Effects of maximal oxygen uptake test and prolonged cycle ergometer exercise

on the quiet standing control

Roger Gomes Tavares Mello
a,b
, Liliam Fernandes de Oliveira
c
, Jurandir Nadal
a,
*
a
Biomedical Engineering Program, COPPE, Federal University of Rio de Janeiro, Brazil
b
Physical Education and Sport Department, Naval School, Brazilian Navy, Brazil
c
Biomechanics Laboratory, EEFD, Federal University of Rio de Janeiro, Brazil
1. Introduction
When a subject stays in quiet standing, his body shows
spontaneous sway that can be monitored by stabilometry. The
center of pressure (COP) is the measured variable, whose dynamics
represents the effects of the central control, in response to afferent
information from visual, proprioceptive and vestibular systems.
Recently, Baratto et al. [1] proposed a body sway control model
based on stabilograms. Three groups of subjects were used to
evaluate this model, which extracts three structural parameters
froma sway density curve (SDC). The discriminative power of each
parameter was evaluated according to the visual condition (eyes
open and closed) and pathology effects. The SDC is constructed by
counting the number of consecutive samples of the COP trajectory
that appears inside a moving circle with 2.5 mm radius, showing a
sequence of peaks and valleys. The peaks correspond to large data
clusters and reect epochs where the COP presents reduced
oscillations, being interpreted as points in which the central
nervous system (CNS) commands are effective [1]. At these points,
the ankle torque and associated motor commands are thus
relatively stable. On the other hand, valleys correspond to time
periods of fast body sway, at which the ankle torque rapidly shifts
from one stable condition to another [1]. Either pathological or
experimental changes, as visual condition or muscle fatigue, are
expected to affect the balance control, therefore changing the
model parameters.
The pioneering evidences about quiet standing point out the
triceps surae muscle as the most important postural muscle [2].
Therefore, the fatigue of this muscle signicantly decreases the
quiet standing control, at both bipedal [3] and unipedal conditions
[4]. Nardone et al. [5] showed that cycle ergometer exercises do not
increase body sways even after exercise intensities above
anaerobic threshold. Since the quadriceps is the main muscle
involved in cycle ergometer exercises, these ndings suggest that
this muscle is not important in quiet standing control [6]. However,
if the same exercise is performed at maximal intensity, the
decrease of the pH can negatively modify neurotransmission and
force production [7]. Besides, the moderate and prolonged exercise
could increase serotonine level, which is responsible for torpor and
Gait & Posture 32 (2010) 220225
A R T I C L E I N F O
Article history:
Received 16 July 2009
Received in revised form 20 April 2010
Accepted 26 April 2010
Keywords:
Body sway control
Stabilometry
Sway density curve
Central fatigue
Peripheral fatigue
A B S T R A C T
This work aims at testing the inuence of peripheral and central fatigue, after maximal oxygen uptake
test (Test1) and prolonged (Test2) cycle ergometer exercises, respectively, on sway density curve (SDC)
parameters of postural control. Sixteen healthy male subjects were submitted to stabilometric tests,
before and after the exercises. The Test1 was started at 12.5 W, with 12.5 W/min increments and 50 rpm
cadence until exhaustion. From the respiratory gas exchange signals, the rst ventilatory threshold was
obtained by the v-slope method. After a minimumof 72 h, the subjects performed the Test2 for 60 min, at
a power output corresponding to 70% of such threshold. Before and just after these exercises, a set of 10
stabilometric trials of 50 s was performed, alternating the eyes open and closed conditions, intercalated
by a 10 s resting period. The resulting signals were used to obtain the SDC. The Test1 caused decrease of
the mean of peaks duration in SDC (p < 0.05), decreasing the stability level, with small changes in the
rates of central nervous system (CNS) and muscular torque controls. Conversely, Test2 increased the
mean of time intervals between peaks in SDC (p < 0.05), thus decreasing the CNS commands rate with
minor changes in the stability level. Visual privation had a greater effect on body sway than these
exercises, which were applied to muscles that are not the main actuators in body sway control.
