Neural adaptations to fatigue: implications for muscle strength and training - implications for muscle strength ...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Neural adaptations to fatigue: implications
for muscle strength and training
DAVID A. GABRIEL, JEFFREY R. BASFORD, and KAI-NAN AN
Biomechanics Laboratory, Brock University, St. Catharines, Ontario, CANADA L2S 3A1; and Departments of Physical
Medicine and Rehabilitation and Orthopedics, Mayo Clinic and Mayo Foundation, Rochester, MN 55901
ABSTRACT
GABRIEL, D. A., J. R. BASFORD, and K.-N. AN. Neural adaptations to fatigue: implications for muscle strength and training. Med.
Sci. Sports Exerc., Vol. 33, No. 8, 2001, pp. 1354 –1360. Purpose: This paper investigates the neural mechanisms responsible for the
increase in strength that occurs during serial isometric contractions. Methods: A three-session design was used. Thirteen subjects (N
⫽ 13) were asked to perform five maximal isometric elbow extension strength trials to serve as baseline. After a 5-min rest, the subjects
were administered a 30-trial fatigue protocol. This process was repeated two more times at 2-wk intervals. Elbow extension torque and
surface electromyography (EMG) of the triceps and biceps brachii were monitored concurrently. The criterion measures were elbow
extension torque, root-mean-square EMG amplitude, and mean power frequency (MPF). Results: Intraclass reliability ranged from
good to excellent. Within each experimental session, the fatigue protocol resulted in a decrease in maximal isometric elbow extension
torque as well as biceps and triceps EMG amplitude and MPF (P ⬍ 0.05). However, the mean of the 30 trials and the magnitude of
the linear decrease in elbow extension torque increased across the three sessions (P ⬍ 0.05). Biceps and triceps EMG amplitude
increased and MPF decreased as the number of sessions increased (P ⬍ 0.05). Conclusions: These findings suggest that the fatigue
protocol served as a training stimulus to down regulate motor-unit firing frequency. Key Words: ELECTROMYOGRAPHY, ELBOW
EXTENSION, ISOMETRIC CONTRACTION, ANTAGONIST COACTIVATION, MOTOR-UNIT ACTIVITY
I
t has been known since the 1960s that strength increases increase with training and are present even when lifting
after a single session of serial isometric contractions loads as light as 10% of their maximum (22,28). The fa-
(17,18). This paper investigates the neural mechanisms tigue-related phenomenon may therefore act as a training
for this phenomenon which was first reported 35 years ago stimulus that with repetition produces a training response
by Kroll (18). Kroll’s protocol consisted of bilateral isomet- (12).
ric wrist flexion strength trials in 20 untrained subjects with The second possible mechanism is based on Basmajian’s
a retest 2 wk later. The contractions were 5 s in duration (2) theory of “progressive inhibition.” This theory suggests
with rest periods of 30 s. The retest session showed an that the nervous system, as it acquires a new motor skill,
improvement relative to the first session in the 20-trial learns to minimize antagonist coactivation and extraneous
mean—a measure of endurance— of about 14% and 12% muscle activity. In this view, optimal isometric strength is
for the right and left wrists, respectively. Additional training obtained when agonist muscle contractions occur without
three times a week over the next 4 wk, however, did not antagonist coactivation (14,19). There is support of this
result in a further increase in strength. concept. For example, the careful work of Carolan and
It is accepted that the increase in strength exhibited by Cafarelli (8) found that biceps femoris EMG activity during
individuals beginning a new activity is due to neural factors knee extension decreased 20% after 8 wk of isometric knee
(16,23). This paper focuses on two possible mechanisms extension strength training whereas knee extension strength
that have some empirical support. The first of these mech- increased 32.5%. The authors emphasized the importance of
anisms involves an increase in synchronization of motor- antagonist coactivation reduction by pointing out that this
unit (MU) firing patterns such as occurs with training in strength increase occurred without a rise in vastus lateralis
normal subjects exerting medium to high levels of force as EMG activity.
