Effects of whole-body vibrations on neuromuscular fatigue: a study with sets of different durations

Study design

In a cross-over repeated measures design, each subject performed three different fatiguing exercise interventions with WBV and three exercise interventions in a sham condition without WBW (SHAM) to determinate the effect of WBV (Fig. 1A). Each intervention comprised a cumulative exercise period with a duration of 9 min divided into different sets (either 18 × 30 s or 9 × 60 s or 3 × 180 s) with 120 s rest between sets (Fig. 1A). The exercise interventions were performed on an activated vibration platform (WBV₃₀, WBV₆₀, WBV₁₈₀) and three on an inactive vibration platform (SHAM₃₀, SHAM₆₀, SHAM₁₈₀). Each intervention was executed on different visits with at least seven days rest in-between. The order was randomised. The subjects were not permitted to undertake explosive strength training or fatiguing workouts for 48 h before each measuring day, in order to eliminate side-effects. The study design, materials and neuromuscular assessments are available for reference in protocols.io (dx.doi.org/10.17504/protocols.io.beadjaa6).

Neuromuscular assessment in the resting position was performed at t₀ (baseline) prior to exercise intervention. The assessment consisted maximum voluntary contraction (MVC) of the knee extensors, interpolated with a high frequency (T_{MV C}) twitch (10 ms interstimuli interval), followed 3 s later by a 100 Hz doublet (T₁₀₀), followed 3 s later by a 20 Hz (50 ms interstimuli interval) doublet (T₂₀), and 3 s later by a potentiated single twitch (TW). The assessment procedure was executed according to (Millet et al., 2011) and repeated at 1 min (t_f), as well as at 15 (t_f15) and 30 min (t_f30) after the final 9-minute intervention. All neuromuscular assessments were performed on the right leg.

Subjects

Ten healthy subjects (6 men and 4 women; age: 21.1 ± 1.41 years, mass: 77.8 ± 11.73 kg, BMI: 22.9 ± 1.25) volunteered to participate in the study. All subjects were recreationally trained athletes, participating in moderate endurance and strength training activities 3 times per week. Exclusion criteria were acute injuries in the upper or lower extremities, locomotor dysfunctions, pregnancy, cardiovascular or neurological conditions. All subjects signed the written informed consent and the study was approved by the Ethics Committee of the Faculty of Sport of the University of Ljubljana 975/2017 and conducted according to the Declaration of Helsinki II.

The sample size was estimated by means of a power analysis aiming to detect large effect sizes (f = 1.2; alpha = 0.05; power = 0.80).

Intervention

The interventions were performed on a side-alternating vibration platform (Galileo Fit, Novotec Medical, Germany) which was running at a frequency of 26 Hz (Rittweger, Mutschelknauss & Felsenberg, 2003; Cochrane et al., 2010) and off, respectively, for WBV and SHAM conditions. Subjects were instructed to maintain a half-squat position with their knees flexed at an angle of 60° (Ritzmann et al., 2010) for several sets with 2-minute rest between sets. Kinematics were controlled with a goniometer. The subjects stood with their feet 40 cm apart at a point where the tilting platform reached peak-to-peak displacement amplitude of five mm (Ritzmann, Gollhofer & Kramer, 2013).

At the beginning of each session, subjects underwent a 6-minute warm-up routine consisting of bench stepping (20 cm high) at a frequency of 0.5 Hz, swapping the leading leg at one minute intervals.

Testing protocols

During the neuromuscular assessment, the subjects remained seated in a custom-built isometric knee extension apparatus equipped with a force transducer (MES, Maribor, Slovenia) (Tomazin, Dolenec & Strojnik, 2008; García-Ramos et al., 2016). The force transducer was calibrated prior to testing sessions. Each subject was seated in an upright position, hip at 100° and trunk leaning against the backrest of the testing apparatus, fixed by straps over the pelvis and a horizontal pad over the distal third of the thigh. The knee joint axis was aligned with the mechanical axis of the dynamometer. The shin pad was placed just superior to the medial malleolus. The right knee joint was fixed at a 60° angle (0° = full extension) (Fig. 1C).

