Altered Triceps Surae Muscle–Tendon Unit Properties after 6 Months of Static Stretching : Medicine & Science in Sports & Exercise

Journal Logo


Altered Triceps Surae Muscle–Tendon Unit Properties after 6 Months of Static Stretching


Author Information
Medicine & Science in Sports & Exercise 53(9):p 1975-1986, September 2021. | DOI: 10.1249/MSS.0000000000002671


Joint flexibility is an important component of human movement, and although it is known that habitual stretching increases joint range of motion (ROM) (1), the physiological mechanisms behind increased ROM are not well understood (2). Potential mechanisms for an increase in ROM include neural properties and mechanical properties of the muscle–tendon units (MTUs) and passive joint structures. Because stretching-induced increases in ROM have been seen without changes in passive resistance, gains in ROM are in general attributed to increased tolerance to stretch or pain (1–5). However, the underlying specific neural mechanisms and the roles of pain threshold or sensitivity, peripheral reflex influence, and neural activation are poorly understood, and the findings remain equivocal: reduced tonic reflex activity has been observed (6), contrasting unchanged EMG amplitude (1,) and unchanged motoneuron excitability (7) after 3–6 wk of stretching. Further to the tolerance theory, some stretching interventions have induced a rightward shift in the passive torque–angle curve (8,9), a reduction in passive torque at standardized joint angles (10–12), or a reduction in passive joint stiffness (6,8,13). Changed passive torque–angle relationships may be a result of neural and/or structural changes. Animal studies have demonstrated that immobilization in lengthened positions may increase the number of serial sarcomeres and muscle fiber length (14), but the loads, intensities, and volumes of such interventions are very different from what is done in functional human studies. Human stretching interventions with indirect measures of structural adaptations are inconsistent, with reports of increased fascicle length in some (15,16) but not all studies (4,12,17,18). To date, there seems to be no consensus on the feasibility of fascicle lengthening with stretching training in humans. Stretching for 3 wk increased ROM and was accompanied by increased fascicle strain at standardized joint angle, with authors suggesting a decrease in muscle stiffness without a concurrent change in tendon stiffness (4). Similarly, increased elongation of the tissue proximal to the musculotendinous junction (MTJ) was indicated after 4 wk of stretching (12), and passive muscle/MTU stiffness was decreased after stretching for 4 (19) and 6 wk (20), suggesting altered muscle behavior with stretching training. Nonetheless, passive muscle–tendon mechanics remain highly complex because of serially coupled structures with different mechanical properties that undergo heterogeneous loading (21).

The effects of stretching training on tendon properties are similarly unclear, as intervention studies report unchanged tendon stiffness (3,8,11,22), reduced stiffness (11,23), and increased stiffness (20), and differences in stretching modality do not seem to fully explain the discrepancies. Theoretically, ROM increases could relate to altered tendon morphology and mechanics including tendon stiffness, Young’s modulus, and/or tendon length, yet one recent study found no changes in tendon length or thickness after 6 wk of stretching (16).

One advantage often attributed to stretching training is the potential influence on contractile function. Angle of peak torque may be changed because of the conditions for myofilament overlap (24), for example, through the addition of serial sarcomeres or increased tendon elongation, or through an overall reduction in passive resistance. The addition of serial sarcomeres would furthermore enable limb rotation at the same angular velocity with reduced sarcomere shortening velocity, perhaps enabling greater torque production (25). Stretching for 6–8 wk shifted angle of peak torque toward extended positions in a consistent manner in a few studies (26,27), and only at some velocities in another study (28). The implications of stretching training on isokinetic work production are also not clear, with increased concentric work in one study (28), whereas this effect was only seen at certain velocities in another (26).

Taken together, previous studies are inconsistent with respect to the role of morphological, mechanical, and neural properties as mechanisms for increased ROM with stretching training. Studies applying more than 12 wk of stretching are scarce and do not examine effects on mechanical properties of force-bearing tissues or morphological adaptation (29,30). Some studies speculate that early adaptations to stretching training occur at a sensory level (2) or are related to nonmuscular structures, which may be richly innervated but contribute marginally to passive torque (31), whereas adaptations altering passive resistance through the structural properties of the MTU may occur with greater stretching durations (2). Such a two-phasic pattern of adaptation is supported by some intervention studies, where maximal passive torque increased after 2–4 wk, but returned to (10) or below (6) the initial level after 4–8 wk. Increased passive torque at the new ROM obtained through stretching could result from increased pain tolerance or reduced sensory input due to adaptations in nonmuscular structures. Nonetheless, stretching interventions of longer duration are required to examine the interplay between ROM and passive resistance to stretch (2).

The purpose of this study was therefore to examine the effects of 24 wk of functional unilateral triceps surae stretching on ROM, on the morphological and mechanical properties of the MTU, on neural activation, and on contractile function. We hypothesized that ROM would increase by 8 wk, whereas by 24 wk, passive torque and corresponding EMG amplitude as well as tendon stiffness would be reduced. Maximal passive torque and passive elongation of muscle and tendon were expected to increase. Muscle thickness, pennation angle, and fascicle length were hypothesized to remain unchanged. Active peak isokinetic torque was expected to occur at a more dorsiflexed joint angle.



A priori sample size calculations were done by estimating PRE–POST changes in passive torque of 3 N·m, with an SD of 3 N·m, based on former stretching intervention studies (8,11). Statistical power was set to 90%. To accommodate an anticipated dropout of 25%, 30 subjects were recruited among recreationally active university students. Exclusion criteria were lower limb injury in the last 6 months, musculoskeletal disease, conditions preventing stretching of the MTU (e.g., ankle impingement, talocrural joint tightness, neuromeningeal tightness, and tingling due to sciatic nerve stretch), and/or a history of systematic stretching (>10 min·wk−1). In accordance with the Declaration of Helsinki, the regional ethics committee approved the study, and each subject signed an informed consent form.

Experimental design

The study was designed as a within-subject design with 24 wk of unilateral stretching. The stretching leg was assigned by stratified randomization with a 50/50 distribution between dominant and nondominant leg. The contralateral leg served as control. Two subjects had an initial side-to-side ROM difference of >10° and were assigned to stretch their least flexible leg. Leg testing order was randomized at the first test session and was replicated during subsequent sessions. Passive dorsiflexion, tendon stiffness, and contractile function tests were completed for one leg before proceeding to the second leg. All measurements and analyses were undertaken by the same investigators, who were always blinded to the leg assignment.

Each subject reported to the laboratory five times: a familiarization session, a preintervention session (PRE), a subset of tests after 8 and 16 wk, and a postintervention session (POST) 24–48 h after the last bout of stretching. The subjects were instructed to refrain from training or stretching 24 h before testing. In female subjects who did not take constant-dose oral contraceptives, all tests (including PRE, 8 wk, 16 wk, POST) were conducted within 14 d from the last menstruation to avoid testing during the luteal phase (32). The subjects were instructed to maintain habitual activity level and diet; to refrain from unaccustomed exercise, anti-inflammatory drugs, and nutritional supplements; and to report illness or injury during the experimental period.

Stretching intervention

Four repetitions of a 60-s self-administered static ankle dorsiflexion stretch were performed daily for 24 wk. On the first 8 wk, stretching was performed with straight knee joint (four repetitions; Fig. 1A), whereas the remaining weeks was undertaken with the knee joint straight for 2 repetitions and flexed (Fig. 1B) for 2 repetitions, to ensure an effect on all compartments of the triceps surae. The subjects were instructed to place the stretching leg as far posteriorly as possible, while pushing the heel down to the ground, the forefoot pointing forward.

