You are on page 1of 12

Anthony J. Blazevich, Dale Cannavan, David R.

Coleman and Sara Horne
J Appl Physiol 103:1565-1575, 2007. First published Aug 23, 2007;

You might find this additional information useful...

This article cites 61 articles, 27 of which you can access free at:

Updated information and services including high-resolution figures, can be found at:

Additional material and information about Journal of Applied Physiology can be found at:

This information is current as of November 18, 2007 .

Downloaded from on November 18, 2007

Journal of Applied Physiology publishes original papers that deal with diverse areas of research in applied physiology, especially
those papers emphasizing adaptive and integrative mechanisms. It is published 12 times a year (monthly) by the American
Physiological Society, 9650 Rockville Pike, Bethesda MD 20814-3991. Copyright © 2005 by the American Physiological Society.
ISSN: 8750-7587, ESSN: 1522-1601. Visit our website at
J Appl Physiol 103: 1565–1575, 2007.
First published August 23, 2007; doi:10.1152/japplphysiol.00578.2007.

Influence of concentric and eccentric resistance training on architectural
adaptation in human quadriceps muscles
Anthony J. Blazevich, Dale Cannavan, David R. Coleman, and Sara Horne
Centre for Sports Medicine and Human Performance, School of Sport and Education, Brunel University, Uxbridge,
United Kingdom
Submitted 30 May 2007; accepted in final form 15 August 2007

Blazevich AJ, Cannavan D, Coleman DR, Horne S. Influence of have performed chronic eccentric work (16, 36, 37), whereas
concentric and eccentric resistance training on architectural adaptation in decreases (16) or a lack of change (36) is shown in those that
human quadriceps muscles. J Appl Physiol 103: 1565–1575, 2007. First worked concentrically. Also, rightward shifts in the force-
published August 23, 2007; doi:10.1152/japplphysiol.00578.2007.— length relationship of muscle have been shown after periods of
Studies using animal models have been unable to determine the eccentric muscle training (37), which has been suggested to be
mechanical stimuli that most influence muscle architectural adapta-
attributable to an increase in serial sarcomere number. These
tion. We examined the influence of contraction mode on muscle
architectural change in humans, while also describing the time course results are strongly indicative of contraction mode being a
of its adaptation through training and detraining. Twenty-one men primary stimulus for fiber or fascicle length change, and there

Downloaded from on November 18, 2007
and women performed slow-speed (30°/s) concentric-only (Con) or is some acceptance of the proposition that eccentric training
eccentric-only (Ecc) isokinetic knee extensor training for 10 wk increases muscle fiber length. Nonetheless, methodological
before completing a 3-mo detraining period. Fascicle length of the difficulties involved with clearly establishing the contraction
vastus lateralis (VL), measured by ultrasonography, increased simi- mode of the measured muscles [chiefly the rat vastus interme-
larly in both groups after 5 wk (⌬Con ⫽ ⫹6.3 ⫾ 3.0%, ⌬Ecc ⫽ ⫹3.1 ⫾ dius (VI)] limit the strength of this evidence (see e.g., Ref. 16).
1.6%, mean ⫽ ⫹4.7 ⫾ 1.7%; P ⬍ 0.05). No further increase was Furthermore, increases in fiber length (or sarcomere number,
found at 10 wk, although a small increase (mean ⬃2.5%; not signif- its surrogate when assuming a fixed sarcomere length) have
icant) was evident after detraining. Fascicle angle increased in both been reported in muscles that were exercised through large
groups at 5 wk (⌬Con ⫽ ⫹11.1 ⫾ 4.0%, ⌬Ecc ⫽ ⫹11.9 ⫾ 5.4%, excursions (34), which is suggestive of muscle excursion range
mean ⫽ 11.5 ⫾ 3.2%; P ⬍ 0.05) and 10 wk (⌬Con ⫽ ⫹13.3 ⫾ 3.0%,
(or absolute length) being a primary stimulus. Alternatively,
⌬Ecc ⫽ ⫹21.4 ⫾ 6.9%, mean ⫽ 17.9 ⫾ 3.7%; P ⬍ 0.01) in VL only
and remained above baseline after detraining (mean ⫽ 13.2%); movement velocity has been implicated. In humans, vastus
smaller changes in vastus medialis did not reach significance. The lateralis (VL) fascicle length has been shown to be highly
similar increase in fascicle length observed between the training predictive of 100-m time in well-trained sprint runners (35),
groups mitigates against contraction mode being the predominant and increases in fascicle length were reported both in a group
stimulus. Our data are also strongly indicative of 1) a close association of athletes who removed heavy resistance training and in-
between VL fascicle length and shifts in the torque-angle relationship creased their high-speed training (9) and in a group of resis-
through training and detraining and 2) changes in fascicle angle being tance trainers who performed the concentric phase of their lifts
driven by space constraints in the hypertrophying muscle. Thus with maximum velocity (3). Contraction velocity has also been
muscle architectural adaptations occur rapidly in response to resis- demonstrated to be an influencing factor in smooth muscle cell
tance training but are strongly influenced by factors other than growth (18). Nonetheless, the influence of muscle contraction
contraction mode. velocity has yet to be studied in animals, and the elucidation of
skeletal muscle; fascicle; muscle adaptation; muscle strength the most significant stimulus for fascicle length adaptation has
not been possible with the use of animal models. Thus the
possibility remains that a range of mechanical stimuli influence
A MUSCLE’S FORCE GENERATING properties are strongly influenced fascicle length.
by the architectural arrangement of its fascicles. Therefore, In humans, the measurement of fascicle length in vivo with
given that muscle architecture is highly plastic, understanding ultrasonography has revealed marked lengthening in both
the mechanical stimuli that influence it is of primary impor- younger (3, 53) and older (47) populations after prolonged
tance for those invested in optimizing muscle function and periods of resistance training, and fascicle shortening has been
movement performance. In particular, fiber, or fascicle, length reported after gastrocnemius recession in spastic diplegia pa-
has a profound effect on muscle excursion range, maximum tients (56). The reliable measurement of fascicle length in
shortening velocity (and thus muscle power), and the force- humans offers the opportunity to use human models, which are
length relationship. Despite the functional importance of fas- associated with few of the methodological limitations inherent
cicle length, the most important mechanical signal required to in animal models. Importantly, contraction mode, muscle ex-
stimulate its adaptation has not been identified. cursion, and contraction velocity during a movement can be
Examination of the effects of a range of mechanical stimuli easily regulated by using isokinetic dynamometry to system-
has typically been performed with animal models. Researchers atically examine the influence of each parameter, and the yet to
have reported increases in fiber length in muscles purported to be described temporal response of fascicle length change can

Address for reprint requests and other correspondence: A. J. Blazevich, Centre The costs of publication of this article were defrayed in part by the payment
for Sports Medicine and Human Performance, School of Sport and Education, of page charges. The article must therefore be hereby marked “advertisement”
Brunel Univ., Uxbridge UB8 3PH, UK (e-mail: in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

http://www. 8750-7587/07 $8.00 Copyright © 2007 the American Physiological Society 1565

be examined since repeated measurements are possible with tric-only (Ecc) knee extension training group. The subjects were
noninvasive methods. In lieu of the above, the primary purpose tested for strength, muscle size, and muscle architecture at weeks 0, 5,
of the present study was to test the hypothesis that fascicle and 10 of the training period and again after 3 mo (14 wk) of
length changes are unique to the training contraction mode. detraining. Tests at 5 and 10 wk of training were conducted 4 (muscle
architecture), 4 –5 (strength testing), and 7 (muscle volume; week 10
The hypothesis would be confirmed if increases in fasci-
only) days after the final training session of that training period; tests
cle length result from eccentric training but a decrease or no were performed at the same time of day as at week 0. Three subjects
change results from concentric training. Given that the time (2 men and 1 woman) withdrew from the training for reasons unre-
course of fascicle length change has yet to be described in lated to the study; 10 and 11 subjects completed Con and Ecc training,
humans, we measured, for the first time, fascicle length respectively.
changes through periods of training (10 wk) and subsequent Resistance training. The subjects performed four (weeks 1–3), five
detraining (3 mo). (weeks 4 –7), or six (weeks 8 –10) sets of six maximal concentric (Con
While fascicle length impacts significantly on the length group) or eccentric (Ecc group) knee extension repetitions three times
range and speed of active force production, a muscle’s fiber, or a week on an isokinetic dynamometer (Biodex system 3, Biodex
fascicle, angle is also of great functional importance. It affects Medical Systems); in the first three sessions the subjects exercised at
50%, 70%, and then 90% of their pretraining maximum to minimize
force production by allowing a greater amount of contractile
muscle damage resulting from the unaccustomed heavy exercise. At
tissue to attach to a given area of tendon or aponeurosis, and by least 1 day separated each session. The Con group trained by extend-
increasing the muscle’s architectural gearing ratio (AGR; ratio ing their knee “as fast and hard as possible,” although the dynamom-
of longitudinal muscle strain to fascicle strain, εx/εf, during eter lever arm was set to move at 30°/s, from the maximum knee
contraction) to allow fascicles the opportunity to produce flexion angle allowed by the dynamometer-seat setup (⬃100°) to

