Documentos de Académico
Documentos de Profesional
Documentos de Cultura
net/publication/8220277
Lower extremity muscle activities during cycling are influenced by load and
frequency
CITATIONS READS
105 378
2 authors:
Brian S. Baum Li Li
Regis University Georgia Southern University
23 PUBLICATIONS 443 CITATIONS 156 PUBLICATIONS 2,360 CITATIONS
Some of the authors of this publication are also working on these related projects:
Received 19 June 2002; received in revised form 31 October 2002; accepted 12 November 2002
Abstract
The purpose of this experiment was to investigate the effects of frequency and inertia on lower extremity muscle activities during
cycling. Electromyographic (EMG) data of seven lower extremity muscles were collected. Sixteen subjects cycled at 250 W across
different cadences (60, 80, and 100 rpm) with different loads (0, 0.5, 1.0, 1.5, and 2.0 kg) attached to distal end of their thighs.
Load and cadence interactions were observed for the offset of the biceps femoris (BF), the active duration of the rectus femoris
(RF), and the peak magnitudes of the vastus lateralis (VL) and the tibialis anterior (TA). Cadence effects were observed in the
onset of the gluteus maximus (GM), RF, BF, VL, and TA; the offset of the GM, RF, BF, VL; the duration of the BF and TA; the
peak magnitude of the RF and gastrocnemius (GAS); and the crank angle at which the peak magnitude was achieved of the BF,
GAS, and soleus (SOL). Load effect was observed from the onset of RF and SOL, the offset of RF, the duration of SOL, and the
peak magnitude of BF. These results indicate that inertial properties influence the lower extremity muscular activity in addition to
the cadence effect.
2002 Elsevier Science Ltd. All rights reserved.
1050-6411/03/$ - see front matter 2002 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S1050-6411(02)00110-4
182 B.S. Baum, L. Li / Journal of Electromyography and Kinesiology 13 (2003) 181–190
decreased from approximately 155 N at 60 rpm to 85 N variables. Discerning this information will help clarify
at 120 rpm. The muscular component of the pedal force the results of previous literature and provide further
may also remain relatively constant [11] as cadence insights into how the body coordinates movement and
increases, but a marked increase in the non-muscular adapts to different demands of speed and inertia. This
component with cadence explains the quadratic trend in knowledge is useful in many fields, including athletics
total forces with the pedaling cadence. Since gravi- and rehabilitation.
tational effects remain fairly constant across cadence Therefore, the purpose of this study was to investigate
conditions at the same body position [1], the increase in the interaction between load and frequency effects on
the non-muscular component reflects an increase of the lower extremity muscle activities during cycling. More
inertial influence on pedal forces at greater cadences. Li specifically, four hypotheses were examined. First, the
[13] reported that muscular coordination changed as cad- influence of non-muscular components of force increases
ence increased. Based on the magnitude of the changes with cadence [11,15,21], therefore, we hypothesized that
in a proximal (greater changes) to distal (lesser changes) loads at the distal end of the thigh, increasing the inertia
order, he suggested that the greater inertial properties of of the thigh, would show increased influences at greater
the proximal limb segments played an important role in cadences. Cadence effects on the lower extremities are
the coordination changes. Combined, these researchers reportedly more considerable proximally than distally,
pointed to the notion that inertia influences the kinetic and the greater inertial properties of the limbs are in this
and muscular changes at greater cadences. However, the order [13]. Therefore, a second hypothesis was that load
effects of inertial properties were compounded with cad- manipulations would produce greater effects proximally.
ence effect in previous studies; consequently, the influ- Furthermore, functional differences were thought to exist
ence of inertia was not investigated independent of cad- between mono- and bi-articular muscles [8,9], so it was
ence, so attributing observed effects to inertia cannot be reasonable to expect the two muscle types to respond
fully supported by these observations. differently to external manipulations. Consequently, our
Influences of altered inertial properties on movement third hypothesis was that mono- and bi-articular muscles
have been reported in the gait literature. For example, would react to load and cadence changes differently: the
in transtibial amputees, lower energy costs and more mono-articular muscles would react by shifts in timing
symmetrical kinematics were observed with prostheses whereas the bi-articular muscles would change in func-
having lesser inertial properties [18]. During walking tion. Finally, some muscles may respond differently to
and running, adding masses to the limbs has been cadence manipulations than others [22], so it was hypo-
reported to increase energy costs [20] and affect joint thesized that there would be coordination changes
kinetics [2,12]. To explain observations of increased between antagonistic muscles due to both cadence and
joint reaction forces and net moments proximally, it was load manipulations.
suggested that the lower extremity muscles increased
their output in response to the inertial loading [12]. The
combination of these reports suggests that changing the 2. Methods
inertial properties of the limbs will affect the muscu-
lature, but EMG data to support this interpretation are 2.1. Participants
scarce.