well as during sustained isometric contractions (25). Hayes Electromyographic studies of these mechanisms have fo-
(12) theorized that the synchronization occurring with fa- cused on initial increases in maximal isometric strength, not
tiguing isometric contractions is closely related to the syn- fatigue resistance. The purpose of this paper is, therefore, to
chronization observed in trained, high-strength subjects. For extend this research and determine whether MU synchroni-
example, resistance-trained individuals have increased rates zation and/or a reduction in antagonist coactivation is re-
of synchronized electromyographic (EMG) activity which sponsible for the initial increase in resistance to isometric
fatigue seen in untrained individuals. Specifically, we mon-
0195-9131/01/3308-1354/$3.00/0 itored elbow extension torque, the root-mean-square (RMS)
MEDICINE & SCIENCE IN SPORTS & EXERCISE® amplitude, and mean power frequency (MPF) of biceps and
Copyright © 2001 by the American College of Sports Medicine triceps EMG activity in 13 untrained subjects who per-
Submitted for publication March 2000. formed 30 maximal isometric elbow extension contractions
Accepted for publication November 2000. on three test sessions at 2-wk intervals. We expected that the
1354fatigue protocol would increase elbow extension strength- fatigue protocol was sufficient to produce the classic shift in
endurance by: 1) a decrease in biceps RMS EMG amplitude mean power frequency toward the lower end of the spectrum
as evidence of reduced antagonist coactivation, and 2) an that is characteristic of muscle fatigue (1,21,24).
increase in triceps RMS EMG amplitude and a decrease in
the MPF of triceps EMG activity, which reflects an increase Recording Force and EMG Activity
in synchronization of MU firing patterns.
The experimental apparatus has been detailed elsewhere
(10) but will be described here briefly. Subjects were seated
MATERIAL AND METHODS at a table designed to isolate the action of the elbow exten-
sors in an isometric contraction. Adjustable supports main-
Approach to the Problem and tained the shoulder and elbow of the arm being tested in 90°
Experimental Design of flexion in the sagittal plane. A wrist-cuff was attached
The focus of the paper was on the neural adaptation to a proximal to the styloid process, and the forearm was in
fatigue regimen. Training was therefore limited to three neutral pronation and supination. Forces were measured by
sessions to avoid metabolic and hypertrophic adaptations. In a load cell (JR3 Inc., Woodland, CA) mounted between a
the event that any metabolic or hypertrophic adaptations did vertical strut and the wrist-cuff to assure that the application
occur, a 2-wk rest interval between each session was given of force was always perpendicular to the load cell. A belt
to allow for detraining. It was assumed that the neural secured the subject to the testing table to increase stability
adaptations identified in this paper would continue with and minimize extraneous movements.
long-term training. However, the changes in neural control Before electrode placement, the skin surface was shaved
could not be separated from other physiological adaptations with a safety razor, lightly abraded, and cleansed with
that would have occurred with additional training (16). The alcohol to reduce the skin-electrode interface impedance to
RMS and MPF of surface EMG activity were monitored below 5 k⍀. After careful preparation of the skin, bipolar
because these measures can be used to make inferences 6-mm active diameter surface electrodes with a center-to-
about changes in MU recruitment (1,6). However, power center interelectrode distance of 6 mm (4) were prepared
spectral analysis can only be applied to signals that have a with adhesive tape and electrolyte gel and placed directly
constant mean and standard deviation (i.e., stationarity). over the belly of the triceps brachii long head and biceps
This assumption has been shown to be true for constant brachii short head muscles, away from the motor points
torque contractions (7). Thus, we analyzed the constant (26). Skin-electrode input impedance was then measured
torque portion of maximal isometric contractions. Use of with the F-EZM5 impedance meter (Grass-Telefactor, As-
isometric contractions also avoided the potential difficulties tro-Med, Inc., West Warwick, RI) to ensure that it was
associated with EMG and torque measurement during dy- below 5 k⍀. It was never necessary to reapply the elec-
namic contractions (33). trodes. Consistent placement of electrodes across test ses-
sions was accomplished by nontoxic-ink pen markings
Subjects maintained by the subjects.
The differential input characteristics of the first stage
After approval of the Mayo Clinic’s Internal Review amplifier (MA100, Motion Lab Systems, Inc., Baton Rouge,
Board, 13 right-handed healthy women with a normal neu- LA) include a common mode rejection ratio (CMRR) of 100
rological exam and history were recruited from the general dB and a bandwidth of 10 Hz to 5 kHz, which was combined
staff to participate in this study. The participants had diverse with a bandpass filter of 20 –300 Hz at the second stage of
physical characteristics; they ranged in age (23–38 yr), amplification (6). The EMG signal was increased with a
height (162–180 cm), and weight (529 – 891 N). Each sub- fixed gain of 325 at the first stage of amplification. An
ject was verbally acquainted with the experimental proce- additional amplifier allowed the signal to be increased fur-
dures and then asked to read and sign an informed consent ther up to 30 times. The EMG gains were adjusted to ensure
document. All subjects were remunerated for their time. maximal A/D converter resolution by monitoring the signals
on the Computer-Based Oscillograph and Data Acquisition
Measurement Schedule System (CODAS, DATAQ Instruments Inc., Akron, OH)
digital oscilloscope which has a ⫾10 V scale. Voltages from
Testing was completed in the Orthopedic Biomechanics
the load cell amplifier and the EMG system were digitized
Laboratory, which was maintained at a room temperature of
at a sampling rate of 2016 Hz per channel by the CODAS
23°C. There were three identical test sessions separated by
12-bit system on an 80486 66 MHz IBM-AT compatible
2-wk intervals. Each session began with the subjects reading
personal computer (Reason Technologies, Inc., Minneapo-
written standardized instructions. A tape recording then
lis, MN) and stored for later off-line processing.