Downloaded from on November 18, 2007
forces at more optimum lengths and shortening speeds (5, 8, maximum knee extension (5–10°) before returning their leg to the
28, 42, 61). Increases in fascicle angle have been reported after start position with a small knee flexion contraction. The Ecc group
periods (14 –16 wk) of heavy resistance training in humans (1, performed a small knee extension contraction (⬃20 Nm) to move the
30), although studies of shorter duration have been unable to leg to an extended (10 –15°) position before maximally extending the
agree on its temporal response (10, 53). It is also not clear how knee to resist the subsequent downward movement of the lever arm of
different muscles within a synergistic group respond to the the dynamometer. The subjects were asked to “stop the dynamometer”
by applying a maximum upward force, although the dynamometer
same stimulus. With respect to the mechanisms that promote lever arm continued to move at 30°/s throughout; the contractions
fascicle angle change, a common theory is that it varies from were completed at a knee angle of ⬃100°. The knee joint ranges of
its “genetic set point” concomitantly with muscle size. The motion during the concentric and eccentric training were greater than
theory is given credence by the strong relationship shown normally encountered during walking (⬃25°; Ref. 7), jogging (⬃55°;
between muscle thickness or cross-sectional area and fascicle Ref. 7), and countermovement jumping (⬃80°; Ref. 11) but only
angle (29, 30) and their concordant change with training (1, slightly greater than stair climbing (⬃88° for 25.5-cm stair; Ref. 4)
30). Alternatively, alterations in AGR resulting from fascicle and so can be considered as being a loading stimulus of a greater than
angle adaptations could influence force generation indepen- normal range of motion. One minute of passive rest was allowed
dently of hypertrophy, so mechanisms unrelated to hypertro- between sets for both training groups, and repetitions within a set were
phy should be considered. A second purpose of the present performed continuously. Knee angles were measured as the angle of
a line joining the lateral epicondyle of the tibia to the lateral malleolus
study, therefore, was to describe the time course of fascicle and a line joining the lateral condyle of the femur to the greater
angle change in two synergist muscles [VL and vastus medialis trochanter. The knee joint center was aligned to the center of rotation
(VM)] and to determine whether it is most strongly related to of the dynamometer, and the dynamometer lever arm was securely
changes in muscle size rather than being an independent attached to the shank at ⬃50 mm above the lateral malleolus. Straps
anatomic adaptation to mechanical loading. were placed firmly across the chest and waist of the subjects to
minimize extraneous movement. Individual positioning of the subject
METHODS with respect to backrest inclination, seat height, and lever arm length
was held constant between test sessions. Loud verbal encouragement
Subjects. Thirty-three recreationally active men and women volun- was given for each repetition, and set-by-set feedback of results was
teered for the study, with 12 men [age ⫽ 24.2 ⫾ 5.7 yr (mean ⫾ SD), provided to aid motivation. These results could be compared with
height ⫽ 178.7 ⫾ 5.9 cm, weight ⫽ 73.6 ⫾ 6.2 kg] and 12 women personal best performances, which were known by the subjects. No
(age ⫽ 21.3 ⫾ 4.3 yr, height ⫽ 160.8 ⫾ 4.3 cm, weight ⫽ 63.5 ⫾ strength training was performed during the 3-mo detraining period,
10.3 kg) being assigned to one of two training groups and 4 men and although the subjects maintained their normal daily activities.
5 women being assigned to a nontraining control group. None of the Fascicle strain during training. To determine the fascicle strain
subjects had a significant lower limb injury within the previous 5 yr, induced by the training contractions, maximum and minimum fascicle
had inflammatory conditions or hypertension, had performed weight lengths were measured at midthigh in the VL in six subjects (3 men
training, had manual lifting-oriented occupations, or performed vig- and 3 women) during concentric and eccentric contractions. An
orous exercise more than four times a week. All subjects gave written ultrasound transducer (45-mm linear array, 10 MHz; Megas GPX,
informed consent before the study, which was approved by the Human Esaote, Italy) was fixed to a foam cast and secured over the midpoint
Research Ethics Committee within the School of Sport and Education of the VL by elastic strapping. Sufficient water-soluble gel was
at Brunel University and was conducted in accordance with the applied to the transducer to aid acoustic coupling and remove the need
Declaration of Helsinki. for dermal contact. The transducer was aligned in the direction of the
Study design. After the subjects were familiarized with the testing muscle fascicles; appropriate orientation was achieved when muscle
procedures, their concentric and eccentric strength was tested, as fascicles could be traced clearly across the ultrasound image. A series
described below. Those allocated to the training groups were paired of 200 images (29-Hz sampling frequency) were taken during each of
with respect to their sex and pretraining torque production (mean of three maximal concentric and eccentric knee extension contractions
peak concentric and eccentric torque measured at 30°/s) and then for each subject. Sonographs containing images of the muscle fasci-
randomly allocated to either the concentric-only (Con) or the eccen- cles at their shortest and longest lengths were used for analysis.

J Appl Physiol • VOL 103 • NOVEMBER 2007 •
Muscle fascicle length was determined by the procedures described in aid acoustic coupling and remove the need to contact the skin; this
Fascicle length measurement below. Strain was calculated as the eliminated deformations of the muscle that can occur when pressure is
percent change from the shortest to longest fascicle length. directly applied to the skin and underlying muscle. Scans were
Strength testing. After warm-up, maximum concentric and eccen- performed on the right leg with the transducer oriented parallel to the
tric knee extension torque was measured at 30°/s with the setup muscle fascicles and perpendicular to the skin. Thus the transducer’s
described above. The subjects performed three consecutive contrac- orientation relative to the longitudinal axis of the thigh was different
tions with 2 min of rest separating the concentric and eccentric tests. between subjects. Transducer alignment was considered appropriate
Torque and position data were calibrated and sampled at a 2,000-Hz when several fascicles could be easily delineated without interruption
analog-digital conversion rate and stored on a computer running across the image. Sonographs were taken in the middle of the VL and
Spike2 software (Cambridge Electronic Design, Cambridge, UK). The the VM at 50% of the distance from the central palpable point of the
data were subsequently filtered with a digital fourth-order, zero-lag greater trochanter to the lateral condyle of the femur. At week 0,
Butterworth filter with a 14-Hz cutoff frequency (62); peak torque scanning locations were mapped onto a malleable plastic sheet, along
produced during each set was taken as the criterion measure. with other distinguishing landmarks (e.g., border of patella, freckles,
Torque-angle relationship. To assess changes in the knee extension scars, etc.) to ensure that repeated scans were taken from the same
torque-angle relationship after training, torque was recorded for each site. One difficulty with repeated scanning is that small changes in the
10° of the range of motion for the best concentric and eccentric trials. orientation of the transducer between test occasions can result in
Changes in both absolute and normalized torque (i.e., expressed significant variation in the muscle architecture recorded on the image.
relative to the peak torque recorded at any angle) were determined at To minimize this, we compared on-screen images at the 5 wk, 10 wk,
each angle. This approach, similar to that used by Butterfield and and detraining testing sessions to those obtained at week 0. Echoes
Herzog (15) and Lynn et al. (37), was used in place of a measurement from interspaces among the fascicles and heterogeneities in the
of the change in angle of peak torque since the torque-angle curves

