We conducted a randomized crossover trial, as illustrated in Figure 1. Prior to the first testing day, all participants attended a familiarization session in which they learned and practiced all stretching methods and measurements. Participants completed measurement sessions on four separate days (1 day for each stretching type). Specifically, they completed four stretching sessions comprising SS alone (SS session), DS alone (DS session), and CS (two CS protocols: 150-sec SS followed by 150-sec DS [SS-DS session] and 150-sec DS followed by 150-sec SS [DS-SS session]) of the right hamstrings for 300 sec. The order of stretching types was randomized. We recorded the ROM of passive knee extension, PPT, passive stiffness, maximum voluntary isometric and concentric knee flexion forces, and electromyographic root mean square (RMS) during force measurements before, immediately after, and 20 and 60 min after stretching (Matsuo et al., 2015). The experiment was performed in a university laboratory where the room temperature was maintained at 26°C (Matsuo et al., 2019; Matsuo et al., 2015). All measurements were performed at the same time of day (± 1 h). The mean testing interval was 17.0 ± 15.0 days (range, 7-63 days); there were no consecutive testing days.

Twenty healthy young men voluntarily participated in this study (mean ± standard deviation: age, 21.8 ± 1.4 years; height, 171.4 ± 6.3 cm; weight, 64.6 ± 10.8 kg; body mass index, 22.0 ± 3.1 kg/m²). All participants provided written informed consent to take part in the study. The study protocol was approved by the Human Research Ethics Committee of our institution. Exclusion criteria were lower extremity joint contractures, history of back or lower extremity surgery, neurological disorders, current treatment involving hormones or muscle-affecting drugs, ability to completely extend the right knee from a sitting position as described below (i.e., exceptional flexibility), engagement in competitive sports, regular resistance, aerobics, and flexibility training. Participants were asked to avoid vigorous physical activity during the experimental period.

As shown in Figure 2a, to perform SS, each participant assumed a standing upright position and placed their right heel (with the leg extended) on a 50-cm-high platform. The participant then reached forward with their arms toward the extended leg while maintaining a proper lordotic curve (Matsuo et al., 2019). SS was performed at a tolerable intensity without pain (Goto et al., 2020; Matsuo et al., 2019; Matsuo et al., 2015). At the SS session, ten 30-sec sets of SS were performed with a 20-sec rest period between each set (Matsuo et al., 2019).

As shown in Figure 2b, to perform DS, each participant assumed a standing upright position beside parallel bars and held a parallel bar with their left hand for stability. To stretch the hamstrings, the participant intentionally contracted the right hip flexors with the knee extended and flexed their right hip joint so that their right leg swung up to the anterior aspect of their body (Hough et al., 2009; Iwata et al., 2019; Matsuo et al., 2019; Yamaguchi and Ishii, 2005). The participant performed this dynamic movement every 2 sec. Each exercise was performed five times slowly for practice, then 10 times as quickly as possible without bouncing (Hough et al., 2009; Iwata et al., 2019; Matsuo et al., 2019; Yamaguchi and Ishii, 2005). At the DS session, ten 30-sec sets of DS (15 repetitions of the DS movement in each set) were performed with a 20-sec rest period between each set (Iwata et al., 2019; Matsuo et al., 2019).

At the two CS sessions, SS and DS were performed as described above. At the SS-DS session, five 30-sec sets of SS followed by five 30-sec sets of DS were performed with a 20-sec rest period between each set. At the DS-SS session, five 30-sec sets of DS followed by five 30-sec sets of SS were performed with a 20-sec rest period between each set.

