The acute effects of higher versus lower load duration and intensity on morphological and mechanical properties of the healthy Achilles tendon: a randomized crossover trial

ABSTRACT The Achilles tendon (AT) exhibits volume changes related to fluid flow under acute load which may be linked to changes in stiffness. Fluid flow provides a mechanical signal for cellular activity and may be one mechanism that facilitates tendon adaptation. This study aimed to investigate whether isometric intervention involving a high level of load duration and intensity could maximize the immediate reduction in AT volume and stiffness compared with interventions involving a lower level of load duration and intensity. Sixteen healthy participants (12 males, 4 females; age 24.4±9.4 years, body mass 70.9±16.1 kg, height 1.7±0.1 m) performed three isometric interventions of varying levels of load duration (2 s and 8 s) and intensity (35% and 75% maximal voluntary isometric contraction) over a 3 week period. Freehand 3D ultrasound was used to measure free AT volume (at rest) and length (at 35%, 55% and 75% of maximum plantarflexion force) pre- and post-interventions. The slope of the force–elongation curve over these force levels represented individual stiffness (N mm−1). Large reductions in free AT volume and stiffness resulted in response to long-duration high-intensity loading whilst less reduction was produced with a lower load intensity. In contrast, no change in free AT volume and a small increase in AT stiffness occurred with lower load duration. These findings suggest that the applied load on the AT must be heavy and sustained for a long duration to maximize immediate volume reduction, which might be an acute response that enables optimal long-term tendon adaptation via mechanotransduction pathways.


Introduction
Tendons are connective tissues comprised of densely packed collagen fibers (solid phase) and water-binding proteoglycans (fluid phase) (Kjaer 2004). Tendons can adapt to a wide range of load requirements via changing their mechanical (i.e., stiffness) and morphological (i.e., thickness and cross-sectional area [CSA]) properties, and this may have implications for performance and injury (Bohm et al. 2015). For example, tendons become more extensible (Kubo et al. 2001b(Kubo et al. , 2009Kay and Blazevich 2009;Burgess et al. 2009;Joseph et al. 2014) and smaller in thickness (Wearing et al. 2007(Wearing et al. , 2014Grigg et al. 2009;Kristiansen et al. 2014) when subjected to acute load, and become stiffer Journal of Experimental Biology • Accepted manuscript (Kubo et al. 2012(Kubo et al. , 2001a(Kubo et al. , 2006 and may hypertrophy (Arampatzis et al. 2007;Couppé et al. 2008;Bohm et al. 2014;Geremia et al. 2018;Kongsgaard et al. 2007) when loaded over long-term periods (> 12-week). The mechanisms responsible for tendon adaptation to load are currently unclear.
It is well accepted that mechanical load raises hydrostatic pressure within the fluid phase causing fluid flow (movement). This may include fluid exudation, that is, fluid movement out of the tendon causing reduced fluid volume which has been demonstrated in ex vivo tendon loading (Butler 1997;Lavagnino et al. 2003;Hannafin and Arnoczky 1994). Tendon undergoing fluid flow/exudation may become thinner or smaller in volume. From a mechanical perspective, this means higher levels of mechanical stress on the solid phase (collagen) and a greater tissue strain. Moreover, load-induced fluid flow provides a mechanical signal for cellular activity (Docking 2013;Wall et al. 2016). This is likely to occur because fluid flows away from the tendon core (it becomes thinner) so tensile tenocyte load may increase (as predicted by Poisson's ratio; Iwanuma et al. (2011)). Shear stresses from the fluid flow may also directly impact tenocyte mechanotransduction and signaling (Archambault et al. 2002;Lavagnino and Arnoczky 2005;Lavagnino et al. 2003Lavagnino et al. , 2008. Therefore, it is reasonable to speculate that load-induced fluid flow/exudation is an important mechanism in tendon mechanotransduction and adaptation.
Fluctuation in fluid content (fluid loss) has been proposed as a potential mechanism responsible for the reduction in tendon thickness/diameter observed post-acute loading (Wearing et al. 2007;Grigg et al. 2012;Grigg et al. 2009). However, reduction in tendon thickness could occur with fluid redistributing within the tendon with no exudation, the latter may well change tendon morphology (volume) (Nuri et al. 2017c,a). The in vivo evidence on AT volume change with acute loading is unclear. There are reports of a small but significant reduction in healthy AT volume immediately after cross-country running (Syha et al. 2014;Grosse et al. 2016), or a lack of acute change after eccentric (Obst et al. 2015) and submaximal isometrics (10 reps. at 50% MVIC) (Nuri et al. 2017b) whilst others found an immediate reduction in tendinopathic Achilles volume with submaximal isometrics (at 50% MVIC) (Nuri et al. 2017c(Nuri et al. , 2018 or an immediate