controlled the procedures that began with the subjects per-
forming five maximal 2-s isometric elbow extension con-
Data Reduction
tractions with 24-s rest periods. After a 5-min rest, subjects
then performed a fatigue protocol that consisted of 30 max- Once collected, the data for each contraction were ex-
imal isometric elbow extension contractions of 2 s with 6 s ported into ASCII format files for data reduction and pro-
between each trial. We observed from pilot data that this cessing using scripts written in MATLAB (The MathWorks,
NEURAL ADAPTATIONS TO FATIGUE Medicine & Science in Sports & Exercise姞 1355FIGURE 1—Representative elbow exten-
sion moment (panel 1), triceps brachii EMG
activity (panel 2), and biceps brachii EMG
activity (panel 3) from one subject.
Inc., Natick, MA). The load cell data was low-pass (100 Hz, repeated measurements (trials) on each subject in each day
3 dB) filtered using a zero phase-lag, fourth-order Butter- constituted a “within-cells” replication of measures. Trials
worth digital filter. The force-time records were then ana- were nested within days, which were also nested within
lyzed for a 0.5-s window in the middle of the 2-s contraction subjects. Subjects were then classified as the main effect
during which the force exerted on the load cell varied by less (between-subjects). The resulting mean squares (MS) were
than ⫾ 2.5% (Fig. 1). The EMG activity associated with this then used to construct the intraclass reliability coefficient
window was assumed to be stationary with a constant mean (R) that is influenced by the true score variance, error
and standard deviation (5,7). The root-mean-square (RMS) variance due to days, and error variance due to trials. The
amplitude was calculated for EMG activity during the 0.5-s reliability was estimated by:
window (3).
The same window was used to calculate the mean power 2 true
frequency (MPF) of EMG activity. The power spectral den- R⫽ (1)
2 e2 2 e1
sity (PSD) function was estimated with the Welch perio- 2 true ⫹ ⫹
a⬘ a⬘䡠n⬘
dogram method using nine successive and overlapping sec-
tions of 512 points over the 0.5-s window (11,27). Each 2 e1 ⫽ MS
Trials (2)
section was detrended and multiplied with a 512-point Han-
ning window before computing its discrete Fourier trans- MS
Days ⫺ MSTrials
2 e2 ⫽ (3)
form (27). The discrete Fourier transforms for all nine n⬘
sections were squared and averaged to compute the PSD MS
Subjects ⫺ MSDays
function (11,27). The MPF was then calculated using the 2true ⫽ (4)
a⬘ 䡠 n⬘
formula described by Bilodeau and colleagues (6). The
frequency resolution was 4 Hz. This frequency resolution In these equations, a' is number of days, n' is number of
is sufficient to demonstrate significant differences across trials, 2e2 is error variance due to days, 2e1 is error
trials within each test session, and across test sessions. variance due to trials, and 2true is the true score variance.
The reason is that the frequency resolution of 4 Hz is Baseline and fatigue measures. Analysis of base-
much smaller than the variation due to multiple trials or line and fatigue data was accomplished using a two-factor
multiple sessions (4). (days ⫻ trials) repeated measures ANOVA with orthog-
onal polynomials for statistical trend testing (15). Post
Statistical Analysis hoc testing was accomplished using Tukey’s honestly
Reliability. The intraclass correlation analysis of vari- significant differences test (15). All statistical procedures
ance (ANOVA) model was used to determine the reliability were performed in SYSTAT (SPSS Inc., Chicago, IL).
of the baseline measures used in this study (17). This The level of significance was established at the 0.05
ANOVA model has two factors (days ⫻ subjects). The probability level.
1356 Official Journal of the American College of Sports Medicine http://www.acsm-msse.orgTABLE 1. Intraclass correlation analysis of variance for the baseline measures.