The following dependent variables were used to assess the acute and prolonged effects of SS, DS, and CS. We first measured the torque–angle relationship (ROM, PPT, and passive stiffness), then recorded the isometric muscle force, concentric muscle force, and electromyographic RMS during force measurements before, immediately after, and 20 and 60 min after stretching. Considering that warm-up activities prior to SS do not enhance decreases in muscle–tendon stiffness compared with SS alone (Fujita et al., 2018), participants did not perform warm-up activities prior to stretching exercises or testing. All dependent variables, except RMS, were obtained using an isokinetic dynamometer (PrimusRS; BTE Technologies, Hanover, MD, USA); electromyographic RMS was determined by surface electromyography (Biomonitor ME6000; Mega Electronics, Kuopio, Finland). Torque and angle signals from the dynamometer and electrical signals from electromyography were subjected to analog-to-digital conversion (PowerLab 8/35; ADInstruments, Dunedin, New Zealand) and stored in a personal computer.

Measurements were performed while the participant assumed a sitting position with their hip joint flexed (Figure 3a), as previously reported (Matsuo et al., 2019; Matsuo et al., 2013; Matsuo et al., 2015). Each participant was seated on a chair with the seat maximally tilted and a wedge-shaped cushion inserted between the trunk and the backrest. The participant’s chest, pelvis, and right thigh were stabilized with Velcro straps. The knee joint was aligned with the dynamometer’s axis of rotation; the lever arm attachment was placed immediately proximal to the malleolus medialis. In this position, the mean hip and knee flexion angles were 107.1° ± 2.5° and 110.6° ± 2.3°, respectively. While the participant was seated in the chair, their knee was passively extended at 5°/s to the point of maximum knee extension immediately before the onset of pain; torque was continuously recorded during passive knee extension (Matsuo et al., 2019; Matsuo et al., 2013; Matsuo et al., 2015). ROM (in degrees) was defined as the maximum knee extension angle from the initial position (0°); PPT (in Nm) was defined as the torque immediately before the onset of pain (Goto et al., 2020; Matsuo et al., 2019; Matsuo et al., 2015). Passive stiffness (in Nm/°) was defined as the slope of the least-squares regression line calculated from the torque–angle relationship (Matsuo et al., 2019; Matsuo et al., 2013; Matsuo et al., 2015). Passive stiffness was calculated using the same knee extension angle range for all time points. The calculated knee extension angle range was defined as the angle from the 50% maximum knee extension angle to the pre-stretching maximum knee extension angle.

Isometric muscle force (in Nm) was measured in the position used to measure the torque–angle relationship (flexed position; Figure 3a). Isometric muscle force was also measured with the seat adjusted parallel to the floor (horizontal position; Figure 3b) while the participant was seated with hip and knee flexion angles of 85° and 90°, respectively (Matsuo et al., 2015). Isometric muscle force was measured in the flexed position, followed by the horizontal position. Two measurement positions were used because stretch-induced strength loss depends on muscle length (McHugh and Nesse, 2008), and the horizontal position is often used to measure isometric muscle force (e.g., Ford-Smith et al., 2001; Kollock et al., 2010). Each participant was instructed to sit with their arms crossed in front of their chest and generate maximum knee flexion force for 3 sec. They performed this exercise three times with a 45-sec rest period between trials (Matsuo et al., 2015). Verbal encouragement was provided during measurements. Peak torque was recorded for each trial; the mean of the three trials was used in subsequent analysis.

After measurement of isometric muscle force in the horizontal position, concentric muscle force (in Nm) was measured in the horizontal position described above (Figure 3b) using an isokinetic dynamometer (Matsuo et al., 2015). Three maximum voluntary concentric knee flexions were continuously performed at an angular velocity of 60°/s for the range of 85° from a knee-extended position (5°) to a knee-flexed position (90°) with the participant’s arms crossed in front of their chest. Verbal encouragement was provided during measurements. Peak torque was recorded for each trial; the mean of the three trials was used in subsequent analysis (Matsuo et al., 2015).