Femoral nerve electrical stimulation

The femoral nerve was stimulated by pressing a monopolar cathode (10-mm in diameter, Ag–AgCl, Type 0601000402, Controle Graphique Medical, Brie-Comte-Robert, France) into the femoral triangle of the iliac fossa (Fig. 1C). A larger (102 mm × 52 mm, Compex, SA, Ecublens, Switzerland) self-adhesive electrode placed over the gluteal fold served as the anode. Electrical impulses (single, square wave, 1-ms duration) elicited by a high voltage constant current electrical stimulator (DS7A; Digitimer, Hertfordshire, UK) were used to trigger the muscle response, which was detected as a change in torque of the knee extensors. The stimulation intensity to elicit maximum knee extensors isometric twitch was determined in each subject at the beginning of each trial and maintained for the entire trial. Starting from an intensity of 10 mA, the stimulation intensity was progressively increased by 10 mA until no further increase in torque was observed despite further increment in electrical current. The current at maximal twitch torque was additionally increased by a factor of 1.5 to obtain a supramaximal stimulus (Verges et al., 2009).

Single twitch

The torque change induced by a single supramaximal femoral nerve stimulus (Place et al., 2007) was analysed to obtain the peak torque value (TW_PT).

High- and low-frequency doublets

The torque change induced by the paired high-frequency (100 Hz, i.e., 10-ms interstimuli interval) and low-frequency (20 Hz, i.e., 50-ms interstimuli interval) supramaximal electrical stimuli (Place et al., 2007; Verges et al., 2009) was analysed to obtain the following parameters: peak torque from 100 Hz doublet (T₁₀₀), peak torque from 20 Hz doublet (T₂₀). In addition, the low- to high-frequency ratio (T_20∕100) was calculated using the following formula: ${\displaystyle T_{20∕100}=\frac{T_{20}}{T_{100}}\ast 100.}$

This ratio was used as a surrogate of low- to high-frequency tetanic stimulation (Verges et al., 2009).

Maximal voluntary contraction with double twitch interpolated techniques

Subjects were asked to perform a 5 s maximal isometric voluntary knee extension (Verges et al., 2009). The signal was smoothed using a 0.5 s window moving average filter and peak torque (MVC) was retained for analysis. The double twitch interpolated technique (Allen, Gandevia & McKenzie, 1995) was performed by superimposing a 100 Hz doublet on the isometric plateau (T_{MV C}). A second analogous stimulation (T₁₀₀) on the relaxed muscle followed after 3 s (Fig. 1B). The ratio of the amplitude of the T_{MV C} over T₁₀₀ was then calculated to obtain the level of voluntary activation (%VA): ${\displaystyle \text{\%}VA=\left(1-\frac{T_{MVC}-MVC}{T_{100}}\right)\ast 100.}$

Statistics

A two-way factorial ANOVA (Type III) was conducted in R(3.5.1) with the afex package (Singmann et al., 2018) to compare the main effects of time (t₀, t_f, t_f15, t_f30) and trial (WBV₃₀, WBV₆₀, WBV₁₈₀, SHAM₃₀, SHAM₆₀, SHAM₁₈₀) and the interaction effect of time × trial. Generalised eta squared (${\eta }_G^2$) effect sizes were calculated for the ANOVA main and interaction effects. In the case of statistically significant interactions, post hoc comparisons with Sidak corrections were applied using the emmeans package (Lenth et al., 2018) in order to compare WBV and SHAM condition. Tukey-corrected pairwise post hoc tests were used to calculate differences to baseline within trials.

In addition to inference statistics, standardised changes in the mean of each measure were used to assess the magnitudes of effect (ES) between WBV and SHAM conditions of the same set duration (e.g., SHAM₃₀-WBV₃₀, SHAM₆₀-WBV₆₀, SHAM₁₈₀-WBV₁₈₀) and were then calculated using Cohen d. The magnitude of ES was interpreted as follows: trivial = <0.20; small = 0.2–0.59; moderate = 0.60–1.19; large = 1.20–1.99; and very large = >2.0 based on recommendations by Hopkins et al. (2009).

Statistical significance was set at the level of p < 0.05. ES results should be interpreted with caution, since negative values imply a larger fatiguing effect of WBV compared to SHAM condition and positive values imply a larger fatiguing effect for SHAM condition compared to WBV.