Extension
Torque Triceps Triceps Biceps BicepsMPF
(Nm) RMS (V) MPF (Hz) RMS (V) (Hz)
Day Mean ⴞ SD Mean ⴞ SD Mean ⴞ SD Mean ⴞ SD Mean ⴞ SD
1 20.6 ⫾ 6.3 123 ⫾ 80 68 ⫾ 10 41 ⫾ 42 85 ⫾ 13
2 21.9 ⫾ 6.9 128 ⫾ 81 70 ⫾ 10 46 ⫾ 35 84 ⫾ 17
3 22.3 ⫾ 5.7 153 ⫾ 67 68 ⫾ 8 57 ⫾ 22 85 ⫾ 10
(df)
MS
Subjects 12 503.210 68173.660 876.195 6908.388 1104.111
MS
Days within subjects 26 26.853 7050.790 62.463 4020.935 318.782
MS
Within cells 156
(2e1 ⫺ Trials) 4.894 694.242 29.922 337.169 96.443
2e2 ⫺ Days 4.392 1271.310 6.508 736.753 44.468
2t ⫺ True 31.757 4074.858 54.249 192.497 45.689
R 0.94 0.90 0.93 0.42 0.68
RESULTS separation between the two fatigue curves (Fig. 2, panel 1).
There was a significant (P ⬍ 0.05) linear decrease in max-
Baseline Reliability
imal isometric elbow extension torque within each experi-
The means and standard deviations for the baseline cri- mental session. Furthermore, the strength decrement in-
terion measures are presented in Table 1. Maximal isometric creased from 3.4 Nm (16%) on session 1 to 4.9 Nm (20%)
elbow extension torque increased 1.7 Nm (8%) with a con- on session 3 (P ⬍ 0.05). This was determined by subtracting
comitant rise of 30 V (24%) in triceps RMS EMG ampli- the mean of the last five (26 –30) trails from the mean of the
tude across the three sessions (P ⬍ 0.05). Biceps RMS EMG
amplitude also increased and changed 16 V (39%) from
session 1 to session 3 (P ⬍ 0.05). In contrast, the MPF of the
triceps and biceps EMG activity failed to reach the 0.05
probability level.
The decrease in stability was offset by a high degree of
consistency within subjects. The intraclass correlation
ANOVA summary for each of the criterion measures is
given in Table 1. The Rs along with the trial and day error
variance components suggest that most of the variability
was between subjects and not due to variances associated
with the multiple trials or days. The Rs ranged from good to
excellent, except for biceps RMS EMG amplitude—further
analysis showed that the low reliability was not due to a
reduction in either stability or consistency. The range de-
creased from 137 V on session 1 to 114 V on session 3.
The standard deviation also exhibited a nearly 50% reduc-
tion. Subjects therefore became more homogeneous with
respect to biceps EMG amplitude by the last test session (6).
Fatigue Patterns
Elbow extension torque. The means and standard
deviations are presented in Table 2. The fatigue patterns for
sessions 1 and 3 are given in Figure 2, panel 1, to illustrate
the direction of change (session 2 was omitted only for the
sake of clarity). The 30-trial mean for maximal isometric
elbow extension torque increased 1.4 Nm (6.9%) from ses-
sion 1 to session 3 (P ⬍ 0.05). This resulted in a distinct
TABLE 2. Mean strength-endurance (30-trial mean) during the fatigue protocol.
Extension
Torque Triceps Triceps Biceps Biceps
(Nm) RMS (V) MPF (Hz) RMS (V) MPF (Hz) FIGURE 2—The 30-trial fatigue patterns for session 1 and session 3
Day Mean ⴞ SD Mean ⴞ SD Mean ⴞ SD Mean ⴞ SD Mean ⴞ SD for elbow extension torque (panel 1); triceps root-mean square (RMS)
EMG amplitude (panel 2); mean power frequency (MPF) of triceps
1 20.4 ⫾ 6 128 ⫾ 70 67 ⫾ 10 41 ⫾ 28 86 ⫾ 15 EMG activity (panel 3); biceps root-mean square (RMS) EMG ampli-
2 21.5 ⫾ 5 150 ⫾ 78 66 ⫾ 9 67 ⫾ 19 82 ⫾ 19 tude (panel 4); and mean power frequency (MPF) of biceps EMG
3 21.8 ⫾ 5 163 ⫾ 65 63 ⫾ 8 88 ⫾ 15 83 ⫾ 11 activity (panel 5). Each curve represents the mean across subjects.