Journal of Experimental Biology • Accepted manuscript
increase with eccentric calf training (Shalabi et al. 2004). Understanding load parameters that may trigger maximal reduction in tendon volume (meaning substantial fluid flow and exudation) may be useful in designing future loading protocols to drive the tendon toward optimal long-term adaptation (i.e., increased stiffness and tendon hypertrophy).
Volume changes of the AT are likely to be linked to the mechanical stress applied (e.g., load duration and intensity), however there is a paucity of supporting evidence. It has been shown that tendons display greater deformation "creep" when loaded slowly or for longer duration because of their viscoelastic properties (Pearson et al. 2007)  reps) and long duration (15 s, 1 set × 17 reps) high-intensity isometric contractions and found an immediate increase in Achilles tendon elongation only after the long-duration contraction. Previous studies by the same group showed that long-duration and highintensity isometric contractions over 12 weeks resulted in a greater increase in tendon stiffness compared to shorter-duration contractions (Kubo et al. 2001a(Kubo et al. , 2006. However, none of these studies has linked changes in mechanical properties (i.e., stiffness) with changes in tendon morphology (i.e., volume). Whether and to what extent long-duration and high-intensity loading could lead to reductions in tendon volume and stiffness in healthy AT, and whether change in material properties (i.e., Young's modulus) occurs in parallel, is still unclear.
This study aimed to investigate the acute effect of long-duration and high-intensity loading on the volume and stiffness of healthy free AT (i.e., pathology free or noninjured tendon). We hypothesized that a higher level of load duration and intensity could result in greater immediate reductions in tendon volume and stiffness compared to lower load duration and intensity.

Trial design
A randomized crossover trial was used to examine whether a higher level of load duration and intensity would result in a greater immediate reduction in tendon volume and stiffness compared to lower level of load duration and intensity. This crossover trial involved three different isometric interventions implemented over a period of three weeks and separated by a one-week washout period between interventions, to ensure recovery of tendon properties (volume and stiffness) to baseline. Crossover design provides the advantage of controlling individual differences (confounders) since the subjects are their own controls, thus, more statistical power and smaller sample size (Portney and Watkins 2014). This crossover trial was reported in accordance with the CONSORT statement for this trial type (Dwan et al. 2019) and was prospectively registered (website: Australian New Zealand Clinical Trials Registry; registration record: ACTRN12620000306910).

Participants
Sixteen healthy participants (12 males, 4 females; age= 24.4 ± 9.4 y; body mass= 70.9 ± 16.1 kg; height= 1.7 ± 0.1 m) volunteered to participate in this study. Participants were recruited from the local university student and staff population and via personal contacts of the researchers. All participants were recreationally active (see Table S1).

