Journal Logo


The Effect of Cycling-specific Vibration on Neuromuscular Performance


Author Information
Medicine & Science in Sports & Exercise: May 2021 - Volume 53 - Issue 5 - p 936-944
doi: 10.1249/MSS.0000000000002565
  • Free


Although surface-induced vibrations are ubiquitous in cycling, their implications for the cyclist’s neuromuscular performance are not yet fully understood. This is of central interest for some of the oldest and most prestigious events in professional cycling, such as the Paris–Roubaix or the Tour of Flanders, as these races typically contain a considerable amount of cobblestone passages or uneven roads.

Surface irregularities as cobblestones cause a primarily vertical oscillating movement of the wheels in a frequency around 40 Hz (1). The bike pedals, saddle, and handlebar transmit the resulting accelerations and forces to the cyclist and create a multicentric mechanical stimulus that can affect the musculoskeletal system. It has been suggested that vibration triggers stretch-sensitive muscle spindles (2,3) and Golgi tendon organs (4), which result in an increased muscle activity. An increase in local muscular activation, which consequently increases the systemic energy demand (5,6), potentially impairs performance in cycling. However, there is no consensus if and how vibration increases oxygen demand or heart rate in cycling (7–10). Studies using a vibration stimulus to the entire body or high exercise intensities tend to report increased oxygen uptake (9,10). Interventions based on a low cranking power (8,10) or an isolated vibration application to the lower extremities (7) reported no increase in oxygen uptake. Considering that the neuromuscular response to vibration depends on the type of vibration stimulus (11,12) and additional loading of the muscle (13), test designs differing in cranking power or the way vibration is applied are reasonable explanations for inconsistent results across studies. Notably, different vibration applications via a decoupled crank (7,10), the front tire only (8), or rear dropout only (9) affect the musculoskeletal system only locally and not comparable with a comprehensive real-world scenario.

Therefore, interpretation and transfer of previous research are difficult and limited when considering the intensity characteristics of a cycle race, which includes close to threshold phases and shorter high-intensity efforts (14). To the best of our knowledge, no experimental design has yet sufficiently considered the specific physiological and mechanical characteristics of a cycling race on cobblestones. In particular, this includes quantifying the vibration exposure of the lower extremities, torso, and upper extremities; the analysis of the muscular response to this stimulus; and the measurement of the resulting physiological energy demand. A cycling-specific loading scheme and a comprehensive approach are steps toward a better understanding of vibration effects in cycling and contribute to bridging the gap between research, sports, technicians, and developers.

Therefore, the purpose of this study was to understand if and to what extent vibration affects the musculoskeletal system and metabolic demands in road cycling. Therefore, vibrations typically imposed in road cycling were added to cycling at low, close to threshold, and submaximum intensity. Based on the literature analysis, we hypothesized (a) that vibration creates a varying mechanical stimulus on the entire musculoskeletal system. Furthermore, we hypothesized (b) that vibration increases the muscular activation of stabilizing muscles of the upper body and propulsive muscles of the lower extremities. We also hypothesized (c) that vibration increases systemic physiological demands such as oxygen consumption or heart rate in cycling. Furthermore, we hypothesized (d) that cranking power modulates vibration-induced mechanical stimuli, muscular responses, and physiological demands.



The study was conducted as a cross-sectional trial. The experimental intervention included a systematic variation of the two independent variables: vibration and cranking power. Dependent variables are transmitted accelerations to the body, muscular activation, heart rate, and oxygen consumption. The local accelerations at the lower and upper body quantified the vibration exposure of the musculoskeletal system. The muscular activation described the local biological response to the mechanical stimuli vibration and cranking power. Oxygen uptake and heart rate estimated the systemic metabolic demand.


Thirty male cyclists (mass = 75.9 ± 8.9 kg, body height = 1.82 ± 0.05 m, V˙O2max = 63 ± 6.8 mL·kg−1⋅min−1) participated in the study. Inclusion criteria were a cycling mileage of more than 4000 km⋅yr−1, no acute or chronic cardiopulmonary or musculoskeletal diseases, and no severe injuries in the year before the intervention. The participants were recruited from local cycling clubs and the university cycling classes. Most of them competed at regional level and were familiar with ergometer-based performance testing. After the explanation of all procedures, each participant gave written informed consent to participate in the study. The study conformed to the World Medical Association Declaration of Helsinki and was approved by the university ethics committee (approval no. 036/2016).

Test protocol

The test protocol aimed to reproduce the intensities and vibration characteristics of a cycling race. Therefore, mean amplitude and frequency for vibration application were derived from the vertical acceleration of the frame dropouts (front: 44 Hz, 4 mm; rear: 38 Hz, 4 mm) when riding on cobblestones (1). The exercise intensities were defined based on the intensity profile of a bicycle race with low intensive, close to threshold and (sub)maximum efforts (14) at 40% (LOW, 137 ± 14 W) and 60% of the participants V˙O2max (MED, 221 ± 18 W) and individual 4-min maximum power output (HIGH, 331 ± 65 W). From a physiological perspective, the cranking power conditions LOW and MED coincide with the aerobic–anaerobic transition and changes in muscle fiber recruitment (15,16). The HIGH power condition refers to the typically highly intensive passage of a cobblestone passage.