Downloaded from on November 18, 2007
subcutaneous adipose tissue are unique to the transducer orientation
generated by the subjects fluctuated slightly throughout the range of (shown in Fig. 1), and these heterogeneities were easily visualized
motion, and particularly during the flatter portion of the curve nearing after the training and detraining periods in ⬎80% of the subjects. Thus
peak torque. These small fluctuations substantially affected the reli- the transducer could be oriented so that these markings were the same
ability of the measure of angle of peak torque. The average shift in the in both sets of images, which provided confidence that transducer
torque-angle relationship for each subject was also calculated by orientation discrepancies were small and were unlikely to account for
averaging the normalized concentric and eccentric torque values at measured pre-/posttraining architectural differences. All measure-
each angle and then calculating their change from the pretraining ments were carried out by the same experienced sonographer.
values; increases at 80 and 90° and decreases at 30, 40, 50, 60, and 70° Fascicle length measurement. Fascicle length was measured only
were taken to contribute to positive torque-angle shifts, so the pro- in the VL, as the complex architecture of the VM ensured that the
portional change was calculated as: reliability of length measurements was unacceptably low (8, 10).
兺共CE 80,90°mid-detrain ⫺ CE80,90°pre兲 Fascicle length was defined as the length of the fascicular path
between the superficial and deep aponeurosis. In most cases, the
⫺ 兺共CE30 – 70°mid-detrain ⫺ CE30 – 70°mid-detrain兲 fascicles extended off the acquired image. The length of the missing
portion was estimated by linear extrapolation. This was done by
where CE is the average of the normalized concentric and eccentric measuring the linear distance from the identifiable end of a fascicle to
torques at a specific movement speed and subscripts 30 –90° refer to the intersection of a line drawn from the fascicle and a line drawn
the joint angles at which the normalized values were calculated; the from the superficial aponeurosis (see, e.g., Refs. 38, 51). The error
subscript pre indicates that the measure was taken at pretraining (i.e., involved with this technique has been shown to be reasonably low
0 wk), whereas the subscript mid-detrain indicates that measures were (⬃2.3%) in contracted tibialis anterior (46), where fascicle curvature
taken at midtraining, posttraining, and detraining (i.e., 5 wk, 10 wk is significant. Given that fascicles in the relaxed VL are relatively
and at detraining). The average shift was then recorded for each straight (see Fig. 1), we estimate that our error would be somewhat
subject and plotted (see Fig. 10 for example). smaller. Repeated measurements yielded a coefficient of variation
Muscle architecture. In vivo muscle architecture was examined (CV) of 1.7% (⬃1.4 mm) and were performed by a single researcher.
with two-dimensional (2D), B-mode ultrasonography (Megas GPX, Fascicle angle measurement. Images were analyzed with publicly
Esaote) with a 45-mm linear array transducer (10 MHz). To obtain the available imaging software (ImageJ 1.36b, National Institutes of
images, subjects lay supine with their legs fully extended and their Health; free to download from Land-
muscles relaxed. A water-soluble gel was applied to the transducer to marks corresponding to the muscle fascicles and aponeurosis were

Fig. 1. Sonographs of vastus lateralis (VL) at
0 wk and 10 wk. Lines delineating the apo-
neurosis and visible fascicle shown on the 0
wk image (left) are used to measure fascicle
length and angle (␪). Arrows on the edges of
the image show the muscle thickness (MT;
mean of distance between superficial and deep
aponeurosis measured at each edge of the
sonograph); a slight increase in muscle thick-
ness is apparent between 0 wk and 10 wk for
this subject. Adipose and connective tissue
marks as well as visible blood vessels are
circled on the 10 wk image (right). Visibility
of these markings is strongly dependent on the
transducer angle and can be used to ensure
repeatability of transducer placement across
time points.

J Appl Physiol • VOL 103 • NOVEMBER 2007 •

Fig. 2. MRI scans of the quadriceps muscles.
T2-weighted images were obtained for the
measurement of both anatomic cross-sectional
area (for later calculation of muscle volume
and VL physiological cross-sectional area) and
the shift in gray scale (image density), which
indicates the level of muscle activity within
days of the tests (see METHODS). Perimeters of
the 4 vastii muscles are shown: VL, vastus
intermedius (VI), rectus femoris (RF), and
vastus medialis (VM). In these images, there is
an increase in the whole quadriceps area but
no change in the gray scale of the image.

digitized (Fig. 1). Two points on each fascicle were digitized, one ⬃3 shifts (38, 43); such shifts have been shown to result both from

Downloaded from on November 18, 2007
mm from the deep aponeurosis and the second at 50% of the distance damage to the contractile apparatus as a result of intense exercise (23,
from the deep to superficial aponeurosis. This allowed accurate 49, 54) and from activation-dependent myocellular metabolism (2, 31,
delineation of the fascicles without incorporating the slightly greater 55). This shift can be easily quantified by examining changes in the
fascicle curvature that can occur at the insertion point of the fascicles image density of T2-weighted MRI images. Although subjects were
on the deep aponeurosis (58). Repeated measurements yielded a CV asked not to exercise before the testing, we examined the image
of 1.6% (⬃0.32°). density (sum of gray scale values of all pixels in the region of interest
Muscle volume and anatomic and physiological cross-sectional divided by total number of pixels) of MRI images by tracing the
area. At weeks 0 and 10, axial T2-weighted MRI scans of the thigh whole quadriceps muscles on images taken at 50% of the muscle
were obtained at 16-mm intervals from the superior border of the length (i.e., at the same point as muscle architecture measurements) to
patella to the greater trochanter [3.0 T Magnetom, Siemens, Berlin, ensure that prior exercise and testing did not influence the cross-
Germany; repetition time (TR) 4,260 ms, echo time (TE) 95 ms, sectional area measures. One subject showed a 12.8% shift in image
averages: 3, field of view 200 ⫻ 200, slice thickness 4 mm, slice density and was omitted from further analysis; the average shift for
separation 14 mm, center-to-center slice distance 16 mm] from the the remaining subjects was ⫺0.49 ⫾ 0.8% (SE), so fluid shifts are
training subjects. Technical difficulties with the scanner at 10 wk unlikely to have influenced the results.
precluded the acquisition of images from the control subjects with Muscle thickness. In addition to the MRI measurements, muscle
sufficient contrast for analysis, although a high stability in MRI-based thickness was also measured at weeks 0, 5, and 10 and after the
cross-sectional area measurements has previously been demonstrated detraining period from the obtained 2D ultrasound images. Muscle
over a 10-wk period [sum of 7 levels of cross section: pre ⫽ 323.7 ⫾ thickness was defined as the mean of the distances between superficial
52.8 (SD), post ⫽ 320.9 ⫾ 53.0 cm2] in subjects who do not change and deep aponeurosis measured at the ends of each 45-mm wide
their exercise patterns (25). For training subjects, the scans were taken sonograph. The method and its reliability have been reported previ-
with a standard body coil, with the subjects lying supine with the legs ously (8, 10).
fully extended; example scans are shown in Fig. 2. From these scans, Statistical analysis. Analyses were performed with SPSS Version
anatomic muscle cross-sectional area was measured for each slice by 13.0 (SPSS, Chicago, IL). Temporal changes in strength, muscle
manually tracing the perimeter of each muscle with Scion Image architecture, and torque-angle variables were examined simulta-
for Windows (free to download at neously by multivariate analysis of variance with repeated measures.
fr_download_now.htm). Care was taken to exclude adipose and Significant main and interaction effects were further examined with
connective tissue incursions. In proximal images, clear delineation of paired (for resolving time-related changes)- or independent (for re-
the vastii muscles was not always possible because of a lack of solving group-related differences)-samples t-tests. Differences in the
observable intermuscular septum. In these cases, a line was drawn change in cross-sectional area measured by MRI at different muscle
from the end of the observable septum to a landmark on the muscle’s locations were tested with paired-samples t-tests. Greenhouse-Geisser
perimeter where the septum had intersected in distal images. Each adjustment was applied on occasions when the assumption of sphe-
slice was measured three times, with the median cross-sectional area ricity was violated (Mauchly’s test of sphericity, P ⬍ 0.05). Unless
being taken as representative; repeated measurements were very otherwise stated, all parameters are reported as means ⫾ SE. Statis-
reliable (CVVL ⫽ 1.5%, CVVM ⫽ 1.9%). Whole quadriceps cross- tical significance was set at P ⬍ 0.05, although the results of statistical
sectional area was calculated as the sum of each of the four muscle tests associated with an ␣-level of ⬍0.1 are discussed.
cross sections, and individual muscle and whole quadriceps volumes
were calculated by summing the product of cross-sectional area and RESULTS
slice thickness of all images. For VL only, physiological cross-
sectional area (PCSA) was also calculated with the equation: Strength changes. Although there was no change in the
control group [⌬Con ⫽ ⫹0.1 ⫾ 2.3 Nm (0.5 ⫾ 1.8%); ⌬Ecc ⫽
PCSA ⫽ ⫹7.5 ⫾ 5.5 Nm (3.0 ⫾ 3.2%)], peak torque recorded in
FL ⫻ cos␪ concentric and eccentric contractions increased significantly
where Vol is the muscle volume, FL is its fascicle length, and ␪ is the after 5 and 10 wk of training in both training groups, as shown
fascicle angle. in Fig. 3. The increase in concentric torque was greater in the
Intense exercise performed within several days of the MRI scan can Con group (24.1 ⫾ 4.2%) than in the Ecc group (16.4 ⫾ 5.1%)
influence muscle volume via osmotically driven intracellular fluid (P ⬍ 0.05), but there was no difference in the increase in
J Appl Physiol • VOL 103 • NOVEMBER 2007 •
no shifts in the torque-angle relationship for the control sub-
Muscle size. Whole quadriceps volume increased from
2,342.9 ⫾ 111.2 to 2,586 ⫾ 114.8 cm3 (10.2%; P ⬍ 0.001)
after 10 wk of training, with no difference between training
groups. VL [825.6 ⫾ 40.9 to 916.9 ⫾ 43.2 cm3 (11.1%); P ⬍
0.001] and VM [486.2 ⫾ 26.8 to 558.3 ⫾ 27.7 cm3 (14.8%);
P ⬍ 0.001] volume also increased significantly. These in-
creases in volume were not related to pretraining muscle
volume. As a result of both volume and architectural changes,
the PCSA of VL increased from 115.1 ⫾ 5.3 to 124.2 ⫾ 4.9
cm2 (7.9%; P ⬍ 0.01).
Although the focus of this study was on VL and VM, we
took the opportunity to examine region-specific changes in all
of the quadriceps muscles by comparing anatomic cross-sec-
tional changes measured at proximal (mean of the 3 slices
nearest to 25% of the distance from the proximal end point of
Fig. 3. Strength increases (both groups combined) after 10 wk of knee the muscle) to distal (75% from proximal end point of the
extension training and 14 wk of detraining. Both concentric and eccentric muscle) locations. Whereas increases in cross section of VL