NEURAL ADAPTATIONS TO FATIGUE Medicine & Science in Sports & Exercise姞 1357first five (1–5) trials. The linear trend component for the Strength Fatigue Patterns
days-by-trials interaction term was significant (P ⬍ 0.05).
We found, consistent with previous reports (19), that
This suggests that the larger strength decrement on session
isometric strength-endurance increased between the first
3 was associated with a steeper fatigue pattern (Fig. 2, panel
and second test session and then stabilized between the
1).
second and third session for a total increase of 6.8%. As
Triceps EMG activity. The means and standard devi- participants increased in strength-endurance, there was a
ations are presented in Table 2. The fatigue patterns for change in the fatigue pattern that agreed with differences
triceps RMS EMG amplitude on sessions 1 and 3 are de- observed between high- versus low-strength individuals
picted in Figure 2, panel 2. Triceps RMS EMG amplitude (19). Thus, the fatigue patterns on session 1 exhibited mild
increased 35 V (27.3%) from the first to last session (P ⬍ linear decreases in strength characteristic of low-strength
0.05). This change was produced by a general increase individuals, whereas a stronger linear decrease similar to
across the 30 trials. Within each experimental session, tri- those that high-strength individuals exhibit after serial iso-
ceps RMS EMG amplitude decreased (P ⬍ 0.05) an average metric contractions (19) was apparent at session 3. The MPF
of 17 V (10%). The fatigue pattern for the MPF of triceps data support these observations as the classic low-frequency
EMG activity was lower on session 3 compared with session shift in EMG activity that occurs with fatigue (1,21,24)
1 (Fig. 2, panel 3). The 30-trial mean decreased 4 Hz (6%) increased from sessions 1 to 3. The following sections
across the three sessions (Table 2). However, this change address the proposed mechanisms that mediate the phenom-
did not reach significance (P ⬍ 0.05) until session 3. There ena described above.
was a decrease in the MPF of triceps EMG activity across
the 30 trials within each experimental session: 1.8 Hz Antagonist Coactivation
(2.7%) on session 1, 2.4 Hz (5%) on session 2, and 4 Hz on
session 3 (6%). These decreases were determined by sub- The increase in agonist muscle strength-endurance was
tracting the mean of the last five (26 –30) trails from the associated with a dramatic rise (115%) in antagonist coac-
mean of the first five (1–5) trials. Again, the 0.05 probability tivation. This observation is consistent with research which
level was not reached until session 3. The significant (P ⬍ suggests that increased antagonist muscle activity provides
0.05) linear trend component for the days-by-trials interac- joint stability in response to increased agonist force output
tion term indicates that the greater frequency shift on session (14). Thus, the data presented in this paper do not support a
3 was produced by a steeper linear decrease in the fatigue role for the theory of progressive inhibition of the antagonist
pattern. (2) in the expression of agonist muscle strength or strength
Biceps EMG activity. The means and standard devia- endurance.
tions are presented in Table 2. The fatigue patterns for Disagreement with the findings of Carolan and Cafarelli
biceps RMS EMG amplitude on sessions 1 and 3 are illus- (8) may be explained by an important methodological dif-
trated in Figure 2, panel 4. Biceps RMS EMG amplitude ference: our research analyzed the absolute amplitude of
increased 47 V (115%) across the three sessions (P ⬍ EMG activity, whereas Carolan and Cafarelli (8) normalized
0.05). The MPF of biceps EMG activity averaged 3 Hz EMG activity by dividing it by the maximal voluntary
(3.4%) less on session 3 than on session 1, but this differ- contraction (MVC). The rationale for this normalization is
that it results in EMG measurements that are less sensitive
ence did not reach the 0.05 probability level (Table 2). The
to differences in day-to-day electrode placement or skin-
fatigue patterns are depicted in Figure 2, panel 5. Within
electrode interface input impedance (30). However, if as we
each session, the MPF of biceps EMG activity decreased 3.1
did, electrode placement, skin preparation, and subject test-
Hz (3%) on session 1, 3.6 Hz (4%) on session 2, and 5 Hz
ing are strictly controlled, absolute EMG amplitude can
(6%) on session 3. This decrease was significant (P ⬍ 0.05)
exhibit excellent reliability across days as demonstrated
only for session 3.
here and in another study (13).
