Strenuous resistance exercise effects on magnetic resonance diffusion parameters and muscle–tendon function in human skeletal muscle




To assess the effects of strenuous exercise on magnetic resonance diffusion parameters and muscle–tendon complex function in skeletal muscle.

Materials and Methods:

Six men performed ankle plantar flexion exercises with eccentric contraction. The fractional anisotropy (FA), λ1, λ2, λ3, mean diffusivity (MD), and T2 values in the triceps surae muscles were measured by magnetic resonance diffusion tensor and spin-echo imaging. Passive torque of plantar flexors, maximal voluntary isometric plantar flexion torques (MVIP), and Achilles tendon stiffness during MVIP were measured by combined ultrasonography and dynamometry. Plasma creatine kinase and muscle soreness were also assessed. These parameters were measured before and 1–8 days postexercise.


The medial gastrocnemius exhibited significantly decreased FA 2–5 days after, increased λ2 3 days after, and increased λ3 2 and 3 days after exercise. This muscle also showed significantly increased MD and T2 values 3 days postexercise. MVIP significantly decreased 2 and 3 days postexercise, while passive torque significantly increased 2 days postexercise. Creatine kinase and muscle soreness increased 3–5 days and 1–5 days postexercise, respectively.


Exercise-induced muscle damage manifested as significant changes in muscle diffusion parameters with muscle–tendon complex dysfunction and delayed-onset muscle soreness. J. Magn. Reson. Imaging 2011;. © 2011 Wiley-Liss, Inc.

Eccentric muscle contraction, in which muscles are lengthened as they generate force, generates high tension in muscle fibers. Muscle biopsy reveals myofibril sarcomere damage, including Z-band streaming and broadening, after eccentric exercise (1). In addition, magnetic resonance (MR) T2-weighted imaging and its coefficient, T2 relaxation time, noninvasively reveal local edema in damaged muscles, the result of inflammation after eccentric exercise (2, 3). Eccentric contraction-induced damage causes temporal dysfunction of the muscle–tendon complex, such as loss of muscle strength (2, 4–6) and decreased joint range-of-motion (increased muscle stiffness) (4, 5), with delayed-onset muscle soreness (DOMS) (3–6). However, several clinical discrepancies are observed between the T2-based evaluation and muscle–tendon complex dysfunction and/or muscle soreness in individual subjects. Other MRI parameters are needed to properly evaluate microscopic muscle damage.

MR diffusion tensor (DT) imaging is used to evaluate the relationship between water movement and the human skeletal muscle microstructure (7–15). The theoretical basis for DT imaging is that cell membranes and other structures restrict water diffusion in tissues, leading to anisotropic diffusion. Skeletal muscle is comprised of highly ordered elongated muscle fibers and has primary (λ1), secondary (λ2), and tertiary (λ3) eigenvalues. It was speculated that λ1, λ2, and λ3 represent diffusive transport along the long axis of a muscle fiber, within the endomysium perpendicular to the long axes of the muscle fibers, and within the cross-section of an individual fiber, respectively (10) (λ1 > λ2 > λ3). The three eigenvalues reflect the muscle microstructure: passive muscle contraction, where the sarcomere length is reduced and muscle fiber diameter is increased, reduces λ1 and increases λ2 and λ3, whereas passive elongation (increased sarcomere length and decreased muscle fiber diameter) has the opposite effect (8, 13).

Fractional anisotropy (FA) and mean diffusivity (MD) are calculated from the eigenvalues. FA reflects the degree of water diffusion anisotropy based on muscle fiber direction (7–15); this value is a robust intravoxel measure that yields values between 0 (perfectly isotropic, a perfect sphere) and 1 (perfectly anisotropic, a cylinder of minimal diameter). MD represents the directionless magnitude of diffusivity, eliminating the effects of directionality from diffusion measurements (15).