Concluding, this study allowed discriminating the effects of exercise intensities on body sway control.
2010 Elsevier B.V. All rights reserved.
* Corresponding author at: Biomedical Engineering Program COPPE/UFRJ,
Federal University of Rio de Janeiro, P.O. Box 68510, 21941-972 - Rio de Janeiro RJ,
Brazil. Tel.: +55 21 2562 8577; fax: +55 21 2562 8591.
E-mail addresses: roger@peb.ufrj.br (R.G.T. Mello), liliam@bridge.com.br
(L.F. de Oliveira), jn@peb.ufrj.br (J. Nadal).
Contents lists available at ScienceDirect
Gait & Posture
j our nal homepage: www. el sevi er . com/ l ocat e/ gai t post
0966-6362/$ see front matter 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.gaitpost.2010.04.016
lethargy [8]. These factors could affect the body sway control due
to systemic changes, independently of the fatigued muscle. The
inuence of localized muscle and whole-body fatigue on body
sway has being investigated [35,9,10], however the role of
peripheral and central fatigue mechanisms is still to be contrasted.
This work aims at discriminating the effects of peripheral and
central fatigue, after maximal oxygen uptake test and prolonged
cycle ergometer exercises, respectively, on body sway control
based on SDC parameters.
2. Materials and methods
2.1. Subjects
Sixteen healthy male subjects were recruited, with age 2032 years, body mass
72.3 11.6 kg (mean standard deviation), and height 1.73 0.10 m. All subjects
were undergraduate students in Physical Education, presenting good physical
conditioning without neurological or orthopedic diseases, and low coronary risk.
Subjects signed a written informed consent according to the protocol approved by the
Local Research Ethics Committee [CAAE0013.0.197.000-05].
2.2. Instrumentation
The instrumentation consisted in a vertical force platform AccuSway Balance
Clinic (AMTI, USA), a mechanically braked cycle ergometer (Monark, Sweden), an
electromyograph (Biovision, Germany), a blood lactate analyzer Accusport (Roche,
Sweden), and a respiratory gas exchange analyzer VO2000 (Medgraphics, USA). The
force platform was previously calibrated and data was sampled at 100 Hz. A
medium ow pneumotachograph was employed, with range 10120 L/min. O
2
and
CO
2
sensors were also calibrated by a gas mixture with 12% of O
2
and 5% of CO
2
,
showing errors below 1%.
2.3. Balance assessment
Stabilometric tests were executed at two different days (Fig. 1). At the
beginning of each session, just after calibration and familiarization, each
individual stood on the force platform, with barefoot, feet comfortably together
and arms relaxed. It was performed ve control trials with eyes open (EO), looking
at a target placed at 50 cmahead at eye level, intercalated by ve with eyes closed
(EC), respecting a 10 s resting period on sitting position between the 50 s trials.
Just after performing cycle ergometer exercises, the subjects repeated these trials.
The total time period was limited to about 10 min to avoid the full recovery after
effort, estimated in 15 min [5].
In the rst day, each subject performed a maximal oxygen uptake test with
cadence 50 rpm (Test1), starting at a power output of 12.5 W, with 12.5 W/min
work increments. During exercise, it was recorded the surface electromyography
(SEMG) of right vastus lateralis muscle, by a differential amplier (gain 2500,
120 dB common mode rejection, 1 TV input impedance, band limited between 10
and 1 kHz), in series with an anti-aliasing Butterworth lter, 2nd order with 500 Hz
cutoff frequency. Two Ag/AgCl electrodes MEDI-TRACE 2000 (Kendall, USA) were
xed on the muscle according to SENIAN[11]. SEMGwas digitized at 2 kHz by a data
acquisition board DacCard 6024E (National Instruments, USA), with 12 bits
resolution, dynamic range of 5 V, and a software built using Labview 5.0 (National
Instruments, USA). Ventilatory gas exchange signals were also monitored for the
identication of the rst ventilatory threshold (VT1) by automatic v-slope method
[12]. Subjects used sport shoes to avoid discomfort. After exhaustion, subjects rapidly
removed shoes and pneumotachograph to the post-exercise stabilometric test.