Consequently, examining both cycling and gait litera- Sixteen male participants of mean (SD) age: 23 (5)
ture reveals a gap. The effects of manipulating inertial years; height: 1.8 (0.2) m; and body mass: 85 (10) kg
properties on both kinetics and kinematics have been were recruited from the Louisiana State University com-
reported, but how this type of manipulation affects mus- munity. Each participant had at least two years of rec-
cular activity and coordination has not been well docu- reational cycling experience. In compliance with the uni-
mented. Since walking and running are complex forms versity Institutional Review Board policy, participants
of locomotion, it is more difficult to isolate the effects gave their informed consent before the experiment. The
of inertia, so cycling may be a more effective mode for experiment was conducted in the Motor Behavior Lab-
this type of investigation. Few data exist related to alter- oratory of Louisiana State University.
ing inertial properties during cycling, and suggestions
that inertial property alterations in gait and prosthetic 2.2. Muscle activity
literature affect the musculature of the lower extremity
lack EMG data confirmation. Keeping cadence constant The muscle activities of gluteus maximus (GM), rec-
and manipulating the inertial component by adding tus femoris (RF), biceps femoris (BF), vastus lateralis
masses to lower extremity segments may provide a win- (VL), tibialis anterior (TA), medial gastrocnemius
dow to investigate this issue. In this way, the effect of (GAS), and soleus (SOL) of the left lower extremity
inertial property manipulation on muscular activity and were monitored with surface electromyography (EMG).
coordination can be investigated independent of other After sites were shaved, lightly abraded, and cleaned
B.S. Baum, L. Li / Journal of Electromyography and Kinesiology 13 (2003) 181–190 183
with alcohol, pairs of silver/silver chloride pre-gelled sidered a point mass 8 cm from the distal end of the
surface electrodes (Marquette Medical Systems, Jupiter, thigh. Again, Eq. (1) was used, but m was the load mass
Florida) of 1.5 cm diameter with a center to center dis- and r was segment length minus 8 cm. The calculated
tance of 2.5 cm were applied along the muscle fibers moments of inertia for the loads were then expressed
over the bellies of the seven muscles for EMG data as a percentage of the estimated moments of inertia of
acquisition. A common reference electrode was placed the thigh.
on a bony site at the distal end of the left ulna. An EMG The loads were created using plastic Ziploc bags
system (Run Technologies, Laguna, CA) that consisted (S.C. Johnson, Racine, WI) reinforced with duct tape and
of bipolar differential amplifier with an input impedance filled with sand. The loads were positioned 8 cm above
of 1 M⍀, a gain up to 10,000, a Common Mode Rejec- the knee joint center at the anterior of the thigh for each
tion Ratio (CMRR) of 90 dB at 60 Hz, and a signal participant. The precise position was difficult to achieve
to noise ratio of ⬍0.8 µV at 60 Hz was used during due to inter-participant anthropomorphic differences and
data collection. the variability of sand as a load. However, since the goal
The seven muscles were chosen to represent lower of the load was simply to change the inertial properties
extremity functional muscle groups based on several cri- at the distal end of the thigh, the precise load location
teria. First, the muscles chosen cross all three of the was not considered a major drawback (e.g. Martin and
major joints in the lower extremity (hip, knee, and Cavanagh [17]). The testing order for each participant
ankle); second, the muscles represent mono- and bi- was randomized to minimize possible order effects, such
articular antagonist muscle groups that cross each of the as fatigue and learning. That order was then altered so
joints; and finally, the seven muscles chosen include that the same load was collected for three consecutive
those most commonly found in literature using cycling trials at the different cadences. The pseudo-random order
as a means of investigation. Based on this selection, all was used to control the load location variability across
the joint flexor and extensor groups (mono- and bi- different cadences. In each condition, the participants
articular) were monitored with the exception of a single pedaled for approximately 1 min to establish a steady
joint hip joint flexor. Single joint hip flexors are deeper state before a 10 s data collection trial. Since participants
muscles and very difficult to monitor by surface elec- pedaled at a power output of 250 W, and this power
trodes. This is a common practice amongst cycling output is relatively small in relation to maximal power
related research projects. output values of 1300 W or more between 60 and 100
rpm [19], fatigue was not considered a major factor.