Experimental setup including timeline on the vertical axis, methods and order of testing on the horizontal axis, and stretching exercises with straight (A) and bent (B) knee on the right side. The method pictograms represent anthropometry, resting EMG measurements, warm-up, and tests of tendon stiffness, passive dorsiflexion, and contractile function. The expected outcomes are presented in the bottom row, where ↓ indicates a reduction, ↑ indicates an increase and == indicates no change to the variable.

Before the intervention, written, verbal, and visual instructions were given. Proper stretching technique was verified at each test session. The subjects were instructed to produce the strongest sensation of stretch possible without being in direct pain. Pain perception during the stretching exercise was recorded on a 10-cm visual analog scale (VAS score) at each session. The daily stretching was self-administered. Adherence throughout the intervention was monitored through daily written journals and telephone calls every 3–4 wk.

Experimental setup

Each test session started with all the resting measurements. Subsequently, warm-up, tendon stiffness, passive dorsiflexion, and tests of contractile function were all completed for one leg, before restarting with warm-up for the second leg. The order of the legs was randomized and was repeated identically at each test session. The experimental setup is illustrated in Fig. 1.

Anthropometry, resting muscle architecture, and morphological properties

Leg length was measured as the distance between the trochanter major and the floor while standing. Calf length was measured from the lateral femoral epicondyle to the calcaneal tuberosity. With subjects lying prone with the foot hanging freely off the examination bed, resting ankle joint angle was measured using a manual goniometer. B-mode ultrasonography was applied (HD11XE with a 50-mm linear array transducer L12-5; Philips, Bothell, WA) to retrieve two image sets for the measurement of fascicle length, pennation angle, and muscle thickness of the gastrocnemius medialis (GM) and Achilles tendon length. The lower half of the transducer frequency (5–12 MHz) was used with a built-in filter to optimize ultrasound penetration while preserving spatial resolution and contrast. All ultrasound still images were analyzed using imaging software (Fiji ImageJ [33]). Muscle variables were obtained from sagittal images recorded at midlength of the muscle belly, as detailed in our previous publication (34). Repositioning of the ultrasonography probe and EMG sensors at the same location on different test days was achieved by marking the positions and the subject’s moles or other distinct marks on acetate sheets. The free Achilles tendon length (calcaneal insertion to soleus (SOL) MTJ) and whole Achilles tendon length (calcaneal insertion to GM MTJ) were measured along the tendon path from sequentially combined images, as detailed in our previous publication (34).


To ensure optimal skin impedance, shaving, gentle skin abrasion, and cleaning with isopropanol were performed in accordance with SENIAM recommendations (35). EMG electrodes (Ambu; Blue Sensor N, Ballerup, Denmark) were placed on SOL, GM, and gastrocnemius lateralis (GL) with an interelectrode distance of 20 mm, and a reference electrode was positioned on the tibial tuberosity. EMG signals were transmitted wirelessly (16-channel TeleMyo 2400 G2 Telemetry System; Noraxon Inc., Scottsdale, AZ) to a receiver (Mini-receiver for TeleMyo G2; Noraxon Inc.). EMG signals were filtered using a bidirectional zero-lag fourth-order Butterworth bandpass filter of 10–500 Hz, rectified, and integrated over 500 ms. EMG amplitudes were normalized to amplitudes recorded during maximal voluntary contractions.

Data sampling, synchronization, and postprocessing

For passive dorsiflexion, tendon stiffness, and tests of contractile function, EMG, dynamometer data, and goniometer data were digitized and sampled at 1500 Hz (MyoResearch XP Master Edition 1.08.17; Noraxon Inc.). A function generator (GwinStec, GFG-8215A; Good Will Instrument Co., Ltd, Tucheng City, Taiwan) and an electric trigger initiated sampling and enabled synchronization of data.

Postprocessing was performed offline using a software package (MATLAB and Statistics Toolbox Release 2015b; The MathWorks, Inc., Natick, MA). Torque, dynamometer angle, and goniometer data were filtered using a bidirectional zero-lag fourth-order Butterworth low-pass filter of 10 Hz. Goniometer data were fitted to a fourth-order polynomial equation. All passive dorsiflexion and tendon stiffness data were resampled to the ultrasound video frequency, whereas contractile function data were resampled to 200 Hz.

Tendon stiffness

After a 5-min bike ergometer warm-up (Monarch, 828E, Varburg, Sweden), the tensile stiffness of the free Achilles tendon and the whole Achilles tendon was examined through ultrasound videos (38–53 Hz) recorded during isometric ramp contractions. The ultrasound probe was placed sagittally over the distal MTJ of SOL and subsequently GM and was fixed to the leg using a custom-made rigid cast enabling stable positioning with minimal tissue compression. Echo-absorptive tape was applied to the skin to allow for the corrections of potential probe displacement relative to the skin.

Subjects were seated in a custom-made isometric ankle dynamometer (Gym2000, Geithus, Norway) instrumented with a load cell (U2A 500 Hottinger Baldwin Messtechnik, Darmstadt, Germany), with 90° hip flexion, with the knee straight and the foot strapped to the dynamometer at anatomical joint angle. The dynamometer settings were individually adjusted to align the axis of the ankle joint with that of the dynamometer. Unwanted joint movement was minimized by careful strapping of the limbs (34).

For the purpose of EMG normalization, the subjects performed two trials of plantar flexion maximal voluntary contraction, separated by 60-s rest. Subsequently, ramped plantar flexion contractions were carried out (constant rate of torque development (45 (N·m)·s−1) with real-time torque displayed to the subject and with standardized verbal encouragement). Each trial was preceded by three brief submaximal contractions serving to precondition the tendon. Three trials were performed for SOL and for GM MTJ, separated by 120-s rest. Additional trials were carried out if the subject did not follow the prescribed rate of torque development or if the ultrasonography plane or contact was lost during recording.

Plantar flexion torque was determined from load cell force and the perpendicular distance to the axis of joint rotation. Because ankle joint rotation during isometric plantar flexion was negligible (2.9° ± 2.3° across groups), the recorded torques were not corrected for antagonist coactivation, as isometric contractions in the applied position do not involve antagonist fascicle shortening and that EMG recordings are primarily due to cross-talk (36). Tendon force was calculated from the plantar flexion torque and the instantaneous tendon moment arm as derived from ankle joint angle and leg length.

MTJ displacement was measured by semiautomated tracking (Tracker 4.11.0; Open Source Physics, Aptos, CA), interpolated at 50-N intervals, and then averaged from three valid trials. Tendon elongation was defined as proximal displacement of the MTJ, corrected for the influence of ankle joint rotation (37) by combining the instantaneous joint angles during ramped contractions with the linear relationship between MTJ displacement and joint angle obtained during a slow, passive ankle plantar flexion (0° to 3°).

Tendon force and tendon elongation data were cut off at 90% of each individual’s maximal force and fitted with a second-order polynomial (R2 = 0.96–0.99) (34). Tendon stiffness was determined as the slope of the upper 20% of the curve. Tendon elongation and strain are reported at common force (the greatest force achieved by all legs across time points) and as individual maximum.