The second session was performed with minimum interval of 72 h, when each
subject was submitted to a moderate and prolonged cycle ergometer exercise
(Test2), with the load corresponding to 70% of his VT1, with a 50 rpm cadence for
60 min.
Both exercises were performed at the same cadence to avoid interference in the
fatigue dynamics [13].
During stabilometric trials, the modied lead II ECG was recorded using the
electromyograph. After QRS detection [14], the R waves amplitudes and time of
occurrences were respectively used to estimate respiratory frequency and heart rate.
The respiratory frequency was estimated by spectral analysis of the R-wave peak
values, after cubic spline interpolation and resampling to 5 Hz. This signal was thus
ltered in direct and reverse order by a high-pass Butterworth lter, 2nd order with
cutoff frequency 0.15 Hz. For each stabilometric trial, this power spectrum was
estimated by Welch periodogram, with 50% overlapping, using Hanning windows
and data segments of 16 s (80 samples) with zero padding to 128 samples. The
maximum spectral value corresponds to the respiratory frequency, with 0.039 Hz
resolution, or 2.35 breaths per minute.
The lactatemia was measured immediately after each exercise.
2.4. Identication of VT1
Ventilatory gas exchange signals were measured breath-by-breath and recorded
by averages of each three breaths, using the Aerograph software (Medgraphics,
USA). This software provided the ventilation minute (

V
E
),

V
O
and

V
CO
signals,
required by the v-slope method [12]. Signals were interpolated by cubic spline to
100 samples/s, and smoothed by a moving average lter with 50 samples applied in
direct and reverse directions to avoid phase distortion. Following, two straight lines
were tted to

V
E

V
O
function, iteratively, being stored the sumsquared error (SSE).
The intersection of straight lines that minimized SSE corresponded to second
ventilatory threshold (VT2) if the slope increased more than 15% after intersection.
The same iterative tting was applied to

V
CO

V
O
function, from its beginning until
VT2, where the straight lines intersection represented VT1 when the slope
increased more than 10%.
2.5. Monitoring muscular fatigue by SEMG
The SEMG was pre-processed for delimitating the frequency band to 10400 Hz
and rejecting the 60 Hz main noise and harmonics [15], and thus decimated to 1 kHz.
During maximal oxygen uptake test, muscular fatigue was conrmed by a
nonlinear increase of the SEMG root mean square values (RMS, 20 ms window).
After the identication of the segments of myoelectric activity in the SEMG, the
time series of the mean RMS from each myoelectric activity epoch was analyzed by
the algorithm similar to the one adopted v-slope [16]. The time series was
smoothed by a moving average lter (25 samples) in direct and reverse directions
and two regression lines were thus tted to the initial and terminal portions of this
time series, with the SSE being stored during iterative changes of the intercept
between these straight lines. In this case, the inexion point represents the fast
twitch muscular bers activation [17].
2.6. Sway density curves
To avoid the initial transient in stabilometric signals, only the last 30 s of the each
trial were considered in the analysis. After mean removal and detrend, stabilograms
referring to each one of eight conditions (before and after the two exercises, either

Fig. 1. Experimental setup indicating familiarization, 10 stabilometric trials before and after 2 different exercises at cycle ergometer, applied 72 h apart. The trials were
performed alternatively with eyes open (EO) and eyes closed (EC). The prolonged exercise was always performed in the second day because the applied load depends on the
rst ventilatory threshold (VT1), identied in maximal oxygen uptake test.