Maximum voluntary contraction

There was a statistically significant time effect (F(3, 27) = 24.40, p < 0.001, ${\eta }_G^2=0.02$), but no significant trial effect (F(5, 45) = 2.13, p = 0.08, ${\eta }_G^2=0.01$) nor trial x time interaction effect (F(15, 135) = 0.60, p = 0.87, ${\eta }_G^2=0.002$) for MVC. Within-trial post hoc tests showed differences between baseline and post-assessments (Fig. 2A).

Level of voluntary activation (%VA)

There was a statistically significant time (F(3, 27) = 3.67, p = 0.024, ${\eta }_G^2=0.02$) and trial (F(5, 45) = 2.52, p = 0.042, ${\eta }_G^2=0.08$) effect, but no trial ×time interaction (F(15, 135) = 1, 21, p = 0.26, ${\eta }_G^2=0.03$) for %VA. Post hoc tests did not reveal significant differences between baseline and post-assessments (Fig. 2B).

Peripheral fatigue

There was a significant time effect (F(3, 27) = 64.43, p < 0.001, ${\eta }_G^2=0.25$) for T₂₀. Trial effects (F(5, 45) = 1.91, p = 0.11, ${\eta }_G^2=0.03$) and trial x time interaction effects (F(15, 135) = 0.90, p = 0.56, ${\eta }_G^2=0.007$) remained statistically insignificant. Post hoc tests revealed significant differences between baseline and post-assessments for each of the trials (Fig. 3A, Table 3).

There was a significant time effect (F(3, 27) = 60.33, p < 0.001, ${\eta }_G^2=0.15$) for T₁₀₀. Trial effect (F(5, 45) = 2.15, p = 0.07, ${\eta }_G^2=0.03$) and trial ×time interaction effect (F(15, 135) = 0.43, p = 0.97, ${\eta }_G^2=0.002$) remained statistically insignificant. Post hoc tests revealed significant differences between baseline and post-assessments for each of the trials (Fig. 3B, Table 3).

There was a significant time effect (F (3, 27) = 46.33, p < 0.001, ${\eta }_G^2=0.17$) for T20/100. Trial effect (F(5, 45) = 1.06, p = 0.40, ${\eta }_G^2=0.02$) and trial × time interaction effect (F(15, 135) = 0.97, p = 0.49, ${\eta }_G^2=0.02$) remained statistically insignificant. Post hoc tests revealed significant differences between baseline and post-assessments for each of the trials (Fig. 3C, Table 3).

Single twitch

There was a significant time effect (F (3, 27) = 48.80, p < 0.001, ${\eta }_G^2=0.23$). Trial effects (F (5, 45) = 0.86, p = 0.52, ${\eta }_G^2=0.006$) and trial x time interaction effect (F(15, 135) = 1.05, p = 0.41, ${\eta }_G^2=0.006$) remained statistically insignificant for TW_PT. Post hoc tests revealed significant differences between baseline and post-assessments for each of the trials (Fig. 2C, Table 3).

Maximal voluntary contraction

Knee extensors MVC torque dropped by 7 to 12% after each fatiguing protocol, which is in line with other WBV induced fatigue studies, where MVC torque decreased by approximately 8% (De Ruiter et al., 2003; Erskine et al., 2007; Colson et al., 2009; Zory et al., 2013). Only Maffiuletti et al. (2013) reported a more substantial decline in MVC (−23%), which is likely associated with the application of additional loads coupled with shorter inter-set rest periods compared to other studies and to our specific experimental setting. This finding is in contrast with our hypothesis that longer set duration exercises superimposed with vibration (WBV₁₈₀) would produce greater fatigue compared to SHAM₁₈₀ condition. However, it has been previously suggested that potentiated electrically elicited supramaximal doublets represent a more suitable indicator of peripheral fatigue and contractile impairments compared to MVC torque (Place et al., 2007).