Surface electromyographic activities during maximal isometric and concentric contractions were recorded from medial and lateral hamstrings by a Biomonitor ME6000 (Mega Electronics) with a sampling frequency of 1 kHz. Before electrode placement, the skin under the electrodes was shaved, abraded, and cleaned. For medial hamstring assessment, a pair of Ag/AgCl sensors (Blue Sensor M-00-s; Ambu, Ballerup, Denmark) was placed halfway between the ischial tuberosity and the medial epicondyle of the tibia. For lateral hamstring assessment, another pair of sensors was placed halfway between the ischial tuberosity and the lateral epicondyle of the tibia (Matsuo et al., 2015). Surface electromyography signals were band-pass filtered (8-500 Hz). RMS values were calculated using Labchart v8 software (ADInstruments). Time windows for RMS calculation were 3 sec during isometric contraction and 1.416 sec during concentric contraction; ROM was 85° (5-90°), and angular velocity was 60°/s during concentric contraction. The mean of three trials for each contraction was used in subsequent analysis.

Test-retest reliabilities for all dependent variables were determined by calculating intra-class correlation coefficients (ICCs) and 95% confidence intervals (CIs) from pre-stretching values for all stretching sessions. The ICC_1,1 results indicated acceptable reliability (ROM [ICC_1,1: 0.845, 95% CI: 0.726-0.927], PPT [ICC_1,1: 0.923, 95% CI: 0.857-0.965], passive stiffness [ICC_1,1: 0.896, 95% CI: 0.810-0.952], isometric muscle force [ICC_1,1: 0.905, 95% CI: 0.852-0.943], concentric muscle force [ICC_1,1: 0.880, 95% CI: 0.783-0.944], and RMS [ICC_1,1: 0.748, 95% CI: 0.685-0.805]) (Koo and Li, 2016).

Sample size estimation was performed using G*Power software (v 3.0.10; Franz Faul, Kiel University, Kiel, Germany). Based on a previous study regarding the acute effects of SS and DS on passive knee extension ROM (Matsuo et al., 2019), the effect size was 0.88; α level was set at 0.05 and power was set at 0.80. Using the Bonferroni post-hoc test to identify a significant difference from the pre-stretching value, the estimated minimum number of participants was 18. Considering the potential for dropout, 20 participants were recruited.

Data normality was assessed using the Shapiro–Wilk test. The results indicated that isometric muscle force in the flexed and horizontal positions, concentric muscle force, and medial hamstring RMS values during concentric muscle contraction were normally distributed. However, the remaining data did not exhibit normal distribution. Parametric tests were applied to normally distributed data; non-parametric tests were applied to the other variables and to the relative changes (in %) for all variables. Changes in dependent variables over time were compared between stretching sessions by two-way repeated measures analysis of variance (RM-ANOVA) (stretching sessions and time) or the Friedman test. When a significant interaction effect or main effect (stretching sessions or time) was identified using two-way RM-ANOVA or the Friedman test, Bonferroni post-hoc analysis was performed to determine significant pairwise differences between sessions at each time point or to reveal differences from pre-stretching values. Analyses were performed using IBM SPSS statistics version 24.0 (IBM Corp., Armonk, NY, USA); the statistical significance threshold was set at p < 0.05. Results are expressed as means ± standard deviations.

ROM and PPT were significantly increased in all stretching sessions immediately and 20 and 60 min post-stretching, compared with pre-stretching (all p < 0.05) (Table 1). There were no significant differences in pre-stretching values, post-stretching values, or relative changes in ROM and PPT between stretching sessions.

Passive stiffness was significantly decreased in SS-DS and DS-SS sessions immediately and 20 and 60 min post-stretching, compared with pre-stretching (all p < 0.05) (Table 1). It was also significantly decreased in SS and DS sessions immediately and 20 min post-stretching, compared with pre-stretching (all p < 0.05). Passive stiffness tended to be decreased in SS and DS sessions 60 min post-stretching (SS; p = 0.09, DS; p = 0.06). There were no significant differences in pre-stretching values, post-stretching values, or relative changes in passive stiffness between stretching sessions.