Galbán et al. (10) used DT imaging to characterize microstructural differences between functionally different muscles in the same region of the human body, also addressing gender differences (11), and age-related changes (12) in water diffusivity. Diffusion parameters are also affected by intramuscular changes induced by muscle contraction or extension (7, 8, 13, 16). Zaraiskaya et al. (9) demonstrated clear differences in diffusion properties between healthy controls and patients with muscle injuries, including intramuscular hematoma or muscle tear. Similar changes in diffusion properties were observed in muscle degeneration and regeneration after femoral artery ligation in mice (17). However, diffusion parameter changes over time have not been determined in human skeletal muscle before and after strenuous resistance exercise.

The purpose of this study was to investigate changes in muscle diffusion properties and muscle–tendon complex function after strenuous resistance exercise involving predominantly eccentric muscle contraction. Eccentric contraction is expected to damage microstructural integrity (1, 18, 19) and cause substantial alterations in diffusion with dysfunction of the muscle–tendon complex. Moreover, the extent of damage should not be uniform within the muscle group recruited by the exercise. DT imaging will provide dynamic information regarding the direction and intensity of water diffusion within damaged muscle, and will differ from conventional T2-based estimations of intramuscular water levels (2, 3).



Six healthy men (age, 21.5 ± 0.8 years; height, 172.7 ± 5.5 cm; weight, 64.7 ± 4.1 kg) were enrolled after obtaining written informed consent. None of the subjects were engaged in any training or exercise programs involving the triceps surae muscles. Subjects were instructed to refrain from participating in physical exercise or undergoing private physical therapeutic activity. Our Institutional Review Board approved the protocol. Before examination, the participants were given a brief description of the purpose, procedures, and potential risks of the study.

Exercise Model

The subjects performed ankle plantar flexion exercises using an exercise machine (Smith Machine, Nautilus, Vancouver, WA) by shouldering the weight bar, straightening the lower extremity at the knee joint, paralleling both sides of the great toe, and putting the metatarsal bone on a stool. The stool was used to elicit full eccentric contraction of the triceps surae muscles (medial gastrocnemius [MG]; lateral gastrocnemius [LG]; and soleus [SOL]). The exercise was performed throughout the full range of motion in the ankle joint using a constant load (kg) equal to the subject's body weight. A metronome paced the exercise at 60 counts per minute. On count 1, the subject raised the weight (concentric muscle contraction phase), on count 2 he held the weight (isometric muscle contraction phase), and on counts 3–5, he lowered the weight through the maximal range of ankle dorsiflexion (eccentric muscle contraction phase). The exercise was performed in five sets of 15 repetitions, with a 1-minute interval between sets.


We used a 1.5-T MR system (Signa EXCITE XI, GE Healthcare, Japan) with a quadrature knee coil. The coil was always placed at the same position on the patient table to ensure reproducibility of the MR images. The supine subjects placed their legs along the z-axis (parallel to the main static magnetic field B0); the axial slice line was perpendicular to the z-axis. The scan position was near the right mid-calf, 75% of the distance between the head of the fibula and the lateral malleolus. Autoshimming was performed before every scan.

Transverse axial DT images were acquired with a single-shot spin-echo echo-planar image sequence: repetition time, 3000 msec; echo time, 81 msec; 128 × 256 matrix; number of excitations, 16; field of view (FOV), 240 mm; rectangular FOV, phase FOV = 0.6; slice thickness, 10 mm; b-value, 400 s/mm2; acquisition time, 5 minutes 42 seconds; and water excitation of a single slice. The choice of a rectangular FOV minimized geometric distortions from susceptibility differences. The motion-probing gradient (MPG) was applied sequentially along each of six noncollinear orthogonal orientations (x-, y-, z-, xy-, xz-, and yz-axes). We obtained one baseline echo-planar T2-weighted image with no MPG and six DW images (b-value, 400 s/mm2) in six directions.