Central fatigue

The level of voluntary activation (%VA) of the knee extensors was not significantly depressed by any intervention utilised in this study, which suggest that mechanisms located in the central nervous system (CNS) were not significantly involved in the decline of MVC. These findings are in line with Colson et al. (2009) and Jordan et al. (2010) but in contrast to De Ruiter et al. (2003) who reported a vibration-induced decline in knee extensors voluntary activation. Despite De Ruiter et al. (2003) reported a similar drop in %VA compared to the present study (approx. 4%), any difference in interpretation between the two studies could be biased by the lack of a control group or control condition in De Ruiter et al. (2003) experiment coupled with the eligibility criteria for volunteers: in our study the population consisted of recreationally trained athletes and De Ruiter et al.’s (2003) enrolled untrained students. Evidence exist for physiological differences and perquisites in motor control between sedentary and trained active subjects (Buford & Manini, 2010). Being hypothesis-driven, our findings indicate that there are no evident superimposed effects of WBV on central fatiguing mechanisms compared to control conditions without WBV. This should be taken into consideration when designing exercise programs or research studies which intend to induce central fatigue. As such, WBV superimposed exercises are unlikely to be more effective than maintaining a static squat alone.

Peripheral fatigue

To the best of our knowledge, this is the first study where electrically elicited supramaximal low- and high-frequency doublets were used to assess the origin and magnitude of peripheral fatigue after WBV exposure. For all protocols, T₂₀ was more affected than T₁₀₀ leading to a decreased T_20∕100 ratio (Fig. 3C). These declines suggest the occurrence of low-frequency fatigue (LFF) in all trials. It is noteworthy that the T_20∕100 ratio for SHAM interventions returned to baseline values 15 min after the intervention, while WBV interventions remained significantly depressed up to 30 min after the intervention. This suggests that LFF is stronger and more long-lasting when a WBV exercise is executed with an emphasis on exposing to longer sets of vibration. The observation favouring LFF as an underlying mechanism can additionally be supported by the findings obtained from single twitch data. Similar to T₂₀ and T₁₀₀, TW_PT progressively decreased as the intervention continued.

Underlying mechanisms

The lack of difference between WBV and SHAM conditions observed in this study suggest that no beneficial effects on neuromuscular function exist when using superimposed WBV. This is particularly true for MVC and the level of voluntary activation. Even though some studies reported that WBV can induce modulation in the neuronal control, which is manifested as a facilitated central drive (Mileva, Bowtell & Kossev, 2009; Krause et al., 2016) this does not translate into central fatigue. Furthermore, the decline in low- and high-frequency doublets, as well as single twitch torque, suggests that a mechanism underlying the decrease in force production in both WBV and SHAM treatments is an impairment in Ca²⁺ handling. This is followed by a gradual recovery of the Ca²⁺ depletion within the 15–30 min following WBV equal to the SHAM intervention. Underlying cellular fatiguing mechanisms explaining the results for SHAM and WBV may refer to three aspects (Westerblad et al., 2000; Allen & Westerblad, 2001; Williams & Ratel, 2009): (a) since doublet peak torques progressively dropped at low- and high-frequencies of stimulation, there could be direct inhibition of inorganic phosphates (P_i) on Ca²⁺, thereby producing an impairment in the cross-bridge force generation (Millar & Homsher, 1990). However, it is unlikely that this mechanism alone accounts for low-frequency fatigue (Allen, Lannergren & Westerblad, 1995). (b) It is likely that the larger drop in T₂₀ compared to T₁₀₀ could indicate a precipitation in Ca²⁺-P_i in the sarcoplasmic reticulum, leading to a decrease in free Ca²⁺ available for release (Allen & Westerblad, 2001). In addition, (c) reduced myofibrillar Ca²⁺ sensitivity can also affect force production (Bruton et al., 2008). Both mechanisms (b and c) have little impact on force production at high frequencies but a large effect on low frequencies (Westerblad et al., 2000).

Limitations

The study might have some limitations. An important limitation of this study (similar to the majority of other vibration studies) is the lack of WBV load normalisation, as this may have considerable side-effects on the results, as was demonstrated by Di Giminiani et al. (2009). Another limiting aspect deals with different work/rest ratios between long sets (180 s work –120 s rest) compared to other shorter set durations (30 s –120 s and 60 s –120 s). There is a great diversity in scientific and practitioner protocols and therefore, future studies should consider the variability in work/rest ratios and duration sets within the experimental design.

Furthermore, the experiment was executed in the right leg only. The leg dominance has thereby not been considered as a variable of influence on fatigue and fatigue mechanisms.