Downloaded from on November 18, 2007
torque increased; the increase in concentric strength was greater in the
concentric-only (Con) group than the eccentric-only (Ecc) group (P ⬍ 0.05;
and rectus femoris were relatively consistent along their
data not shown). Torque production decreased significantly in eccentric con- lengths, as shown in Table 1, there was a trend (P ⫽ 0.08)
tractions after the detraining period (dashed line) only (P ⬍ 0.05); torque toward a greater increase distally in VM (⌬VMprox ⫽ ⫹1.1 ⫾
produced after detraining was significantly greater than at 0 wk (P ⬍ 0.01). 0.4 cm2, ⌬VMdist ⫽ ⫹3.0 ⫾ 0.9 cm2). There was also a trend
Significant difference from 0 wk to 10 wk: **P ⬍ 0.01, ***P ⬍ 0.001. toward a greater increase distally in VI when calculated as an
absolute change (⌬VIprox ⫽ 0.0 ⫾ 0.5, ⌬VIdist ⫽ ⫹2.2 ⫾ 0.8
eccentric strength between Con and Ecc groups (⌬Con ⫽ cm2; P ⫽ 0.055), and a significantly greater increase distally
⫹35.9 ⫾ 12.7%; ⌬Ecc ⫽ ⫹38.9 ⫾ 14.2%). when calculated as a relative change (⌬VIprox ⫽ ⫹1.0 ⫾ 2.1%,
After 3 mo of detraining, there was a nonsignificant decrease ⌬VIdist ⫽ ⫹13.6 ⫾ 4.8%; P ⬍ 0.05). While individual variation
in concentric torque of 5.6 ⫾ 2.7% (P ⫽ 0.084) from that
recorded at 10 wk for the training subjects (pooled data) and a
statistically significant decrease in eccentric torque of 11.4 ⫾
2.3% (P ⬍ 0.05). There was no difference between the training
groups. After detraining, concentric and eccentric torque re-
mained above pretraining levels [⌬Con ⫽ ⫹13.8 ⫾ 3.5% (P ⬍
0.01); ⌬Ecc ⫽ ⫹17.7 ⫾ 7.1% (P ⬍ 0.05)].
Torque-angle relationship. At 0 wk, greater torque was
produced in eccentric, compared with concentric, actions by
the training subjects (Figs. 3 and 4). Both peak concentric and
eccentric torque were produced at a knee angle of ⬃70° (note:
data were taken at each 10° of range of motion, so the optimum
knee angle was not specifically determined; see METHODS), with
less torque being produced at angles further from optimum.
Both absolutely and proportionally, peak eccentric torques
exceeded concentric torques at larger knee angles, although
normalized torque at smaller angles was the same between
contraction modes, so there was a small difference in the
stimulus applied to the muscles of the subjects during training.
Despite this, there were no between-group differences in
either absolute or normalized (to peak) torque-angle relations
at any time point in either concentric or eccentric actions.
Pooled analyses revealed both an increase in torque produc-
tion, largely around the angle of peak torque (Fig. 4), and a
shift in the normalized torque curve (Fig. 5), which occurred
largely within the first 5 wk of training. There was a small and
nonsignificant trend toward a reverse shift from 5 to 10 wk.
The shift in the torque-angle curve was most prominent in
eccentric contractions where greater absolute changes were Fig. 4. Increases in knee angle-specific concentric (top) and eccentric (bottom)
found. Interestingly, there was a further shift in the torque- torque with training (pooled data). Increases were greatest near the optimum
angle and were not different between the training groups. 0° ⫽ full extension.
angle curve to more acute knee angles (i.e., longer muscle a
Significant difference between 0 and 5 wk, P ⬍ 0.05; bsignificant difference
lengths) with detraining (see Fig. 6). These shifts were only between 0 and 10 wk, P ⬍ 0.05. To improve readability, SE bars are shown for
statistically significant for eccentric contractions. There were 90° angle only.

J Appl Physiol • VOL 103 • NOVEMBER 2007 •

moderately longer (⬃14 mm) at maximum knee flexion. Rel-
ative to the overall length change, however, this difference was
small and was not statistically significant (P ⫽ 0.30) for the six
subjects tested.
There was no change in VL fascicle length in control
subjects [⌬ ⫽ ⫺0.3 ⫾ 0.7 mm (⫺0.3 ⫾ 0.9%)] over the 10-wk
testing period. A pooled analysis of the training groups re-
vealed a significant increase after 5 wk in the training subjects
(4.7 ⫾ 1.7%; P ⬍ 0.05), but no further change to 10 wk
(⌬pre-post ⫽ ⫹4.2 ⫾ 1.4%; see Fig. 8). However, there was no
difference in the change in fascicle length between the training
groups [⌬Con ⫽ ⫹6.3 ⫾ 3.0% (P ⫽ 0.039); ⌬Ecc ⫽ ⫹3.1% ⫾
1.6% (P ⫽ 0.056)]. The small mean increase in fascicle length
after detraining (⌬post-detraining ⫽ ⫹2.5 ⫾ 2.2%) did not reach
statistical significance and was not different between training
groups. VM fascicle length was not estimated because its
complex architecture rendered the estimates unreliable.
There was no change in VL or VM fascicle angle over 10 wk
in the control subjects [⌬VL ⫽ ⫺0.24 ⫾ 0.52° (0.4 ⫾ 3.6%);

Downloaded from on November 18, 2007
⌬VM ⫽ ⫹0.40 ⫾ 0.31° (2.5 ⫾ 2.1%)]. A pooled analysis
revealed statistically significant VL fascicle angle increases of
11.5 ⫾ 3.2% (P ⬍ 0.05) and 17.9 ⫾ 3.7% (P ⬍ 0.01) after 5
and 10 wk, respectively (mean data are shown in Fig. 9). There
was a nonsignificant reduction after detraining such that fasci-
Fig. 5. Changes in normalized angle-specific concentric (top) and eccentric cle angle remained greater after detraining than at pretraining
(bottom) torque with training (pooled data). Subjects were able to produce (13.2 ⫾ 4.2%). Smaller changes in VM fascicle angle of 4.8 ⫾
relatively greater torque at acute knee angles (i.e., longer muscle lengths) after 2.5% and 8.2 ⫾ 3.4% after 5 and 10 wk were more variable by
the training, whereas less torque was produced at larger knee angles (i.e., comparison with VL and did not reach statistical significance.
shorter muscle lengths). Greater differences were seen for eccentric contrac-
tions where peak torque production was higher. There were no between-group
There were also no statistically significant training group-
differences. aSignificant change from 0 to 5 wk, P ⬍ 0.05; bsignificant change
from 0 to 10 wk, P ⬍ 0.05. To improve readability, SE bars are shown for 90°
angle only.

