. 2011;39(6):
36(14):1663–72. 1226–32.

1904 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

 9. Ekstrand J, Waldén M, Hägglund M. Hamstring injuries have in- 29. Balshaw TG, Funnell MP, McDermott EJ, et al. The effect of specific
creased by 4% annually in men's professional football, since 2001: bioactive collagen peptides on tendon remodeling during 15 wk of lower
a 13-year longitudinal analysis of the UEFA elite Club injury study. body resistance training. Med Sci Sports Exerc. 2023;55(11):2083–95.
Br J Sports Med. 2016;50(12):731–7. 30. Massey GJ, Balshaw TG, Maden-Wilkinson TM, Folland JP. Tendi-
10. Liu Y, Sun Y, Zhu W, Yu J. The late swing and early stance of sprinting nous tissue properties after short- and long-term functional overload:
are most hazardous for hamstring injuries. J Sport Health Sci. 2017; differences between controls, 12 weeks and 4 years of resistance
6(2):133–6. training. Acta Physiol (Oxf). 2018;222(4):e13019.
11. Chumanov ES, Schache AG, Heiderscheit BC, Thelen DG. Ham- 31. Craig CL, Marshall AL, Sjostrom M, et al. International physical ac-
strings are most susceptible to injury during the late swing phase of tivity questionnaire: 12-country reliability and validity. Med Sci Sports
sprinting. Br J Sports Med. 2012;46(2):90. Exerc. 2003;35(8):1381–95.
12. Schache AG, Dorn TW, Blanch PD, Brown NA, Pandy MG. Me- 32. Bouchard C, An P, Rice T, et al. Familial aggregation of VO(2max)
chanics of the human hamstring muscles during sprinting. Med Sci response to exercise training: results from the HERITAGE Family
Sports Exerc. 2012;44(4):647–58. Study. J Appl Physiol (1985). 1999;87(3):1003–8.
13. Chumanov ES, Heiderscheit BC, Thelen DG. The effect of speed and 33. Balshaw TG, Massey GJ, Maden-Wilkinson TM, Tillin NA, Folland
influence of individual muscles on hamstring mechanics during the JP. Training-specific functional, neural, and hypertrophic adaptations
swing phase of sprinting. J Biomech. 2007;40(16):3555–62. to explosive- vs. sustained-contraction strength training. J Appl Physiol
14. Cuthbert M, Ripley N, McMahon JJ, Evans M, Haff GG, Comfort P. (1985). 2016;120(11):1364–73.
The effect of Nordic hamstring exercise intervention volume on ec- 34. Vickers AJ. The use of percentage change from baseline as an out-
centric strength and muscle architecture adaptations: a systematic re- come in a controlled trial is statistically inefficient: a simulation study.
view and meta-analyses. Sports Med. 2020;50(1):83–99. BMC Med Res Methodol. 2001;1:6.
15. Maeo S, Huang M, Wu Y, et al. Greater hamstrings muscle hypertro- 35. Zebis MK, Skotte J, Andersen CH, et al. Kettlebell swing targets