R.G.T. Mello et al. / Gait & Posture 32 (2010) 220225 221
with EO and EC) were analyzed. For each condition, the ve stabilograms were
ltered in direct and reverse direction by a low-pass Butterworth lter, 2nd order
and cutoff frequency 12.5 Hz, and decimated to 50 Hz. Each value of the SDC was
thus obtained by counting consecutive samples of stabilograms that remained
inside a circle with radius 2.5 mm, around each sample. Then, the SDC were ltered
by a low-pass Butterworth lter, 4th order and cutoff frequency 2.5 Hz, and
multiplied by the sampling period to be dimensioned in seconds [1].
The mean duration of the peaks (MP) and the mean time interval between two
consecutive peaks (MT) were extracted from the SDC. The mean distance between
two consecutive peaks (MD) was measured directly from the statokinesigram. The
effects of learning on quiet standing control were evaluated by the mean velocity
(mVel) and the elliptic sway area.
2.7. Statistical analysis
A three-way single-group repeated-measures ANOVA was used to assess
changes in body sway control: the rst factor was the stance trials before versus
after exercise, the second was the visual condition, and the third one the exercise
(Test1 versus Test2). A two-way repeated-measures ANOVA was used to assess the
effect of the fatigue and exercise type on respiratory frequency and heart rate. Both
ANOVA were followed by post hoc Tukey tests. Students paired one-tailed t test
was employed to compare blood lactate after the two exercises with the value of
4 mmol/L, considered as the onset value of blood lactate accumulation [18]. The
signicance level was always 0.05.
3. Results
Whilst all subjects showed lactatemia greater than 4 mmol/L
after Test1 (p < 10
3
), in Test2 the subjects showed lactate lower
than 4 mmol/L (p < 10
3
).
All subjects showed nonlinear increases in SEMG RMS values
during Test1 (Fig. 2A). Consequently, the fatigable fast twitch
muscular bers were activated and this assures the muscular
fatigue by anaerobic metabolites accumulation.
The heart rate showed interaction between fatigue and exercise
type (p < 10
3
), indicating signicant increase only after Test1
(Fig. 2B). Although the respiratory frequency did not show
interaction between fatigue and exercise type (p = 0.37)
(Fig. 2C), the main effects of these conditions were signicant
(p < 10
3
and p = 0.03, respectively)
The statokinesigram shows the existence of the time clusters
(Fig. 3A) that correspond to epochs at which CNS commands are
more stable [1]. When computing the consecutive points inside the
moving circle, as shown in 2D representation (Fig. 3B), the
corresponding SDC is obtained (Fig. 3C). The Test1 caused a
decrease in MP (Fig. 3D).
Three-way ANOVA showed signicant interactions between
fatigue condition, and exercise type (p < 0.03 for MP and MT;
p < 10
3
for mVel) (Table 1). MP decreased more after Test1, whist
MT and mVel increased more after Test2 and Test1, respectively
(Fig. 4AC). mVel also showed interaction among fatigue, visual
condition and exercise type (p < 0.03) (Table 1). Therefore, mVel
increased signicantly either with EO and EC only after Test1. MD
and mVel also showed interaction between visual condition and
exercise order (p < 0.02 and p < 0.002, respectively). Additionally,
Test2 showed mVel signicantly lower than Test1 only with EC
(p < 0.05, Fig. 4D and E).
The ANOVA also showed main effects of fatigue on MP
(p < 10
3
), MT (p < 0.02), MD (p < 0.003), mVel (p < 10
3
), and
area (p < 0.003). All stabilometric parameters were strongly
inuenced by visual condition (p < 10
3
). The signicant increase
of MT in Test2 (p < 0.04), without change of the MD (p = 0.56),
caused a signicant decrease of mVel (p < 0.007). However, area
signicantly increased in Test2 (p < 0.001, Table 1).