2.3. Experiment conditions Nonetheless, each participant rested for at least 1 min
between trials with an extra 5 min of rest after every
The participants rode a Monarch cycling ergometer three trials to minimize possible fatigue effect.
(model 868, Stockholm, Sweden) at a constant power
output of 250 W. Resistances were changed according
2.4. Data collection and processing
to the desired pedaling cadences to meet this require-
ment. Seat height was adjusted such that the distance
between the seat and the crank center was 100% of the An internally synchronized motion analysis system
participant’s greater trochanter length at standing posture was employed to collect both kinematic and EMG data.
[6,7,10]. Cadence conditions of 60, 80, and 100 rpm rep- The system included a 2-D 60 Hz camera
resented low, medium, and high cadence conditions, (MotionAnalysis, Santa Clara, CA) to capture kinematic
respectively, and participants maintained these cadences data with passive reflective markers and a 16-channel
using a cadence monitor attached to the ergometer. Mean surface EMG system to capture EMG activity.
cadences were then calculated from kinematic data. Five Reflective markers were placed on the crank center
sand bags with different loads of 0, 0.5, 1.0, 1.5, and and pedal spindle of the bicycle to represent the crank
2.0 kg were attached approximately 8 cm from the distal arm in order to identify crank cycles and to calculate the
end of the thigh via Flexi-Wrap (Cramer, Gardener, KS) mean cadences for each participant. The coordinates of
without compromising the location of the electrodes. these markers were recorded by the camera and auto-
Moments of inertia of the thigh were estimated via matically digitized by the MotionAnalysis system. One
Eq. (1): complete cycle (Fig. 1) was defined by consecutive
occurrences of crank top-dead-center (TDC), with the
I ⫽ mr2 (1)
crank arm vertical and the pedal at its highest position.
where I was the moment of inertia, m was the segment The EMG data were sampled at 960 Hz. Raw data
mass, and r, the radius of gyration, was 0.54×segment were unbiased, full-wave rectified, and then smoothed
length [3]. The thigh segment length (hip center to knee with a low pass, fourth order, zero lag Butterworth filter
center) was estimated as 0.245×body height [25]. For at 7 Hz to create the linear envelope. The lab configur-
estimated added moments of inertia, the loads were con- ation was free of high frequency noises, and the motion
184 B.S. Baum, L. Li / Journal of Electromyography and Kinesiology 13 (2003) 181–190
3. Results
Fig. 2. Examples of EMG linear envelopes with onset and offset threshold criterion. In (a) 10% of the maximum value across conditions was
considered appropriate for the onset and offset threshold criterion. In (b) 20% of the maximum value was considered appropriate for the threshold
criterion. The 10% threshold value was deemed inappropriate because it created a problematic determination of when the muscle was considered
“on” or “off.”
Fig. 5. Mean ensemble curves of EMG activity for each muscle across (a) cadence and (b) load conditions. The crank angle range represents
TDC to next TDC, 0–360°. EMG curves for each subject were scaled to the maximum value observed across all 15 conditions. The mean curves
here were calculated from the scaled individual curves of gluteus maximus (GM), rectus femoris (RF), biceps femoris (BF), vastus lateralis (VL),
tibialis anterior (TA), gastrocnemeus (GAS) and soleus (SOL).
Table 1
Mean peak of EMG activity per cycle across cadences expressed as a percentage of the maximum value of each muscle
Values are mean (±SD). ∗Significant linear trend; #significant quadratic trend. a,b,c
Indicate significantly different heterogeneitic groups.