in the hypertrophic response precluded the greater distal in-
crease in whole quadriceps cross-sectional area from reaching
statistical significance (⌬prox ⫽ 7.3%, ⌬dist ⫽ 12.6%), in-
creases were greater in the distal compared with proximal
region (3-fold) in 75% of subjects.
To more closely examine the temporal hypertrophic re-
sponse, muscle thickness measures were also obtained.
Whereas no change in muscle thickness was found for control
subjects [⌬VL ⫽ ⫹0.56 ⫾ 0.96 mm (2.6 ⫾ 3.6%); ⌬VM ⫽
⫹0.33 ⫾ 0.57 mm (2.5 ⫾ 3.8%)], there was a statistically
significant increase in muscle thickness of VL at 5 and 10 wk
(P ⬍ 0.05) and of VM at 10 wk (P ⬍ 0.05) (Fig. 7) in the
training subjects (both groups combined). Thus the temporal
response of the muscles was not the same. The slight decrease
in both muscles after detraining was not significant; however,
the detraining values were also not different from values at 0
wk. There was no statistical difference in the changes in
muscle thickness between Con and Ecc groups [⌬Con ⫽ ⫹25.4 ⫾
12.6 mm (11.8%); ⌬Ecc ⫽ ⫹18.4 ⫾ 10.4 mm (8.3%)].
Muscle architecture. The mechanical strain of VL fascicles
measured at midthigh in three men and three women before
training was 105.1 ⫾ 9.0% (62.4 ⫾ 7.5 to 128.3 ⫾ 15.5 mm) Fig. 6. Changes in normalized angle-specific concentric (top) and eccentric
and 90.1 ⫾ 12.3% (60.0 ⫾ 4.4 to 114.5 ⫾ 12.0 mm) for (bottom) torque with detraining (pooled data). The subjects were able to
concentric and eccentric contractions, respectively. Thus there produce relatively greater torque at acute knee angles (i.e., longer muscle
lengths) after detraining, but less torque at greater knee angles. These differ-
was a tendency for subjects to produce tension through a ences were significant during concentric contractions, but the small differences
slightly longer fascicle length range during concentric contrac- in normalized eccentric strength were not statistically significant. *P ⬍ 0.05.
tions, which largely resulted from their VL fascicles being To improve readability, SE bars are shown for 90° angle only.

J Appl Physiol • VOL 103 • NOVEMBER 2007 •
Table 1. Changes in anatomic cross-sectional area measured proximally (25% from proximal end point) and distally
(75% from proximal end point)
Proximal Distal

Muscle 0 wk 10 wk Absolute Change % Change 0 wk 10 wk Absolute Change % Change

Quad 76.6⫾3.8 82.2⫾3.9 5.6⫾1.6 7.3 61.0⫾3.0 68.7⫾3.9 7.7⫾2.5 12.6
VL 29.3⫾1.4 32.5⫾1.4 3.2⫾0.7 10.9 17.7⫾1.2 20.1⫾1.6 2.4⫾0.8 13.6
VM 6.14⫾0.5 7.24⫾0.6 1.1⫾0.4 17.9 24.4⫾1.2 27.4⫾1.3 3.0⫾0.9* 12.3
VI 22.6⫾1.5 22.6⫾1.4 0.0⫾0.5 0.0 16.4⫾0.8 18.6⫾1.1 2.2⫾0.8† 13.4‡
RF 18.2⫾1.2 19.4⫾1.2 1.1⫾0.6 6.6 6.0⫾0.5 6.5⫾0.5 0.5⫾0.2 8.3
Values are means ⫾ SE. Quad, quadriceps femoris; VL, vastus lateralis; VM, vastus medialis; VI, vastus intermedius; RF: rectus femoris. Whereas the
increases in cross-sectional area were generally greater in the distal, compared to the proximal, quadriceps, statistically significant region-specific hypertrophy
was only seen in VI. *Mean difference from proximal, P ⫽ 0.08; †mean difference from proximal, P ⫽ 0.055; ‡significantly different from proximal, P ⬍ 0.05.

dependent differences in either VL or VM fascicle angle the increase in Con was statistically greater than that in Ecc. To
change over the training period. For VL, fascicle angle of the our knowledge, such a result has not been reported previously.
Ecc group increased by 21.4 ⫾ 7.3% (P ⫽ 0.03), whereas Concomitant with the strength change was a significant
changes in the Con group did not reach statistical significance increase in muscle volume and VL PCSA. By obtaining T2-

Downloaded from on November 18, 2007
(13.3 ⫾ 3.0%; P ⫽ 0.06) after 10 wk. The smaller changes in weighted MRI images and measuring the image density in both
VM fascicle angle for Con (7.0 ⫾ 4.9%) and Ecc (10.2 ⫾ 0 wk and 10 wk images we were able to confirm that the
5.1%) groups over the same period were not significant. increase in muscle size was not a residual effect of the training
or testing contractions performed by the subjects, although the
MRI data for one subject were omitted because of a significant
We have examined the relative contribution of the training shift in image density. Thus the training induced significant
contraction mode to muscle architectural adaptations in human strength and muscle size gains in both training groups. Inter-
VL and VM. While there were no changes in the strength of a estingly, cross-sectional changes were not consistent at differ-
control group measured across a 10-wk nontraining period, the ent muscle locations. Most notably, there was a greater in-
training induced significant increases in concentric and eccen- crease in cross section of distal, compared with proximal, VI.
tric strength of both concentric-only training (Con) and eccen- There was also a trend (P ⫽ 0.08) toward a greater distal
tric-only training (Ecc) subjects. The knee extensor strength increase in VM. While there was not a statistically different
responses to high-resistance training reported in the literature regional change shown for the quadriceps as a whole, a set of
vary widely [e.g., compare results from Higbie et al. (25), 16 “responders” (i.e., 75% of the sample) showed a signifi-
Hortobágyi et al. (26) and Seger et al. (52)], although the cantly greater (⬎3-fold) increase distally. Thus our data con-
magnitude of the change measured in the present study is not firm those of others who have reported a heterogeneous hyper-
out of line with these reports, and our finding of a greater trophic response after knee extension training (17, 24, 27, 40),
increase in eccentric strength after training has been commonly although there is little agreement as to which regions generally
reported (1, 10, 20, 25, 52). Interestingly, training mode spec- show the greatest response. Given that different regions of the
ificity was only observed under concentric conditions, where quadriceps muscles are architecturally disparate and the mus-
cles would encounter varying stresses during contraction, it is
not unexpected that hypertrophy would vary according to

Fig. 7. Changes in VL and VM muscle thickness (pooled data). Muscle Fig. 8. Changes in VL fascicle length (pooled data). There was a significant
thickness increased significantly after 5 wk in VL and after 10 wk in VL and increase from 0 wk to 5 wk (4.7%; P ⬍ 0.05) but no further change in the 2nd
VM in training subjects. The decrease after detraining was not significant, 5-wk training period in training subjects. The mean increase in fascicle length
although these values also were not statistically different from 0 wk. Muscle in detraining (2.5 ⫾ 2.2%) was not statistically significant. Changes in fascicle
thickness changes were not different between the training groups. *P ⬍ 0.05. length were not different between the training groups. *P ⬍ 0.05.