Fig. 2. (A) Typical example of the mean RMS (mRMS) signal with inexion point, which represents the electromyographic threshold that indicates the onset of fast twitch
muscular bers activation. (B) Heart rate and (C) respiratory frequency (mean standard error) before (values 110) and after (1120) exercises. Whilst heart rate signicantly
increased only after maximal exercise and maintained until the last stabilometric trial, respiratory frequency increased also only after maximal exercise but recovered in the third
stabilometric trial. *Signicantly different from values before exercise (p < 0.05).
R.G.T. Mello et al. / Gait & Posture 32 (2010) 220225 222

Fig. 3. (A) Typical example of three-dimensional statokinesigram visualization, showing local clusters, which represent instants of greater postural stability; (B)
statokinesigram with the superimposed moving circle with radius 2.5 mm, used for calculating the sway density curve; and sway density curves obtained (C) before and (D)
after the maximal oxygen uptake test with eyes closed. Peaks (*) represent stability moments and the valleys between peaks showshifts in the center of pressure due to CNS
control commands. The mean duration of peaks (MP) is smaller in D (0.79 s) than C (0.81 s). All data are from subject #1.
}
}
}
}
}
}
}
}
}
}
}
}
}
}
}
}
}
}
}
} }
}
}
}
} }
} } } } }
Table 1
Values of SDC and traditional stabilometric parameters before and after each exercise, for both visual conditions (meanstandard error).
MP (s) MT (s) MD (mm) mVel (mm/s) Area (mm
2
)
BE-ME-EO 1.080.04 0.5820.006 3.620.13 14.730.33
a
55.144.63
AF-ME-EO 1.010.05 0.5790.005 3.730.14 15.830.44
a
*
59.174.56
BE-ME-EC 0.700.04 0.5990.005 6.210.39 19.680.93
a
88.914.66
AF-ME-EC 0.630.03
*
0.6030.004 6.700.33
*
21.30 0.76
a
*
101.396.08
BE-PE-EO 1.080.05 0.5840.005 3.790.20 15.080.50 61.585.18
AF-PE-EO 1.030.05 0.5970.005 4.030.19 15.600.40 68.496.39
BE-PE-EC 0.680.04 0.5970.003 6.270.34 19.770.86 100.666.57
AF-PE-EC 0.700.04 0.6080.006 6.360.37 19.670.69 107.259.59
BE-EO 1.080.03 0.5830.004 3.710.12 14.910.30 58.363.47
AF-EO 1.020.03
*
0.5880.004 3.880.12 15.720.29
*
63.833.95
BE-EC 0.690.03 0.5980.003 6.240.26 19.730.63 94.784.10
AF-EC 0.660.03 0.6060.004 6.530.25
*
20.490.53
*
104.325.61
*
BE-ME 0.890.02
a,
0.5900.003
a
4.920.15 17.210.31
a
72.032.59
AF-ME 0.820.02
a
*
0.5910.003
a
*
5.220.16
*
18.570.32
a+
80.283.22
BE-PE 0.880.02
a,
0.5900.003
a
5.030.15 17.430.31
a *
81.123.38
AF-PE 0.860.02
a,
0.6030.003
a+
5.190.15 17.640.26
a
87.873.86
ME-EO 1.040.03
+
0.5800.004 3.680.09
a
15.280.29
a
57.163.22
PE-EO 1.060.04
+
*
0.5900.004
*
3.910.14
a
*
15.340.32
a
65.034.09
ME-EC 0.660.03 0.6010.003
+
6.450.26
a+
20.490.61
a+ *
95.153.93
+
*
PE-EC 0.690.03 0.6020.004
+
6.310.25
a+
19.720.54
a
103.955.75
+
BE 0.880.02 0.5900.002 4.970.10 17.320.22 76.572.14
AF 0.840.02
*
0.5970.002
*
5.210.11
*
18.100.21
*
84.082.52
*
ME 0.850.02 0.5910.002 5.070.11 17.890.22 76.152.08
PE 0.870.02 0.5960.002
*
5.110.10 17.530.20
*
84.492.57
*
EO 1.050.02 0.5850.003 3.790.08 15.310.21 61.102.63
EC 0.680.02
*
0.6020.002
*
6.380.18
*
20.110.41
*
99.553.50
*
Abbreviations: BE, before exercise; AF, after exercise; ME, maximal oxygen uptake test exercise (Test1); PE, prolonged exercise (Test2); EO, eyes open; EC, eyes closed.
