Participants were asked to visit the laboratory on two separate sessions to complete both interventions (VFR vs NVFR) within 2–7 days. Prior to the first measurements, participants had a familiarization session to introduce them to the foam rolling procedure. The intervention was randomized by the participants choosing a hidden card. At both appointments, participants performed a 10-min warm-up on a stationary bike (Monark, Ergomedic 874 E, Sweden) at 60 rev.min−1 and a resistance of 90 W. Functional parameters (i.e. ROM, MVIC, PRT) and muscle mechanical properties (i.e. shear modulus) of the right quadriceps muscles were examined pre and post the VFR and NVFR interventions. The duration between the warm-up and the first measurement was about 5 min. The functional parameters were MVIC, PRT, and hip extension ROM. The mechanical properties of the quadriceps muscle were tested in the VL, VM, and RF by shear wave elastography (SWE). The surface electromyography (Myon 320, Myon AG, Zurich, Switzerland) was measured on the VL muscle during MVIC, PRT, and SWE testing, before and after the intervention. Tests were performed in the order and time frame listed in Fig. 1. The order for the SWE was VL, VM, and RF.
An a priori sample size calculation (primary outcome variable range of motion) for a repeated-measures ANOVA based on the literature (Phillips et al. 2018) suggests a necessary group size of at least 15 participants (alpha = 0.05, beta = 0.8, f = 0.4). Thus, 21 physically active male participants (age: 25.2 ± 3.8 years; weight: 77.6 ± 8.8 kg; height: 182.5 ± 6.9 cm) volunteered in this study. Participants were free of any injuries of the lower extremities. Participants were informed about the test procedure before they signed a written informed consent form. The ethical approval was obtained by the ethical commission of the university and conformed to the standards of the Declaration of Helsinki.
Maximum voluntary isometric contraction (MVIC) peak torque
The MVIC measurements were performed with an isokinetic dynamometer (Con Trex MJ, CMV AG, Dübendorf, Switzerland). The participant was seated on the dynamometer, with the hip and knee angle of the right leg (test leg) being 110° (180° = full hip and knee extension) (Jakobsen et al. 2012; Kaya et al. 2019). A custom-made laser device was used to align the center of rotation of the dynamometer with the knee joint axis in a relaxed state right at 110° knee angle. To ensure the same sitting position during all assessments on the dynamometer, we recorded the exact position of the participant during the first MVICs, and we placed the participant in the same position in the following measurements. The trunk and test leg were fixed with straps to minimize the possibility of evasive movement. The lever arm fixation was set about 2 cm above the medial malleolus (Morales-Artacho et al. 2017). Each participant was asked to cross their arms in front of their chest and to perform three MVICs for 5 s each. The participant was asked to push as hard as possible, and received strong verbal encouragement during the measurements. Maximum torque value was communicated after each attempt and participants were motivated to try to exceed the previous result. Between the three MVICs, the participant rested for 1 min. The attempt with the highest torque value was considered for further analysis.
Passive resistive torque (PRT)
PRT measurement was done in the same sitting position as previously described for the MVIC measurement. The knee joint was passively moved for five cycles at an angular velocity of 5°/s from 90° to 60°. According to previous studies (Kubo et al. 2002; Mahieu et al. 2009), the velocity of the dynamometer was set to 5°/s to exclude any reflexive muscle activity. Participants were asked to relax completely. The lowest torque value of the last three circles in the extension phase was taken for further analysis.
Muscle shear modulus
Muscle shear modulus was measured on the VL, RF, and VM by SWE with an ultrasound scanner (Aixplorer V6, Supersonic Imaging, Aix-en-Provence, France) coupled with a linear transducer array (4–15 MHz, SuperLinear 10–2; Vermon, Tours, France). The machine was used in shear wave elastography mode (musculo-skeletal preset, penetration mode, smoothing level 5, persistence off, scale 0–300 kPa). The measuring system generates a two-dimensional map of the shear modulus of the measured tissue at 1 Hz, with a spatial resolution of 1 × 1 mm. Muscles were scanned using a handheld technique, based on previous studies that allowed a reliable measure for muscle stiffness (Bercoff et al. 2004; Lacourpaille et al. 2012; Hug et al. 2015). The participant was seated on the dynamometer with a hip angle of 110° and knee angle of 70° to achieve a slightly stretched position of the quadriceps muscles (Lacourpaille et al. 2017). To ensure similar placement of the probe in all measurements, reusable foil marked with the scars and birthmarks of the participant’s skin was used, which was also marked with the probe placement. Moreover, to facilitate reproduction during the following measurements, a B-Mode ultrasound image of the measured muscle part was recorded. SWE was performed in the same order at all measurements: VL, VM, and RF. VL was measured at about half way between the trochanter mayor and the lateral epicondyle of the knee (Coombes et al. 2018), VM at about one-third of the way between the medial epicondyle and anterior iliac crest (Coombes et al. 2018), and RF in the first-third distal between the proximal edge of the patella and anterior iliac crest (Ham et al. 2020). Care was taken to not put pressure on the skin, to avoid deformation of structures and muscle tissue (according to Kot et al. 2012).The ROI was maximized as much as possible, but excluding any aponeurosis. The transducer was aligned in plane with the fascicles and held in the same position during the whole process (according to Le Sant et al. 2017). The PRT test (as previously described) prior to the SWE was used as conditioning for the shear modulus testing, to guarantee the same muscle conditions. Participants were asked to remain completely relaxed during the measurements. Three videos of 15 s each were collected for each muscle. The mean of the five consecutive frames with the lowest standard deviation of the shear modulus averaged over the Range of interest (ROI) within a video was considered for further analysis. The two closest mean values per muscle from the three videos taken for each muscle were used to calculate the mean passive stiffness per muscle (Morales-Artacho et al. 2017).