This difference in procedure may not be merely aca-
demic. Normalization is not an appropriate data transforma-
DISCUSSION tion because factoring the data by a different value for each
subject changes the rank order of the distribution and nar-
We evaluated maximal isometric elbow extension torque,
rows the distribution of scores (20,29). The exception to this
EMG amplitude, and MPF of the elbow extensors and rule would be if there were a perfect correlation (r ⫽ 1.0)
flexors during a fatigue protocol to help us study the neural between the factor and the dependent variable. This would
mechanisms responsible for an increase in strength-endur- be the same as factoring all the data by the same constant,
ance (30-trial mean) after only one training session. We which is highly unlikely for force and EMG data as anthro-
hypothesized that an increase in strength-endurance would pometric and other quantities vary between subjects (29).
be associated with an increase in MU sychnronization
and/or a reduction in antagonist coactivation. In the follow-
Motor-Unit Recruitment
ing paragraphs, we will discuss the phenomena observed
and theoretical aspects of changes in EMG activity and Within experimental sessions. This study demon-
strength-endurance curves. strated a shift in the power spectrum toward lower frequencies
1358 Official Journal of the American College of Sports Medicine http://www.acsm-msse.organd decreases in EMG amplitude within each experimental A comparison of the MPF of triceps EMG activity for the
session. Explanation of the EMG data within each experimen- first five (1–5) trials of session 1 versus session 3 provides
tal session is based on two hypotheses. The first, a peripheral indirect support for this hypothesis (see Fig. 2, panel 3). The
hypothesis, includes two issues that are difficult to disentangle. MPF of triceps EMG activity was depressed for the first five
Fast-twitch muscle fibers have high twitch tensions, large po- (1–5) trials of the fatigue curve for maximal isometric
tentials, and fast conduction velocities. These fibers dominate extension torque on session 3, before such a decrease could
force production in maximal isometric contractions but are theoretically be caused by an accumulation of metabolites
easily fatigued and drop out, leaving fatigue resistant slow- (Fig. 2, panel 1). It is reasonable to argue that the first five
twitch muscle fibers to govern force production. These latter (1–5) trials were insufficient to promote either neuromus-
muscle fibers have low twitch tensions, small potentials, and cular transmission failure or impaired membrane conduction
slow conduction velocities. This firing sequence would explain because maximal isometric elbow extension torque was still
the decreases in EMG amplitude and in MPF (1,32). However, increasing. It is also important to recall that triceps RMS
at the same time, it is well known that fatiguing contractions EMG amplitude increased whereas the MPF of triceps EMG
can impair conduction of the action potential across the muscle activity remained unchanged across sessions during baseline
membrane (1,21,24,32). Slowed muscle fiber conduction ve- testing. The lack of change in baseline supports the hypoth-
locity may be brought about by an accumulation of either lactic esis that the decrease in the 30-trial mean for the MPF of
acid in the muscle or K⫹ in the extracellular space (1,21,24,32). triceps EMG activity was a neural adaptation specific to the
This second phenomenon could explain a decrease in both fatigue protocol.
EMG amplitude and MPF equally well as fast-twitch fiber drop To summarize, our fatigue protocol resulted in a decrease
out. in maximal isometric elbow extension torque, RMS EMG
Moritani (24) proposed a second centrally mediated hy- amplitude, and the MPF of EMG activity within each ex-
pothesis that involves an attempt by the nervous system to perimental session. However, there was an increase in max-
generate muscle force while avoiding peripheral neuromus- imal isometric elbow extension strength-endurance (30-trial
cular transmission failure. Continued high MU firing rates mean) across sessions that was associated with an increase
would eventually impair excitation and contraction coupling in RMS EMG amplitude and a decrease in the MPF of EMG
through either a depletion of Na⫹ or adenosine triphsophate activity. Thus, our results support neither a reduction in
(ATP). Therefore, the nervous system decreases MU firing antagonist coactivation nor an increase in MU synchroniza-
as a protective mechanism. Moritani (24) cites the observa- tion as the primary mechanism for an increase in strength-
tion of a reduction in MU firing during sustained maximal endurance (30-trial mean) after only one training session.
isometric contractions before any evidence of neuromuscu- We suggest that the fatigue protocol was a training stimulus
lar transmission failure in support of his hypothesis. for a down regulation of MU firing frequency, which was
Between experimental sessions. The increase in combined with increased MU recruitment across the sessions.