+
and

, signicantly greater or lower than other conditions inside the respective brace.
a
Interaction among conditions.
*
Signicant difference (p<0.05).
R.G.T. Mello et al. / Gait & Posture 32 (2010) 220225 223
4. Discussion
The present study showed that SDC parameters were sensitive
to cycle ergometer exercises, agreeing with Gauchard et al. [19],
who also observed increased body sway after maximal-effort
cycling exercise. Nardone et al. [5] showed decreases in COP area
and sway path after cycle ergometer exercise. These results,
opposite to the present study, are partially explained by protocol
differences. Whereas those authors performed submaximal tests
(25 min) with intensities above anaerobic threshold, in the present
work the Test1 assured that subjects reached the quadriceps
exhaustion. Even at Test2, the exercises series was sustained for
60 min, to assure some level of central fatigue.
In the present study, the most recruited muscular group was the
quadriceps, whose concentric contractions are expected to cause a
smaller proprioceptive disturbance than eccentric contractions
[20], and are not primarily related to quiet standing control [6]. The
cycle ergometer exercises generated different changes in SDC
parameters: MP reduced more after Test1, whilst MT and mVel
increased more after Test2 and Test1, respectively. The increased
MT after Test2 indicates decreased rate of CNS commands [1],
reecting some level of central fatigue, whilst the changes in MP
and mVel reect a decreased stability level of the postural control
system [1], not related to CNS. Thus, the use of SDC allowed
discriminating changes in body sway related to different exercise
intensities.
The adopted exercise protocols caused different physiological
responses. Test1 showed lactatemia signicantly above 4 mmol/L,
the threshold of anaerobic metabolism and the fast twitch
muscular bers activation, as conrmed by the nonlinear increase
in RMS SEMG values, whilst Test2 did not increase the lactatemia
neither increased the SEMG intensity (not shown).
The predominant physiological changes can explain the
different changes of SDC parameters. In Test1 the lactic anaerobic
metabolism is predominant, causing mainly peripheral fatigue
[21]. Among the principal causes of peripheral fatigue, there is the
accumulation of H
+
, due to ATP hydrolysis and the reduction of
NAD
+
coenzyme during glycolysis [7]. The decreased pH: (1)
depresses the muscular function by inhibiting the Ca
++
release by
sarcoplasmic reticulum, interfering on binding capacity of Ca
++
on
troponin [22]; (2) inhibits the activity of the enzyme rate-limiting
phosphofructokinase (PFK), a key enzyme of the anaerobic
glycolysis [23,24]; and (3) decreases the conduction velocity of
the myoelectric signal [25], which may affect the proprioceptive
control. Consequently, the changes in MP and mVel may be
reecting the decreased capability in force production and delays
in myoelectric conduction and neurotransmission. However, it is
not discharged the effects of increased heart and respiratory
frequencies, which are also related to decreased pHand may affect
the stabilometric parameters [26].
The central fatigue happens with larger probability in tasks
with force of low magnitude and long duration [8], as the protocol
of Test2. Prolonged muscular efforts are associated with an
increase of the plasmatic serotonin that is related to motivation
and lethargy, because the serotoninergic neurons are disperse
thoroughly for several areas related with these sensations, as the
limbic system [8]. This hypothesis is reinforced by Huffman et al.