Interventions
Each participant attended the laboratory on three occasions, at the same time of the day, with a 7-day interval between sessions. The participants were asked to refrain from strenuous physical activities 24 h prior to each testing session. One isometric intervention was performed per testing session in a random sequence (according to the allocated sequence for each participant). The isometric interventions varied in load duration (contraction duration) and load intensity but consisted of a similar volume (10 reps × 4 sets) and a similar rest time between the repetitions (10 s) and sets (3 min.) (Table 1). This allowed direct comparison amongst the interventions for the effects of different load duration and load intensity on the AT morphological and mechanical properties. To ensure that target load duration and load intensity for each intervention were consistently attained, a VBScript (Visual Basic Scripting) was created for each intervention using LabChart 8 software scripting (ADInstruments, Spechbach, Germany) and displayed on real-time visual feedback.
All interventions were undertaken with the participants seated in a custom-built dynamometer (Fig. 1A). The hip was in 120 o of flexion, the knee was in full extension (knee angle= 0 o ), and the tested ankle (dominant side) was in 0 o plantar flexion (PF) with the lateral malleolus aligned to the axis of rotation of the dynamometer. The tested ankle was firmly strapped to the footplate of the dynamometer using a ladder strap to prevent downward foot slide during the intervention and was repositioned or tightened if movement was detected (this was done at the end of each set). To prevent knee flexion during PF contractions, a non-elastic strap was positioned above the knees.
Participants were instructed to position both arms on the chest and to rest the nontested leg down to the floor during the intervention to prevent compensation patterns.
In the first session and before commencing the isometric intervention, participants performed 3-5 trials of maximal PF isometric contractions (2-3 s) while seated in the dynamometer to establish target load intensity for the allocated intervention and subsequent interventions (i.e., 35% and 75%). Standardised instructions were provided during testing to motivate participants and ensure achieving maximal PF torque. To Journal of Experimental Biology • Accepted manuscript prevent fatigue, participants were given 1-2 minutes rest between maximal PF isometric contractions test and commencing the first intervention.

Preconditioning and MVIC test
Tendon preconditioning was performed at each test session, followed by AT volume and stiffness assessment, the intervention, and then a repeat of the AT volume and stiffness assessment (Fig. 2). To precondition the tendon, participants performed five 80% maximal voluntary isometric contractions (MVICs) of PF (3-5 s) whilst positioned prone on the same dynamometer ( Fig. 1B). Both hips were secured to the dynamometer using a nonelastic strap to prevent forward body shift during PF trials and the ankle was kept in 0 o plantar flexion and strapped to the footplate using a ladder strap. In the first session and with employing an identical positioning, tendon preconditioning was modified to include PF MVIC testing to establish torque levels for AT stiffness assessment (i.e., staged protocol; see below). The participants performed between 3 to 5 trials of ramped isometric PF contraction to a maximum over 3-5 s. duration. Maximal PF torque was confirmed to be attained when participants performed the two highest trials within 10%. Standardised instructions were provided prior to testing (push toward the footplate and build up your force to maximum within 2-3 s. and hold at maximum for 2-3 s) and during testing (Push as hard as possible and hold for 2-3 s) to motivate participants and ensure achieving maximal PF torque. All torque data were displayed on real-time visual feedback, recorded at 1000 Hz with a PowerLab 26T, and analyzed using LabChart 8 software (all ADInstruments, Spechbach, Germany).

Freehand-3D ultrasound measures of Achilles tendon volume and stiffness
Using freehand three-dimensional ultrasound (3DUS), the free AT volume and length were measured pre and post each intervention ( Fig. 2) whilst participants were positioned prone on the same dynamometer (Fig. 1B). The hips and knees were extended, and the tested ankle was in 0 o plantar flexion with the lateral malleolus aligned to the axis of rotation of the force transducer. The freehand 3DUS system that Journal of Experimental Biology • Accepted manuscript was utilized in this study has been described in previous studies investigating the mechanical and morphological properties of musculoskeletal tissues including AT (Barber et al. 2009;Farris et al. 2013;Merza et al. 2021;Obst et al. 2014a) and the setup in the current study was identical to those we used in a previous study (Merza et al. 2021). The freehand 3DUS includes a B-mode ultrasound device (ArtUS EXT-1H, TELEMED, Vilnius, Lithuania, EU) and a four-camera optical tracking system (OptiTrack

Journal of Experimental Biology • Accepted manuscript
To obtain measurements of free AT morphology (i.e., resting length and cross-sectional area), two single sweep 3DUS scans were performed at rest (pre and post intervention).
Scans were performed at a steady speed from calcaneal notch to soleusmusculotendinous junction (MTJ). The duration of single scanning lasted between 15 to 20 s with approximately 0.1 mm. distance between acquired US frames.
To estimate the free AT stiffness, two 3DUS scans from calcaneal notch to soleus-MTJ were performed during sustained PF contractions at 35%, 55%, and 75% MVICs in a random order (staged protocol). The PF torque levels for the staged protocol were standardized across the three testing sessions based on the values achieved in the first testing session. The participants were requested to hold each contraction until completion of scanning while maintaining the target force level and were given 30-60 s rest between each trial to minimize fatigue. To minimize heel lift during the sustained submaximal PF contractions (also during PF MVIC test), the ankle was firmly strapped to the footplate of the dynamometer using a ladder strap (Arya and Kulig 2010).
Infrequently, re-scanning was done in case the participant was unable to achieve or steadily maintain the target torque level or in case of excessive heel lift (the strap was tightened prior to re-scanning). To reduce the influence of contraction duration (time under tension) on tendon viscoelastic behaviour and prevent tendon creep, an effort was made to standardise scanning duration at 8 s for all MVIC trials.