Slow, passive ankle dorsiflexion stretch

Passive torque, ankle joint angle, EMG (GM, GL, and SOL), and ultrasound videos (SOL MTJ, GM MTJ, GM fascicles) were obtained during passive dorsiflexion to maximal joint angle in an isokinetic dynamometer (HUMAC NORM 770; Computer Sports Medicine Inc., Stoughton, MA). Subjects were seated with 65° hip flexion and extended knee. The mediolateral axis of the ankle joint was aligned with the dynamometer axis. Unwanted joint movement was minimized by strapping of the limbs (34). Subjects confirmed no sensation of stretch in the calf or hamstrings in the starting position. To prevent influence of visual perception, the tests were performed with eyes closed.

To determine ankle dorsiflexion ROM, an investigator manually rotated the foot at approximately 2°·s−1 from a resting position to the maximally tolerated dorsiflexion. The dynamometer angle at this point was then recorded as the end point of ROM alongside with self-perceived pain (VAS score).

To assess passive resistance to stretch, the ankle joint was rotated by the dynamometer from 10° plantar flexion to maximal joint angle and back at 2°·s−1 while subjects were fully relaxed. The procedure was repeated six times, separated by 120-s rest, to secure six sets of EMG data along with two sagittal ultrasound videos (17–19 Hz) of the distal SOL MTJ, two of the distal GM MTJ, and two of GM mid-belly fascicles. Echo-absorptive tape was applied to the skin to allow postprocessing corrections for potential probe displacement relative to the skin. Additional trials were performed if torque or EMG signals indicated muscle contraction (intermediate peak instead of a curvilinear torque development), or if the ultrasonography plane or contact was lost during recording.

Ankle joint angle was concurrently obtained with a 2D electrogoniometer (Noraxon Inc.), secured to the medial part of the first metatarsal and distal tibia. For analyses of passive dorsiflexion, joint angles were obtained from the electrogoniometer, rather than from the dynamometer, to avoid the error induced by misalignment of the foot and dynamometer. Data are reported at anatomically neutral joint angle (0°), at standardized joint angle (defined as the maximal dorsiflexion angle that was common to each leg across time points), and at maximal joint angle (defined as the maximal dorsiflexion angle achieved by each leg at the separate time points).

GM fascicle lengths and pennation angles were measured by automatic tracking using optical flow algorithms (38). Displacement of the MTJ was measured by semiautomated tracking (Tracker 4.11.0; Open Source Physics), as detailed in our previous publication (34). Elongation and strain of fascicles are reported relative to resting length. MTU elongation was estimated from the average goniometer joint angles and calf length (39). Elongation of any structure proximal to the MTJ, represented by the displacement of the MTJ, is hereafter referred to as muscle elongation. Elongation occurring distally to the MTJ, calculated by subtracting distal MTJ displacement from MTU elongation, is hereafter referred to as tendon elongation. Muscle and tendon elongation are reported based on lengths measured at anatomical joint angle as absolute values and as percent contribution to total MTU elongation. All data were interpolated at 0.05° intervals using spline functions, and the valid trials were averaged.

Contractile function

Contractile function during isometric and isokinetic contractions was assessed using the same isokinetic dynamometer and position as for passive dorsiflexion. Isometric maximal plantar flexion torque was determined as peak torque over a 5-s contraction. Two trials were performed at 10° plantar flexion and at 0°, 5°, 10°, and 15° of dorsiflexion, all trials separated by 60-s rest. Isokinetic concentric dorsiflexion and plantar flexion torque was determined between 10° of dorsiflexion and 30° of plantar flexion, at 30°·s−1 (dorsiflexion and plantar flexion) and at 45°·s−1, 60°·s−1, and 90°·s−1 (plantar flexion). Three trials were performed at each angular velocity separated by 60-s rest. Warm-up consisting of three submaximal contractions at 30°·s−1 preceded the tests. Standardized verbal encouragements and visual feedback of the instantaneous torque were provided. The order of the trials was repeated identically PRE–POST, from least to greatest stretch and velocity, so that an eventual effect of stretch would not affect subsequent trials.

Isokinetic peak torque, angle of peak torque, and work were determined from the trial with the greatest peak torque. Positive work was calculated as the area below the torque–angle curve after interpolating torque at 0.1° intervals.


The muscle–tendon ultrasonography measurements at rest had excellent internal consistency (Cronbach’s α > 0.98–0.99). For the remaining variables, reliability was established from the literature, and original reliability data were not collected during the study. The methods have been used in previous work at our and others’ laboratories. Their satisfactory reliability has consistently been demonstrated, with r = 0.98–0.99, coefficient of variation = 6%–9% for the passive stretch methods (40), and with intraclass correlation coefficient > 0.90 for ROM, passive stretch, EMG, tendon stiffness, and muscle architecture (41).

Statistical analysis

Resting length of the free Achilles tendon, EMG amplitudes, and VAS scores during passive dorsiflexion tests were not normally distributed (D’Agostino and Pearson normality test) and were hence log-transformed to normality. Baseline characteristics were analyzed using paired, two-tailed Student’s t-test. One-way ANOVA for repeated measures was used to identify changes in VAS scores during stretching exercise. Two-way (leg–time) ANOVA for repeated measures was used to identify main effects and interactions, in which case post hoc, Sidak’s multiple comparisons tests were performed, using multiplicity adjusted P values. The level of significance was set to α = 0.05. All data are presented as mean ± SD. Cohen’s d effect sizes (ES) were calculated to estimate the magnitude of the treatment effects, with the magnitude of effects considered either small (0.20–0.49), medium (0.50–0.79), and large (>0.80) (42).


Four subjects withdrew from the study because of unrelated injury or personal reasons. The final data set comprised 26 subjects, with 17 women and 9 men (age, 22.0 ± 1.6 yr, height and mass, 169 ± 7 cm and 61 ± 11 kg for women 184 ± 5 cm and 80 ± 12 kg for men) reporting a weekly training volume of 2 ± 2 h of endurance activities and 1 ± 1 h of strength training during the intervention period.

For technical reasons, one subject was excluded from passive dorsiflexion analyses, one from the passive dorsiflexion EMG analyses, and one from the passive dorsiflexion ultrasound analyses. Because of insufficient ultrasound video quality or goniometer malfunctioning, five subjects were excluded from free Achilles tendon stiffness analyses and three from whole Achilles tendon stiffness analyses. Thus, the sample size for passive dorsiflexion was n = 25, that for passive dorsiflexion EMG and ultrasound was n = 24, and that for free and whole Achilles tendon stiffness was n = 21 and 23, respectively. For isometric strength, the sample size was n = 19 and 8 at 10° and 15° dorsiflexion, respectively, due to restricted ROM in the remaining subjects. Adherence to the training was 89% ± 10%, resulting in 10.9 ± 1.5 h or approximately 40,000 ± 5000 s of static stretching exercise.

ROM and pain during passive stretching

Stretching resulted in a bilateral increase in dorsiflexion ROM (interaction, P < 0.005; side, P = 0.66; time, P < 0.001; ES = 1.12; Fig. 2). Perceived pain during the passive ROM test was unchanged (interaction, P = 0.58; side, P = 0.81; time, P = 0.06; ES = 0.15; stretching, 4.2 ± 2.1 cm for PRE and 4.0 ± 2.9 cm for POST; control, 4.2 ± 2.2 cm for PRE and 4.3 ± 2.7 cm for POST). Self-perceived pain intensity during stretching exercise was reduced from baseline to subsequent time points (intervention, P < 0.001; ES = 0.95; PRE, 4.6 ± 2.6 cm; 8 wk, 2.2 ± 2.2 cm; 16 wk, 2.6 ± 2.0 cm; POST, 2.0 ± 1.7 cm).