For the definition of LOW and MED exercise intensities, the participants completed a performance test on a day preceding the vibration interventions. Thereby, maximum oxygen consumption (V˙O2max) was determined (Zan600 USB, nSpire Health GmbH, Oberthulba, GER), based on a resistance protocol starting at 100 W and increasing cranking resistance by 20 W every 30 s. The test was performed until maximum volitional exhaustion. Gas exchange was measured breath by breath, and V˙O2max was defined as the highest oxygen uptake averaged over 20 s. Test day 2 included the vibration interventions (NoVib–Vib) and the definition of the HIGH power condition (Fig. 1). Because of methodological reasons (e.g., EMG electrode placement for motion analysis and fatigue), vibration testing was subdivided into (i) physiological testing and (ii) motion analysis (Fig. 1). During all interventions, cadence was defined between 80 and 90 rpm.

Schematic representation of the testing protocol with a V˙O2max performance test on test day one. Vibration interventions at LOW, MED, and HIGH cranking power on test day two consist of physiology testing (heart rate and oxygen consumption) and movement analysis (randomized order of vibration and power within movement analysis), including acceleration and EMG measurements. Black bars indicate vibration.

Test setup

The vibration was applied via two vibration plates (VTE 5/5-2NEG 50300, Netter Vibration, Germany) to the front (44 Hz, 4 mm) and rear dropouts (38 Hz, 4 mm) of the bicycle frame (Fig. 2). The direct and simultaneous vibration initiation replicated the real-world scenario, where uneven road surfaces are transferred via the wheels to the front and rear frame dropouts. A cycle trainer (Satori Smart, Tacx, Wassernaar, Netherlands), attached to the posterior vibration platform, ensured bike fixation and cranking resistance. A powermeter (SRM 5th Gen, SRM, Jülich, Germany) monitored the mechanical power at the crank of the bike. All interventions were performed on Specialized Tarmac Pro Race 2015 carbon bikes. The participants set up their habitual saddle position and used their own cycling shoes with the corresponding clipless pedal.

(I) Vibration test setup, including vibration plates for direct vibration application at the front (B: 44 Hz, 4 mm) and rear dropout (A: 38 Hz, 4 mm) and the cyclo trainer providing driving resistance (C). (II) Positioning of the acceleration sensors at the 1) lower leg, 2) thigh, 3) lower back, 4) shoulder, 5) neck, and 6) forearm.

Physiological testing

Physiological testing was designed based on a stepwise increasing resistance profile. To establish a physiological steady state, each power level (LOW and MED) was maintained for 12 min. The vibration was applied during the last 6 min of each power level. After visual confirmation of a steady state by constant values, heart rate (HRM-Dual; Garmin, Schaffhausen, Switzerland) and oxygen consumption (Zan600 USB, nSpire Health GmbH; sampled breath by breath) were recorded and averaged over the last 3 min of every condition. Immediately after the power LOW and power MED interventions, a 4-min maximum effort time trial with vibration defined HIGH cranking power for the subsequent motion analysis (Fig. 1).

Motion analysis

After a 45-min recovery period, EMG and acceleration measurements were collected. To avoid fatigue and familiarization effects, vibration interventions and cranking power conditions were randomized in order and shortened to 2 min each (Fig. 1). Local accelerations of the body were recorded with six acceleration sensors (275 Hz; Aktos-T, Myon AG, Schwarzenberg, Switzerland) placed at the medial shin, the medial distal thigh, the lower back at the height of L5, the neck at the height of c7, the mid-forearm, and the acromion (Fig. 2). Signal components caused by low frequent voluntary movements were separated from the high-frequency vibration signal component (filter: Butterworth, 5 Hz, high pass, second order, recursive). Because of the unknown and changing sensor orientations, acceleration was calculated as the resultant acceleration of the x, y, and z components and represented as the root-mean-square (RMS) acceleration over 20 s. The RMS of accelerations has been used previously in cycling to measure vibration exposure (8,17).

As a representation of propulsive and stabilizing muscles, the activation of the triceps surae (gastrocnemius medialis, gastrocnemius lateralis, soleus), quadriceps femoris (vastus lateralis, vastus medialis, rectus femoris), triceps brachii, flexor carpi ulnaris, and lumbar erector spinae was recorded (1000 Hz; Aktos-T, Myon AG). Skin preparation, electrode placement (BlueSensor; Ambu®, Ballerup, Denmark), and signal processing followed the SENIAM guidelines (18). Signal processing included band-pass filtering (Butterworth, 5–500 Hz, second order, recursive) to reduce nonphysiological signal components and voluntary movement artifacts. Notch filtering at the vibration input frequencies and their harmonics reduced possible vibration artifacts. EMG envelopes were generated with a 15-Hz low-pass filter (Butterworth, second order, recursive). Fifteen crank cycles were recorded, separated, time normalized, and averaged for each condition. EMG envelopes were amplitude normalized relative to the averaged peak activation of the MED–NoVib baseline condition (19,20). Muscular activation was expressed as the mean activation over the crank cycle.