Table 2
Mean peak of EMG activity per cycle across loads expressed as a percentage of the maximum value of each muscle
Table 3
Mean crank angle, in degrees, at which the peak EMG activity per cycle occurred across cadences
Table 4
Mean crank angle, in degrees, at which the peak EMG activity per cycle occurred across loads
included the offset of the biceps femoris (BF), the dur- SOL, and the peak magnitude of BF. These results indi-
ation of the rectus femoris (RF), and the peak magni- cate that load and cadence can influence lower extremity
tudes of the vastus lateralis (VL) and the tibialis anterior muscular activity either independently (main effects) or
(TA). Moreover, more than 45% of the tested variables jointly (interaction).
(16) reacted to the pedaling frequency with significant Li [13] and Li and Caldwell [14] proposed that cad-
alteration (significant cadence effect). These parameters ence effects observed on neuromuscular coordination
included the onset of the gluteus maximus (GM), RF, stemmed from a systematic increase of the inertial
BF, VL, and TA; the offset of the GM, RF, BF, VL; the influence with increasing pedal speeds. Little empirical
duration of the BF and TA; the peak magnitude of the evidence exists related to physically altering the inertial
RF and gastrocnemius (GAS); and the crank angle at properties of the lower extremity, but several researchers
which the peak magnitude was achieved of the BF, GAS, have discussed the inertial influences on both the kinetics
soleus (SOL). In addition to the abovementioned 20 and muscular activities during cycling. For example,
parameters, five parameters (more than 14%) exhibited Kautz and Hull [11] reported increased magnitudes in
significant influence of the load effect, including the the non-muscular component (specifically the inertial
onset of RF and SOL, the offset of RF, the duration of subcomponent) of the pedal force with cadence. Using
B.S. Baum, L. Li / Journal of Electromyography and Kinesiology 13 (2003) 181–190 189
a loading method, Martin and Cavanagh [17] reported timing was not observed in the present study. This may
that kinematic patterns of gait did not change after mani- be attributed, however, to the lesser range of cadences
pulating inertial properties of the lower extremity, but used in this study. However, the data presented here still
increases in net joint moments and reaction forces did support that changing movement speed affects the mus-
occur. They supported these findings by suggesting that cular activity timing. We also observed onset and offset
the inertial loading would increase muscular activity, as timing changes due to cadence alterations. Neptune et
seen by Lestienne [12] during a simple elbow flexion– al. [22] hypothesized that EMG muscle burst onset
extension task. The present experiment examined the values must shift earlier in the crank cycle as cadences
effect of inertia in addition to the effect of cadence. Iner- increase in order to develop pedal forces in the same
tial influences were observed in the onset timing, offset relative area of the crank cycle. Although their results
timing, duration, peak magnitudes, as well as peak mag- indicated muscle burst onset timing changes, different
nitude timing. muscles responded differently to the cadence changes,
A greater load influence was observed at greater cad- thus suggesting a coordination change between muscles.
ences; for example, see the duration of RF and peak They observed linear trends shifting the EMG burst
magnitude of VL and TA. However, at even greater cad- onset to earlier in the crank cycle for GM, RF, BF, and
ences, perhaps a more pronounced load effect would be vastus medialis (VAS). Their SOL shifted later in the
seen. Neptune and Herzog [21] noticed negligible crank cycle, GAS exhibited a quadratic trend, and TA
amounts of negative muscular crank torque created at 90 had no trend for EMG onset due to cadence. In this study
rpm, but greater pedaling rates (105 and 120 rpm) pro- similar trends were observed for GM, RF, BF, and VL;
duced substantial negative crank torque and that torque no trend for GAS or SOL; and a linear trend for TA
increased with the increasing pedaling rates. These onset timing due to cadence. Differences between TA,
authors also noticed a similar trend in average pedal GAS, and SOL trends between the studies may be attri-
forces due to the non-muscular (inertial and buted to the use of different onset and offset threshold
gravitational) component. Consequently, if the effects criteria [5,15]. Furthermore, Neptune et al. [22] used a
observed by Neptune and Herzog [21] only became greater range of cadences, from 45–120 rpm, than the
prevalent at cadences over 100 rpm, especially relating range used in this study. Unique to this study, timing
to non-muscular components, perhaps the inertial effects changes were also observed with load manipulations.