Ankle dorsiflexion ROM in the stretching and control leg before (PRE), during, and after (POST) 24 wk of stretching. †An interaction effect (P < 0.005). *A difference from PRE (P < 0.001). §A difference from 8 wk (P < 0.001).

Resting properties of the MTU

Resting ankle joint angle, Achilles tendon morphological properties, and GM fascicle length were unchanged, whereas GM thickness and pennation angle increased with time in both legs (Table 1). Fascicle length and tendon lengths normalized to resting MTU length showed the same statistical outcome as absolute values and are not reported.

TABLE 1 - Resting properties of the ankle joint, GM, and the AT before (PRE) and after (POST) 24 wk of stretching.
PRE POST P Inter. P Side P Time Post Hoc P ES
Resting ankle angle, °
 Stretching 21.5 ± 3.5 23.1 ± 4.2 0.09 0.92 0.26 0.09 0.48
 Control 22.6 ± 4.4 22.2 ± 3.6 0.89
 Thickness, mm
  Stretching 19.7 ± 2.1 21.2 ± 2.0 0.44 0.56 <0.001 <0.001 0.02
  Control 19.8 ± 2.3 21.7 ± 2.7 <0.001
 Pennation angle, °
  Stretching 22.0 ± 2.9 23.9 ± 3.5 0.99 0.88 <0.001 <0.005 0.00
  Control 21.8 ± 3.3 23.8 ± 3.5 <0.001
 Fascicle length, mm
  Stretching 53.8 ± 6.1 54.5 ± 5.8 0.52 0.89 0.48 0.57 0.18
  Control 53.9 ± 7.5 53.9 ± 7.2 1.00
 Free AT length, mm
  Stretching 55.3 ± 20.2 57.6 ± 20.7 0.34 0.59 0.47 0.42 0.27
  Control 53.5 ± 19.3 53.8 ± 20.8 0.98
 Whole AT length, mm
  Stretching 178.8 ± 22.1 184.0 ± 22.4 0.19 0.98 0.18 0.13 0.36
  Control 181.5 ± 21.7 181.6 ± 22.2 1.00
AT, Achilles tendon; Inter., interaction.

Mechanical properties and EMG amplitude during passive dorsiflexion

Passive torque (Figs. 3A, B) decreased with time at anatomical joint angle in the stretching leg (interaction, P = 0.29; side, P = 0.48; time, P < 0.005; ES = 0.31; Fig. 3E) and at standardized joint angle in both legs (interaction, P = 0.40; side, P = 0.34; time, P < 0.001; ES = 0.24; Fig. 3F), whereas passive torque at maximal joint angle increased in the stretching leg (interaction, P < 0.01; side, P = 0.55; time, P < 0.001; ES, = 0.81; Fig. 3G).

Passive torque (A, B) and normalized (C, D; maximal voluntary contraction) EMG amplitude during passive dorsiflexion for the stretching leg (A, C) and the control leg (B, D) at anatomical, standardized, and maximal joint angle, before (PRE) and after (POST) 24 wk of stretching. In the stretching leg, standardized joint angle = PRE maximal joint angle. Error bars are left out for legibility. Statistical analyses of passive torque (E–G) and SOL EMG amplitude (H): †An interaction effect (P < 0.001–0.01). ‡A time effect (P < 0.001–0.005). *Significant post hoc tests (P < 0.001–0.05). Statistical analyses of GM and GL EMG are identical to SOL and are not shown.

EMG amplitudes of GM, GL, and SOL (Figs. 3C, D) at standardized joint angle decreased with time in both legs (interaction, P = 0.43–0.80; side, P < 0.07–0.30; time, P < 0.001; ES = 0.07–0.23; Fig. 3H). GM pennation angle change during passive dorsiflexion was negligible (CON PRE, 14.4° ± 3.5° at 0° and 14.9° ± 3.8° at maximal joint angle; STR POST, 13.2° ± 3.5° at 0° and 13.4° ± 3.4° at maximal joint angle) and was not altered by the intervention (interaction, P = 0.35–0.57; side, P = 0.30–0.35; time, P = 0.09–0.65; ES = 0.20–0.27). (For absolute values and detailed statistics, see Table, Supplemental Digital Content 1,

No interaction effect was seen for GM fascicle elongation, but there was a time effect in the control leg at standardized joint angle and in both legs at maximal joint angle (Table 2). Elongation of the SOL and GM MTU at maximal joint angle (Figs. 4B, D) was changed in both legs, with increased elongation of tendon in both legs, increased elongation of muscle only in the stretching leg, and with increased contribution from tendon elongation (except control leg GM, where P = 0.08 for contribution at maximal joint angle). At standardized joint angle (Figs. 4A, C), the control leg SOL MTU showed decreased elongation of muscle and increased elongation and contribution of tendon, whereas the GM MTU displayed no changes. Elongations normalized to resting MTU length give the same statistical outcomes as absolute values and are not reported. (For absolute values and detailed statistics, see Table, Supplemental Digital Content 2,

TABLE 2 - Length, elongation, and strain of GM fascicles during passive dorsiflexion, before (PRE) and after (POST) 24 wk of stretching.
PRE POST P Inter. P Side P Time Post Hoc P ES
GM fascicle length, mm
 At anatomical angle
  Stretching 58.4 ± 8.3 59.9 ± 8.6 0.46 0.69 <0.005 0.15 0.21
  Control 57.0 ± 8.0 59.4 ± 7.7 <0.05
 At standardized angle
  Stretching 61.7 ± 8.7 64.4 ± 9.5 0.15 0.88 <0.001 <0.05 0.41
  Control 60.3 ± 7.3 65.1 ± 9.1 <0.001
 At maximal angle
  Stretching 61.7 ± 8.7 67.4 ± 8.7 0.76 0.65 <0.001 <0.001 0.09
  Control 60.4 ± 7.2 66.5 ± 9.7 <0.001
GM fascicle elongation, mm
 At anatomical angle
  Stretching 4.4 ± 7.2 5.6 ± 7.2 0.40 0.95 <0.05 0.48 0.25
  Control 3.7 ± 5.9 6.1 ± 5.4 0.06
 At standardized angle
  Stretching 7.8 ± 7.5 10.1 ± 8.1 0.15 0.81 <0.001 0.13 0.43
  Control 6.9 ± 5.6 11.8 ± 6.3 <0.001
 At maximal angle
  Stretching 7.8 ± 7.5 13.1 ± 7.5 0.56 0.89 <0.001 <0.001 0.17
  Control 7.0 ± 5.6 13.3 ± 6.9 <0.001
GM fascicle strain, %
 At anatomical angle
  Stretching 8.7 ± 13.7 10.7 ± 13.2 0.40 0.97 <0.05 0.55 0.24
  Control 7.3 ± 11.4 11.8 ± 10.0 0.07
 At standardized angle
  Stretching 14.9 ± 14.3 19.0 ± 14.9 0.17 0.78 <0.001 0.16 0.41
  Control 13.5 ± 10.9 22.3 ± 10.8 <0.005
 At maximal angle
  Stretching 14.9 ± 14.3 24.6 ± 14.3 0.62 0.92 <0.001 <0.001 0.15
  Control 13.7 ± 10.8 25.1 ± 12.1 <0.001
Elongation and strain are reported based on fascicle resting length, as measured with subjects resting in prone position.
Inter., interaction.