Statistical analysis

Discrete values are presented as mean and SD. Before statistical analysis, the assumption of normality was tested based on quantile–quantile plots and a Kolmogorov–Smirnov test. To analyze how vibration (VIB) and cranking intensity (POWER) affect the vibration stimulus (resultant acceleration) at different body parts (LOCATION), a three-way repeated-measures ANOVA (VIB–POWER–LOCATION) was carried out. A two-way repeated-measures ANOVA (VIB–POWER) was run for the selected lower extremity and upper body muscles to analyze the effect of vibration and cranking intensity on mean muscular activation. The effect of vibration and cranking intensity on systemic physiological demands was analyzed based on heart rate and oxygen consumption with two-way repeated-measures ANOVA (VIB–POWER). Pairwise comparisons were based on Bonferroni corrected post hoc tests. Significant effects were accepted with a P value <0.05. Violations of the self-paced cranking power during physiology steady-state testing and dropouts due to discomfort or vibration nausea led to the exclusion of several data points. Therefore, the sample size was reduced for the heart rate and oxygen consumption data sets to N = 23, for the EMG data set to N = 25, and for the acceleration data set to N = 25. The strength of the observed effects was described by calculating effect sizes (partial eta squared, μ2). A post hoc sensitivity power analysis (α error probability = 0.05, power = 0.08) resulted for the present study design with 23 (heart rate and O2) and 25 participants (EMG and ACC), respectively, in minimal detectable effect sizes of f = 0.211 for EMG and acceleration analysis and f = 0.221 for heart rate and oxygen consumption analysis. This represents small to medium effect sizes (21). Signal processing, descriptive statistics, and inferential statistics were done with Matlab (Matlab R2018B; MathWorks, Natick, MA).


Local accelerations

A significant three-way interaction between vibration (Vib–NoVib), cranking power (LOW–MED–HIGH), and sensor position (lower leg–thigh–lower back–neck–shoulder–forearm) was found for local accelerations (P = 0.045, μ2 = 0.119) (Fig. 3). Vibration increased the RMS of acceleration for all sensor positions and all cranking power levels compared with the corresponding NoVib condition significantly. With vibration, the RMS of acceleration was in a range between 3.37g ± 1.70g at the forearm and 0.21g ± 0.08g at the neck. Without vibration, values lay within 0.04g ± 0.01g at the neck and 0.30g ± 0.06g at the lower leg. With induced vibration, accelerations decreased significantly from distal to proximal for the upper extremities (forearm vs shoulder) and the lower extremities (lower leg vs thigh).

RMS of acceleration for all participants, displayed as mean and SD for the lower leg, thigh, lower back, neck, arm, and shoulder (acromion) for LOW, MED, and HIGH cranking power. Accelerations at all body parts are with vibration significantly higher in comparison with the NoVib baseline. NoVib values can be found in Tables 1–3. *Significant increase of local accelerations with vibration (P < 0.05). Curly braces indicate differences in all cranking power conditions.

There was a significant simple two-way interaction between cranking power and sensor position with (P = 0.026, μ2 = 0.137) and without (P < 0.001, μ2 = 0.480) induced vibration. With induced vibration, accelerations transferred to the lower leg (P < 0.001, μ2 = 0.628) and thigh increased (P < 0.001, μ2 = 0.421) significantly with a higher cranking power; for the lower back (P < 0.001, μ2 = 0.530) and the neck, this effect was inverse (P < 0.001, μ2 = 0.570). The highest mean increase of transmitted acceleration with cranking power was found for the lower leg (+33%), and the most pronounced decrease was found at the lower back (−18%). No effect of cranking power on transmitted vibration was found for the lower arm (P = 0.254) and shoulder (P = 0.181). Although local accelerations at all measurement points increased significantly with cranking power when no vibration was applied, in absolute numbers, the highest increase was still not bigger than 0.09g (thigh LOW–NoVib vs thigh HIGH–NoVib). Therefore, the interactions of vibration transmission and cranking power at the torso and lower extremities remained present even after a trial basis baseline correction (Vib–NoVib). Discrete values and results of simple simple comparisons are summarized in Table 1 for the lower extremities, Table 2 for the hand–arm system, and Table 3 for the torso.