Electromyography and muscle co-contraction
To account for the co-contraction of tibialis anterior (TA) muscle during submaximal PF contractions and to obtain a true PF torque, maximal dorsiflexion (DF) contractions were performed (Arya and Kulig 2010) (the position of DF MVIC test was identical to PF MVIC test -see above). The participants were asked to perform up to 3 trials of ramped maximal isometric DF contraction until the two highest were within 10%. The participants were instructed to flex their foot against the ladder strap (i.e., isometric DF of the ankle) as hard as possible and hold for 2-3 s. Antagonist DF torque = P(1) × Up1 + P (2) Where P(1) is the slope of the gradient of the relationship between TA EMG activity and recorded torque during the DF contractions, Up1 is the envelope of the TA EMG signal, and P(2) is the regression constant.

Tendon moment arm
To obtain PF force, the resultant true PF torque was divided by the internal moment arm represented the AT moment arm (i.e., the perpendicular distance from COR to the LOA) (Geremia et al. 2018;Kongsgaard et al. 2011 Maganaris et al. 1998).

Freehand 3D-ultrasound image analysis and reconstruction of free AT volume
Stradwin software was used to perform image analysis. To measure free AT length (at rest and MVICs) two anatomical sites: calcaneal notch and soleus-MTJ were manually identified on sagittal and transverse images, then marked using a landmark tool (twopoint method) (Fig. 3A, B, C). The soleus-MTJ was the first visible cross-section of muscle tissue and was determined using sagittal, frontal, and transverse, image planes.
The distance between the calcaneal notch and soleus-MTJ was defined as the free AT and the length was measured using the length tool within the Stradwin software.
Tendon cross-sections were manually contoured from the transverse B-mode images from the calcaneal notch to soleus-MTJ at 5-10 mm interval (Fig. 3D). The 3D tendon image reconstruction was performed on the segmented cross sections using the in-built interpolation algorithm in Stradwin software (Fig. 3E).

Determination of free AT stiffness and Young's modulus
To calculate the free AT stiffness (N/mm), all torque and EMG data obtained at maximal DF and submaximal PF trials, as well as data for the free AT length (i.e., at rest and during MVICs), were exported to MATLAB software R2019b (MathWorks, Natick, MA, USA). To measure the amount of tendon elongation at 35%, 55%, and 75% MVIC, the mean resting length was subtracted from the mean length obtained at each corresponding MVIC level. The slope of the line fitted to the force-elongation curve (i.e., PF force at 35%, 55%, and 75% MVIC and corresponding tendon elongation) across all force levels represented the individual stiffness. Tendon stiffness was normalised to Journal of Experimental Biology • Accepted manuscript CSA and length to determine Young's modulus which was calculated by dividing tendon force with the average resting CSA (i.e., stress MPa) and tendon elongation with the resting length (i.e., strain %). The average CSA was determined from the contoured cross sections using a custom MATLAB script. Young's modulus was the slope of the line fitted to the stress-strain curve across the entire stress range.

Sample size calculation
The sample size was calculated using our pilot data (unpublished). We found a Cohen's standardised effect of ≥ 0.5 between the study interventions. Therefore, to achieve an effect size of ≥ 0.5 with 5% significance level and power of 80%, the sample size required for this crossover trial was n=11. This assumed a correlation between repeated measures of 0.2.

Randomization
Each participant performed three different isometric interventions in a random sequence using a computer-based random sequence generation. An off-site researcher (P.M) not involved in data collection informed the researcher who performed data collection (E.M) of the allocation sequence. This occurred immediately prior to the first session. Due to the nature of the interventions, the participants were aware of the exercise intervention they performed each week. However, we were very careful to ensure that the participants were not aware of the hypothesis of the study, and at no time was this disclosed in any written materials or verbal interactions.

