Sports Biomechanics

ISSN: 1476-3141 (Print) 1752-6116 (Online) Journal homepage: [Link]

Cycling: joint kinematics and muscle activity

during differing intensities

Wendy Holliday, Raymond Theo, Julia Fisher & Jeroen Swart

To cite this article: Wendy Holliday, Raymond Theo, Julia Fisher & Jeroen Swart (2019): Cycling:
joint kinematics and muscle activity during differing intensities, Sports Biomechanics, DOI:
10.1080/14763141.2019.1640279

To link to this article: [Link]

Published online: 02 Sep 2019.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

[Link]
SPORTS BIOMECHANICS
[Link]

Cycling: joint kinematics and muscle activity during differing

intensities
Wendy Holliday, Raymond Theo, Julia Fisher and Jeroen Swart
Division of Exercise Science and Sports Medicine, Department of Human Biology, Faculty of Health
Sciences, University of Cape Town, Cape Town, South Africa

ABSTRACT ARTICLE HISTORY

Full body kinematics and electromyographic (EMG) patterns may Received 8 March 2019
alter based on the workloads that are encountered during cycling. Accepted 1 July 2019
Understanding the effect of differing intensities on the cyclist can KEYWORDS
guide clinicians and bike fitters in improving specific muscle Biomechanics;
strength and cycling posture to optimise training and racing. We electromyography; bike
aimed to assess changes in lower limb EMG magnitudes and full fitting
body 3D kinematics of 17 well-trained cyclists at three different
exercise intensities: 60%, 80% and 90% of maximum heart rate.
Significant results were demonstrated for all the joints except the
hip and shoulder. Cyclists’ ankle dorsiflexion and knee extension
increased between 6% and 9% with higher intensities. The elbow
adopted a significantly more flexed position, increasing flexion by
39% from 60% to 90% intensity, whilst the lumbar and thoracic
flexion increased by 7% at the higher intensity. There were sig-
nificant increases in EMG signal amplitude at higher intensities for
all muscle groups measured. These results will guide clinicians in
strengthening specific muscles at specific ranges of the cycling
pedal revolution. Guidelines for optimal bicycle configuration
should take into account the full body position of the cyclist as
well as the training and racing intensity when assessing
kinematics.

Introduction
Optimal static bicycle configuration has been the topic of numerous studies (Bini,
Hume, & Croft, 2011; Peveler, 2008; Peveler, Bishop, Smith, Richardson, &
Whitehorn, 2005; Peveler, Pounders, & Bishop, 2007). The freely chosen bicycle con-
figuration and subsequent cyclist kinematics, muscle activity and physiological
responses can be influenced by adjusting any of the contact points on the bicycle
(Burt, 2014). Previous research on the correct positioning of the handlebars, pedal
crank arm length and saddle fore-aft position is based on personal perspectives and
comfort (de Vey Mestdagh, 1998; Silberman, Webner, Collina, & Shiple, 2005; Burt,
2014), whereas static saddle height recommendations have been based on scientific
methods. Currently, there are three main methods used in clinical practice to set the
saddle height: anthropometrics (inseam length and trochanteric leg length), static knee

CONTACT Wendy Holliday HLLWEN005@[Link]

© 2019 Informa UK Limited, trading as Taylor & Francis Group
2 W. HOLLIDAY ET AL.

flexion angle methods and dynamic methods (during pedalling). The recommended
static method is the Holmes method (Peveler et al., 2005). The cyclist is in a stationary
seated position with the crank arm in the lowest or 6 o’clock position and the pedal
surface in a horizontal orientation. Knee flexion angle (KFA) is measured with
a goniometer and recommended to be in a range between 25° and 35° (where full
knee extension is equal to 0°). It has been demonstrated that setting the saddle at this
KFA range statically is optimal for injury prevention and performance (Peveler, 2008).
More recently it has been recommended that bike fitting be conducted in a dynamic
functional manner, as kinematics can be influenced by cycling workload (Ferrer-Roca,
Roig, Galilea, & Garcia-Lopez, 2012; Peveler, Shew, Johnson, & Palmer, 2012). With the
advancement of technology, we are now able to record the cyclists’ position in full
three-dimensional motion capture, however as yet there are limited scientific recom-
mendations for optimal joint ranges for dynamic bicycle configuration. Static recom-
mendations for optimal bicycle configuration cannot be transferred to dynamic
methods as the difference between static and dynamic lower limb angles has been
highlighted (Bini et al., 2016; Fonda, Sarabon, & Li, 2014; Holliday, Fisher, Theo, &
Swart, 2017; Peveler et al., 2012). The range of knee flexion recommended during static
assessment using the Holmes method (25–35°) increases by ~ 5–8° (to approximately
30–40° KFA) depending on the study and the relative workload intensity (Farrell,
Reisinger, & Tillman, 2003; Fonda et al., 2014; Peveler et al., 2012). Increased knee
and hip extension were demonstrated at maximal workloads (Bini & Diefenthaeler,
2010; Bini, Diefenthaeler, & Mota, 2010) and could be linked to a shift in forward
position on the bicycle (Bini, Senger, Lanferdini, & Lopes, 2012). This forward position
on the bicycle was demonstrated during sustained high intensity cycling (Sayers &
Tweddle, 2012). Increased sagittal plane thoracic angle (i.e., the thoracic segment
moving anteriorly relative to the crank arm) occurred towards the end of the protocol
and was suggested to be linked to cyclists shifting their body forwards as they fatigued,
enabling more weight to be exerted onto the pedals. This was further confirmed where
a greater trunk lean angle was demonstrated during a fatiguing protocol, with all
participants also displaying an increase into dorsiflexion at the ankle joint (Dingwell,
Joubert, Diefenthaeler, & Trinity, 2008). There was a positive association, such that an
increase in EMG median frequency signal preceded an increase in movement kine-
matics. These suboptimal positional changes with fatigue may lead to maladaptive joint
loading and thus may result in an increased risk of repetitive strain or long-term
injuries. It was concluded that as fatigue occurs, cyclists changed their body position
and muscle activation patterns to maintain performance (Dingwell et al., 2008).
The typical muscle activation pattern displayed during cycling has been studied in
more depth due to the recent advances in technology (Hug & Dorel, 2009). Likewise,
numerous studies have investigated muscle recruitment patterns in the final stages of
exhaustion during cycling, and it is known that EMG patterns change with the onset of
fatigue (So, Ng, & Ng, 2005). Lower limb muscle coordination during an all-out sprint
cycling task displayed a significant change between the submaximal and maximal
cycling exercises (Dorel, Guilhem, Couturier, & Hug, 2012). The increase in the
duration of all muscle activity during the sprint is suggestive of a strategy to enhance
the work generated by each of the muscle groups. During the all-out sprint, there was
a large increase in hip flexor activity, a lesser extent to the knee flexor activity, whereas
 SPORTS BIOMECHANICS 3