TABLE 1 - Effects of vibration at LOW, MED, and HIGH cranking power for the lower extremities.
Lower Extremity
Power LOW Power MED Power HIGH
Measures NoVib Vib NoVib Vib NoVib Vib
RMS of acceleration (g)
 Lower leg 0.22 ± 0.04 1.26 ± 0.41 a 0.23 ± 0.05 1.46 ± 0.47 a,b 0.30 ± 0.06 b,c 1.67 ± 0.48 a,b,c
 Thigh 0.15 ± 0.03 d 0.56 ± 0.15 a,d 0.18 ± 0.04 b,d 0.60 ± 0.16 a,d 0.24 ± 0.05 b,c,d 0.68 ± 0.14 a,b,c,d
Muscular activation (relative to peak baseline)
 Gast Lat 0.33 ± 0.06 0.36 ± 0.07 a 0.38 ± 0.06 b 0.41 ± 0.08 a,b 0.47 ± 0.08 b,c 0.50 ± 0.09 a,b,c
 Gast Med 0.28 ± 0.04 0.33 ± 0.05 a 0.31 ± 0.05 b 0.35 ± 0.06 a,b 0.38 ± 0.10 b,c 0.42 ± 0.10 a,b,c
 Soleus 0.27 ± 0.06 0.37 ± 0.10 a 0.33 ± 0.05 b 0.45 ± 0.11 a,b 0.42 ± 0.10 b,c 0.51 ± 0.10 a,b,c
 Rec Fem 0.31 ± 0.13 0.32 ± 0.17 0.40 ± 0.07 b 0.45 ± 0.17 b 0.69 ± 0.25 b,c 0.74 ± 0.36 b,c
 Vast Lat 0.24 ± 0.03 0.25 ± 0.06 0.35 ± 0.04 b 0.36 ± 0.06 b 0.47 ± 0.08 b,c 0.50 ± 0.11 a,b,c
 Vast Med 0.24 ± 0.05 0.24 ± 0.07 0.34 ± 0.04 b 0.35 ± 0.05 b 0.47 ± 0.09 b,c 0.50 ± 0.09 b,c
The table includes local accelerations at the lower leg and thigh and mean muscular activation of the gastrocnemii (gast. lat., gast. med.), soleus, rectus femoris (rec. fem.), and the vastii (vast. lat., vast. med.). Significant effects were accepted with a P value <0.05.
aSignificant increase with vibration.
bSignificant increase in comparison with LOW.
cSignificant increase in comparison with MED.
dThe RMS of acceleration values also include a column-wise comparison of sensor position.

TABLE 2 - Effects of vibration at LOW, MED, and HIGH cranking power for the hand–arm system.
Hand–Arm System
Power LOW Power MED Power HIGH
Measures NoVib Vib NoVib Vib NoVib Vib
RMS of acceleration (g)
 Forearm 0.05 ± 0.01 2.99 ± 1.31 a 0.06 ± 0.02 b 3.06 ± 1.61 a 0.09 ± 0.03 b,c 3.30 ± 1.63 a
 Shoulder 0.04 ± 0.01 d 0.47 ± 0.21 a,d 0.05 ± 0.01 b,d 0.44 ± 0.19 a,d 0.07 ± 0.02 b,c,d 0.41 ± 0.14 a,d
Muscular activation (relative to peak baseline)
 Flexor Carp. Ul. 0.70 ± 0.35 1.44 ± 0.64 a 0.74 ± 0.12 1.47 ± 0.61 a 1.14 ± 0.43 b,c 1.60 ± 0.64 a
 Triceps Brac. 0.67 ± 0.14 0.95 ± 0.21 a 0.70 ± 0.06 0.91 ± 0.20 a 0.75 ± 0.16 b 0.92 ± 0.21 a
The table includes local accelerations at the forearm and shoulder and mean muscular activation of the flexor carpi ulnaris (forearm) and the triceps brachii (triceps br.). Significant effects were accepted with a P value <0.05.
aSignificant increase with vibration.
bSignificant increase in comparison with LOW.
cSignificant increase in comparison with MED.
dThe RMS of acceleration values also include a column-wise comparison of sensor position.

TABLE 3 - Effects of vibration at LOW, MED, and HIGH cranking power for the torso.
Power LOW Power MED Power HIGH
Measures NoVib Vib NoVib Vib NoVib Vib
RMS of acceleration (g)
 Lower back 0.04 ± 0.01 0.72 ± 0.19 a 0.04 ± 0.01 b 0.67 ± 0.17 a,b 0.06 ± 0.01 b,c 0.59 ± 0.14 a,b,c
 Neck 0.04 ± 0.01 0.32 ± 0.13 a,d 0.04 ± 0.01 b 0.28 ± 0.12 a,d 0.06 ± 0.01 b,c 0.21 ± 0.07 a,b,c,d
Muscular activation (relative to peak baseline)
 Erector spinae 0.43 ± 0.15 0.46 ± 0.18 0.51 ± 0.12 b 0.54 ± 0.19 b 0.63 ± 0.23 b,c 0.64 ± 0.26 b,c
The table includes local accelerations at the lower back and neck and mean muscular activation of erector spinae. Significant effects were accepted with a P value <0.05.
aSignificant increase with vibration.
bSignificant increase in comparison with LOW.
cSignificant increase in comparison with MED.
dThe RMS of acceleration values also include a column-wise comparison of sensor position.