EMG amplitude and a decrease in MPF across sessions
could be interpreted as an increase in MU synchronization Implications for Muscle Strength and Training
(12). MU synchronization appears sporadically in surface
EMG recording as large periodic waveforms (12,25), and it The theoretical implications of the present findings are
is produced by increased recruitment and firing frequency at directed at the specificity of training. Strength-endurance
the same time (3,12). The increased probability for temporal training is associated with specific metabolic adaptations.
overlap then results in an increase in EMG amplitude as We showed that the specific adaptations also extend to
observed here (3,12). However, intramuscular recordings neural control: decreases in motor unit firing frequency
fail to demonstrate MU synchronization as a dominant form were present only during the fatigue protocol, not during
of force gradation (9), and there are methodological limita- baseline strength trials. Changes in neural control that re-
tions with the work quantifying its existence (31). Equally duced muscular fatigue occurred within three sessions with
important, because the fatigue protocol did not result in MU serial muscular contractions that resulted in a 16 –20% dec-
synchronization with each session, it is uncertain how it rement in strength. We recommend that exercise prescrip-
would serve as a training stimulus for MU synchronization tion for strength-endurance training include a 20% decre-
across sessions. ment in strength as a minimal training stimulus. The classic
A more parsimonious explanation that connects the low-frequency shift in the power spectrum of muscle activ-
within and between sessions adaptations is based on Mori- ity that is associated with muscle fatigue did not reach
tani’s (24) idea that the nervous system down regulates MU significance until a 20% strength decrement had been
firing frequency to prevent neuromuscular transmission fail- achieved.
ure. Our fatigue protocol served as a training “stimulus” for
the down regulation of MU firing frequency that with rep- The authors would like to thank Diana Hanson and Eric Growney
for their technical assistance. This work was supported by a grant
etition produced a training “response.” Increased MU re- from the Mayo Clinic/Mayo Foundation and the NIH training grant
cruitment (16) then combined with the down regulation of (HD07447).
MU firing frequency across sessions. This would result in Address for correspondence: David A. Gabriel, Ph.D., Depart-
ment of Physical Education, Brock University, 500 Glenridge
our observation of increased EMG amplitude in conjunction Avenue, St. Catharines, Ontario, Canada L2S 3A1; E-mail:
with a decreased MU firing frequency. dgabriel@arnie.pec.brocku.ca.
NEURAL ADAPTATIONS TO FATIGUE Medicine & Science in Sports & Exercise姞 1359REFERENCES
1. ARENDT-NELSON, L., K. R. MILLS, and A. FORSTER. Changes in and Sport Sciences Reviews (Vol. 24), J. O. Holloszy (Ed.).
muscle fiber conduction velocity, mean power frequency, and Baltimore: Williams & Wilkins, 1996, pp. 363–397.
mean EMG voltage during prolonged submaximal contractions. 17. KROLL, W. P. Reliability variations of strength in test-retest situ-
Muscle Nerve 12:493– 497, 1989. ations. Res. Q. Exerc. Sport 34:50 –55, 1963.
2. BASMAJIAN, J. V. Motor learning and control: a working hypoth- 18. KROLL, W. P. Measurement schedule effects upon isometric en-
esis. Arch. Phys. Med. Rehabil. 58:38 – 41, 1977. durance. Int. Z. Angew. Physiol. Einschl. Arbeitsphysiol. 21:224 –
3. BASMAJIAN, J. V., and C. J. DELUCA. Muscles Alive: Their Func- 229, 1965.
tions Revealed by Electromyography. Baltimore: Williams & 19. KROLL, W. P. Analysis of local muscular fatigue patterns. Res. Q.
Wilkins, 1985, pp. 94 –222. Exerc. Sport 52:523–539, 1981.
4. BILODEAU, M., A. B. ARSENAULT, D. GRAVEL, and D. BOURBONNAIS. 20. LINDQUIST, E. F. Design and Analysis of Experiments in Psychol-
The influence of an increase in the level of force on the EMG ogy and Education. Boston: Houghton Mifflin, 1953, pp. 149 –
power spectrum of elbow extensors. Eur. J. Appl. Physiol. 61: 151.
461– 466, 1990. 21. LUTTMANN, A. Physiological basis and concepts of electromyogra-
5. BILODEAU, M., A. B. ARSENAULT, D. GRAVEL, and D. BOURBONNAIS. phy. In: Electromyography in Ergonomics, S. Kumar and A. Mital
EMG power spectra of the elbow extensors during ramp and step (Eds.). Bristol, PA: Taylor and Francis, 1996, pp. 51–95.
isometric contractions. Eur. J. Appl. Physiol. 63:24 –28, 1991. 22. MILNER-BROWN, H. S., R. B. STEIN, and R. G. LEE. Synchronization
6. BILODEAU, M., A. B. ARSENAULT, D. GRAVEL, and D. BOURBONNAIS. of human motor units: possible roles of exercise and supraspinal
EMG power spectrum of elbow extensors: a reliability study. reflexes. Electroencephogr. Clin. Neurophsiol. 38:245–254, 1975.