Surface electromyography (EMG)
Muscle activity was monitored by EMG (myon 320, myon AG, Zurich, Switzerland) during the MVIC, PRT, and SWE measurements. After standard skin preparation, surface electrodes (Blue Sensor N, Ambu A/S, Ballerup, Denmark) were placed on the muscle belly of the VL, according to SENIAM recommendations (Hermens et al. 1999). The sample rate was 2000 Hz. EMG signals of the MVIC measurements were high-pass filtered (10 Hz, Butterworth) and the root-mean square (RMS, 50 ms window) values were calculated. The mean of 500 ms (± 250 ms around the peak value) was calculated around the maximum value. During the passive measurements we monitored the live EMG signal for activity. If signal changes were observed, the trial was repeated. Furthermore, a post-hoc analysis was performed for the PRT and SWE to ensure that the subject was relaxed, i.e. did not show EMG activity exceeding 5% of MVIC, if we could detect changes in the raw EMG signal during analyzing process (Gajdosik et al. 2005; Kato et al. 2010). In these cases, the EMG signal was high-pass filtered (10 Hz, Butterworth) and the root-mean square (RMS, 50 ms window) values were calculated.
Passive hip extension range of motion (ROM)
For the ROM measurements, a 3D-motion capture system (Qualisys, Göteborg, Sweden) was used. Eight cameras were used in fixed positions, and the system was calibrated with a standardized L-frame and wand at the beginning of each day. Reflective markers were added according the Qualisys Gait module (type: Cast) to the participant’s hip (with two extra markers on the lateral iliac crest to ensure a proper measurement in a supine position) and test leg. The participant was then asked to perform three modified Thomas tests of the test leg for 5 s each on a medical treatment bed. The participant lay supine, with the ischial tuberosity close to the edge of the bed (Younis Aslan et al. 2018). The participant was asked to hold the knees by hand, with the arm extended to ensure the same hip angle between measurements, and a flat lumbar spine. The legs were completely relaxed. While holding the contralateral leg in position, the test leg was lowered toward the floor and the participant was asked to remain as relaxed as possible in the end position. After processing the Qualisys data with Visual3D Professional (C-Motion, Inc., Germantown, USA), the angles of the joints were assessed. The attempt with the lowest hip extension angle was taken for further analysis.
Foam rolling intervention
The same foam roller (Blackroll Booster Set in combination with a Blackroll Standard foam roll, Bottighofen, Switzerland) was used throughout the intervention. The vibration booster is an additional vibrating cylinder, which is positioned along the longitudinal hole in the middle of the foam roll. If switched on, the whole foam roll vibrates. The intensity of the vibration can be set to between 12 and 56 Hz in 15 different levels. The rolling was applied for 1 min per muscle, with a frequency of 30 repetitions per minute, including a break of 30 s between sets, resulting in an overall rolling duration of 180 s. The duration of 60 s per muscle of the thigh (VM, VL, RF) was chosen since Baumgart et al. (2019) reported a significant decrease of RF stiffness following 60 s of foam rolling on the RF muscle only. A metronome provided auditory signals to pace the movement, and the participant was asked to reach the starting position every 2 s (1 s from distal to proximal and 1 s from proximal to distal). Start position was always proximal to the knee (Fig. 2). The muscles in the right thigh only (test leg) were rolled in the following order: (1) VL (rolled on the lateral side of the thigh); (2) VM (rolled on the medial side of the thigh); and (3) RF (rolled on the anterior part of the thigh). During the VFR, the vibration booster was switched on with a vibration intensity of 32 Hz. During the NVFR, the same foam roll was used but the vibration mode was switched off. In both conditions, participants rolled with their own bodyweight and were asked to put as much pressure on the tissue as possible, to the point of discomfort.
SPSS (version 26.0, SPSS Inc., Chicago, Illinois) was used for all the statistical analyses. To determine the intra-rater and inter-day reliability of the shear wave elastography measurements, intraclass correlation coefficients (ICC, 2-way mixed-effects model, absolute agreement definition) were used. The standard error of the mean of the shear modulus values was calculated as the standard deviation multiplied by the square root of one minus the ICC.
The variables tested were hip extension ROM, PRT, MVIC, and SWE of the VL, VM, and RF. A Shapiro–Wilk test was used to verify the normal distribution of all the variables. If the data were normally distributed, we performed a two-way repeated-measures ANOVA [factors: time (pre vs. post) and rolling modality (VFR vs. NVFR)]. Otherwise, we performed a Friedman test to test the effects of the foam rolling protocols (NVFR and VFR). If ANOVA with repeated measures or the Friedman test was significant, we performed a t test or a Wilcoxon test, respectively. To verify that the baseline conditions in VFR and NVFR were similar, paired t-tests or Wilcoxon tests were performed. To test possible differences between the rolling conditions (VFR vs. NVFR), paired t-tests or Wilcoxon tests of the delta values (post−pre) of each parameter were performed. The effect sizes d (for the t test) and r (for the Wilcoxon test) were established following the suggestions of Cohen (1988). Thus, the effect sized was defined as 0.2, 0.5, and 0.8 for a small, medium, and large effect, respectively. Moreover, the effect size r was defined as < 0.3, 0.3–0.5, and > 0.5 for a small, medium, and large effect, respectively. The alpha level was set to 0.05.