The imaging sequence (spin-echo-type) for calculating muscle T2 relaxation time was as follows: repetition time, 3000 msec; echo time, 25, 50, 75, and 100 msec; 256 × 160 matrix; number of excitations, 1; FOV, 240 mm; slice thickness, 10 mm; and acquisition time, 8 minutes 36 seconds.


We drew regions of interest (ROIs) around each of the triceps surae muscles, taking care to avoid the inclusion of subcutaneous fat, blood vessels, and bone. FA, λ1, λ2, λ3, and MD in the MG, LG, and SOL were calculated from each ROI with the diffusion Tensor Visualizer (dTV.II SR, Image Computing and Analysis Laboratory, Department of Radiology, University of Tokyo Hospital, Japan; http://www.ut-radiology. and Volume-One software v. 1.72 (

The FA value is a dimensionless index that expresses the anisotropic ratio of water diffusion for each imaging voxel, calculated as follows:

equation image(1)
equation image(2)

The six independent elements of DT (Dx, Dy, Dz, Dxy, Dxz, and Dyz) were calculated using one baseline T2-weighted image and DW images obtained from six directions. From the DT, three eigenvalues (λ1 > λ2 > λ3) were derived by diagonalizing the DT at each voxel.

To construct an FA map, MR image data were processed using Functool2 (GE Healthcare, Japan), built into the MR device. Echo-planar imaging distortions (scaling, translation, and shearing) from raw images were removed during FA map construction. An FA map was calculated from one baseline echo-planar T2-weighted image and six DW images in six orthogonal orientations. Each pixel on the FA map corresponds to the absolute FA value in tissue.

ROIs surrounded the MG, LG, and SOL on an image with a 25-msec echo time, and were then copied onto the 50, 75, and 100 msec echo time images. T2 relaxation time was calculated by least-squares analysis, fitting the signal intensity (SI) at each echo time to a monoexponential decay

Passive and Maximal Voluntary Torques of Plantar Flexors and Achilles Tendon Stiffness

Each subject was seated on the chair of a specially designed dynamometer (VINE, Japan), and the right foot was firmly attached to the footplate (Fig. 1a). Passive torque (PTQ) of plantar flexors at 90° ankle joint angle with the knee joint fully extended was measured prior to any voluntary torque exertions. During the PTQ measurement, subjects were asked to remain completely relaxed. The ankle joint was moved from the plantar flexed position to 90°. To exclude PTQ fluctuations, plantar flexion torque was averaged over 1 second when torque relaxation had appreciably subsided after the ankle angle was fixed to 90°.

Figure 1.

Measurements of MVIP and MTJ displacement. a: The right foot was fixed to the footplate of a dynamometer with the ankle joint at 90°. A 7.5-MHz linear array probe was attached to the dermis to measure MTJ displacement. b: The MTJ of the medial gastrocnemius (MG) was displayed on a B-mode ultrasound image before and after muscle contraction. The point of MTJ moved to the proximal side during MVIP generation.

Before measuring the maximal voluntary isometric plantar flexion torques (MVIP), a standardized warm-up and submaximal contractions were performed to accustom subjects to the tests. Each subject was asked to gradually increase the torque from 0 (relaxed) to MVIP within 5 seconds (20). The ankle joint was fixed at 90° with the knee joint fully extended. For the pre-exercise MVIP measurement, the subject exerted MVIP twice or more, unless the difference between torque values was confirmed to be within 10%. On postexercise days, MVIP was measured once. The maximum torque values were recorded as MVIP.

We measured displacement of the distal myotendinous junction (MTJ) of the MG during voluntary contractions to estimate Achilles tendon stiffness. A longitudinal ultrasound image of the MG was obtained by real-time ultrasound with a 7.5-MHz linear array probe (SSD1000 and UST-579T, Aloka, Japan) (Fig. 1b) (20), which was attached to the dermis with double-sided adhesive tape to prevent sliding over the mediolateral center of the MG (Fig. 1a). Slight ankle joint rotation during isometric plantar flexion is inevitable despite firm fixation (21) and influences MTJ displacement (22). The ankle joint rotation angle was measured using an electrogoniometer (SG110/A, Biometrics, UK) with its ends attached to the tibia and calcaneus.