Reliability of tendon volume and stiffness measurements
Freehand 3DUS has been shown to provide accurate measures of phantom volumes (± 0.5 ml) and reliable measures for in vivo free AT volume (intra-class correlation coefficient (ICC) ≥ 0.98) (Obst et al. 2014a) and stiffness (ICC= 0.994) (Merza et al. 2021

Reliability of plantar flexor force
Between-sessions (using pre-interventions data) and within-session reliability of plantar flexor force at 35%, 55%, and 75% MVICs were assessed using the intra-class correlation coefficient [ICCs (3,1), two-way mixed-effects model with absolute agreement] and 95% confidence intervals (CIs), coefficient of variation (CV), and the minimal detectable change (MDC). The between-sessions and within-session ICCs for PF force at all MVIC levels were greater than 0.97 (lower band of 95% confidence interval was ≥ 0.845). The CV and MDC ranged between 0% to 9% and between 68.5 to 179 N, respectively.

Statistical methods
A (2 × 3) repeated measures ANOVA with time (pre-post intervention) and type of isometric intervention (1, 2, and 3) as within-subjects factors, was used to investigate the interaction and simple effects of the time and type of intervention on the following dependent variables i) free AT volume, and ii) free AT stiffness. Data normality was evaluated with the Kolmogorov-Smirnov test. Effect size (ES) was reported for betweengroup differences (ES= difference between change scores/pooled change score standard deviation) (Middel and van Sonderen 2002

Free AT volume
There was a significant time-by-intervention interaction effect (F=25.098, df= (2,30), P≤ 0.001) on the free AT volume. Interaction (within-subjects) contrasts analysis shows that the acute effect of the 8 s, 75% MVIC intervention on the free AT volume was significantly greater than the 2 s, 75% MVIC (ES= 2) and 8 s, 35% MVIC (ES= 1.4) interventions but there was no significant difference between the 2 s, 75% MVIC and 8 s, 35% MVIC interventions (ES= 0.5) on the free AT volume (Fig. 4C). The simple effects analysis shows that the 2 s, 75% MVIC intervention had no significant acute effects on the free AT volume (P= 0.106) which changed only by an average of < -0.1 ml (2.3%), whilst the 8 s, 75% MVIC and 8 s, 35% MVIC interventions significantly reduced the free AT volume by an average of 0.3 ml (13.6%) (P≤ 0.001) and 0.1 ml (5.3%) (P= 0.004), respectively (Fig. 4A). The same test indicates no significant differences in the free AT volume at baseline (pre-intervention) across interventions (P ≥ 0.05). The 8 s, 75% MVIC intervention had the greatest acute effect on free AT volume. The pre-and postintervention mean (s.d.) for the free AT volume are shown in Table 2.

Free AT Stiffness
There was a significant time-by-intervention interaction effect (F= 60.503, df= (2,30), P≤ 0.001) on the free AT stiffness. Interaction (within-subjects) contrasts analysis shows that the acute effect of the 8 s, 75% MVIC intervention on the free AT stiffness was significantly greater than the 2 s, 75% MVIC (ES= 0.9) and 8 s, 35% MVIC (ES= 0.8) interventions but there was no significant difference between the 2 s, 75% MVIC and 8 s, 35% MVIC interventions (ES= 0.0) on the free AT stiffness (Fig. 4D).  Table 2.

Discussion
The present study investigated the acute effect of long-duration and high-intensity loading on the morphological and mechanical properties of the free AT, in healthy humans. Consistent with our hypothesis, the main finding was that a long-duration and high-intensity loading resulted in greater immediate reductions in the volume and stiffness of the free AT compared to lower levels of load duration and intensity.