J Appl Physiol • VOL 103 • NOVEMBER 2007 •

uphill running training (it was not possible to measure VI
length change). They confirmed that VL actively lengthens
during downhill running and largely shortens during uphill
running but found a similar decrease in sarcomere number in
the VL of both training groups [downhill group ⫽ ⫺4.0%, not
significant (NS); uphill group ⫽ ⫺5.3%, P ⬍ 0.05]. Interest-
ingly, their Figures 1 and 2 (Ref. 16) showed that VL is active
at short lengths during the period of EMG activity in both
uphill and downhill running, which might explain the similar
loss of sarcomeres in both training groups. In VI, sarcomere
number increased in the downhill-trained rats and decreased in
the uphill-trained rats, but there is no evidence (in their study
or other studies) that VI operates similarly to VL. Thus some
caution must be exercised in interpreting these data as evidence
for contraction mode-related alterations in sarcomere number.
In another study by the same research group (32), lengthening
Fig. 9. Changes in VL and VM fascicle angle (pooled data). VL fascicle angle contractions of the tibialis anterior and extensor digitorum
increased significantly after 5 wk of training (11.5%; P ⬍ 0.05) and continued longus of rabbits did not result in significant increases in
to increase through to 10 wk (17.9%; P ⬍ 0.01) in training subjects. The slight
sarcomere number. To further confirm the finding that the

Downloaded from on November 18, 2007
decrease in fascicle angle after detraining was not statistically significant, and
it remained above 0 wk levels. The mean increases in VM fascicle angle did training range of motion, rather than contraction mode or
not reach statistical significance; changes in VM fascicle angle (not shown) velocity, has the greater effect on fascicle length change,
were more variable than changes in VL fascicle angle. Fascicle angle changes examination of fascicle length changes to training using a less
were not different between the training groups. *P ⬍ 0.05; **P ⬍ 0.01.
than normal contraction length range is required. Furthermore,
our data are unable to determine whether the maximum length
location (8), although the reason for the different locations of achieved by the muscle or the total range of lengthening it
hypertrophic gain reported in these studies requires more experiences is the primary stimulus for range of motion-
detailed study. dependent fascicle lengthening, since the fascicle strains en-
Stimulus for fascicle length adaptation. A main purpose of countered during the training were not appreciably different.
the present research was to test the hypothesis that muscle Time course of fascicle length adaptation. We also sought to
contraction mode strongly influences the magnitude of fascicle describe the temporal response of fascicle length change for the
length adaptation. Our finding that Con and Ecc groups showed first time in humans. An important finding was that fascicle
similar increases (rather than opposite responses) is not sup- length adapted within the first weeks of training, during which
portive of it, so factors other than contraction mode must period neural adaptations have traditionally been thought to
influence fascicle length. Using our study design, we were able dominate. Our results confirm the rapid adaptation in fascicle
to gain some insight into the possible factors underpinning length reported by Seyennes et al. (53) and by others using
fascicle length change. We were able to examine the possible animal models (16, 36, 37). We found increases of ⬃4.7%
influence of the training range of motion (or muscle excursion) compared with increases of 9.9% reported by Seyennes et al.
by having the subjects train through a knee joint range of (53). The lesser increase found in the present study can prob-
motion of 95–100°, which is substantially greater than that ably be explained by our measurement of fascicle length with
normally encountered during walking (⬃25°; Ref. 4), jogging the knee fully extended, and the muscle shortened, compared
(⬃55°; Ref. 7), and countermovement jumping (⬃80°; Ref. with Seyennes et al. (53), who measured fascicle length with
11), although only slightly greater than stair climbing (⬃88° 80° of knee flexion. Also, their subjects trained three times a
for 25.5-cm stair; Ref. 4), and so can be considered as being a week, with testing being conducted at short intervals (10, 20,
loading stimulus of greater than normal range of motion. The and 35 days). Thus training and testing were possibly separated
finding of a similar increase in fascicle length between the groups by only a few days, and osmotic fluid shifts resulting from the
is supportive of the training range of motion (or muscle training stress might have had some impact on the measured
excursion range) being the dominant stimulus for fascicle change. We took sonographs at least 4 days after the last
length adaptation. It is unlikely that shortening velocity had a training session to minimize the impact of this phenomenon,
strong influence because we would have expected that slow yet our results confirm the rapid changes they reported, al-
training velocities would cause a reduction (or no change) in though ours were smaller in magnitude. We have shown that
fascicle length in both training groups, since muscle loading at these adaptations do not continue past ⬃5 wk in humans, and
higher shortening speeds has been associated with fascicle they therefore have a different temporal response to the fascicle
length increases in skeletal muscle (3, 9). angle adaptations (see below). The implications of this are that
Our results might be somewhat surprising given the data only relatively short periods of training are required to induce
suggestive of contraction mode being an important mediating the necessary adaptation, so alterations in muscle function
factor for VI sarcomere addition when muscle loading in rats is imposed by the fascicle length changes should also occur
provided by downhill running training (16, 36, 37). However, rapidly.
in those studies the excursion range of VI was not explicitly One such functional alteration associated with fascicle
confirmed. Butterfield et al. (16) analyzed the operating length length change is the shift in the knee extension torque-angle
range of the VL in rats by sonomicrometry before measuring relationship. Significant shifts toward longer muscle lengths
sarcomere number adaptations to 10 days of either downhill or (i.e., rightward shift) have been shown previously after periods
J Appl Physiol • VOL 103 • NOVEMBER 2007 •
strong evidence that fascicle length adaptations directly influ-
ence force-length relations of humans muscles, and that the
range of motion of the training exercises is the strongest
influencing factor for fascicle length adaptations. It also con-
firms the association between fascicle length changes and the
force-length relation shown rarely in humans (e.g., Ref. 47) but
consistently shown between sarcomere addition (i.e., fiber
lengthening) and changes in force-length relation in animals
(14, 15).
Fascicle angle adaptation. To test the hypothesis that fas-
cicle angle adaptations would be strongly related to muscle
Fig. 10. Temporal patterns of normalized fascicle length (FL) and torque- size, we measured muscle thickness of both VL and VM during
angle relationship change (pooled data). Fascicle length change was calculated
as a proportion of pretraining values. The shift in the torque-angle relationship
training and detraining and then compared its temporal re-
was calculated by averaging the normalized concentric and eccentric torques at sponse to that of fascicle angle. In both muscles, there was a
each angle for each subject and then calculating the change from pretraining tendency for muscle thickness and fascicle angle to increase
values; increases at 80 and 90° and decreases at 30, 40, 50, 60, and 70° with training and then decrease slightly with detraining. As
contributed to positive torque-angle relationship shifts. The average shift was shown in Fig. 11, the responses in VL, for which we obtained
then recorded for each subject and plotted. There was a similar temporal
response of fascicle length and the torque-angle relationship through both measures of the highest reliability, are almost identical. We
training and detraining. found a constant ratio of adaptation such that a 1% increase in

Downloaded from on November 18, 2007
fascicle angle was associated with a ⬃0.5% increase in muscle
thickness. These data are indicative of a strong relationship
of knee extension training (17, 47, 50), although in one study between muscle size and fascicle angle, as has been previously
an increase in tendon stiffness after training in older subjects reported (29, 30). As expected, the changes in fascicle angle,
was suggested to have prevented a shift in the torque-angle which were measured in a single plane, and those of the 2D
relationship despite there being a clear shift in the muscle (cross-sectional area) and 3D (volume) measures of muscle
length-force relation (47). The idea that architectural adapta- size were not as well related. Our alternative hypothesis was
tions might underpin changes in torque-angle relationships has that fascicle angle increases were a separate anatomic adapta-
gained momentum since there has been a lack of association of tion to the stress imposed on the muscle during training and
neural and torque-angle adaptations, with some authors show- might be unrelated to muscle hypertrophy. Our finding of a
ing a concomitant change (57) and others not (47, 59, 60). similar response in the Con group, who produced lesser torque
Other authors have also examined relationships between vol- in training, and the Ecc group, who produced greater torque in
untary and electrically elicited contractions and found little training, mitigates against this. The slightly greater increase in
similarity (32). This has been taken as good evidence that fascicle angle obtained by Ecc subjects was influenced by a
mechanical, rather than neural, factors underpin the torque- strong “responder,” and therefore was associated with a greater
angle relationship. Given that increases in fiber length resulting SE. Therefore, we are not confident that the greater training
from the addition of sarcomeres are typically associated with load imposed on Ecc subjects can account for the small (and
the rightward shift of the torque-angle curve in animals (14, nonsignificant) between-group difference. Thus our data are
15), we expected to see a relationship between fascicle length more suggestive that fascicle angle adapts in response to
change and the torque-angle shift, which would therefore be muscle hypertrophy, and might therefore be driven by space
the same in both training groups. As shown in Figs. 4 and 5, the constraints in the hypertrophying muscle.
torque-angle relationship shifted toward the right in the first 5 The temporal response of fascicle angle to training has not
wk of training but did not shift (in fact, there was a small trend been previously described. We found that in VL fascicle angle
toward a negative shift) from 5 to 10 wk of training; these increased significantly (11.0%; Fig. 9) after 5 wk and contin-
changes were not group dependent. Unexpectedly, we found a ued to increase throughout the 10-wk training period (17.9%).
further small shift to the right in detraining (Fig. 6). The Previously, Seyennes et al. (53) reported a significant (7.7%)
magnitude of increase (2.5%; NS) was more than half of the increase in VL fascicle angle after 5 wk. The smaller magni-
increase attributable to the training, although there was some
interindividual variability in this response. The reasons for this
increase are not clear, although evidence for an increase in
fascicle length in VL after the removal of heavy resistance
training has been previously documented (9); interestingly, the
changes reported in that study also occurred within 5 wk. We
take this as evidence that learning responses cannot be solely
responsible for the shifts because we would not then have seen
a change in the detraining period. Regardless, the temporal
response of the torque-angle curve matched the fascicle length
response very closely, even to the point of them having a
similar response to detraining (Fig. 10). The finding also of a
constant ratio of adaptation, such that a 1% increase in fascicle Fig. 11. Temporal patterns of normalized VL fascicle angle (FA) and muscle
length was associated with approximately a 1% shift in the thickness (MT) change (pooled data). There was a similar response of fascicle
normalized torque-angle relationship, can also be considered as angle and muscle thickness through both training and detraining.