Torque and angle signals were A–D converted at a sampling rate of 1 kHz (Powerlab 16SP, AD Instruments, Australia) and stored on a personal computer. The ultrasound images were recorded on videotape at 30 Hz and synchronized with the torque and joint angle data using an electronic timer. To define MTJ displacement due to ankle joint rotation, the joint was passively rotated by the tester and MTJ displacements during passive rotation were measured from plantar flexed 30° to dorsiflexed 10°. The effect of joint rotation was thus taken into consideration to estimate MTJ displacement during MVIP torque generation. MTJ displacements were digitized at each daily 10% MVIP increase (ImageJ 1.41b; National Institutes of Health, Bethesda, MD) (20). The measured torque during MVIP was converted to muscle force (Fm) as follows:

equation image(3)

where k is the relative contribution of the physiological cross-sectional area of the gastrocnemius within the plantar flexor muscles (23) and MA is the moment arm length of the Achilles tendon. The moment arm length as a function of ankle joint angle was derived from a previous report (24). Fm and MTJ displacements above 50% MVIP on each day were fitted to a linear regression equation, the slope of which indicated stiffness (25).

Plasma Creatine Kinase and Muscle Soreness

Whole blood samples were extracted from the finger after cleansing with an alcohol-impregnated swab. A small puncture in the skin was made with an Autoclix lancet. Arterialized capillary blood samples (30 μL) were collected in capillary tubes and analyzed for plasma creatine kinase (CK) using a Reflotron Plus biochemical analyzer (Roche Diagnostics K.K., Japan) that was calibrated prior to each day's use.

Right calf muscle soreness was assessed on the same measurement schedule using a 100-mm visual analog scale, on which 0 indicates no pain and 100 indicates “pain preventing one from walking unassisted.” Subjects were instructed to rate their soreness during walking.

Statistical Analysis

All parameters were evaluated before and 1, 2, 3, 5, and 8 days after exercise. Means and standard deviations were calculated, and significant changes from pre-exercise values were analyzed with repeated-measures analysis of variance (ANOVA) followed by Dunnett's post-hoc test. Statistical significance was set at P < 0.05 for all analyses.


MR Parameters

Table 1 shows the time course of changes in the FA, λ1, λ2, λ3, MD, and T2 in the triceps surae muscles before and after exercise. FA values decreased significantly in the MG at 2 days (P = 0.003), 3 days (P < 0.001), and 5 days (P = 0.011), and in the LG at 2 days (P = 0.029) and 3 days (P = 0.01). Decreased postexercise FA values in a representative subject are evident from color-coded images on FA maps (Fig. 2). Regions of decreased FA were prominent in the MG at 3 days. The MG showed significantly increased λ2 at 3 days (P = 0.004) and increased λ3 at 2 days (P = 0.022) and 3 days (P = 0.001). The MG showed significantly increased MD (P = 0.004) and T2 (P = 0.024) values at 3 days. Figure 3 shows transverse axial T2-weighted images of the right leg before and after exercise of the same subject shown in Fig. 2. Regions of increased SI were observed in the MG and LG at 2, 3, 5, and 8 days postexercise. The SOL showed no significant change.

Figure 2.

Changes in transverse axial FA maps of the right leg before and after exercise. The white arrow indicates the regions of decreased FA values in the MG.

Figure 3.

Changes in transverse axial T2-weighted images (repetition time = 3000 msec, echo time = 75 msec) of the right leg before and after exercise. The white arrow indicates increased SI in the MG and LG on postexercise T2-weighted images. SI elevations were observed in the intermuscular space between MG/LG and SOL, particularly 2 and 3 days after exercise (arrowhead).