the plantar flexors and knee extensors displayed an even smaller increase. It is possible
that alternative muscles are recruited as fatigue accumulates in working muscles, as
demonstrated by a decrease in Rectus Femoris EMG activity during all-out cycling
sprints (Kay et al., 2001).
These studies were investigated at maximal power or to exhaustion and it is known
that the body position on the bicycle and the muscle recruitment patterns are altered
compared to riding at low intensities. Knowledge of how the muscles adapt to differing
intensities, in conjunction with the position the cyclist is in, would help clinicians and
bike fitters to strengthen those muscles in that range, at that cycling intensity. Racing at
a workload of 55–60% VO2max has been suggested as a strategic way to maximise power
output while minimising the risk of early fatigue (Blake, Champoux, & Wakeling, 2012).
In order to maximise the use of muscle coordination patterns learned during
training, it has been recommended that the cyclist train in similar conditions that
they race in (Blake et al., 2012). The research published to date explores the adaptations
of the lower limb kinematics and muscle activity with maximal effort or fatigue. The
cyclist will, however, spend only a fraction of the race or training at absolute fatigue
and/or maximal effort greater than 90% heart rate intensity, with the majority of the
ride shifting between 60% and 80% heart rate intensity (Padilla et al., 2001; Palmer,
Hawley, Dennis, & Noakes, 1994).
It is beneficial for clinicians and bike fitters to understand how the full body
kinematics and lower limb muscles are affected by differing intensities encountered in
cycling training and racing, not only with fatigue or maximal efforts. The only study to
date that has assessed the relationship between workload intensity and 3D kinematics,
demonstrated a small to moderate difference in lateral spine inclination and spine
rotation between recreational and competitive cyclists (Bini et al., 2016). Currently,
there are no studies investigating full body 3D kinematics simultaneously as well as
lower limb muscle activity at differing intensities. The aim of this study was therefore to
assess how the full body kinematics and specific muscle magnitude is affected by
different intensities that are encountered in cycling. Furthermore, we aim to guide
clinicians and bike fitters with recommendations for which joints or body segments to
focus on during dynamic bike fitting and how cycling intensity during a bike fit may be
of importance.
It was hypothesised that the upper body would adopt a more flexed position with
intensity, whilst the ankle would move into a more dorsiflexed position and the knee
into a more extended position, and that individual muscle activity would also increase
proportionally. Furthermore, we hypothesise that the spinal segments and upper limb
joints will demonstrate significant changes and that cycling intensity will have an
impact on the bike fitting process.

Methods
Participants
Seventeen well-trained male road cyclists (31.2 ± 9.1 years, 75.5 ± 7.5 kg, 178.4 ±
4.4 cm) conforming to Level 2 or greater (De Pauw et al., 2013) were recruited for this
study. Level 2 is described as having a relative VO2max between 45 and 54.9 ml/kg/min,
4 W. HOLLIDAY ET AL.

Table 1. General characteristics of cyclists (n = 17).

Variable Mean ± SD
Age (years) 31.2 ± 9.1
Body mass (kg) 75.5 ± 7.5
Stature (cm) 178.4 ± 4.4
Percentage body fat (%) 8.4 ± 2.8
Trochanteric Leg Length (cm) 95.0 ± 2.7
Skinfolds (mm) 57.6 ± 15.4
PPO (W) 354.0 ± 34.5
PPO (W/kg) 4.7 ± 0.4
VO2max (ml/kg/min) Relative 55.2 ± 6.4

and a relative Peak Power Output (PPO) between 3.6 and 4.5 W/kg. The general
characteristics and performance parameters of the 17 cyclists are shown in Table 1.
Prior to testing, each participant was informed of the risks and stresses associated with
participation in the research trial, were personally interviewed about their training
history, completed a Physical Activity Readiness Questionnaire (PAR-Q) (Whaley,
Brubaker, & Otto, 2007) and signed an informed consent form. The study was approved
by the Human Research Ethics Committee of the Faculty of Health Sciences of the
University of Cape Town, and conformed to the principles of the World Medical
Association Declaration of Helsinki (World Medical Association, 2013).