Electromyogr. Clin. Neurophysiol. 34:149 –158, 1994. 23. MORITANI, T., and H. A. DEVRIES. Neural factors versus hypertro-
7. BILODEAU, M., M. CINCERA, A. B. ARSENAULT, and D. GRAVEL. phy in the time course of muscle strength gains. Am. J. Phys. Med.
Normality and stationarity of EMG signals of elbow flexors mus- 58:115–130, 1979.
cles during ramp and step isometric contractions. J. Electromyogr. 24. MORITANI, T. Muscle energetics and electromyography. In: Elec-
Kinesiol. 7:87–96, 1997.
tromyography in Ergonomics, S. Kumar and A. Mital (Eds.).
8. CAROLAN, B., and E. CAFARELLI. Adaptations after isometric resis-
Bristol, PA: Taylor and Francis, 1996, pp. 127–161.
tance training. J. Appl. Physiol. 73:911–917, 1992.
25. PERSON, R. S., and L. P. KUDINA. Cross correlations of electro-
9. DELUCA, C., Z. ERIM, and A. M. ROY. Synchronization of motor-
myograms showing interference patterns. Electroencephogr. Clin.
unit firing in several human muscles. J. Neurophysiol. 70:2010 –
Neurophsiol. 25:58 – 68, 1968.
2023, 1993.
10. GABRIEL, D. A., J. BASFORD, and K.-A. An. Effects of the reversal 26. QUARING, D. P., and J. H. WARFEL. The Extremities. Philadelphia:
of antagonists upon isometric elbow extension strength and en- Lea & Febiger, 1963, pp. 28 –30.
durance. Arch. Phys. Med. Rehabil. 78:1191–1195, 1997. 27. SHIAVI, R. Introduction to Applied Statistical Signal Analysis.
11. HARY, D., M. J. BELMAN, B. J. PROPST, and S. LEWIS. A statistical Homewood, IL: Irwin and Aksen Associates, 1991, pp. 266 –340.
analysis of the spectral moments used in EMG tests of endurance. 28. STEPANOV, A. S., and M. L. BURLAKOV. Electrophysiological in-
J. Appl. Physiol. 53:779 –783, 1982. vestigation of fatigue in muscular activity. Sechnov Physiol.
12. HAYES, K. C. A theory of the mechanism of muscular strength J. USSR 47:43– 47, 1961.
development based upon EMG evidence of motor unit synchro- 29. TANNER, J. M. Fallacy of per-weight and per-surface area standards
nization. In: Biomechanics of Sports and Kinanthropometry, F. and their relation to spurious correlation. J. Appl. Physiol. 2:1–15,
Landry and W.A.R. Orban (Eds.). Miami, FL: Miami Symposia 1949.
Specialists, 1978, pp. 69 –77. 30. YANG, J. F., and D. A. WINTER. Electromyographic reliability in
13. HEINONEN, A., H. SIEVÄNEN, J. VIITASALO, P. PASANEN, P. OJA, and maximal and submaximal isometric contractions. Arch. Phys.
I. VUORI. Reproducibility of computer measurement of maximal Med. Rehabil. 64:417– 420, 1983.
isometric strength and electromyography in sedentary middle- 31. YUL, G., A. J. FUGLEVAND, M. A. NORDSTROM, and R. ENOKA.
aged women. Eur. J. Appl. Physiol. 68:310 –314, 1994. Limitations of the surface electromyographic technique for esti-
14. KELLIS, E., and V. BALTZOPOULIS. The effects of antagonist moment mating motor unit synchronization. Biol. Cyber. 73:223–233,
on the resultant knee joint moment during isokinetic testing of the 1995.
knee extensors. Eur. J. Appl. Physiol. 76:253–259, 1997. 32. WINTER, D. A. EMG interpretation. In. Electromyography in Er-
15. KIRK, R. E. Experimental Design: Procedures for the Behavioral gonomics, S. Kumar and A. Mital (Eds.). Bristol, PA: Taylor and
Sciences. Belmont, CA: Brooks/Cole Publishing, 1968, pp. 171– Francis 1996, pp.109 –125.
244. 33. WINTER, D. A. Biomechanics and Motor Control of Human Move-
16. KRAEMER, W. J., S. J. FLECK, and W. J. EVANS. Strength and power ment (2nd Ed.). New York: John Wiley & Sons, 1990, pp. 191–
training: physiological mechanisms of adaptation. In: Exercise 212.
1360 Official Journal of the American College of Sports Medicine http://www.acsm-msse.orgYou can also read