Table 1. Time course of changes in MR parameters in the triceps surae muscles before and after exercise
  Pre1 d2 d3 d5 d8 d
  1. Unit: FA (unitless), λ1 – λ3 and MD (× 10−3mm2/s), and T2 (ms).

  2. Astarisk (*) indicates a significant difference from the pre- exercise value.

MGFA0.32 ± 0.010.29 ± 0.030.26 ± 0.06*0.23 ± 0.06*0.27 ± 0.04*0.29 ± 0.03
λ12.2 ± 0.082.22 ± 0.112.29 ± 0.132.36 ± 0.192.24 ± 0.142.22 ± 0.06
λ21.57 ± 0.061.63 ± 0.121.73 ± 0.211.83 ± 0.21*1.66 ± 0.161.59 ± 0.07
λ31.15 ± 0.051.25 ± 0.131.39 ± 0.26*1.51 ± 0.27*1.34 ± 0.221.25 ± 0.09
MD1.64 ± 0.061.7 ± 0.121.81 ± 0.191.9 ± 0.21*1.75 ± 0.171.69 ± 0.06
T232.5 ± 2.334.2 ± 2.740.3 ± 14.844.4 ± 10.4*42.5 ± 11.142.1 ± 9.0
LGFA0.33 ± 0.020.32 ± 0.030.3 ± 0.03*0.29 ± 0.02*0.32 ± 0.040.32 ± 0.04
λ12.19 ± 0.112.16 ± 0.112.16 ± 0.092.17 ± 0.112.17 ± 0.132.24 ±0.14
λ21.56 ± 0.081.55 ± 0.081.59 ± 0.091.6 ± 0.091.55 ± 0.111.59 ± 0.09
λ31.1 ± 0.071.13 ± 0.081.17 ± 0.111.2 ± 0.091.14 ± 0.131.14 ± 0.1
MD1.62 ± 0.091.61 ± 0.081.64 ± 0.091.66 ± 0.091.62 ± 0.111.66 ± 0.09
T232.9 ± 1.132.8 ± 1.033.7 ± 2.635.6 ± 4.235.5 ± 4.934.8 ± 3.7
SOLFA0.29 ± 0.020.28 ± 0.020.27 ± 0.030.28 ± 0.010.29 ± 0.020.3 ± 0.05
λ12.26 ± 0.122.28 ± 0.162.27 ± 0.092.35 ± 0.182.26 ± 0.122.29 ± 0.11
λ21.7 ± 0.071.73 ± 0.141.73 ± 0.091.77 ± 0.141.67 ± 0.071.68 ± 0.07
λ31.26 ± 0.061.29 ± 0.11.31 ± 0.111.31 ± 0.111.24 ± 0.051.23 ± 0.1
MD1.74 ± 0.081.76 ± 0.131.77 ± 0.091.81 ± 0.141.73 ± 0.071.73 ± 0.06
T233.4 ± 1.233.1 ± 1.035.0 ± 2.035.2 ± 1.234.3 ± 1.334.0 ± 1.8

PTQ, MVIP, and Achilles Tendon Stiffness

Table 2 displays the changes in PTQ of plantar flexors, MVIP, and Achilles tendon stiffness. PTQ significantly increased 2 days (P = 0.019) after exercise. MVIP significantly decreased at 2 days (P = 0.041) and 3 days (P = 0.033), but Achilles tendon stiffness did not change significantly.

Table 2. Time course of changes in MVIP, passive torque, the stiffness of Achilles tendon before and after exercise
 Pre1 d2 d3 d5 d8 d
  1. Astarisk (*) indicates a significant difference from the pre- exercise value.