Muscular activation

The neuromuscular response to vibration and cranking power was not homogenous for all muscles (Fig. 4). A significant main effect of vibration was present for all parts of the triceps surae. The activation of the gastrocnemius medialis (P < 0.001, μ2 = 0.646), gastrocnemius lateralis (P = 0.001, μ2 = 0.352), and soleus (P < 0.001, μ2 = 0.694) increased with vibration significantly. A significant interaction effect of vibration and power (P = 0.048, μ2 = 0.110) was found for the vastus lateralis. Vibration significantly increased the activation at HIGH intensity (HIGH–NoVib vs HIGH–Vib, P < 0.015, μ2 = 0.208) but not at LOW and MED intensity. A significant interaction between vibration and cranking power was present for the arm muscles triceps brachii (P = 0.004, μ2 = 0.209) and flexor carpi ulnaris (P = 0.002, μ2 = 0.234). Without induced vibration, muscular activation increased significantly at HIGH cranking power (LOW–NoVib vs HIGH–NoVib) for the triceps brachii (P < 0.001, μ2 = 0.376) and forearm (P < 0.001, μ2 = 0.376). With vibration, the muscular activity of triceps brachii (P < 0.001, μ2 = 0.694) and flexor carpi ulnaris (P < 0.001, μ2 = 0.659) increased significantly but was not affected systematically by cranking power. Erector spinae, vastus medialis, and rectus femoris did not respond to vibration systematically. A main effect of cranking power manifested in a significantly increased muscular activation for all muscles except triceps brachii and flexor carpi ulnaris. Discrete values for muscular activation are summarized in Table 1 for the lower extremities, Table 2 for the hand–arm system, and Table 3 for the torso.

Mean activation of selected lower leg, thigh, and upper body muscles for LOW (L), MED (M), and HIGH (H) cranking power. Values are displayed as mean ± SD. Muscular activation is expressed relative to the MED–NoVib peak activation. *Significant increase in activation (P < 0.05). Solid black bars indicate the NoVib condition and white bars the Vib conditions. The asterisks within the bars mark the Vib–NoVib comparisons. Curly braces indicate differences of all cranking power conditions included. Solid curly braces refer to the NoVib conditions and dashed curly braces to the Vib conditions.

Heart rate and oxygen consumption

Heart rate increased significantly with cranking power (P < 0.001, μ2 = 0.953) in a range of 123 ± 16 bpm (LOW) and 161 ± 17 bpm (MED) (Table 4). An interaction effect between vibration and cranking power (P = 0.004, μ2 = 0.296) points toward a more prominent vibration-induced increase of heart rate at higher cranking power. At LOW cranking power, the heart rate increased with vibration by 5% from 123 ± 15 to 129 ± 16 bpm. At MED cranking power, the increase was 7% from 151 ± 16 bpm to 161 ± 16 bpm. Relative oxygen uptake increased significantly with cranking power (P < 0.001, μ2 = 0.956) from 28.8 ± 4.3 mL·kg−1⋅min−1 (LOW) to 41.9 ± 4.9 mL·kg−1⋅min−1 (MED) (Table 4). Also, there was a significant main effect of vibration (P = 0.001, μ2 = 0.427). In the presence of vibration, oxygen uptake increased by 2.7% on average.

TABLE 4 - Effects of vibration on relative oxygen uptake and heart rate at steady-state LOW and MED cranking power.
Power LOW Power MED
Measures NoVib Vib NoVib Vib
Oxygen uptake (mL·kg−1⋅min−1) 28.31 ± 4.10 29.34 ± 4.49a 41.48 ± 5.08b 42.38 ± 4.85 a,b
Heart rate (bpm) 123 ± 15 129 ± 16 a 151 ± 16 b 161 ± 16 a,b
Significant effects were accepted with a P value <0.05.
aSignificant increase with vibration.
bSignificant increase in comparison with LOW.


In this paper, vibration was added to cycling at low, close to threshold, and submaximum intensity. The main findings are that vibrations typically imposed in road cycling create varying mechanical stimuli on the entire musculoskeletal system. This implies the acceptance of hypothesis A. The neuromuscular response to these stimuli was muscle specific. Muscular activation increased at the lower limb and hand–arm system, but not at the lower back. The main propulsive muscles as the quadriceps femoris were little affected. Therefore, hypothesis B can be accepted only partwise. The demands on the cardiopulmonary and respiratory system increased slightly in the presence of vibration, which implies the acceptance of hypothesis C. Cranking power modulated the vibration stimulus at most segments and vibration-induced muscular response at the thigh. The increase of heart rate with vibration was more pronounced at higher cranking power. This implies the partial acceptance of hypothesis D.

Hand–arm system

The accelerations of more than 3g RMS on the forearm indicate a considerable vibration exposure of the hand–arm system. The magnitude of forearm vibration exposure was within the hand–arm system’s range of accelerations during test rides with a road bike on rough roads (22) and cobblestones (17). The vibration exposure of the forearm and shoulder remained comparable at all cranking power levels. A high intersubject variability of accelerations at the forearm indicates a different transmission of vibrations to the hand–arm system, most likely due to a looser or firmer handlebar grip. The activation of the flexor carpi ulnaris on the forearm and the triceps brachii on the upper arm increased at all cranking power levels comparably with vibration. Similarly, Arpinar-Avsar and colleagues (22) found increased muscular activation of wrist extensors and flexors on the forearm while riding on rough surfaces. Regardless of whether the increased muscular activation at the forearm relates to a firmer grip on the handlebars or vibration-induced reflex activity, vibration created a considerable mechanical stimulus at the hand–arm system, increasing muscular demands and was linked previously to overuse symptoms (17,22).