Testing procedure
The participants reported to the laboratory on three separate occasions (one week apart,
over three weeks) with their own cycling shoes and pedals. A CycleOps 400 Indoor Pro
Cycle (Power Tap: Saris Cycling Group®. Madison, WI. USA) was used for all trials.
Saddle height, saddle setback, handlebar reach and handlebar height were set to match
the configuration of the participant’s own bicycle as previously described (Holliday
et al., 2017).
On the first visit to the laboratory, the participant’s anthropometric measurements
were taken, followed by an incremental exercise test to volitional exhaustion. The
Cycleops VirtualTraining app (VirtualTraining, version 1.7.3, Czech Republic) was
used to control the ergometer and was set according to the participant’s individual
characteristics of age, mass and stature. Heart rate for all sessions was captured by
a Suunto® T6C heart rate monitor (Suunto Oy, Vanata, Finland). The participant
completed a PPO and Peak Oxygen Consumption test to determine the required
workload for the experimental trials. The gas analysis was monitored over 15-s intervals
using an on-line breath-by-breath gas analyser and pneumotach (Oxycon, Viasis,
Hoechberg, Germany). Participants started exercising at a workload of 100 Watts and
resistance was increased by continuous ramp protocol at a rate of 20 W every 60 s until
the participant was exhausted and could not sustain a cadence of at least 60 revolutions
per minute (rpm). PPO was calculated by averaging the power output for the
final minute of the VO2peak test. VO2peak was recorded as the highest VO2 reading
recorded for 30 s during the test. The maximum heart rate (MHR) of each participant
was calculated during the peak power output test, and was used to calculate the target
heart rates for the intensity protocol.
 SPORTS BIOMECHANICS 5

On the second and third visit to the laboratory, the researcher attached the EMG and
3D motion capture markers to the participant (Hermens et al., 1999; “Plug-in Gait
model details”, 2008). This was followed by a static calibration of the motion capture
system before the participant was seated on the CycleOps ergometer.
Each participant performed a 15-min exercise protocol at three different workload
intensities based on the Lamberts Submaximal Cycle Test (Lamberts et al., 2009), which
was previously demonstrated to be highly reliable, with an ICC of R = 0.96 and typical
error of measurement less than 2 beats per minute (bpm). The first stage of the protocol
involved cycling for 6 min at 60% MHR, followed immediately by 6 min at 80% MHR
and a further 3 min at 90% MHR. Cyclists had to elicit and maintain their heart rate
with resistance increased or decreased to avoid their heart rate deviating by more than 2
bmp. Participants were requested to maintain a cadence as close to 90 rpm as possible
throughout the trial. Participants were instructed to remain seated and not to alter their
riding position during the trial, i.e., no standing whilst pedalling or changing the
handgrip position. The riding position was standardised with the cyclists hands on
the brake hoods in order to avoid changes in metabolic cost due to modification of the
trunk angle (Heil, Derrick, & Whittlesey, 1997).
Rating of Perceived Exertion (RPE) was recorded at the end of each intensity stage
using the Borg 6–20 RPE scale (Borg, 1982). Power output, heart rate, speed, cadence
and distance were recorded continuously for later analysis.
The participants repeated this procedure on a third visit to the laboratory one week
later. By doing a repeat session, the reliability of the study’s data was increased,
suggesting that the hypothesised changes in kinematics may be reported with confi-
dence and would also help to reduce the risk of errors in the data interpretation
(Hopkins, 2013; Lamberts et al., 2009).

Instrumentation and analysis

Three-dimensional motion capture and EMG were recorded simultaneously during
the second and third testing sessions. Data were recorded for 15 s during the second
minute of each intensity interval, specifically at 2, 8 and 13 min. The 3D motion capture
markers and EMG electrodes were placed by the primary investigator by measuring and
recording the placement of each marker and electrode to increase the accuracy of
placement in subsequent trials (Tsushima, Morris, & McGinley, 2003).

3D kinematics
An eight-camera motion capture system (Oxford Metric Vicon, Oxford, UK) was used
to capture kinematic data and was recorded at a sampling rate of 250 Hz. The Vicon full
body plug-in gait marker set allows for the measurement of all joint locations and
angles of rotation as well as the calculation of joint moments. Plug-in Gait is
a biomechanical model based on the Newington–Helen Hayes gait model that calculates
joint kinematics and kinetics from the XYZ marker positions and specific subject
anthropometric measurements. The standard full marker set was modified by placing
the tenth thoracic (T10) vertebra marker over the fifth thoracic (T5) vertebra instead.
This was done to more closely approximate static methods used to measure shoulder
6 W. HOLLIDAY ET AL.

flexion angle. All other joint angles and segments were defined as per the manual
(“Plug-in Gait model details”, 2008). Reflective markers were also placed on the pedal
spindle and crank axis, to define a local coordinate system.
Analysis of the 3D kinematic data was performed using MATLAB (The Mathwork®,
USA). The 3D kinematic data were low-pass filtered using a fourth-order Butterworth
filter with a cut-off frequency of 12 Hz. Analysis of 10 revolutions from each intensity
stage was performed on the range of the ankle, knee, hip, shoulder and elbow joint
angles, as well as the lumbar and thoracic spine. Ankle and knee angles for each trial
were reported at bottom dead centre (BDC) pedal position, determined as the point at
which the pedal reflective marker reached its minimal vertical position, i.e., 180°. Full
knee extension is equal to 0°. Ankle neutral is equal to 90° (dorsiflexion ≤90°, plantar-
flexion >90°). The hip flexion angle for each trial was reported at top dead centre (TDC)
pedal position, determined as the point at which the pedal reflective marker reached its
maximal vertical position, i.e., 360°. Thoracic flexion was calculated relative to the local
coordinate system, indicating a forward thoracic tilt or lean on the bicycle. Shoulder,
elbow and spinal angles were taken as an average over the 360° cycle.