MVIP (Nm)227.4 ± 22.7212.3 ± 19.6200.2 ± 31.1*199.3 ± 31.5*223.2 ± 28.6226.1 ± 23.9
Passive torque (Nm)13.7 ± 2.617.4 ± 3.118.4 ± 4.6*12.6 ± 2.612.8 ± 2.112.8 ± 2.8
The stiffness of Achilles tendon (N/cm)611.7 ± 222.6833.0 ± 353.0873.5 ± 280.8915.1 ± 484.1893.4 ± 261.5876.8 ± 309.5

Plasma CK and Muscle Soreness

Changes in plasma CK are presented in Fig. 4. Plasma CK levels were significantly higher 3 days (P < 0.001) and 5 days (P < 0.001) after exercise; the peak value occurred at 3 days.

Figure 4.

Changes in plasma CK level after exercise (mean ± SD). Asterisks (*) indicate significant increases.

Figure 5 depicts changes in muscle soreness. All subjects complained of soreness within 2 days following exercise. Muscle soreness was significantly greater at 1 day (P < 0.001), 2 days (P < 0.001), 3 days (P < 0.001), and 5 days (P = 0.001); the peak value occurred at 2 days postexercise.

Figure 5.

Muscle soreness before and after exercise (mean ± SD). Asterisks (*) indicate significant increases.


The results suggest that strenuous resistance exercise, including repetitive eccentric contractions, caused microscopic muscle damage, indirectly indicated by increased plasma CK. This damage was closely associated with muscle dysfunction and significant changes in the level, diffusivity, and directionality of intramuscular water. Damage was not uniform within the gastrocnemius–soleus complex; changes in MR parameters were most prominent in MG muscle. This nonuniformity is attributed to the degree of myofibrillar disruption by eccentric muscle contraction.

Blood CK level serves as an indirect index of eccentric contraction-induced muscle damage (2, 3). Plasma CK levels significantly increased 3 days and 5 days after exercise, indicating muscle fiber damage. Muscle fibers are subject to greater mechanical stress during eccentric contraction, which leads to Z-band streaming and broadening, and destroys myofibril sarcomeres (1). Although CK level reflects neither qualitative nor quantitative damage to a specific muscle, the MR parameters suggest a higher contribution of the MG to the increased CK levels observed here.

The MG showed prominent changes in FA, λ2, λ3, MD, and T2 values after exercise. Considering that the MG is most recruited by dynamic ankle plantar flexion exercise with the knee straightened (26), it would have been subject to greater mechanical stress than the LG and SOL in this study. Fast twitch fibers are more susceptible to damage during high-force exercise (1, 18). Disturbances in the Z-band are predominantly observed in fast twitch fibers immediately and 3 days after eccentric exercise (1). The proportion of fast twitch fibers is higher in the gastrocnemius than in the SOL (27). The LG also showed significantly decreased FA values, which also may be attributable to relatively high recruitment and fiber composition of the LG.

Elevated T2 in the MG reflects increased intramuscular water due to muscle damage and inflammation. T2 elevation is closely associated with increased CK in the days after eccentric exercise (2). Similar changes between T2 values in the MG and plasma CK levels were observed in our study. CK efflux into the extracellular space may result from the loss of sarcolemmal integrity by repetitive eccentric muscle contractions, and the elevated interstitial osmotic pressure may draw blood plasma into the MG interstitial space. Other inflammatory substances, such as histamine and prostaglandin E2, might cause vasodilation and elevated vascular permeability, increasing interstitial fluid in the injured area. Intracellular water may also have increased because of some impairment of cell metabolism related to muscle fiber damage (17, 19). We speculate that intra- and extracellular water were increased in the damaged area.