Muscle and tendon elongation during passive dorsiflexion. GM and the whole Achilles tendon (AT; A and B), and SOL and the free Achilles tendon (C and D) in the stretching leg (STR) and control leg (CON) at standardized and maximal joint angle before (PRE) and after (POST) 24 wk of stretching. Numbers on the bars represent percent contribution to MTU elongation. †An interaction effect for elongation (P < 0.001–0.01). ‡A time effect for elongation and contribution (P < 0.001–0.05), ↓↑Significant post hoc tests for contribution (P < 0.001–0.05).

Tendon mechanical properties

Neither free Achilles tendon stiffness (interaction, P = 0.22; side, P = 0.52; time, P = 0.20; ES = 0.39; stretching, 554 ± 218 N·mm−1 for PRE and 552 ± 190 N·mm−1 for POST; control: 547 ± 184 N·mm−1 for PRE and 488 ± 144 N·mm−1 and for POST, Fig. 5A) nor whole Achilles tendon stiffness (interaction, P = 0.54; side, P = 0.89; time, P = 0.42; ES = 0.19; stretching, 349 ± 84 N·mm−1 for PRE and 347 ± 100 N·mm−1 for POST; control, 361 ± 84 N·mm−1 for PRE and 341 ± 76 N·mm−1 for POST; Fig. 5B) changed after the intervention. Tendon elongation and strain at common force level and at maximal elongation increased with time, in both legs for the whole Achilles tendon (elongation: interaction, P = 0.49–0.88; side, P = 0.23–0.37; time, P < 0.001; strain: interaction, P = 0.63–0.98; side, P = 0.23–0.24; time, P < 0.001; ES = 0.00–0.21; Fig. 5B), and in the control leg for the free Achilles tendon (elongation: interaction, P = 0.22–0.30; side, P = 0.76–0.87; time, P < 0.01; strain: interaction, P = 0.36–0.47; side, P = 0.75–0.99; time, P < 0.005; ES = 0.23–0.40; Fig. 5A).

Force–elongation relations during isometric plantar flexion. Free Achilles tendon (AT; A) and (B) whole Achilles tendon in the stretching leg (STR) and control leg (CON) before (PRE) and after (POST) 24 wk of stretching. Continuous lines represent group mean up to common force level, and symbols represent individual maximal force and elongation. ‡A time effect for tendon elongation (P < 0.001–0.01).

Contractile function

Maximal isometric plantar flexor torque was unchanged at all joint angles (interaction, P = 0.06–0.64; side, P = 0.72–0.83; time, P = 0.32–0.84; ES = 0.13–0.56; Fig. 6A). Concentric plantar flexor peak torque was unchanged across angular velocities (interaction, P = 0.32–0.64; side, P = 0.75–0.99; time, P = 0.14–0.74; ES = 0.18–0.29; Fig. 6B), whereas a time effect was seen for concentric dorsiflexion peak torque (interaction, P = 0.67; side, P = 0.97; time, P < 0.001; ES = 0.13; stretching, 16.1 ± 5.1 N·m for PRE and 17.6 ± 5.8 N·m for POST (P = 0.26); control, 16.2 ± 5.7 N·m for PRE and 17.4 ± 5.8 N·m for POST). The angle of peak torque of the plantar flexor muscles shifted toward more dorsiflexed position in both legs (Fig. 6C), at 30°·s−1, 45°·s−1, and 60°·s−1 (interaction, P = 0.62–0.84; side, P = 0.69–0.91; time, P < 0.005–0.05; ES = 0.06–0.15), but not at 90°·s−1 (interaction, P = 0.56; side, P = 0.88; time, P = 0.11; ES = 0.17). Work at 30°·s−1 was unchanged (interaction, P = 0.41; side, P = 0.80; time, P = 0.31; ES = 0.24; stretching, 156 ± 45 J for PRE and 160 ± 8 J for POST; control, 155 ± 35 J for PRE and 156 ± 38 J for POST).

Torque–angle and torque–velocity relations. Isometric plantar flexion torque–angle relation (A), isokinetic plantar flexion torque–velocity relation (B), isokinetic torque–angle of peak torque relation (C) in the stretching leg (STR) and control leg (CON) before (PRE) and after (POST) 24 wk of stretching. Note that the y axes differ and are broken. ‡A time effect (P < 0.005–0.05).


The purpose of this study was to examine the effects of 24 wk of functional unilateral triceps surae stretching on ROM, on the morphological and mechanical properties of the MTU, on neural activation, and on contractile function. Comparing the intervention and control legs, ROM, passive torque at maximal joint angle, and passive MTU elongation at maximal joint angle increased, supporting our hypotheses. Also as hypothesized, no indications of altered morphological properties were observed. Further to our hypotheses, significant changes were seen in the nonstretching control leg, suggesting a cross-education effect of habitual unilateral stretching.

When comparing studies, it must be noted that the intervention stimuli may be highly different, with respect to both study duration and stretching volume—time under stretch. In the present study, time under stretch was approximately 40,000 s for a duration of 24 wk, whereas for example, Magnusson et al. (1) used 9000 s over 3 wk and Ben and Harvey (5) used 56,000 s over 6 wk. The present study aimed to provide conclusions applicable to sports and exercise by applying doses that can realistically be maintained on a daily basis in sports and exercise settings over a duration that approaches habitual stretching.

ROM and cross-education

After 24 wk of stretching, ankle dorsiflexion ROM increased by 11° in the stretching leg and by 4° in the control leg. ROM continued increasing between 8 and 24 wk (Fig. 2), although the perceived pain during stretching exercise was significantly reduced after 8 wk. These data indicate that, although stretching may seem less intense after the initial weeks, ROM may still be affected. ROM did not change significantly between 16 and 24 wk. Possibly, the most flexible subjects were at this point approaching a level of ROM where the anatomical constraints of the ankle joint begin opposing further motion.

To the best of our knowledge, a bilateral increase in ROM after unilateral stretching training has not previously been reported. Although reports of acute bilateral gains in ROM following unilateral stretching exist (43,44), previous unilateral stretching training studies found unchanged control leg ROM (5,6,8,45,46) and unchanged control leg passive torque (8,45). Previous studies were limited to 3–10 wk of stretching, whereas in the present study, increases in control leg ROM did not reach significance until 16 wk. The present findings may be explained physiologically or biomechanically. Considering the proposed role of tolerance and sensory factors for increasing ROM (1,2), the bilateral responses to stretching training may likely be ascribed to adaptations in the neural pathways, somewhat similar to the cross-education effects reported with strength training (47). An alternative possible explanation is cross-education effects involving adaptations in the motor areas (48).

EMG amplitude and pain during passive dorsiflexion

When passively stretched to a standardized joint angle, the EMG amplitude of GM, GL, and SOL decreased after long-term stretching with no difference between legs (Figs. 3C, D), which is in contrast to previous reports of unchanged EMG amplitude after 3 wk of stretching (1). It is therefore possible that such neural adaptation may require greater durations or volumes of stretching. However, decreased EMG amplitude seems consistent with the reduced tonic reflex reported after 6 wk of stretching (6). Combined with the unchanged EMG amplitude and pain at the increased maximal joint angle, these findings support the notion that reflex activity and pain are important mechanisms for ROM.