The mechanical vibration stimulus acting on the torso was in comparison with the lower and upper extremities weaker. The transmitted accelerations at the pelvis and neck decreased significantly with increasing cranking power. This result can be explained by a load redistribution from the saddle toward the pedals at higher cranking power. With higher cranking power and the same cadence, the pedal forces increase and the saddle loading decreases (23). It seems plausible that because of more translational movement of the pelvis to the saddle, the transmission of vibrations to the torso is reduced. Reduced transmission of vibrations via the saddle is also supported by anecdotal reports of increased riding comfort at high speed on cobblestones compared with slow rinding.

The activation of the lumbar erector spinae increased with cranking power but was not affected systematically by vibration. This agrees with Hansson and colleagues (24), who found no immediate increased activation of the lumbar part of the erector spinae with vibration for seated participants. Nevertheless, they reported increased activation of the thoracic part of the erector spinae with vibration and accelerated and higher muscle fatigue. Therefore, fatigue-related muscular effects are a future research perspective, especially because lower back pain is common among cyclists (25).

Lower extremity

The second highest vibration exposure of the body was found at the lower leg. Following previous findings, the activation of all lower extremity muscles increased with increasing cranking power (26). Vibration led to significantly increased muscular activation of all triceps surae parts. This is in agreement with previous findings for soleus and gastrocnemius medialis (8). Interestingly, the soleus appears to respond stronger to vibration than the gastrocnemii. This is assumed to relate to a higher density of muscle spindles in the soleus (27), which causes a more pronounced reflex response. Similar behavior was demonstrated for standing on vibration plates (28).

The muscular activation of vastus medialis and rectus femoris did not increase with vibration. The comparatively weak vibration stimulus at the thigh might explain this because lower vibration amplitudes and lower vertical accelerations are associated with a weaker muscular response (11,12,29). Opposite to this, the muscular activation of the vastus lateralis increased with vibration at high cranking power. This aligns logically with previous results, which indicate an additional external loading of the muscle, which manifests in cycling as cranking power increases the sensitivity to vibration (13). Also, the vibration exposure of the lower leg and thigh increased with cranking power. This is an inverse trend to the upper body but can be explained by the same load redistribution mechanism described previously, which might not only result in looser coupling at the saddle but tighten the coupling at the pedals. Therefore, one may argue that at high cranking power, a stronger vibration stimulus acts on a more vibration-sensitive muscle.

Findings in the scientific literature on the effects of vibration on knee extensor muscular activation are adverse. In contrast to our findings, Munera and colleagues (8) found a slight vibration-induced increased activation of the vastus medialis and rectus femoris for cycling. Avelar and colleagues (30) found no vibration influence on the EMG of the vastus lateralis during squats. Comparable with the present results, Rønnestad and colleagues (9) found a vibration-induced increase in the activation of the vastus lateralis for cycling at high intensity. Differences in the experimental designs can explain conflicting results. It has been suggested earlier that the characterization of the vibration stimulus through frequency, amplitude, duration and direction (11,13,31); the external load (13); or even the vibratory plates used (32) affect the muscular response to vibration. In particular, because a dose–response relationship between mechanical stimulus and neuromuscular response during cycling is only documented to a limited extent (8), the comparison of previous results is complicated.

Out of a functional perspective, calf and thigh muscles contribute differently to propulsion in cycling. The hip and knee extensors generate a large amount of the propulsive forces. These forces are transmitted to the crank by a stiffened ankle joint, which produces in comparison with hip and knee far less power (33–35). Because the power-contributing knee extensors are less affected by vibration than the power-transferring plantar flexors, vibration does not substantially influence propulsion. This suggestion does not take into account possible antagonizing effects due to co-contraction. However, previous studies showed only a slight increase in the activation of the biceps femoris when vibration stimuli were applied (8,11).


Vibration increased oxygen uptake and heart rate. Although the 2.7% increase in oxygen consumption is close to the range of measurement precision, heart rate supports the existence of an elevated metabolic demand. This is consistent with previous studies that reported a 7% increase in heart rate (5) and an increase in oxygen consumption of 1.12 mL·kg−1⋅min−1 during activities on a vibration plate (5,6). Similarly, Sperlich and colleagues (10) reported an increase in oxygen uptake by 3.3% and 4.4%, respectively, for cycling at 250 and 300 W. The small increase in oxygen uptake is reasonable, as vibration did only affect some muscles in cycling and in particular big muscle groups as the quadriceps femoris were not affected systematically. The findings align with other studies of full-body vibration or high exercise intensities that report increased oxygen uptake (9,10).

Practical implications and limitations

In the context of a cycling race, an increase in heart rate and oxygen demand with vibration indicates a performance loss on cobblestones, although the mainly propulsion-relevant muscles only partially responded to vibration. Compared with low-intensity recreational cycling, vibration exposure and the response of the propulsive muscles are more pronounced at high cranking power levels as they occur in racing. Because reductions in vibration magnitude of 15% at the hands and lower back are already perceptible (36), technical solutions such as damping systems are getting more and more attention in sports and industry. Previous studies raised the question of the extent to which the muscles of the upper body also contribute to increased oxygen consumption with vibration (9). In this context, the increased muscular activation of the hand–arm system due to vibration exposure indicates a possible starting point for technical interventions as bike damping, which potentially affects comfort and cycling performance.