Electromyography
The EMG activity during the testing sessions was recorded using an 8-channel EMG
system (Telemyo 2400 G2, Noraxon, USA, Inc., Arizona, USA) Two electrodes (Blue
Sensor, Medicotest, Denmark) were placed on the belly of the right Gluteus Maximus
(GMax), Vastus Medialis Oblique (VMO), Vastus Lateralis Oblique (VLO), Tibialis
Anterior (TA), Rectus Femoris (RF), Medial Gastrocnemius (MG) and Biceps Femoris
(BF) muscles. Prior to placing the electrodes on the skin, the skin over the muscle was
shaved and cleaned with ethanol. The placement and location of the electrodes were
according to the recommendations by SENIAM (Surface EMG for Non-invasive
Assessment of Muscles) (Hermens et al., 1999).
All EMG activity was sampled at 1984 Hz, thus providing raw data at a high enough
frequency for reliable data collection and quantitative data analyses. A 50 Hz notch
filter was applied to filter out the power line noise. The signal was filtered using
a 15–500 Hz band-pass filter to allow movement artefact below 15 Hz and non-
physiological signals above 500 Hz to be removed. The data were smoothed using
root mean squared analysis (RMS), which was calculated for a 50 ms window. Ten
revolutions from each data set were used for EMG analysis, which was performed using
MATLAB (The Mathwork®, USA). The processed EMG data were further analysed into
each quadrant of the cycle revolution, where quadrant 1 represents 0–90°, quadrant 2:
90–180°, quadrant 3: 180–270° and quadrant 4: 270–360°. The average magnitude from
each intensity level, from each quadrant, was expressed as a percentage of the average
magnitude obtained during 10 full revolutions from the first intensity level.
For example, the average magnitude during the 80% intensity stage, in quadrant 3
was calculated as follows:

average of 10 revolutions during f80%g intensity stage in quadrant f3g

avg magnitude ¼ %
average of10 revolutions during 60% intensity stage over a full revolution
 SPORTS BIOMECHANICS 7

Statistical methods
All joint kinematic and EMG magnitude data are expressed as means ± standard
deviation (mean ± SD). The data were statistically tested using a one-way ANOVA
with repeated measures. When significant main effects were found, a Tukey test was
used for post-hoc analysis. Significance was accepted when p value <0.05. The statistical
analyses were performed using GraphPad Prism v7.0a (GraphPad Software, San Diego,
CA, USA).

Results
The mean ± SD and p values for all the joint kinematics can be found in Table 2 and
Figure 1. There was a significant change in all joints across all intensities, except for the
hip and the shoulder joint. The ankle joint progressively moved into dorsiflexion with
the increased intensity with a decrease in mean from 100 ± 5° at 60%, 97 ± 5° at 80%
and 94 ± 6° at 90%, (F(1.215, 27.95) = 26.79). The knee flexion decreased progressively
with an increase in intensity, with a decrease in mean from 37 ± 7° at 60%, 35 ± 6° at
80% and 34 ± 6° at 90%, (F(1.75, 40.19) = 17.45). The spinal flexion increased with an
increase in intensity, with an increase from 45 ± 9° at 60%, 47 ± 11° at 80% and 48 ± 11°
at 90%, (F(1.68, 36.94) = 17.80). The thoracic angle increased with an increase in
intensity, with an increase in mean from 60 ± 5° at 60%, 62 ± 5° at 80% and 64 ± 5°
at 90%, (F(1.37, 30.16) = 21.59). Elbow flexion increased progressively with increased
intensity with an increase in mean from 31 ± 5° at 60%, 36 ± 5° at 80% and 43 ± 10° at
90%, (F(1.23, 29.45) = 35.50).
The mean ± SD and p values for all muscle EMG magnitudes can be found in Table 3
and Figure 2. There were significant changes in all muscle groups with increasing
intensity. The change was most visible in the quadrant that the muscle has been shown
to be most active in, and between 60% and 80% and 60% to 90% intensity. For example,
VLO magnitude increased mostly in quadrant 1 which is the period during which knee
extension is used to generate pedalling power.

Table 2. Mean ± standard deviation and p values for joint kinematics at different intensities and
RPE.
60% 80% 90% p
Ankle BDC 100 ± 5 97 ± 5 94 ± 6 <0.0001*†±
Knee BDC 37 ± 7 35 ± 6 34 ± 6 <0.0001*†
Hip TDC 122 ± 6 122 ± 6 122 ± 6 0.856
Lumbar flexion 45 ± 9 47 ± 11 48 ± 11 <0.0001*†±
Thoracic lean 60 ± 5 62 ± 5 64 ± 5 <0.0001*†±
Shoulder 103 ± 9 104 ± 10 104 ± 8 0.815
Elbow 31 ± 5 36 ± 5 43 ± 10 <0.0001*†±
RPE 9±1 13 ± 2 16 ± 2 <0.0001*†±
Average HR (bpm) 109 ± 6 144 ± 6 164 ± 7 <0.0001*†±
Average cadence (rpm) 88 ± 4 92 ± 2 92 ± 4 0.358
Average speed (km/hr) 36 ± 3 38 ± 4 35 ± 4 0.002*±
Average power (W) 133 ± 17 240 ± 35 303 ± 45 <0.0001*†±
Significant change between 60 and 80% maximal heart rate (MHR), †significant change from 60% to 90% MHR,
±
significant change between 80 and 90% MHR. RPE = Rate of perceived exertion. BDC = bottom dead centre. TDC =
top dead centre.
8 W. HOLLIDAY ET AL.

. B.
120 * 60

*
110
50

100

Degrees

Degrees
40
90

30
80
±

70 20

%
%

60

80

90
60

80

90
p <0.0001 p <0.0001

Ankle BDC Knee BDC

C. D.
140
*
80

130
60
Degrees

Degrees
120
40

110
20
±

100 0
%

%
60

80

90

60

80

90
p = 0.856 p <0.0001

Hip TDC Lumbar flexion

E. F
F.
80 140
*
70 120
Degrees
Degrees

60
100

50
±
80

40
%

%
%

60

80

90
60

80

90

p <0.0001 p = 0.815

Thoracic lean Shoulder

G.
*
60
Degrees

40

20 ±

0
%

%
60

80

90

p <0.0001

Elbow

Figure 1. Joint angles over the different intensities with p values.