on the muscular activity would also be amplified at These timing differences were observed for both mono-
greater cadences. and bi-articular muscles for each of the tested para-
Previous researchers have reported significant proxi- meters. However, the hypothesis that cadence and load
mal to distal differences of muscular activities, manipulations would alter bi-articular muscle func-
especially in mono-articular muscles [13,22], and these tioning was not supported by our results. Suzuki et al.
differences were explained by the greater inertial proper- [24] reported that BF and RF began to exhibit double
ties proximally in the lower extremities [13]. Conse- bursting patterns as subjects approached maximum peda-
quently, the muscles acting on the knee and ankle joints ling velocity. This corresponded to contributions of mus-
reacted to the lesser inertial properties of the leg and cle activation at the hip and knee. This double-bursting
foot, respectively. Our results provided support for these activity was not observed in the current study. Instead,
suggestions. Changes in timing and magnitude between the bi-articular muscles exhibited one burst but remained
mono- and bi-articular antagonists do indicate a proxi- active so as to act at both the traversed joints.
mal to distal order of load effects when observed as a The present results demonstrated that muscles
whole. More specifically, when examining individual responded to cadence and load manipulations differently,
muscle responses, including both the significant differ- so muscular coordination changes were observed. If the
ences and significant trends, EMG burst offset timing antagonistic muscles reacted differently as cadence or
clearly demonstrated a proximal to distal hierarchy of loads were increased (e.g. antagonist 1 duration
change. decreases with increased cadence and antagonist 2 dur-
Mono-articular muscle timing changes due to cadence ation does not change or increases), then a coordination
and load manipulations were observed in the present change between the muscles was said to have occurred.
study. Peak timing shifts due to cadence in this experi- These coordination changes were observed in at least
ment are similar to those reported by Marsh and Martin one antagonistic muscle pair in each of the five tested
[16]. Those authors reported significant timing differ- parameters after both cadence and load changes. Further-
ences due to cadence in VL, RF, BF, GAS, and SOL, more, examining Fig. 4a reveals distinct differences
and linear trends for the peak magnitude to occur earlier regarding the ankle musculature reaction to cadence
in the crank cycle as cadence increased from 50–110 compared with the reactions of the hip and knee muscu-
rpm for each of the muscles except RF. The only dis- lature. Here the muscles that act only on the hip and/or
crepancy between results of this study and those of knee (GM, RF, BF, and VL) all display similar trends
Marsh and Martin [16] is that a difference in RF peak for the onset and offset timing to decrease as cadence
190 B.S. Baum, L. Li / Journal of Electromyography and Kinesiology 13 (2003) 181–190
increased. However, when examining the ankle muscu- [14] L. Li, G.E. Caldwell, The effect of cycling cadences on the coor-
lature (TA, GAS, and SOL), those clear trends are no dination of mono- and bi-articular muscles, in: Proceedings of
the Sixteenth Southern Biomedical Engineering Conference,
longer apparent and these values sometimes increased Biloxi, MS, 1997.
with cadence. These differences illustrated that the coor- [15] L. Li, G.E. Caldwell, Coefficient of cross correlation and the time
dination between these muscle groups has changed. domain correspondence, J. Electromyogr. Kinesiol. 9 (1999)
385–389.
[16] A.P. Marsh, P.E. Martin, The relationship between cadence and
lower extremity EMG in cyclists and noncyclists, Med. Sci.
5. Conclusion Sports Exerc. 27 (1995) 217–225.
[17] P.E. Martin, P.R. Cavanagh, Segment interactions within the
In conclusion, our results indicate that changing swing leg during unloaded and loaded running, J. Biomech. 23
motion speed and altering the inertia of the thigh will (1990) 529–536.
[18] S.J. Mattes, P.E. Martin, T.D. Royer, Walking symmetry and
affect the muscular activity and coordination in the lower energy cost in persons with unilateral transtibial amputations:
extremity during cycling. Coordination changes were matching prosthetic and intact limb inertial properties, Arch.
observed after both cadence and load changes. Cadence Phys. Med. Rehabil. 81 (2000) 561–568.
had greater effects proximally than distally for onset tim- [19] N. McCartney, J.F. Heigenhauser, N.L. Jones, Power output and
ing, offset timing, and some antagonist pair coordination fatigue of human muscle in maximal cycling exercise, J. Appl.