Torque–angle properties during passive dorsiflexion

Passive torque and joint stiffness were reduced at comparable dorsiflexion angles after the present intervention (Figs. 3A, B). Previous intervention studies are inconsistent, with no change (1,3,5), or reduced torque at standardized joint angles (10–12) or reduced passive joint stiffness (6,8,13) after stretching. The discrepancies may be related to methodological differences, including intervention duration and/or stretch volume. In the present study, the reduced passive torque in both legs at standardized joint angle corresponds with the reduced EMG amplitude. Although the magnitude of the potential influence from reduced EMG amplitude cannot presently be assessed, these data support an association between neural activation and active muscle stiffness and passive resistance to stretch. However, reduced passive torque at anatomical joint angle (Fig. 3E), where almost no EMG was recorded, suggests that changes in neural activation do not fully explain the observed changes in passive torque. Because structural connective tissue properties contribute to passive resistance (49), this finding may indicate that stretching has induced a change in structural properties of connective tissue, although the present data set cannot ascertain whether such change occurred.

The present 35% increase in passive torque at maximal joint angle in the stretching leg is relatively similar to previous studies with short interventions (e.g., 28% after 3 wk [4]), despite large differences in ROM gains. This suggests that sensory adaptations enabling the subject to tolerate a greater maximal joint angle and torque may occur in the first weeks of stretching, whereas ROM is subsequently increased by factors that also reduce passive resistance, such as neural activation or structural adaptations.

Morphological properties

Unexpectedly, a small but consistent increase in GM thickness and resting pennation angle was observed bilaterally after the stretching intervention (Table 1). Limited explanation can be offered, other than perhaps seasonal variation due to PRE data being collected after summer and POST data being collected during spring. However, the self-reported activity level during the intervention period did not vary between subjects. On the other hand, no change was seen in muscle strength, and thus, any functional relevance is likely negligible. GM fascicle length and Achilles tendon length remained unchanged. A few human intervention studies have reported increased fascicle length after stretching (15,16), and longer GM fascicles have been observed in professional ballet dancers compared with nonstretching controls (34). To standardize tension in the present study, resting fascicle length was acquired with the foot hanging freely, and no difference was seen in resting ankle angle. Furthermore, fascicle length at 0° was 4–6 mm greater than at resting length and increased with time. The susceptibility of fascicle length measurements to low levels of tension demonstrated here may explain the discrepancy between the present and previous studies regarding fascicle length at 0°. For more direct measures of the effect of stretching training on muscle adaptations, invasive methods and/or more advanced imaging techniques, for example, second harmonic generation microendoscopy (50), could be applied. Given the current methods, no stretching-induced changes in muscle or tendon length were observed.

Tendon mechanical properties

The unchanged tendon stiffness observed in the present study is in line with previous studies applying static stretching training (3,8,11,22). However, reduced tendon stiffness has been reported after proprioceptive neuromuscular facilitation (23) or ballistic (11) stretching training, and in a cross-sectional study, ballet dancers had less Achilles tendon stiffness compared with controls (34). It is thus possible that potential tendon adaptations require stimuli of greater amplitude, for example, as achieved with strength training (51), than was attained by the present stretching protocol.

The increased tendon elongation and strain at common and maximal force levels of ramped isometric contractions (Fig. 5) remain compatible with unchanged tendon stiffness, as an increase in toe-region strain would allow for greater tissue elongation at low force levels without altering the force–elongation slope (stiffness) near maximal force. Thus, a potential effect of stretching training on tendon material properties and toe limit strain should be examined in further studies. Increased tendon elongation in the control leg is surprising because tendon tissue is not expected to respond to central neural adaptations; however, this finding is consistent with observations in the control leg during passive dorsiflexion.

Passive MTU elongation

At standardized joint angle, elongation of the GM muscle and whole Achilles tendon was unchanged in both legs. At the maximal joint angle, which was increased after training, the whole Achilles tendon elongation increased in both legs, whereas GM muscle elongation increased only in the stretching leg. In the stretching leg, the tendon contribution to total MTU elongation was increased at maximal joint angle; the relative contribution of tendon (35%) versus muscle (65%) changed to 51% and 49% from PRE to POST. In the control leg, the change in contribution from 39% and 61% PRE to 47% and 53% POST was not significant (P = 0.08). These values compare well to a recent report of 50/50 contribution between the muscle and the tendon during submaximal acute stretches (52). GM fascicle elongation and strain increased in the control leg at standardized joint angle and increased similarly in both legs at maximal joint angle (Fig. 4). The differences between the methods of calculating muscle and fascicle elongation (MTJ displacement vs midbelly scans) do not allow for direct comparisons between the two variables. Furthermore, muscle elongation may be slightly underestimated because of stretching of the skin when approaching large joint angles. These factors may explain the different contribution ratios reported in a previous cross-sectional study where GM fascicles accounted for 72% and tendon 28% at 30° dorsiflexion (53).

In the present study, the SOL MTU responded similarly to the GM MTU, but with increased contribution from the tendon in both legs at maximal joint angle and with decreased muscle elongation and increased tendon elongation in the control leg at standardized joint angle. The lack of change in the stretching leg at standardized joint angle contrasts with another intervention study that showed increased muscle and fascicle strain and decreased tendon strain after 3 wk of stretching (4). Differences between studies may be related to study duration and the unchanged passive torque at standardized joint angle in the study by Blazevich et al. (4). In the present study, the increased elongation of the free Achilles tendon in the control leg is surprising. However, the finding of increased tendon elongation during passive trials does match the findings of increased tendon elongation during ramped contraction, both at maximal elongation and at common force. Furthermore, a matching tendency was seen for both legs and muscles at standardized joint angle of passive dorsiflexion. Increases in tendon toe limit strain, as speculated previously, could potentially facilitate the increased tendon elongation and contribution at lower torque levels.

The present data suggest that when maximal MTU elongation increased after stretching training, elongation of SOL and GM muscle bellies contributed only in the stretching leg, indicating that stretching may have modified muscle stiffness. However, control leg ROM increased less than stretching leg ROM, constituting a smaller increase in total MTU elongation, which perhaps, owing to the sensitivity of muscle elongation measures, may explain why muscle elongation did not increase significantly in the control leg. Greater muscle elongation with greater ROM is in line with cross-sectional studies showing greater maximal muscle or fascicle elongation in flexible subjects (17,34,53). On the other hand, after the present intervention, tendon contribution to total MTU elongation increased. These findings match our former observation that both muscle and the series elastic element contribute to the greater MTU elongation in flexible ballet dancers (34). Potentially, at low tension, muscle/fascicle compliance may be large relative to tendon compliance (52), whereas at greater tension, muscle/fascicle compliance may be similar or lower than tendon compliance, letting the tendon take up an increased portion of the elongation. The relationship between tension, toe strain limit, muscle compliance, and tendon compliance entails closer examination in future studies.

Contractile function

Despite a small increase in GM thickness, the present intervention did not increase isometric or isokinetic strength, matching most (3,4,6,8,13,45) but not all (27,28) stretching intervention studies. Plantar flexion work was also unchanged, matching again some previous data (26) but not all (28). After the intervention, the angle of peak torque shifted toward a more dorsiflexed position at the lower angular velocities, with no difference between legs (Fig. 6), matching some (26,27) but not all studies (28). Taken together, unchanged torque and work production along with the architecture data suggest no addition of serial sarcomeres with the present stretching intervention.