The acute effect of different isometric interventions on the free AT volume
The present study found that long-duration contraction at 75% MVIC reduced free AT volume by 60% more than long-duration contraction at 35% MVIC and 82.3% more than short-duration contraction at 75% MVIC. The reduction in the free AT volume found in the current study is likely a manifestation of transient fluid flow and exudation. This is in light of the ex vivo investigations that demonstrated immediate fluid flow and reduced fluid volume as a response to cyclic and static load (Butler 1997;Lavagnino et al. 2003;Hannafin and Arnoczky 1994), and in vivo investigations using MRI (i.e., off-resonance saturation combined with ultrashort echo time sequence) that show reductions in AT volume and concomitant hydration state after cross-country running (Grosse et al. 2016;Syha et al. 2014). Fluid flow driven by the mechanical load might be attributed to increased hydrostatic pressure associated with the reorganisation of collagen fibers (Connizzo et al. 2013;Atkinson et al. 1997). On the other hand, the relatively constant tendon volume observed with short-duration contraction likely reflects a preserved fluid content.

Journal of Experimental Biology • Accepted manuscript
The 13.5% reduction in AT volume following the long-duration contraction at 75% MVIC is higher than reductions reported following cross-country running (4% and 5.8%, respectively) (Syha et al. 2014;Grosse et al. 2016). This may be because running exposes the tendon to high mechanical stress but lasting only for a short duration (0.2 -0.25 s) (Obst et al.). Our findings suggest that the change in tendon volume is occurring in a time-dependent manner (i.e., a greater volume reduction is occurring with 8 s than 2 s duration contraction) which may likely be attributed to the viscoelasticity of tendinous structure (Pearson et al. 2007). We also investigated the effect of long-duration contraction performed at a low load intensity (35% MVIC). Interestingly, we found a significant reduction in tendon volume by 5.4% (> 0.1 ml, P= 0.004) with long-duration contraction at 35% MVIC. The significant reduction in tendon volume following longduration contractions was greater than our minimal detectable change [(MDC)= standard error of measurement × 1.96 ×√2] which was ± 0.1ml, therefore, was likely a real change, on average.
In contrast to our findings of reduced AT volume by 5.4% with 35% MVIC contractions, Nuri et al. (2017b) found that the AT volume of healthy humans, measured by freehand Our free AT volume (2.5 ml) is comparable to MRI data reported by Syha et al. (2014) and Grosse et al. (2016) (2.4 cm 3 and 2.3 ml, respectively), and freehand 3DUS data reported by Devaprakash et al. (2020) (< 3 ml for trained runners and 3 ml for healthy controls). However, our free AT volume measurement is smaller than 3DUS data reported by Obst et al. (2014b) and Nuri et al. (2017b) for healthy young males (ranged between 3.9 ml to 4.9 ml). The average CSA value reported in the present study was Journal of Experimental Biology • Accepted manuscript relatively comparable to 3DUS data reported by Nuri et al. (2017b) (52 mm 2 ) for the midportion of the free AT but smaller than MRI data reported by Devaprakash et al.

Comparison of volume and diameter changes
Previous in vivo studies had investigated the acute effect of different loading interventions on tendon structural properties (i.e., thickness and diameter) and demonstrated immediate thickness/diametral reductions post-acute loading (Wearing et al. 2007;Grigg et al. 2012;Grigg et al. 2009). For example, Wearing et al. (2007) reported a reduction of 15% in Achilles anterior-posterior diameter following resistive PF exercise and had attributed such reduction to transient fluid loss. Importantly, the diametral reduction could likely occur with fluid redistributing within the tendon without the fluid leaving the tendon core (i.e., exudation). Therefore, it is unknown to what extent the reduction in tendon diameter reflects a change in tendon volume or fluid flow/exudation.

The acute effects of different isometric interventions on the free AT stiffness
The present study found that long-duration contraction at 75% MVIC changed the free AT stiffness by 46% more than long-duration contraction at 35% MVIC and 44% more than short-duration contraction at 75% MVIC. While there were no changes in force across intervention conditions, there was a significant increase in corresponding tendon elongation following the interventions involving the long-duration contractions (8 s, 75% MVIC and 8 s, 35% MVIC) (Table 2). Therefore, the observed reduction in the free AT stiffness was also associated with an increased tendon elongation for a given force. Our stiffness value for the free AT (between 325 and 331 N/mm) is comparable to the value reported by Devaprakash et al. (2020) (410 ± 164 N/mm) for the free AT, using 3DUS.