*Significant difference between 60% and 80% maximum heart rate (MHR), ±significant difference between 80% and
90% MHR, †significant difference between 60% and 90% MHR.
 SPORTS BIOMECHANICS 9

Table 3. Mean ± standard deviation in percentages for each muscle in each quadrant, and p value
during each different intensity.
Muscle Quadrant 60% 80% 90% p value
GMax 0-90 232.60 ± 54.51 486.44 ± 165.64 624.37 ± 289.36 <0.0001*†±
90-180 60.60 ± 24.34 138.42 ± 67.38 205.08 ± 93.85 <0.0001*†±
180-270 40.31 ± 28.79 50.23 ± 36.69 59.17 ± 32.91 0.0007†
270-360 57.63 ± 22.93 68.45 ± 35.53 82.22 ± 52.82 0.015†±
VMO 0-90 281.55 ± 26.97 387.97 ± 84.33 430.42 ± 114.39 <0.0001*†±
90-180 37.16 ± 18.74 47.12 ± 23.55 54.00 ± 24.50 0.001*†
180-270 9.92 ± 5.62 15.48 ± 9.52 18.99 ± 12.07 <0.0001*†±
270-360 74.30 ± 25.20 104.06 ± 41.75 120.46 ± 69.23 0.006*†
VLO 0-90 280.88 ± 26.29 396.08 ± 90.17 471.79 ± 130.03 <0.0001*†±
90-180 32.92 ± 16.27 43.86 ± 22.36 61.27 ± 47.42 0.009*†
180-270 8.61 ± 3.06 11.86 ± 4.81 13.14 ± 5.02 <0.0001*†
270-360 79.80 ± 21.61 112.33 ± 47.44 140.76 ± 71.31 0.0011*†
TA 0-90 61.85 ± 30.47 92.68 ± 59.60 106.75 ± 66.44 <0.0001*†±
90-180 68.07 ± 39.44 87.41 ± 51.46 92.95 ± 54.12 0.051
180-270 62.95 ± 32.32 78.30 ± 35.82 103.16 ± 63.33 0.029
270-360 199.37 ± 52.00 267.89 ± 94.84 272.06 ± 132.52 0.017*
RF 0-90 146.17 ± 54.65 228.44 ± 142.76 296.10 ± 181.83 <0.0001*†±
90-180 31.78 ± 16.95 37.76 ± 19.14 52.86 ± 26.60 <0.0001*†±
180-270 64.42 ± 25.55 63.37 ± 29.74 68.19 ± 35.12 0.711
270-360 156.95 ± 53.29 214.05 ± 89.06 244.66 ± 118.12 0.003*†
MG 0-90 77.12 ± 34.67 81.49 ± 34.07 85.20 ± 41.59 0.440
90-180 259.81 ± 27.19 269.59 ± 43.28 259.77 ± 45.78 0.309
180-270 52.93 ± 28.90 66.54 ± 33.32 75.24 ± 40.37 0.011†
270-360 10.73 ± 4.44 12.71 ± 5.64 13.07 ± 5.45 0.025†
BF 0-90 121.07 ± 48.40 210.98 ± 73.27 293.15 ± 129.19 <0.0001*†±
90-180 206.17 ± 57.91 325.20 ± 83.63 440.58 ± 145.11 <0.0001*†±
180-270 45.14 ± 29.46 87.53 ± 62.95 129.06 ± 75.81 <0.0001*†±
270-360 27.18 ± 7.40 43.93 ± 19.77 48.13 ± 23.06 <0.0001*†
Significant change between 60 and 80% MHR, †significant change from 60% to 90% MHR, ± significant change between
80 and 90% MHR.

The rating of perceived exertion increased progressively and linearly in keeping with
the increased intensity from a score of 9 ± 1 for 60%, 13 ± 2 for 80% and 16 ± 2 for 90%
intensity (Table 2).

Discussion and implications

The ankle and knee normative values for static bike fitting have been well researched
(Bini et al., 2011; Bini, Hume, Croft, & Kilding, 2011; Hamley & Thomas, 1967;
Holmes, Pruitt, & Whalen, 1994; Peveler, 2008; Peveler et al., 2005, 2007) and more
recently the difference between static and dynamic angles, specifically the ankle and
knee, have been highlighted (Bini et al., 2016; Fonda et al., 2014; Holliday et al., 2017;
Peveler et al., 2012). As yet there are no normative data to describe the dynamic values
for full body joint angles recommended for cycling at differing intensities.
The muscle recruitment pattern during cycling has also been well researched (So
et al., 2005), including intramuscular EMG (Chapman, Vicenzino, Blanch, Knox, &
Hodges, 2010; Da Silva et al., 2016). The purpose of this study was not to report the
activation patterns, but to assess the changes in the lower limb EMG magnitudes during
differing intensities as it is not clear how these may change with the natural adoption of
10 W. HOLLIDAY ET AL.

Figure 2. EMG magnitudes over different intensities, in each quadrant.