[27], which showed that the free tryptophan, precursor of the
serotonin, has signicantly increased concentration after 75 min of
running in 60% of the aerobic power. Besides, the low ratio
between dopamine and serotonin during the prolonged exercise
might have reduced the motivation [28]. Even a small increase of
the ammonemia might have happened during Test2 [29]. These
factors contributes to central fatigue, thus may negatively affect
the rate of CNS commands on body sway control, as indicated by
increased MT parameter.
The present protocol does not allow discharging small changes
in mVel in Test2 and EC condition related to learning effects on

Fig. 4. (A) MP, (B) MT and (C) mVel before and after maximal oxygen uptake test (Test1) and prolonged exercise (Test2); parameters showed interaction between fatigue
condition and exercise type: MP decreased more after Test1, MT and mVel increased more after Test2 and Test1, respectively. (D) MD and (E) mVel for the rst and second
days (Test1 and Test2, respectively) with EO and EC; these parameters showed interaction between exercise order and visual condition: only with eyes closed, MD tends to
decrease and mVel presents signicant decrease in the second day. *Signicantly different fromthe other condition (p < 0.05). **Signicantly greater than prolonged exercise
with EC (p < 0.05).
R.G.T. Mello et al. / Gait & Posture 32 (2010) 220225 224
body sway control [30], however an increased elliptic area
occurred in the second day.
Finally, the SDC parameters suffered greater effect of visual
privation than exercise, corroborating the dominance of visual
feedback over proprioception [5].
As a summary, the different cycle ergometer exercise intensities
cause different changes in the quiet standing control, which are
measurable by the sway density curves. Maximal oxygen uptake
test, related to peripheral fatigue, showed a major decrease in MP
parameter, indicating decreased stability. After the prolonged
aerobic effort employed to induce central fatigue, conversely, a MT
increase suggests a decreased control rate, with minor changes in
the stability.
Acknowledgements
This study was partially supported by the Brazilian Research
Council (CNPq) and Jose Bonifa cio University Foundation (FUJB).
The rst author wishes also to thank the scholarships given by
CAPES Foundation, and to Dr. P. Morasso by the scripts for SDC
calculation.
Conict of interest statement
Neither of the authors has any nancial or personal relationship
that could inappropriately inuence the work submitted for
publication.
References
[1] Baratto L, Morasso PG, Re C, Spada G. A new look at posturographic analysis in
the clinical context: sway-density vs. other parameterization techniques.
Motor Control 2002;6:24670.
[2] Basmajian JV, Bentzon JW. An electromyographic study of certain muscles of
the leg and foot in the standing position. Surg Gynecol Obstet 1954;98:6626.
[3] Mello RGT, Oliveira LF, Nadal J. Anticipation mechanism in body sway control
and effect of muscle fatigue. J Electromyogr Kinesiol 2007;17:73946.
[4] Vuillerme N, Nougier V, Prieur JM. Can vision compensate for a lower limbs
muscular fatigue for controlling posture in humans? Neurosci Lett 2001;308:
1036.
[5] Nardone A, Tarantola J, Giordano A, Schieppati M. Fatigue effects on body
balance. Electroencephalogr Clin Neurophysiol 1997;105:30920.
[6] Joseph J, Nightingale A. Electromyography of muscles of posture: thigh mus-
cles in males. J Physiol 1954;126:815.
[7] Robergs RA, Ghiasvand F, Parker D. Biochemistry of exercise-induced meta-
bolic acidosis. Am J Physiol Regul Integr Comp Physiol 2004;287:R50216.
[8] Blomstrand E. Amino acids and central fatigue. Amino Acids 2001;20:2534.
[9] Nelson JK, Johnson BL. Effects of local and general fatigue on static balance.
Percept Mot Skills 1973;37:6158.
[10] Springer BK, Pincivero DM. The effects of localized muscle and whole-body
fatigue on single-leg balance between healthy men and women. Gait Posture
2009;30:504.