Although the present 24-wk within-subjects design has advantages, there are also limitations to such a design. Muscle or tendon tissue properties may change because of external factors such as dietary status or seasonal changes in activity, affecting both legs similarly. It cannot be excluded that the changes seen in the control leg may be related hereto. Potential crossover effects may influence control–intervention comparisons, thus challenging the control design of the study. The lack of original reliability data for some variables limits the interpretation of the present findings. For logistic reasons, we did not collect a complete set of reliability data. The reliability of the methods is established in the literature, and the study statistical power was deemed sufficient a priori. However, the changes in the control group were not anticipated, and with hindsight, a complete set of reliability data specific to this study would have helped mitigating the risk of type I error. Additional studies are therefore required to ascertain that the present changes in control leg ROM as well as passive torque, elongation, and EMG amplitude during passive trials originate from physiological factors.

Learning effects in, for example, tests of contractile function can potentially bias results; however, the test sessions were separated by 8–16 wk, so such effects seem unlikely. Investigator bias cannot be ruled out, but during passive tests, investigators and subjects had no active role, and during post processing, investigators were blinded. It seems that for the automated torque–angle analyses, the changes in passive torque and EMG likely reflect the intervention. The potential changes to be observed by ultrasonography are small, and some measurements may be approaching current resolution limits. Finally, although the triceps surae muscles are well suited for ultrasonography and investigations of passive resistance, there may be joint constraints (e.g., ligamentous constraints or bony contact) that restrict ROM from approaching extreme joint angles. The most flexible subjects may have approached such angles within the 24 wk, leading to an underestimation of the effects of stretching.


Twenty-four weeks of functional unilateral static stretching increased dorsiflexion ROM, maximal passive dorsiflexion torque, and MTU elongation at maximal joint angles in the stretching leg compared with the control leg, affirming the role of central neural factors. Passive dorsiflexion torque and EMG amplitude at standardized joint angles decreased, suggesting a role of peripheral factors. By 16 wk, ROM also increased in the control leg. The significant and conforming PRE–POST changes both in the training leg and in the control leg are taken to indicate a crossover effect of habitual stretching.

The greater passive torque tolerated by the stretching leg indicates that pain threshold or pain sensitivity was affected, which may have influenced muscle stiffness through reduced sensory input. This corresponds with reduced passive torque in both legs at comparable PRE–POST joint angles. Reflex activity may contribute to passive torque via its contribution to muscle stiffness. This is supported by the corresponding reductions in EMG amplitude and passive torque at standardized joint angle. However, reduced passive torque at anatomical joint angle, where almost no EMG amplitude was recorded, suggests that neural activation does not entirely explain the change in passive torque.

In conclusion, stretching training increases ROM through central neural adaptations and possibly structural adaptations within the MTU or within passive tissues that span the joint, and has the potential to modify passive and active torque–angle relations.

Further research is required to understand the relations between pain, reflexes, neural activation, resistance of passive structures, passive torque, and ROM with stretching training. Additional insight may be gained, for example, by investigating central responses and connective tissue composition or by applying more advanced imaging techniques during passive motion.

The authors would like to acknowledge Sarang Arayesh, Christian Marin, and Olaug Gullbrå for assistance in pilots, testing, or analysis.

The authors declare no conflicts of interest and no external funding. The results of the present study do not constitute endorsement by the American College of Sports Medicine. The authors declare that the results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.