The testing interventions duration, intensity, and vibration stimulus reproduced the passage of a representative cobblestone section. Key characteristics of the vibrations on cobblestones can be reproduced with some limitations with the present laboratory setting. The closed link of bike and vibration plate does not allow a purely stochastical signal as the nonconstrained situation outdoors. Given the duration of a cycle race of up to 6 h with repeated passages of cobblestone sections, race performance may also be compromised by additional biomechanical, physiological, environmental, mechanical, and psychological factors (37). Therefore, this paper focuses on short-term performance, not on fatigue-related effects. Fatigue-oriented testing protocols are in our understanding the next logical step toward a detailed understanding of how vibration affects the musculoskeletal system in cycling. Another interesting question is to what extent the reactions to vibration are influenced by accompanying secondary effects. For example, the vibrating handlebar may cause a firmer handgrip, which additionally modulates the activation of the forearm muscles. The same applies to more demanding line selection on cobblestones and disruptions because of unplanned driving maneuvers. Secondary effects were not considered for the lab-based testing interventions because of a focus on the primarily vibration-related neuromuscular effects. The transmission of vibrations also depends on arm or body posture (38,39). An unintentional optimization of the seating position, such as changing the trunk posture to minimize the transmission of vibrations, can also influence muscular activation. Because the hand positions were standardized, cycling pedals were used, and the seat position settings remained unchanged for each participant, possible degrees of freedom for position adjustment were limited. Nevertheless, an analysis of the full-body kinematics would help to distinguish secondary vibration-induced responses from primary vibration effects as reflex responses.


To the best of our knowledge, this study was the first to use cranking power and vibration characteristics of a cobblestone cycling race to analyze the implications of vibration on neuromuscular performance. Accelerations increased with power output on the lower extremities, whereas they decreased on the upper body. At submaximal intensities, the lower extremities muscular response to vibration was more pronounced. This indicates that vibration-related demands on the propulsive musculoskeletal system are higher in bicycle racing than that in low-intensity recreational sports. Also, the demands on the cardiopulmonary and respiratory system increased slightly in the presence of vibration at all cranking power levels. An interesting perspective is to what extent technical modifications to the bicycle can modulate the vibration exposure and contribute to comfort or performance.

The authors gratefully acknowledge the support and participation of Specialized Bicycle Components, Inc. in this study.

Specialized Bicycle Components, Inc. provided funding for this study. None of the authors had any conflict of interest associated with the study. The results of the present study do not constitute endorsement by the American College of Sports Medicine. The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.