J Appl Physiol • VOL 103 • NOVEMBER 2007 •

tude of increase shown in that study might have resulted from In summary, we have confirmed previous findings of rapid
their testing of muscle architecture with the knee flexed to 80°, muscle architectural adaptation in humans and show for the
where fascicle angle would be smaller and changes possibly first time that the temporal responses of fascicle angle and
harder to detect. We have also shown for the first time that length are not the same: changes in fascicle angle continue
fascicle angle continues to increase as training progresses, at relatively linearly for the first few months of training, whereas
least to the 10th week, where our measurements were com- fascicle length adapts rapidly and does not continue past the
pleted. Previous studies have reported significant increases first few weeks. We have also shown that adaptations in
after 14- and 16-wk training periods (1, 30), but the pattern of fascicle length are not influenced by contraction mode, but are
adaptation had not been examined. Together, our data and most likely influenced by the training range of motion (i.e.,
those of others strongly suggest that fascicle angle increases fascicle strain); contraction velocity probably has little effect.
continually, at least for the first few months of training. Given In contrast, fascicle angle seems to be inextricably linked to
that increases in fascicle angle allow a greater amount of muscle hypertrophy. Interestingly, 3 mo of detraining was not
contractile tissue to attach to the tendon and aponeurosis and sufficient for muscle thickness, fascicle angle, or fascicle
that it promotes an increase in the AGR, which allows the length to return to baseline levels, and in fact there was a small
rotating fibers the opportunity to work closer to their optimum increase (not significant) in fascicle length in detraining. Inter-
(from both force-velocity and force-length perspectives), this estingly, the changes in fascicle length were closely associated
increase in fascicle angle most probably contributes to the with shifts in the torque-angle curve, and because the increases
continuing increase in strength over the first months of resis- were essentially the same in both training groups it is likely
tance training.

Downloaded from on November 18, 2007
that the training range of motion was the most significant
Interestingly, changes in VM were typically of lesser mag- influencing factor. Whether the length range of force produc-
nitude (8.2%) and more variable. The VM has a complex tion or the absolute muscle length is the greater stimulus cannot
architecture (8, 51) and is probably activated differently from
be determined from our data; however, we can confirm that the
VL, and its activation adaptation progresses differently with
genetic and/or signaling basis for fascicle length adaptations is
training (21, 22, 44, 45). Our data show that the adaptations in
different from that for muscle hypertrophy/fascicle angle ad-
fascicle angle are not always consistent between the two
aptation in humans.
muscles and that the response of VM has more interindividual
variability. This is the first study examining changes in syner-
gist muscles with training in humans, so it is not clear whether REFERENCES
such intermuscular differences occur routinely. Nonetheless, 1. Aagaard P, Andersen JL, Dyhre-Poulsen P, Leffers AM, Wagner A,
we have also found that the fascicle angle adaptations in both Magnusson P, Halkjaer-Kristensen J, Simonsen EB. A mechanism for
VM (nonsignificant) and VL are largely retained after a 3-mo increased contractile strength of human pennate muscle in response to
strength training: changes in muscle architecture. J Physiol 534: 613– 623,
detraining period. To our knowledge, no studies have exam-
ined the response of fascicle angle to a period of detraining 2. Adams GR, Harris RT, Woodard D, Dudley GA. Mapping of electrical
after a previous program of training in humans or other muscle stimulation using MRI. J Appl Physiol 74: 532–537, 1993.
animals. The lack of significant decrease in fascicle angle in 3. Alegre LM, Fernando J, Gonzalo-Orden JM, Martin-Acero R,
detraining has important consequences for the retention of Aguado X. Effects of dynamic resistance training on fascicle length and
strength after detraining (note that knee extension torque did isometric strength. J Sports Sci 24: 501–508, 2006.
4. Andriacchi TP, Andersson GBJ, Fermier RW, Stern D, Galante JO. A
not return to baseline after the detraining period, either) and is study of lower-limb mechanics during stair-climbing. J Bone Joint Surg
suggestive of the possibility that advantageous adaptations to Am 62: 749 –757, 1980.
training can be largely retained after lengthy periods of injury 5. Azizi E, Brainerd EL. Architectural gear ratio and muscle fiber strain
or illness in previously untrained subjects who complete a homogeneity in segmented musculature. J Exp Zool 307A: 145–155, 2007.
long-enough period of resistance exercise. 6. Bandy WD, Hanten WP. Changes in torque and electromyographic
activity of the quadriceps femoris muscles following isometric training.
Given that the likely primary stimuli for fascicle length and Phys Ther 73: 455– 467, 1993.
angle changes are not the same, and the time courses of 7. Biewener AA, Farley CT, Roberts TJ, Temaner M. Muscle mechanical
adaptation for fascicle length and angle are obviously different, advantage of human walking and running: implications for energy cost.
it is clear that there are separate genetic and/or signaling bases J Appl Physiol 97: 2266 –2274, 2004.
underpinning fascicle angle and length changes. Recent ad- 8. Blazevich AJ, Gill ND, Zhou S. Intra- and intermuscular variation in
human quadriceps femoris architecture assessed in vivo. J Anat 209:
vances in the muscle molecular and genetic sciences have 289 –310, 2006.
revealed a complex system of mechanotransduction in skeletal 9. Blazevich AJ, Gill ND, Bronks R, Newton RU. Training-specific muscle
muscle and a mesh of transcription and translation events architecture adaptation after 5-wk concurrent training in athletes. Med Sci
leading to muscle hypertrophy [see reviews by Rennie et al. Sports Exerc 35: 2013–2022, 2003.
(48), Chin (19), and Burkholder (13)]. Muscle stretch, or 10. Blazevich AJ, Gill ND, Deans N, Zhou S. Lack of human muscle
architectural adaptation after short-term strength training. Muscle Nerve
simultaneous stretch and activation, is commonly used to study 35: 78 – 86, 2007.
these processes. It is perhaps not surprising that the specific 11. Bobbert MF, Mackay M, Schinkelshoek D, Huijing PA, van Ingen
sequence of events leading to fiber hypertrophy vs. lengthening Schenau GJ. Biomechanical analysis of drop and countermovement
have not been elucidated since simultaneous stress and strain jumps. Eur J Appl Physiol 54: 566 –573, 1986.
seem to result in both of these adaptations. Models used to 12. Burkholder TJ. Age does not influence muscle fiber length adaptation to
explore the genetic and/or signaling phenomena that are unique increased excursion. J Appl Physiol 91: 2466 –2470, 2001.
13. Burkholder T. Mechanotransduction in skeletal muscle. Front Biosci 12:
to either muscle hypertrophy or fiber lengthening should per- 174 –191, 2007.
haps incorporate a stimulus in which stress is applied at shorter 14. Burkholder TJ, Lieber RL. Sarcomere number adaptation after retinac-
vs. longer muscle lengths. ulum transection in adult mice. J Exp Biol 201: 309 –316, 1998.