Physiol. 55 (1983) 218–224.
changes, while greater load effects were observed proxi- [20] M.J. Myers, K. Steudel, Effect of limb mass and its distribution
mally for peak EMG offset and the antagonist pair coor- on the energetic cost of running, J. Exp. Biol. 116 (1985) 336–
dination changes as a whole. Mono-articular discrete 373.
event timing shifts due to loads for VL and SOL were [21] R.R. Neptune, W. Herzog, The association between negative
observed. muscle work and pedaling rate, J. Biomech. 32 (1999) 1021–
1026.
[22] R.R. Neptune, S.A. Kautz, M.L. Hull, The effect of pedaling rate
on coordination in cycling, J. Biomech. 30 (1997) 1051–1058.
References [23] R. Redfield, M.L. Hull, On the relation between joint moments
and pedalling rates at constant power in bicycling, J. Biomech.
[1] D.A. Brown, S.A. Kautz, C.A. Dairaghi, Muscle activity patterns 19 (1986) 317–329.
altered during pedaling at different body orientations, J. Biomech. [24] S. Suzuki, S. Watanabe, S. Homma, EMG activity and kinematics
29 (1996) 1349–1356. of human cycling movements at different constant velocities,
[2] Y.H. Chang, H.W.C. Huang, C.M. Hamerski, R. Kram, The inde- Brain Res. 240 (1982) 245–258.
pendent effects of gravity and inertia on running mechanics, J. [25] D.A. Winter, Biomechanics and Motor Control of Human Move-
Exp. Biol. 203 (2000) 229–238. ment, 2nd ed., Wiley, New York, 1984.
[3] W.J. Dempster, Space requirements of the seated operator,
Brian Baum received his Bachelor’s degree in
Wright Patterson Air Force Base, WADC-TR-55-159, 1955.
Kinesiology from Louisiana State University in
[4] B.J. Fregly, F.E. Zajac, A state-space analysis of mechanical 1999. In 2001 he earned his Master’s degree in
energy generation, absorption, and transfer during pedaling, J. Biomechanics from the same institution. He cur-
Biomech. 29 (1996) 81–90. rently works as a research kinesiologist in the
[5] P.W. Hodges, B.H. Bui, A comparison of computer-based Department of Orthopaedics at Baylor Univer-
methods for the determination of onset muscle contraction using sity Medical Center in Dallas, Texas.
electromyography, Electroencephal. Clin. Neurophysiol. 101
(1996) 511–519.
[6] M.L. Hull, H. Gonzalez, Bivariate optimization of pedalling rate
and crank arm length in cycling, J. Biomech. 21 (1988) 839–849.
[7] M.L. Hull, M. Jorge, A method for biomechanical analysis of
bicycle pedalling, J. Biomech. 18 (1985) 631–644. Li Li, PhD, received his BS degree in Physics
[8] G.J. van Ingen Schenau, P.J.M. Boots, G. de Groot, R.J. Snack- from Peking University (1982), and his PhD in
ers, W.W.L.M. van Woensel, The constrained control of force Biomechanics from the University of Massachu-
and position in multi-joint movements, Neuroscience 46 (1992) setts (1999). Dr Li currently serves as an assist-
197–207. ant professor in the Department of Kinesiology
[9] G.J. van Ingen Schenau, C.A. Pratt, J.M. Macpherson, Differen- at Louisiana State University, Baton Rouge. Dr
Li is a Research Consortium Fellow of Amer-
tial use and control of mono- and bi-articular muscles, Hum.
ican Alliance for Health, Physical Education,
Mov. Sci. 13 (1994) 495–517. Recreation and Dance. He is on the editorial
[10] M. Jorge, M.L. Hull, Analysis of EMG measurements during board of the Journal of Electromyography and
bicycle pedalling, J. Biomech. 19 (1986) 683–694. Kinesiology.
[11] S.A. Kautz, M.L. Hull, A theoretical basis for interpreting the
force applied to the pedal in cycling, J. Biomech. 26 (1993)
155–165.
[12] F. Lestienne, Effects of inertial load and velocity on the braking
process of voluntary limb movements, Exp. Brain Res. 35 (1979)
407–418.
[13] L. Li, The influence of the inertial properties of the human body:
cycling at different pedaling speeds, Doctoral dissertation, Uni-
versity of Massachusetts, Amherst, 1999.