[11] Hermens HJ, Freriks B, Merletti R, Stegeman D, Blok J, Rau G, et al. SENIAM 8:
European recommendations for surface electromyography. Enschede: Roes-
singh Research and Development; 1999.
[12] Beaver WL, Wasserman K, Whipp BJ. A new method for detecting anaerobic
threshold by gas exchange. J Appl Physiol 1986;60:20207.
[13] Gardner AS, Martin DT, Jenkins DG, Dyer I, Van Eiden J, Barras M, et al. Velocity-
specic fatigue: quantifying fatigue during variable velocity cycling. Med Sci
Sports Exerc 2009;41:90411.
[14] Pan J, Tompkins WJ. A real-time QRS detection algorithm. IEEE Trans Biomed
Eng 1985;32:2306.
[15] Mello RGT, Oliveira LF, Nadal J. Digital Butterworth lter for subtracting noise
from low magnitude surface electromyogram. Comput Methods Programs
Biomed 2007;87:2835.
[16] Mello RGT, Oliveira LF, Nadal J. Detection of the anaerobic threshold
by surface electromyography. Conf Proc IEEE Eng Med Biol Soc 2006;1:
618992.
[17] Hug F, Faucher M, Kipson N, Jammes Y. EMG signs of neuromuscular fatigue
related to the ventilatory threshold during cycling exercise. Clin Physiol Funct
Imaging 2003;23:20814.
[18] Heck H, Mader A, Hess G, Mu cke S, Mu ller R, Hollmann W. Justication of the
4 mmol/l lactate threshold. Int J Sports Med 1985;6:11730.
[19] Gauchard GC, Gangloff P, Vouriot A, Mallie JP, Perrin PP. Effects of exercise-
induced fatigue with and without hydration on static postural control in adult
human subjects. Int J Neurosci 2002;112:1191206.
[20] Givoni NJ, Pham T, Allen TJ, Proske U. The effect of quadriceps muscle fatigue
on position matching at the knee. J Physiol 2007;584.1:1119.
[21] Gandevia SC. Spinal and supraspinal factors in human muscle fatigue. Physiol
Rev 2001;81:172589.
[22] Nakamura Y, Schwartz S. The inuence of hydrogen ion concentration on
calcium binding and release by skeletal muscle sarcoplasmic reticulum. J Gen
Physiol 1972;59:2232.
[23] Trivedi B, Danforth WH. Effect of pH on the kinetics of frog muscle phospho-
fructokinase. J Biol Chem 1966;241:41102.
[24] Spriet LL, Soderlund K, Bergstrom M, Hultman E. Skeletal muscle glycogenoly-
sis, glycolysis, and pH during electrical stimulation in men. J Appl Physiol
1987;62:61621.
[25] Brody LR, Pollock MT, Roy SH, De Luca CJ, Celli B. pH-induced effects on median
frequency and conduction velocity of the myoelectric signal. J Appl Physiol
1991;71:187885.
[26] Conforto S, Schmid M, Camomilla V, DAlessio T, Cappozzo A. Hemodynamics
as a possible internal mechanical disturbance to balance. Gait Posture
2001;14:2835.
[27] Huffman DM, Altena TS, Mawhinney TP, Thomas TR. Effect of n-3 fatty
acids on free tryptophan and exercise fatigue. Eur J Appl Physiol 2004;92:
58491.
[28] Nybo L, Secher NH. Cerebral perturbations provoked by prolonged exercise.
Prog Neurobiol 2004;72:22361.
[29] Banister EW, Cameron BJC. Exercise-induced hyperammonemia: peripheral
and central effects. Int J Sports Med 1990;11:S12942.
[30] Tarantola J, Nardone A, Tacchini E, Schieppati M. Human stance stability
improves with the repetition of the task: effect of foot position and visual
condition. Neurosci Lett 1997;228:758.
R.G.T. Mello et al. / Gait & Posture 32 (2010) 220225 225