J Appl Physiol • VOL 103 • NOVEMBER 2007 •
15. Butterfield TA, Herzog W. Effect of altering starting length and activa- 38. Meyer RA, Prior BM. Functional magnetic resonance imaging of muscle.
tion timing of muscle on fiber strain and muscle damage. J Appl Physiol Exerc Sport Sci Rev 28: 89 –92, 2000.
100: 1489 –1498, 2006. 39. Morse CI, Thom JM, Birch KM, Narici MV. Changes in triceps surae
16. Butterfield TA, Leonard TR, Herzog W. Differential serial sarcomere muscle architecture with sarcopenia. Acta Physiol Scand 183: 291–298,
number adaptations in knee extensor muscles of rats is contraction type 2005.
dependent. J Appl Physiol 99: 1352–1358, 2005. 40. Narici MV, Hoppeler H, Kayser B, Landoni L, Claassen H, Gavardi
17. Carey Smith R, Rutherford OM. The role of metabolites in strength C, Conti M, Cerretelli P. Human quadriceps cross-sectional area, torque
training. I. A comparison of eccentric and concentric contractions. Eur and neural activation during 6 months strength training. Acta Physiol
J Appl Physiol 71: 332–336, 1995. Scand 157: 175–186, 1996.
18. Cheng GC, Brigs WH, Gerson DS, Libby P, Grondzinsky AJ, Gray 41. Narici MV, Maganaris CN, Reeves ND, Capodaglio P. Effect of aging
ML, Lee RT. Mechanical strain tightly controls fibroblast growth factor-2 on human muscle architecture. J Appl Physiol 95: 2229 –2234, 2003.
release from cultured human vascular smooth muscle cells. Circ Res 80: 42. Otten E. Concepts and models of functional architecture in skeletal
28 –36, 1997. muscle. Exerc Sport Sci Rev 16: 89 –137, 1988.
19. Chin ER. Role of Ca2⫹/calmodulin-dependent kinases in skeletal muscle 43. Patten C, Meyer RA, Fleckenstein JL. T2 mapping of muscle. Semin
plasticity. J Appl Physiol 99: 414 – 423, 2005. Musculoskelet Radiol 7: 297–305, 2003.
20. Colliander EB, Tesch PA. Responses to eccentric and concentric resis- 44. Pincivero DM, Salfetnikov Y, Campy RM, Coelho AJ. Angle- and
tance training in females and males. Acta Physiol Scand 141: 149 –156, gender-specific quadriceps femoris muscle recruitment and knee extensor
1990. torque. J Biomech 37: 1689 –1697, 2004.
21. Cramer JT, Housh TJ, Weir JP, Johnson GO, Berning JM, Perry SR, 45. Rabita B, Pérot C, Lensel-Corbeil G. Differential effect of knee exten-
Bull AJ. Gender, muscle, and velocity comparisons of mechanomyo- sion isometric training on the different muscles of the quadriceps femoris
graphic and electromyographic responses during isokinetic muscle ac- in humans. Eur J Appl Physiol 83: 531–538, 2000.
tions. Scand J Med Sci Sports 14: 116 –127, 2004. 46. Reeves ND, Narici MV. Behavior of human muscle fascicles during
22. Ebersole KT, Housh TJ, Johnson GO, Evetovich TK, Smith DB, Perry shortening and lengthening contractions in vivo. J Appl Physiol 95:

Downloaded from on November 18, 2007
SR. MMG and EMG responses of the superficial quadriceps femoris 1090 –1096, 2003.
muscles. J Electromyogr Kinesiol 9: 219 –227, 1999. 47. Reeves ND, Narici MV, Maganaris CN. In vivo human muscle structure
23. Fleckenstein JL, Shellock FG. Exertional muscle injuries: magnetic and function: adaptations to resistance training in old age. Exp Physiol 89:
resonance imaging evaluation. Top Magn Reson Imaging 3: 50 –70, 1991. 675– 689, 2004.
24. Häkkinen K, Pakarinen A, Kraemer WJ, Häkkinen A, Valkeinen H, 48. Rennie MJ, Wackerhage H, Spangenburg EE, Booth FW. Control of
Alen M. Selective muscle hypertrophy, changes in EMG and force, and the size of the human muscle mass. Annu Rev Physiol 66: 799 – 828, 2004.
serum hormones during strength training in older women. J Appl Physiol 49. Rybak LD, Torriani M. Magnetic resonance imaging of sports related
muscle injuries. Top Magn Reson Imaging 14: 209 –219, 2003.
91: 569 –580, 2001.
50. Schott J, McCully K, Rutherford OM. The role of metabolites in
25. Higbie EJ, Cureton KJ, Warren GL, Prior BM. Effects of concentric
strength training. II. Short versus long isometric contractions. Eur J Appl
and eccentric training on muscle strength, cross-sectional area, and neural
Physiol 71: 337–341, 1995.
activation. J Appl Physiol 81: 2173–2181, 1996.
51. Scott SJ, Engstrom CM, Loeb GE. Morphometry of human thigh
26. Hortobágyi T, Barrier J, Beard D, Braspennincx J, Koens P, Devita P,
muscles. Determination of fascicle architecture by magnetic resonance
Dempsey L, Lambert J. Greater initial adaptations to submaximal muscle
imaging. J Anat 182: 249 –257, 1993.
lengthening than maximal shortening. J Appl Physiol 81: 1677–1682,
52. Seger JY, Arvidsson B, Thorstensson A. Specific effects of eccentric
1996. and concentric training on muscle strength and morphology in humans.
27. Housh DJ, Housh TJ, Johnson GO, Chu WK. Hypertrophic response to Eur J Appl Physiol Occup Physiol 79: 49 –57, 1998.
unilateral concentric isokinetic resistance training. J Appl Physiol 73: 53. Seyennes OR, de Boer M, Narici MV. Early skeletal muscle hypertrophy
65–70, 1992. and architectural changes in response to high-intensity resistance training.
28. Kaufman KR, An KN, Chao EY. Incorporation of muscle architecture J Appl Physiol 102: 368 –373, 2007.
into the muscle length-tension relationship. J Biomech 8: 943–948, 1989. 54. Sesto ME, Radwin RG, Block WF, Best TM. Anatomical and mechanical
29. Kawakami Y, Abe T, Fukunaga T. Muscle-fiber pennation angles are changes following repetitive eccentric exertions. Clin Biomech (Bristol, Avon)
greater in hypertrophied than in normal muscles. J Appl Physiol 74: 20: 41– 49, 2005.
2740 –2744, 1993. 55. Shellock FG, Fleckenstein JS. Muscle physiology and pathophysiology:
30. Kawakami Y, Abe T, Kuno SY, Fukunaga T. Training-induced changes magnetic resonance imaging evaluation. Semin Musculoskelet Radiol 4:
in muscle architecture and specific tension. Eur J Appl Physiol Occup 459 – 479, 2000.
Physiol 72: 37– 43, 1995. 56. Shortland A, Fry NR, Eve LC, Gough M. Changes to medial gastroc-
31. Kinugasa R, Kawakami Y, Fukunaga T. Muscle activation and its nemius architecture after surgical intervention in spastic diplegia. Dev
distribution within the triceps surae muscles. J Appl Physiol 99: 1149 – Med Child Neurol 46: 667– 673, 2004.
1156, 2005. 57. Thépaut-Mathieu C, Van Hoecke J, Maton B. Myoelectrical and
32. Koh TJ, Herzog W. Evaluation of voluntary and elicited dorsiflexor mechanical changes linked to length specificity during isometric training.
torque-angle relationships. J Appl Physiol 79: 2007–2013, 1995. J Appl Physiol 64: 1500 –1505, 1988.
33. Koh TJ, Herzog W. Eccentric training does not increase sarcomere 58. Van Leeuwen JL, Spoor CW. Modelling mechanically stable muscle
number in rabbit dorsiflexor muscles. J Biomech 31: 499 –501, 1998. architectures. Phil Trans R Soc Lond B Biol Sci 336: 275–292, 1992.
34. Koh TJ, Herzog W. Excursion is important in regulating sarcomere 59. Weir JP, Housh TJ, Weir LL. Electromyographic evaluation of joint
number in the growing rabbit tibialis anterior. J Physiol 508: 267–280, angle specificity and cross-training after isometric training. J Appl Physiol
1998. 77: 197–201, 1994.
35. Kumagai K, Abe T, Brechue WF, Ryushi T, Takano S, Mizuno M. 60. Weir JP, Housh TJ, Weir LL, Johnson GO. Effects of unilateral
Sprint performance is related to muscle fascicle length in male 100-m isometric strength training on joint angle specificity and cross-training.
sprinters. J Appl Physiol 88: 811– 816, 2000. Eur J Appl Physiol Occup Physiol 70: 337–343, 1995.
36. Lynn R, Morgan DL. Decline running produces more sarcomeres in rat 61. Wickiewicz TL, Roy RR, Powell PL, Perrine JJ, Edgerton VR. Muscle
vastus intermedius muscle fibers than does incline running. J Appl Physiol architecture and force-velocity relationships in humans. J Appl Physiol 57:
79: 1439 –1444, 1994. 435– 443, 1984.
37. Lynn R, Talbot JA, Morgan DL. Differences in rat skeletal muscles after 62. Winter DA. Biomechanics and Motor Control of Human Movement (2nd
incline and decline running. J Appl Physiol 85: 98 –104, 1998. ed.). New York: Wiley Interscience, 1990.

J Appl Physiol • VOL 103 • NOVEMBER 2007 •