*Significant difference between 60% and 80% maximum heart rate (MHR), ±significant difference between 80% and
90% MHR, †significant difference between 60% and 90% MHR.
 SPORTS BIOMECHANICS 11

different body positions as intensities increase, nor whether the joint kinematics change
significantly.
This study produced similar results to that of Blake et al. (2012), showing that there
was a general increase in muscle activation across muscle groups as the intensity
increased. Blake et al. (2012) analysed nine male cyclists at a low (25–55% VO2max)
and a high intensity (60–90% VO2max). In keeping with this previous study, our study
also demonstrated an increase in EMG activity of TA and RF across the top and early
part of the pedal cycle (0–90°) with increasing intensity. TA was shown to work
predominately in the fourth quadrant, thus suggesting an increase in ankle dorsiflexion
near the TDC. There was a significant change in RF from 60% to 90% MHR in the
first, second and fourth quadrants which correspond to hip flexion, driving the knee
over the TDC of the pedal revolution and knee extension in the push phase. The
significant changes in TA and RF may indicate that these are the muscles responsible
for driving the pedal across the TDC, an area where the major muscle groups are unable
to exert effective force to drive crank rotation.
The role of GMax is to extend the hip joint, and there were significant increases in
EMG signal through all three intensities in the pushing phase of the pedal revolution
(from 0° to 180°). Similarly, the VMO and VLO extend the knee joint in the same push
phase of the pedal revolution, and there were significant increases in EMG magnitude
through all three intensities in the first quadrant. The MG worked predominantly in
the second quadrant, corresponding to the second half of the push phase of the pedal
revolution; however, the magnitude remained constant with only minor significant
changes in the third and fourth quadrants between 60% and 90%. Even though
Soleus was not examined in this study, it has been demonstrated that Soleus and MG
work together from 340° through to 270° in the pedal revolution to stabilise the ankle
and to transfer force to the pedal exerted by the relatively large GMax and quadriceps
muscles (Fonda & Sarabon, 2010; Jorge & Hull, 1986). As such, even at lower workloads
the force applied by MG in order to stabilise the ankle may be relatively higher. Similar
results have been reported by Blake et al. (2012) where GMax had the largest increase in
activity from a low to a high intensity, VMO and VLO were both highly active in the
push phase with increasing intensity and the MG showed very little change with
increasing workloads.
Numerous studies have shown an increase in hip and knee extension, as well as ankle
dorsiflexion, with incremental cycling (Bini et al., 2010; Bini & Diefenthaeler, 2010).
The previous knee and ankle findings are consistent with our study, suggesting
a movement into dorsiflexion to increase stability around the ankle joint in order to
transfer force effectively to the pedals to maintain the power output. The movement
into dorsiflexion may increase the efficiency of MG or increase passive tension in the
muscle tendon unit to assist with force transfer. The ankle increased into dorsiflexion
by 6° between 60% and 90% intensity, which is a greater difference than the reported
TEM of 3.5° (Holliday et al., 2017). As the bicycle contact points are fixed, this increase
in ankle dorsiflexion requires an increase in knee extension (Peveler et al., 2012).
Dynamic bike fitting systems recommend a dynamic KFA of 30–40°, however there is
no research validating this specific range. Previous research has demonstrated
a difference in KFA of between 5° and 8° in static relative to dynamic measures
(Farrell et al., 2003; Fonda et al., 2014; Holliday et al., 2017). The results from this
12 W. HOLLIDAY ET AL.

study also demonstrated a difference in KFA at low and high intensities, and it is,
therefore, possible to infer that optimal KFA at BDC position using dynamic measure-
ments should range from 33° to 43° at low intensity and 30–40° at high intensity.
Although statistically significant, from a clinical and practical perspective, it is recom-
mended that the use of dynamic 2D and 3D kinematic data should interpret knee
flexion in relation to the relative intensity during data capture.
There were no significant hip joint angle changes in any of the quadrants, at any of
the intensities. This differs from previous studies that have shown hip extension
increases with incremental cycling (Bini et al., 2010; Bini & Diefenthaeler, 2010;
Sanderson & Black, 2003). The hip angles in previous studies were measured as an
angle bisecting the length of the femur and a line parallel to the floor or as an angle
bisecting the length of the femur and a line from the hip joint centre to the shoulder
centre. These measures exclude the spinal segments and do not measure the indepen-
dent hip joint angle (long axis of femur and lumbar spine-sacrum), as was done in this
study.
Similar to the hip, the shoulder angle is often determined as an angle between the
elbow, mid-shoulder and hip joint centre. A clinical shoulder angle will take the
thoracic spine into account, as was done in this study. There were no significant
changes in the shoulder angle, at any of the different intensities, yet the elbow and
thoracic lean angle changed significantly between all three intensities. This is consistent
with research where there was a significant change in forward body position on the
bicycle at maximal power output (Bini et al., 2012; Sayers & Tweddle, 2012). It was
suggested that cyclists increased their trunk lean angle in response to muscular fatigue,
and that changes in EMG preceded changes in mean trunk lean angle (Dingwell et al.,
2008). It was hypothesised that the increase in trunk lean angle was in order to focus on
increasing hip extensor muscle length and reducing knee flexor moment (Bini et al.,
2012; Dingwell et al., 2008).
Our findings that the hip joint position remained unchanged while significant
lumbar flexion did occur, indicate that the previous basic methods of measuring the
angles of the body, without taking into consideration the spine, should be discarded.
The spine consists of 33 bones and each joint has varying degrees of movement. It is
clear from this study that movement occurs in the lumbar, thoracic and elbow joints
with increased intensity, not at the hip or shoulder. A possible rationale for this change
in position may relate to the transfer of force across the hip joint. GMax demonstrated
the largest change in EMG magnitude from low to high workloads. Increased GMax
activity may aid the transfer of the increased force across the hip joint by stabilising the
pelvis (Li & Caldwell, 1998). The increase in lumbar flexion and elbow flexion may,
therefore, be a compensatory mechanism to stabilise the pelvis through the contact
points at the hands as the forces across the hip joint increase (Grant, Watson, & Baker,
2015). Future research on more detailed spinal segment kinematics as well as spinal and
upper limb EMG analysis with increasing cycling intensity should be considered.