1. Magnusson SP, Simonsen EB, Aagaard P, Sorensen H, Kjaer M. A mechanism for altered flexibility in human skeletal muscle. J Physiol. 1996;497(Pt 1):291–8.
2. Freitas SR, Mendes B, Le Sant G, Andrade RJ, Nordez A, Milanovic Z. Can chronic stretching change the muscle–tendon mechanical properties? A review. Scand J Med Sci Sports. 2018;28(3):794–806.
3. Konrad A, Tilp M. Increased range of motion after static stretching is not due to changes in muscle and tendon structures. Clin Biomech. 2014;29(6):636–42.
4. Blazevich AJ, Cannavan D, Waugh CM, et al. Range of motion, neuromechanical, and architectural adaptations to plantar flexor stretch training in humans. J Appl Physiol. 2014;117(5):452–62.
5. Ben M, Harvey LA. Regular stretch does not increase muscle extensibility: a randomized controlled trial. Scand J Med Sci Sports. 2010;20(1):136–44.
6. Guissard N, Duchateau J. Effect of static stretch training on neural and mechanical properties of the human plantar–flexor muscles. Muscle Nerve. 2004;29(2):248–55.
7. Hayes BT, Harter RA, Widrick JJ, Williams DP, Hoffman MA, Hicks-Little CA. Lack of neuromuscular origins of adaptation after a long-term stretching program. J Sport Rehabil. 2012;21(2):99–106.
8. Kubo K, Kanehisa H, Fukunaga T. Effect of stretching training on the viscoelastic properties of human tendon structures in vivo. J Appl Physiol. 2002;92(2):595–601.
9. Toft E, Espersen GT, Kalund S, Sinkjaer T, Hornemann BC. Passive tension of the ankle before and after stretching. Am J Sports Med. 1989;17(4):489–94.
10. Chan SP, Hong Y, Robinson PD. Flexibility and passive resistance of the hamstrings of young adults using two different static stretching protocols. Scand J Med Sci Sports. 2001;11(2):81–6.
11. Mahieu NN, McNair P, De Muynck M, et al. Effect of static and ballistic stretching on the muscle–tendon tissue properties. Med Sci Sports Exerc. 2007;39(3):494–501.
12. Nakamura M, Ikezoe T, Takeno Y, Ichihashi N. Effects of a 4-week static stretch training program on passive stiffness of human gastrocnemius muscle–tendon unit in vivo. Eur J Appl Physiol. 2012;112(7):2749–55.
13. Marshall PW, Cashman A, Cheema BS. A randomized controlled trial for the effect of passive stretching on measures of hamstring extensibility, passive stiffness, strength, and stretch tolerance. J Sci Med Sport. 2011;14(6):535–40.
14. Tabary JC, Tabary C, Tardieu C, Tardieu G, Goldspink G. Physiological and structural changes in the cat’s soleus muscle due to immobilization at different lengths by plaster casts. J Physiol. 1972;224(1):231–44.
15. Freitas SR, Mil-Homens P. Effect of 8-week high-intensity stretching training on biceps femoris architecture. J Strength Cond Res. 2015;29(6):1737–40.
16. Simpson CL, Kim BDH, Bourcet MR, Jones GR, Jakobi JM. Stretch training induces unequal adaptation in muscle fascicles and thickness in medial and lateral gastrocnemii. Scand J Med Sci Sports. 2017;27(12):1597–604.
17. Blazevich AJ, Cannavan D, Waugh CM, Fath F, Miller SC, Kay AD. Neuromuscular factors influencing the maximum stretch limit of the human plantar flexors. J Appl Physiol. 2012;113(9):1446–55.
18. e Lima KMM, Carneiro SP, Alves Dde S, Peixinho CC, de Oliveira LF. Assessment of muscle architecture of the biceps femoris and vastus lateralis by ultrasound after a chronic stretching program. Clin J Sport Med. 2015;25(1):55–60.
19. Nakamura M, Ikezoe T, Umegaki H, Kobayashi T, Nishishita S, Ichihashi N. Changes in passive properties of the gastrocnemius muscle–tendon unit during a 4-week routine static-stretching program. J Sport Rehabil. 2017;26(4):263–8.
20. Kay AD, Richmond D, Talbot C, Mina M, Baross AW, Blazevich AJ. Stretching of active muscle elicits chronic changes in multiple strain risk factors. Med Sci Sports Exerc. 2016;48(7):1388–96.
21. Herbert RD, Bolsterlee B, Gandevia SC. Passive changes in muscle length. J Appl Physiol. 2019;126(5):1445–53.
22. Mahieu NN, Cools A, De Wilde B, Boon M, Witvrouw E. Effect of proprioceptive neuromuscular facilitation stretching on the plantar flexor muscle–tendon tissue properties. Scand J Med Sci Sports. 2009;19(4):553–60.
23. Konrad A, Gad M, Tilp M. Effect of PNF stretching training on the properties of human muscle and tendon structures. Scand J Med Sci Sports. 2015;25(3):346–55.
24. Gordon AM, Huxley AF, Julian FJ. The variation in isometric tension with sarcomere length in vertebrate muscle fibres. J Physiol. 1966;184(1):170–92.
25. Lieber RL, Friden J. Functional and clinical significance of skeletal muscle architecture. Muscle Nerve. 2000;23(11):1647–66.
26. Ferreira GN, Teixeira-Salmela LF, Guimaraes CQ. Gains in flexibility related to measures of muscular performance: impact of flexibility on muscular performance. Clin J Sport Med. 2007;17(4):276–81.
27. Chen CH, Nosaka K, Chen HL, Lin MJ, Tseng KW, Chen TC. Effects of flexibility training on eccentric exercise-induced muscle damage. Med Sci Sports Exerc. 2011;43(3):491–500.
28. Handel M, Horstmann T, Dickhuth HH, Gulch RW. Effects of contract–relax stretching training on muscle performance in athletes. Eur J Appl Physiol Occup Physiol. 1997;76(5):400–8.
29. Simao R, Lemos A, Salles B, et al. The influence of strength, flexibility, and simultaneous training on flexibility and strength gains. J Strength Cond Res. 2011;25(5):1333–8.
30. Donti O, Papia K, Toubekis A, Donti A, Sands WA, Bogdanis GC. Flexibility training in preadolescent female athletes: acute and long-term effects of intermittent and continuous static stretching. J Sports Sci. 2018;36(13):1453–60.
31. Nordez A, Gross R, Andrade R, et al. Non-muscular structures can limit the maximal joint range of motion during stretching. Sports Med. 2017;47(10):1925–9.
32. Herzberg SD, Motu’apuaka ML, Lambert W, Fu R, Brady J, Guise JM. The effect of menstrual cycle and contraceptives on ACL injuries and laxity: a systematic review and meta-analysis. Orthop J Sports Med. 2017;5(7):2325967117718781.
33. Schindelin J, Arganda-Carreras I, Frise E, et al. Fiji: an open-source platform for biological-image analysis. Nat Methods. 2012;9(7):676–82.
34. Moltubakk MM, Magulas MM, Villars FO, Seynnes OR, Bojsen-Moller J. Specialized properties of the triceps surae muscle–tendon unit in professional ballet dancers. Scand J Med Sci Sports. 2018;28(9):2023–34.
35. Hermens HJ, Freriks B, Disselhorst-Klug C, Rau G. Development of recommendations for SEMG sensors and sensor placement procedures. J Electromyogr Kinesiol. 2000;10(5):361–74.
36. Raiteri BJ, Cresswell AG, Lichtwark GA. Ultrasound reveals negligible cocontraction during isometric plantar flexion and dorsiflexion despite the presence of antagonist electromyographic activity. J Appl Physiol. 2015;118(10):1193–9.
37. Arampatzis A, Monte GD, Karamanidis K. Effect of joint rotation correction when measuring elongation of the gastrocnemius medialis tendon and aponeurosis. J Electromyogr Kinesiol. 2008;18(3):503–8.
38. Cronin NJ, Carty CP, Barrett RS, Lichtwark G. Automatic tracking of medial gastrocnemius fascicle length during human locomotion. J Appl Physiol. 2011;111(5):1491–6.
39. Hawkins D, Hull ML. A method for determining lower extremity muscle–tendon lengths during flexion/extension movements. J Biomech. 1990;23(5):487–94.
40. Magnusson SP, Simonsen EB, Aagaard P, Gleim GW, McHugh MP, Kjaer M. Viscoelastic response to repeated static stretching in the human hamstring muscle. Scand J Med Sci Sports. 1995;5(6):342–7.
41. Ce E, Longo S, Rampichini S, et al. Stretch-induced changes in tension generation process and stiffness are not accompanied by alterations in muscle architecture of the middle and distal portions of the two gastrocnemii. J Electromyogr Kinesiol. 2015;25(3):469–78.
42. Cohen J. Statistical Power Analysis for the Behavioral Sciences. Hillsdale: Lawrence Erlbaum Associates; 1988.
43. Chaouachi A, Padulo J, Kasmi S, Othmen AB, Chatra M, Behm DG. Unilateral static and dynamic hamstrings stretching increases contralateral hip flexion range of motion. Clin Physiol Funct Imaging. 2017;37(1):23–9.
44. Killen BS, Zelizney KL, Ye X. Crossover effects of unilateral static stretching and foam rolling on contralateral hamstring flexibility and strength. J Sport Rehabil. 2019;28(6):533–9.
45. Akagi R, Takahashi H. Effect of a 5-week static stretching program on hardness of the gastrocnemius muscle. Scand J Med Sci Sports. 2014;24(6):950–7.
46. Nelson AG, Kokkonen J, Winchester JB, et al. A 10-week stretching program increases strength in the contralateral muscle. J Strength Cond Res. 2012;26(3):832–6.
47. Scripture EW, Smith TL, Brown EM. On the education of muscular control and power. Studies from the Yale Psychological Laboratory, 2. 1894. Available from: Accessed May 4, 2018.
48. Lee M, Carroll TJ. Cross education: possible mechanisms for the contralateral effects of unilateral resistance training. Sports Med. 2007;37(1):1–14.
49. Meyer GA, Lieber RL. Elucidation of extracellular matrix mechanics from muscle fibers and fiber bundles. J Biomech. 2011;44(4):771–3.
50. Sanchez GN, Sinha S, Liske H, et al. In vivo imaging of human sarcomere twitch dynamics in individual motor units. Neuron. 2015;88(6):1109–20.
51. Arampatzis A, Karamanidis K, Albracht K. Adaptational responses of the human Achilles tendon by modulation of the applied cyclic strain magnitude. J Exp Biol. 2007;210(Pt 15):2743–53.
52. Herbert RD, Heroux ME, Diong J, Bilston LE, Gandevia SC, Lichtwark GA. Changes in the length and three-dimensional orientation of muscle fascicles and aponeuroses with passive length changes in human gastrocnemius muscles. J Physiol. 2015;593(2):441–55.
53. Abellaneda S, Guissard N, Duchateau J. The relative lengthening of the myotendinous structures in the medial gastrocnemius during passive stretching differs among individuals. J Appl Physiol. 2009;106(1):169–77.


Supplemental Digital Content

Copyright © 2021 by the American College of Sports Medicine