Results were similar between flexed and horizontal positions. Two-way RM-ANOVA revealed no significant interaction effects but showed a significant main effect of time (p < 0.05). Isometric muscle force was significantly decreased in all stretching sessions immediately and 20 and 60 min post-stretching, compared with pre-stretching (all p < 0.05) (Table 2 and Table 3). There were no significant differences in pre-stretching values, post-stretching values, or relative changes in isometric muscle force between stretching sessions.

Two-way RM-ANOVA revealed no significant interaction effects but showed significant main effects of stretching session and time (p < 0.05). Concentric muscle force was significantly decreased in SS and SS-DS sessions immediately and 20 and 60 min post-stretching, compared with pre-stretching (all p < 0.05) (Table 4). It was also significantly decreased in DS-SS sessions 60 min post-stretching (p < 0.05), and it tended to be decreased immediately (p = 0.06) and 20 min post-stretching, compared with pre-stretching (p = 0.07). Concentric muscle force in DS sessions did not significantly change.There were no significant differences in pre-stretching values, immediately and 60 min post-stretching values, or relative changes in concentric muscle force between stretching sessions. However, concentric muscle force was higher in DS sessions than in DS-SS sessions 20 min post-stretching (p < 0.05). Additionally, concentric muscle force tended to be higher in DS sessions than in SS-DS sessions 20 min post-stretching (p = 0.07).

Medial hamstring RMS values were significantly decreased in SS and DS sessions immediately and 20 and 60 min post-stretching, compared with pre-stretching (p < 0.05) (Table 2). They were also significantly decreased in SS-DS and DS-SS sessions immediately and 60 min post-stretching, compared with pre-stretching (all p < 0.05); they tended to be decreased 20 min post-stretching (SS-DS; p = 0.08, DS-SS; p = 0.05).

Lateral hamstring RMS values were significantly decreased in DS sessions 20 and 60 min post-stretching (both p < 0.05), and in DS-SS sessions immediately and 20 min post-stretching (both p < 0.05), compared with pre-stretching. In SS-DS sessions, they tended to be decreased 20 min post-stretching (p = 0.06). There were no significant differences in pre-stretching values, post-stretching values, or relative changes in RMS values between stretching sessions.

Medial hamstring RMS values were significantly decreased in SS-DS sessions 60 min post-stretching, compared with pre-stretching (p < 0.05) (Table 3). In DS sessions, they tended to be decreased immediately and 20 min post-stretching, compared with pre-stretching (immediately; p = 0.08, 20 min; p = 0.05). Lateral hamstring RMS values were significantly decreased in SS and DS sessions 20 min post-stretching, compared with pre-stretching (both p < 0.05). Moreover, RMS values in SS sessions 60 min post-stretching and in DS sessions immediately post-stretching tended to be decreased, compared with pre-stretching (SS; p = 0.08, DS; p = 0.08). There were no significant differences in pre-stretching values, post-stretching values, or relative changes in RMS values between stretching sessions.

Two-way RM-ANOVA revealed no significant interaction effects but showed a significant main effect of time for medial hamstring RMS values (p < 0.05). Medial hamstring RMS values were significantly decreased in DS, SS-DS, and DS-SS sessions immediately and 20 and 60 min post-stretching, compared with pre-stretching (all p < 0.05) (Table 4). They were also significantly decreased in SS sessions immediately and 60 min post-stretching (both p < 0.05); they tended to be decreased 20 min post-stretching (p = 0.07). Moreover, lateral hamstring RMS values were significantly decreased in SS sessions 60 min post-stretching (p < 0.05) and in DS sessions immediately and 20 and 60 min post-stretching (all p < 0.05), compared with pre-stretching. RMS values in SS sessions 20 min post-stretching (p = 0.07) and in DS-SS sessions immediately and 20 min post-stretching (immediately; p = 0.08, 20 min; p = 0.07) tended to be decreased. There were no significant differences in pre-stretching values, post-stretching values, or relative changes in RMS values between stretching sessions.