1. Viellehner J, Potthast W. Road to lab: cobblestone cycling vibrations transferred to the lab. In: ISBS—Conference Proceedings Archive: 38 International Society of Biomechanics in Sports; 2020.
2. Burke D, Hagbarth KE, Löfstedt L, Wallin BG. The responses of human muscle spindle endings to vibration of non-contracting muscles. J Physiol. 1976;261(3):673–93.
3. Dindar F, Verrier M. Studies on the receptor responsible for vibration induced inhibition of monosynaptic reflexes in man. J Neurol Neurosurg Psychiatry. 1975;38(2):155–60.
4. Fallon JB, Macefield VG. Vibration sensitivity of human muscle spindles and golgi tendon organs. Muscle Nerve. 2007;36(1):21–9.
5. Avelar NC, Simão AP, Tossige-Gomes R, et al. Oxygen consumption and heart rate during repeated squatting exercises with or without whole-body vibration in the elderly. J Strength Cond Res. 2011;25(12):3495–500.
6. Cochrane DJ, Sartor F, Winwood K, Stannard SR, Narici MV, Rittweger J. A comparison of the physiologic effects of acute whole-body vibration exercise in young and older people. Arch Phys Med Rehabil. 2008;89(5):815–21.
7. Jemni M, Gu Y, Hu Q, et al. Vibration cycling did not affect energy demands compared with normal cycling during maximal graded test. Front Physiol. 2019;10:1083.
8. Munera M, Bertucci W, Duc S, Chiementin X. Analysis of muscular activity and dynamic response of the lower limb adding vibration to cycling. J Sports Sci. 2018;36(13):1465–75.
9. Rønnestad BR, Moen M, Gunnerød S, Øfsteng S. Adding vibration to high-intensity intervals increase time at high oxygen uptake in well-trained cyclists. Scand J Med Sci Sports. 2018;28(12):2473–80.
10. Sperlich B, Kleinoeder H, Quarz D, Linville J, Haegele M. Physiological and perceptual responses of adding vibration to cycling. JEPonline. 2009;12(2):7.
11. Lienhard K, Vienneau J, Nigg S, Meste O, Colson SS, Nigg BM. Relationship between lower limb muscle activity and platform acceleration during whole-body vibration exercise. J Strength Cond Res. 2015;29(10):2844–53.
12. Pollock RD, Woledge RC, Mills KR, Martin FC, Newham DJ. Muscle activity and acceleration during whole body vibration: effect of frequency and amplitude. Clin Biomech. 2010;25(8):840–6.
13. Ritzmann R, Gollhofer A, Kramer A. The influence of vibration type, frequency, body position and additional load on the neuromuscular activity during whole body vibration. Eur J Appl Physiol. 2013;113(1):1–11.
14. Sanders D, Heijboer M. Physical demands and power profile of different stage types within a cycling grand tour. Eur J Sport Sci. 2019;19(6):736–44.
15. Lucia A, Sanchez O, Carvajal A, Chicharro JL. Analysis of the aerobic–anaerobic transition in elite cyclists during incremental exercise with the use of electromyography. Br J Sports Med. 1999;33(3):178–85.
16. Vøllestad NK, Blom PCS. Effect of varying exercise intensity on glycogen depletion in human muscle fibres. Acta Physiol Scand. 1985;125(3):395–405.
17. Chiementin X, Rigaut M, Crequy S, Bolaers F, Bertucci W. Hand–arm vibration in cycling. J Vib Control. 2013;19(16):2551–60.
18. Hermens HJ, Freriks B, Merletti R, et al. European Recommendations for Surface ElectroMyoGraphy. Roessingh Research and Development.” 1999. Available from:
19. Albertus-Kajee Y, Tucker R, Derman W, Lambert M. Alternative methods of normalising EMG during cycling. J Electromyogr Kinesiol. 2010;20(6):1036–43.
20. da Silva JCL, Tarassova O, Ekblom MM, Andersson E, Rönquist G, Arndt A. Quadriceps and hamstring muscle activity during cycling as measured with intramuscular electromyography. Eur J Appl Physiol. 2016;116:1807–17.
21. Cohen J. Statistical Power Analysis for the Behavioral Sciences. Routledge; 2013. p. 689.
22. Arpinar-Avsar P, Birlik G, Sezgin ÖC, Soylu AR. The effects of surface-induced loads on forearm muscle activity during steering a bicycle. JSSM. 2013;1(12):9.
23. Holliday W, Fisher J, Swart J. The effects of relative cycling intensity on saddle pressure indexes. J Sci Med Sport. 2019;22(10):1097–101.
24. Hansson T, Magnusson M, Broman H. Back muscle fatigue and seated whole body vibrations: an experimental study in man. Clin Biomech. 1991;6(3):173–8.
25. Schwellnus M, Derman E. Common injuries in cycling: prevention, diagnosis and management. S Afr Fam Pract. 2005;47(7):14–9.
26. Holliday W, Theo R, Fisher J, Swart J. Cycling: joint kinematics and muscle activity during differing intensities. Sports Biomech. 2019;1–15.
27. Banks RW. An allometric analysis of the number of muscle spindles in mammalian skeletal muscles. J Anat. 2006;208:16.
28. Mildren RL, Peters RM, Carpenter MG, Blouin J-S, Inglis JT. Soleus single motor units show stronger coherence with Achilles tendon vibration across a broad bandwidth relative to medial gastrocnemius units while standing. J Neurophysiol. 2019;122(5):2119–29.
29. Lienhard K, Cabasson A, Meste O, Colson SS. Determination of the optimal parameters maximizing muscle activity of the lower limbs during vertical synchronous whole-body vibration. Eur J Appl Physiol. 2014;114(7):1493–501.
30. Avelar NCP, Ribeiro VGC, Mezêncio B, et al. Influence of the knee flexion on muscle activation and transmissibility during whole body vibration. J Electromyogr Kinesiol. 2013;23(4):844–50.
31. Di Giminiani R, Masedu F, Tihanyi J, Scrimaglio R, Valenti M. The interaction between body position and vibration frequency on acute response to whole body vibration. J Electromyogr Kinesiol. 2013;23(1):245–51.
32. Abercromby AFJ, Amonette WE, Layne CS, Mcfarlin BK, Hinman MR, Paloski WH. Variation in neuromuscular responses during acute whole-body vibration exercise. Med Sci Sports Exerc. 2007;39(9):1642–50.
33. Aasvold LO, Ettema G, Skovereng K. Joint specific power production in cycling: the effect of cadence and intensity. PLoS One. 2019;14(2):e0212781.
34. Mornieux G, Guenette JA, Sheel AW, Sanderson DJ. Influence of cadence, power output and hypoxia on the joint moment distribution during cycling. Eur J Appl Physiol. 2007;102(1):11–8.
35. Zajac FE, Neptune RR, Kautz SA. Biomechanics and muscle coordination of human walking. Part I: introduction to concepts, power transfer, dynamics and simulations. Gait Posture. 2002;16(3):215–32.
36. Ayachi FS, Drouet J-M, Champoux Y, Guastavino C. Perceptual thresholds for vibration transmitted to road cyclists. Hum Factors. 2018;60(6):844–54.
37. Abbiss CR, Laursen PB. Models to explain fatigue during prolonged endurance cycling. Sports Med. 2005;35(10):865–98.
38. Tsukahara Y, Iwamoto J, Iwashita K, Shinjyo T, Azuma K, Matsumoto H. What is the most effective posture to conduct vibration from the lower to the upper extremities during whole-body vibration exercise? OAJSM. 2016;5.
39. Xu XS, Dong RG, Welcome DE, Warren C, McDowell TW, Wu JZ. Vibrations transmitted from human hands to upper arm, shoulder, back, neck, and head. Int J Ind Ergon. 2017;62:1–12.


Copyright © 2020 by the American College of Sports Medicine