Journal of Experimental Biology • Accepted manuscript
Similar to our findings for volume, mechanical loading seems to influence change in stiffness. We found that long-duration contraction at 35% and 75% MVICs reduced the free AT stiffness but the reduction at 35% was significantly smaller than at 75% MVIC (11.3% vs 20.4%, respectively). In agreement with our finding, there is evidence that stiffness of vastus lateralis tendon significantly reduced after long-duration contractions but not after repetitive drop jumps (Kubo et al. 2005)  Theoretically, reduced fluid volume (i.e., reduction in the viscous element) would make the tendon thinner and may deprive the elastic matrix components of protection against tensile loading causing greater stress and strain (i.e., reduced stiffness). An unexpected finding was that the short-duration contraction at 75% MVIC induced a significant but small increase in the free AT stiffness by 11.5%. It is noteworthy that the changes in free AT stiffness following the 2 s, 75%, and 8 s, 35% interventions were smaller than our MDC value (≤ 45 N/mm), therefore, within measurement error, on average. Notably, the between-group differences support our hypothesis of greater changes in the intervention with the highest duration and intensity.

Implications for tendon adaptation
Our findings may provide evidence that the applied load on the tendon must be heavy and sustained for a long duration to maximize immediate volume reduction. This acute tendon response may be important to shift the tendon toward optimal long-term adaptation, however, this remains speculative and requires further investigation.
Load-induced fluid flow/exudation may be a remodeling signal for tendon cells. This is likely to occur because radial fluid flow to the periphery (exudation) may temporarily reduce the fluid volume within the tendon core, enabling tenocytes to perceive loading.
At the same time, the flow of tendon fluid generates shear stresses on the tenocytes membrane, all of which stimulate cellular activity (Wall et al. 2016;Docking 2013;Ingber 2006). Therefore, the immediate reduction in free AT volume observed after the long-Journal of Experimental Biology • Accepted manuscript duration high-intensity loading might be a physiological response that stimulates tenocytes mechanotransduction and signaling. We speculate that chronic exposure to long-duration load at high load intensity should lead to greater changes in tendon matrix and ultimately greater tendon adaptation (i.e., increased stiffness and tendon hypertrophy) compared to lower levels of load duration and intensity. In this context, Kubo et al. (2001a) found that tendon stiffness remarkably increased after a 12-week of high-intensity isometric training with a longer-duration contraction compared to a shorter duration.

Limitations
The present study has some limitations that need to be considered. First, the level of PF force in the present study was not maximal (the highest force was exerted at 75% MVIC). The AT viscoelastic behavior would likely differ under higher/maximal absolute tendon forces, that is, a higher force level would have caused the free AT to display higher stiffness (Rigby et al. 1958). However, with freehand 3DUS scanning, sustaining a maximal contraction for long enough to attain high quality images is extremely challenging. Second, we examined changes in volume and stiffness at the level of the free AT, which according to the evidence exhibits distinct material properties (i.e., higher longitudinal strain) compared to the proximal gastrocnemii tendon (Magnusson et al. 2003;Iwanuma et al. 2011). Therefore, our findings may not be generalisable to the gastrocnemius portion of the AT. Third, our analysis only presents volumetric changes across the entire free AT, and without conducting regional analysis, it is not possible to determine if there are any region-specific changes in tendon volume. Fourth, although the direction of loading response for each intervention was the same in every subject, there was an inter-subject variability in the magnitude of change. Various levels of activity and plantar flexor strength amongst the participants may explain this variability.
Future studies with specific inclusion criteria may be needed to examine whether there is a differential adaptive response between athletic and recreationally active populations and between men and women, separately. Finally, we had a small sample size of healthy participants (n=16) with a greater percentage of men (n = 12) than women (n = Journal of Experimental Biology • Accepted manuscript 4) (M: 75%). Therefore, we suggest caution in generalizing our results to other populations.

Conclusion
The present study demonstrates that long-duration and high-intensity loading resulted in greater immediate reductions in the volume and stiffness of the free AT compared to lower levels of load duration and intensity. Journal of Experimental Biology • Accepted manuscript   Football, basketball, and gym 10 8 Basketball 2  9 Soccer, weightlifting, and basketball 6 10 Soccer and volleyball 10 11 Power lifting 8-10 12 Skating 30  13 Weightlifting 5  14 None -15 None -16 Swimming 2