Decreased FA values indicate that intramuscular water diffusion became more isotropic. The eigenvalues reflect the diffusion behavior of water in the intracellular space (7–15); thus, increased λ2 and λ3 in the MG imply elevated diffusivity perpendicular to the muscle fiber long axis, suggesting that the muscle fibers were swollen by muscle damage (17, 19). Increased λ2 and λ3 would lead to reduced FA and elevated MD in the MG. FA and MD changes also would be associated with disrupted muscle fiber microstructure (the water diffusion pathways) resulting from repetitive eccentric contractions. This damage is expected to cause local edema (presumably increased intra- and extracellular water), resulting in more isotropic water diffusion in the injured area. Zaraiskaya et al. (9) also observed increased λ1, λ2, λ3, and reduced FA values in patients with muscle tear or hematoma. Moreover, it is well known that the inflammation response to tissue injuries causes the elevation of tissue temperature. Given that changes in water diffusion are closely related to tissue temperature (28), the elevated intramuscular temperature due to inflammation might have influenced the eigenvalues and MD in the MG. The significantly decreased FA in the LG may be attributed to slight LG muscle damage. FA may be so sensitive to muscle damage that it cannot be sufficiently detected by other MR parameters.

MVIP tended to decrease 1 day after exercise and was significantly reduced at 2 and 3 days. Disrupted muscle filaments may have reduced the overlap between actin and myosin filaments in sarcomeres; fewer cross-bridges may inhibit MVIP generation (1, 2, 29). MG should be most recruited in MVIP generation with the knee fully extended (26); thus, MG damage should be associated with decreased MVIP. Muscle edema in the MG could affect MVIP generation 2 and 3 days after exercise. The peak muscle soreness observed 2 days postexercise also might inhibit MVIP generation. Although the mechanical property of the Achilles tendon likely contributes to MVIP generation, it was minimally affected by the ankle plantar flexion exercise; Achilles tendon stiffness during MVIP generation did not change throughout the study.

The PTQ value tended to increase 1 day after exercise, and significantly increased at 2 days, likely due to increased passive tension of the ankle plantar flexors resulting from muscle damage. Although local edema is responsible for some increase in passive muscle tension (5), it cannot entirely account for the increased PTQ after exercise. PTQ of plantar flexors increased significantly 2 days postexercise while T2 values of the triceps surae muscles showed no significant increase at that time. At 3 days, the MG T2 value increased significantly while PTQ of plantar flexors returned to pre-exercise values. Thus, local edema had little influence on the increased PTQ after exercise. Another possible cause for the rise in passive muscle tension is membrane damage and calcium homeostasis disturbance, causing injury contractures in the damaged fibers (5, 6). Increased PTQ of plantar flexors also may be attributed to the shortening of noncontractile elements (connective tissue elements) in parallel with muscle fibers (5). If PTQ of plantar flexors had been measured at a more dorsiflexed position than the 90° ankle joint, that is, in the direction of longer muscle length, we may have observed a higher PTQ value. However, some subjects could not dorsiflex the ankle joint beyond 90° because of muscle soreness.

Muscle soreness significantly increased at 1 day and peaked at 2 days postexercise. Numerous studies have observed similar DOMS peak times (3, 5, 6). Subjects reported marked muscle soreness with mechanical stimulus, such as muscle extension during walking, while they claimed almost no soreness during relaxation. Myofibrillar damage likely underlies muscle soreness, and mechanoreceptor stimulation may contribute to the perception of soreness (30).

DT imaging can be used to evaluate muscle injuries, particularly for evaluating such microstructural muscle damage, as it cannot be sufficiently assessed by T2-weighted imaging, and for ongoing monitoring of recovery. Our results suggest the clinical feasibility of DT tractography to visualize the direction of muscle fibers in the evaluation of microtears (7, 9, 14).

In conclusion, microscopic muscle damage after repetitive eccentric contractions caused significant changes in MR diffusion parameters in skeletal muscle with dysfunction of the muscle–tendon complex and DOMS, and muscle damage was not uniform within the recruited muscle group. Muscle damage was not sufficiently reflected by T2 relaxation time. DT imaging provided dynamic information regarding the directionality and intensity of water diffusion within the specific damaged muscle, unlike conventional T2-based estimations of intramuscular water.