Conclusion
It is clear from this study that the magnitudes of muscles used during cycling increase
with increasing intensity. The ankle adopts a more dorsiflexed position and the knee
 SPORTS BIOMECHANICS 13

moves into a more extended position with an increase in cycling intensity. The elbow
and lumbar and thoracic spinal segments also adopt a more flexed position as intensity
increases. Previous recommendations for optimal cycling position have been suggested
for the lower limb, however from these results, it is essential that lumbar and thoracic
spinal segments are also taken into account. Guidelines for optimal bicycle configura-
tion should, therefore, consider the full body kinematics as well as conducting the bike
fit at an intensity applicable to the cyclist’s individual training and racing goals.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
This work was supported by the National Research Fund of South Africa [101413].

References
Bini, R., Dagnese, F., Rocha, E., Silveira, M., Carpes, F., & Mota, C. (2016). Three-dimensional
kinematics of competitive and recreational cyclists across different workloads during cycling.
European Journal of Sport Science, 16, 553–559. doi:10.1080/17461391.2015.1135984
Bini, R., & Diefenthaeler, F. (2010). Kinetics and kinematics analysis of incremental cycling to
exhaustion. Sports Biomechanics, 9, 223–235. doi:10.1080/14763141.2010.540672
Bini, R., Diefenthaeler, F., & Mota, C. B. (2010). Fatigue effects on the coordinative pattern
during cycling: Kinetics and kinematics evaluation. Journal of Electromyography and
Kinesiology, 20, 102–107. doi:10.1016/[Link].2008.10.003
Bini, R., Hume, P., & Croft, J. (2011). Effects of bicycle saddle height on knee injury risk and
cycling performance. Sports Medicine, 41, 463–476. doi:10.2165/11588740-000000000-00000
Bini, R., Hume, P., Croft, J., & Kilding, A. (2011). Effects of saddle position on pedalling technique
and methods to assess pedalling kinetics and kinematics of cyclists and triathletes (PhD thesis).
Auckland University of Technology.
Bini, R., Senger, D., Lanferdini, F., & Lopes, A. L. (2012). Joint kinematics assessment during
cycling incremental test to exhaustion. Isokinetics and Exercise Science, 20, 99–105.
doi:10.3233/IES-2012-0447
Blake, O. M., Champoux, Y., & Wakeling, J. M. (2012). Muscle coordination patterns for efficient
cycling. Medicine and Science in Sports and Exercise, 44, 926–938. doi:10.1249/
MSS.0b013e3182404d4b
Borg, G. (1982). Psychophysical bases of perceived exertion. Medicine & Science in Sports &
Exercise, 14, 377–381. doi:10.1249/00005768-198205000-00012
Burt, P. (2014). Bike fit: Optimise your bike position for high performance and injury avoidance
(pp. 37–67). London: Bloomsbury Publishing Plc.
Chapman, A., Vicenzino, B., Blanch, P., Knox, J., & Hodges, P. (2010). Intramuscular fine-wire
electromyography during cycling: Repeatability, normalisation and a comparison to surface
electromyography. Journal of Electromyography and Kinesiology, 20, 108–117. doi:10.1016/j.
jelekin.2008.11.013
Da Silva, J., Tarassova, O., Ekblom, M. M., Andersson, E., Ronquist, G., & Arndt, A. (2016).
Quadriceps and hamstring muscle activity during cycling as measured with intramuscular
electromyography. European Journal of Applied Physiology, 116, 1807–1817. doi:10.1007/
s00421-016-3428-5
14 W. HOLLIDAY ET AL.