BASIC SCIENCES
phy but similar damage protection after training at long versus short semitendinosus and supine leg curl targets biceps femoris: an EMG study
muscle lengths. Med Sci Sports Exerc. 2021;53(4):825–37. with rehabilitation implications. Br J Sports Med. 2013;47(18):1192–8.
16. Petersen J, Thorborg K, Nielsen MB, Budtz-Jorgensen E, Holmich P. 36. Keerasomboon T, Soga T, Hirose N. Influence of altered knee angle
Preventive effect of eccentric training on acute hamstring injuries in on electromyographic activity of hamstring muscles between Nordic
men's soccer: a cluster-randomized controlled trial. Am J Sports Med. hamstring exercise and Nordic hamstring exercise with incline slope
2011;39(11):2296–303. lower leg board. Int J Sports Phys Ther. 2022;17(5):832–40.
17. van der Horst N, Smits DW, Petersen J, Goedhart EA, Backx FJ. The 37. Vigotsky AD, Halperin I, Lehman GJ, Trajano GS, Vieira TM. Inter-
preventive effect of the Nordic hamstring exercise on hamstring inju- preting signal amplitudes in surface electromyography studies in
ries in amateur soccer players: a randomized controlled trial. Am J sport and rehabilitation sciences. Front Physiol. 2017;8:985.
Sports Med. 2015;43(6):1316–23. 38. Bourne MN, Opar DA, Williams MD, Al Najjar A, Shield AJ. Mus-
18. Arnason A, Andersen TE, Holme I, Engebretsen L, Bahr R. Preven- cle activation patterns in the Nordic hamstring exercise: impact of
tion of hamstring strains in elite soccer: an intervention study. Scand prior strain injury. Scand J Med Sci Sports. 2016;26(6):666–74.
J Med Sci Sports. 2008;18(1):40–8. 39. Bourne MN, Duhig SJ, Timmins RG, et al. Impact of the Nordic ham-
19. Kellis E, Blazevich AJ. Hamstrings force-length relationships and string and hip extension exercises on hamstring architecture and mor-
their implications for angle-specific joint torques: a narrative review. phology: implications for injury prevention. Br J Sports Med. 2017;
BMC Sports Sci Med Rehabil. 2022;14(1):166. 51(5):469–77.
20. Maeo S, Wu Y, Huang M, et al. Triceps brachii hypertrophy is substan- 40. Jadhav SP, More SR, Riascos RF, Lemos DF, Swischuk LE. Com-
tially greater after elbow extension training performed in the overhead prehensive review of the anatomy, function, and imaging of the
versus neutral arm position. Eur J Sport Sci. 2023;23(7):1240–50. popliteus and associated pathologic conditions. Radiographics. 2014;
21. Oranchuk DJ, Storey AG, Nelson AR, Cronin JB. Isometric training 34(2):496–513.
and long-term adaptations: effects of muscle length, intensity, and 41. Nyland J, Lachman N, Kocabey Y, Brosky J, Altun R, Caborn D.
intent: a systematic review. Scand J Med Sci Sports. 2019;29(4): Anatomy, function, and rehabilitation of the popliteus musculotendinous
484–503. complex. J Orthop Sports Phys Ther. 2005;35(3):165–79.
22. Schmitt B, Tim T, McHugh M. Hamstring injury rehabilitation and 42. Timmins RG, Bourne MN, Shield AJ, Williams MD, Lorenzen C,
prevention of reinjury using lengthened state eccentric training: a Opar DA. Short biceps femoris fascicles and eccentric knee flexor
new concept. Int J Sports Phys Ther. 2012;7(3):333–41. weakness increase the risk of hamstring injury in elite football (soccer):
23. Tyler TF, Schmitt BM, Nicholas SJ, McHugh MP. Rehabilitation af- a prospective cohort study. Br J Sports Med. 2016;50(24):1524–35.
ter hamstring-strain injury emphasizing eccentric strengthening at 43. Jorgenson KW, Phillips SM, Hornberger TA. Identifying the struc-
long muscle lengths: results of long-term follow-up. J Sport Rehabil. tural adaptations that drive the mechanical load-induced growth of
2017;26(2):131–40. skeletal muscle: a scoping review. Cells. 2020;9(7):1658.
24. Evangelidis PE, Massey GJ, Pain MTG, Folland JP. Biceps femoris 44. Lee JWY, Mok KM, Chan HCK, Yung PSH, Chan KM. Eccentric
aponeurosis size: a potential risk factor for strain injury? Med Sci hamstring strength deficit and poor hamstring-to-quadriceps ratio are
Sports Exerc. 2015;47(7):1383–9. risk factors for hamstring strain injury in football: a prospective study
25. Kellis E. Intra- and inter-muscular variations in hamstring architec- of 146 professional players. J Sci Med Sport. 2018;21(8):789–93.
ture and mechanics and their implications for injury: a narrative re- 45. Lanza MB, Balshaw TG, Folland JP. Is the joint-angle specificity of
view. Sports Med. 2018;48(10):2271–83. isometric resistance training real? And if so, does it have a neural ba-
26. Rehorn MR, Blemker SS. The effects of aponeurosis geometry on sis? Eur J Appl Physiol. 2019;119(11–12):2465–76.
strain injury susceptibility explored with a 3D muscle model. J Biomech. 46. Maeo S, Shan X, Otsuka S, Kanehisa H, Kawakami Y. Neuromuscu-
2010;43(13):2574–81. lar adaptations to work-matched maximal eccentric versus concentric
27. Fiorentino NM, Epstein FH, Blemker SS. Activation and aponeurosis training. Med Sci Sports Exerc. 2018;50(8):1629–40.
morphology affect in vivo muscle tissue strains near the myotendinous 47. Maeo S, Shan X, Otsuka S, Kanehisa H, Kawakami Y. Single-joint ec-
junction. J Biomech. 2012;45(4):647–52. centric knee extension training preferentially trains the rectus femoris
28. Wakahara T, Ema R, Miyamoto N, Kawakami Y. Increase in vastus within the quadriceps muscles. Transl Sports Med. 2018;1(5):212–20.
lateralis aponeurosis width induced by resistance training: implica- 48. Maeo S, Yoshitake Y, Takai Y, Fukunaga T, Kanehisa H. Neuromus-
tions for a hypertrophic model of pennate muscle. Eur J Appl Physiol. cular adaptations following 12-week maximal voluntary co-contraction
2015;115(2):309–16. training. Eur J Appl Physiol. 2014;114(4):663–73.

NORDIC HAMSTRING VS LENGTHENED STATE TRAINING Medicine & Science in Sports & Exercise® 1905