De Pauw, K., Roelands, B., Cheung, S., de Geus, B., Rietjens, G., & Meeusen, R. (2013).
Guidelines to classify subject groups in sport-science research. International Journal of
Sports Physiology and Performance, 8, 111–122. doi:10.1123/ijspp.8.2.111
de Vey Mestdagh, K. (1998). Personal perspective: In search of an optimum cycling posture.
Applied Ergonomics, 29, 325–334. doi:10.1016/S0003-6870(97)00080-X
Dingwell, J., Joubert, J., Diefenthaeler, F., & Trinity, J. (2008). Changes in muscle activity and
kinematics of highly trained cyclists during fatigue. IEEE Transactions on Bio-Medical
Engineering, 55, 2666–2674. doi:10.1109/TBME.2008.2001130
Dorel, S., Guilhem, G., Couturier, A., & Hug, F. (2012). Adjustment of muscle coordination
during an all-out sprint cycling task. Medicine and Science in Sports and Exercise, 44,
2154–2164. doi:10.1249/MSS.0b013e3182625423
Farrell, K. C., Reisinger, K. D., & Tillman, M. D. (2003). Force and repetition in cycling: Possible
implications for iliotibial band friction syndrome. Knee, 10, 103–109. doi:10.1016/S0968-
0160(02)00090-X
Ferrer-Roca, V., Roig, A., Galilea, P., & Garcia-Lopez, J. (2012). Influence of saddle height on
lower limb kinematics in well-trained cyclists: Static versus dynamic evaluation in bike fitting.
Journal of Strength and Conditioning Research, 26, 3025–3029. doi:10.1519/
JSC.0b013e318245c09d
Fonda, B., & Sarabon, N. (2010). Biomechanics of cycling. Sports Science Review, XIX, 187–210.
doi:10.2478/v10237-011-0012-0
Fonda, B., Sarabon, N., & Li, F. (2014). Validity and reliability of different kinematics methods
used for bike fitting. Journal of Sports Sciences, 32, 940–946. doi:10.1080/
02640414.2013.868919
Grant, M., Watson, H., & Baker, J. (2015). Assessment of the upper body contribution to
multiple-sprint cycling in men and women. Clinical Physiology and Functional Imaging, 35,
258–266. doi:10.1111/cpf.12159
Hamley, E., & Thomas, V. (1967). Physiological and postural factors in the calibration of the
bicycle ergometer. Journal of Physiology, 191, 55–56.
Heil, D., Derrick, T., & Whittlesey, S. (1997). The relationship between preferred and optimal
positioning during submaximal cycle ergometry. European Journal of Applied Physiology and
Occuptional Physiology, 75, 160–165. doi:10.1007/s004210050141
Hermens, H. J., Freriks, B., Merletti, R., Stegeman, D., Blok, J., Rau, G., . . . Hägg, G. (1999).
European recommendations for surface electroMyoGraphy. Enschede, the Netherlands:
Roessingh Research and Develpoment.
Holliday, W., Fisher, J., Theo, R., & Swart, J. (2017). Static versus dynamic kinematics in cyclists:
A comparison of goniometer, inclinometer and 3D motion capture. European Journal of Sport
Science, 17, 1129–1142. doi:10.1080/17461391.2017.1351580
Holmes, J., Pruitt, A., & Whalen, N. (1994). Lower extremity overuse in bicycling. Clinics in
Sports Medicine, 13, 187–205.
Hopkins, W. (2013). A new view of statistics. Retrieved from [Link]
stats/[Link]
Hug, F., & Dorel, S. (2009). Electromyographic analysis of pedaling: A review. Journal of
Electromyography and Kinesiology, 19, 182–198. doi:10.1016/[Link].2007.10.010
Jorge, M., & Hull, M. (1986). Analysis of EMG measurements during bicycle pedalling. Journal of
Biomechanics, 19, 683–694. doi:10.1016/0021-9290(86)90192-2
Kay, D., Marino, F. E., Cannon, J., St Clair Gibson, A., Lambert, M. I., & Noakes, T. D. (2001).
Evidence for neuromuscular fatigue during high-intensity cycling in warm, humid conditions.
European Journal of Applied Physiology, 84, 115–121. doi:10.1007/s004210000340
Lamberts, R. P., Swart, J., Richard, W., Hons, M., Noakes, T. D., & Lambert, M. I. (2009).
Measurement error associated with performance testing in well- trained cyclists : Application
to the precision of monitoring changes. International SportMed Journal, 10, 33–44.
Li, L., & Caldwell, G. (1998). Muscle coordination in cycling: Effect of surface incline and
posture. Journal of Applied Physiology, 85, 927–934. doi:10.1152/jappl.1998.85.3.927
 SPORTS BIOMECHANICS 15

Padilla, S., Mujika, I., Orbananos, J., Santisteban, J., Angulo, F., & Goiriena, J. (2001). Exercise
intensity and load during mass-start stage races in professional road cycling. Medicine &
Science in Sports & Exercise, 33, 796–802. doi:10.1097/00005768-200105000-00019
Palmer, G., Hawley, J., Dennis, S., & Noakes, T. (1994). Heart rate responses during a 4-d cycle
stage race. Medicine and Science in Sports and Exercise, 26, 1278–1283. doi:10.1249/00005768-
199410000-00016
Peveler, W. (2008). Effects of saddle height on economy in cycling. Journal of Strength and
Conditioning Research, 22, 1355–1359. doi:10.1519/JSC.0b013e318173dac6
Peveler, W., Bishop, P., Smith, J., Richardson, M., & Whitehorn, E. (2005). Comparing methods
for setting saddle height in trained cyclists. Journal of Exercise Physiology Online, 8, 51–55.
Peveler, W., Pounders, J., & Bishop, P. (2007). Effects of Saddle Height on anaerobic power
production in cycling. Journal of Strength and Conditioning Research, 21, 1023–1027.
doi:10.1519/R-20316.1
Peveler, W., Shew, B., Johnson, S., & Palmer, T. (2012). A kinematic comparison of alterations to
knee and ankle angles from resting measures to active pedaling during a graded exercise
protocol. Journal of Strength and Conditioning Research, 26, 3004–3009. doi:10.1519/
JSC.0b013e318243fdcb
Plug-in Gait model details. (2008, March). Retrieved from Vicon website [Link]/down
loads/documentation/plug-in-gait-model-details
Sanderson, D. J., & Black, A. (2003). The effect of prolonged cycling on pedal forces. Journal of
Sports Sciences, 21, 191–199. doi:10.1080/0264041031000071010
Sayers, M., & Tweddle, A. (2012). Thorax and pelvis kinematics change during sustained cycling.
International Journal of Sports Medicine, 33, 314–390. doi:10.1055/s-0031-1291363
Silberman, M., Webner, D., Collina, S., & Shiple, B. (2005). Road bicycle fit. Clinical Journal of
Sport Medicine, 15, 271–276.
So, R., Ng, J., & Ng, G. (2005). Muscle recruitment pattern in cycling: A review. Physical Therapy
in Sport, 6, 89–96. doi:10.1016/[Link].2005.02.004
Tsushima, H., Morris, M. E., & McGinley, J. (2003). Test-retest reliability and inter-tester
reliability of kinematic data from a three-dimensional gait analysis system. Journal of the
Japanese Physical Therapy Association, 6, 9–17. doi:10.1298/jjpta.6.9
Whaley, M., Brubaker, P., & Otto, R. (2007). Preparticipation health screening and risk stratifica-
tion. In M. Whaley, P. Brubaker, & R. Otto (Eds.), ACSM’s guidelines for exercise testing and
prescription (pp. 26). Philadelphia: Lippincott Williams & Wilkins.
World Medical Association. (2013). World medical association declaration of Helsinki: Ethical
principles for medical research involving human subjects. Journal of the American Medical
Association, 310, 2191–2194. doi:10.1001/jama.2013.281053