Behavioral/Systems/Cognitive
Responses of Human Motoneurons to Corticospinal
Stimulation during Maximal Voluntary Contractions
and Ischemia
Jane E. Butler, Janet L. Taylor, and Simon C. Gandevia
Prince of Wales Medical Research Institute and University of New South Wales, Randwick, New South Wales, 2031 Australia
The discharge frequency of human motoneurons declines during a sustained isometric maximal voluntary contraction (MVC) of elbow
flexor muscles, but the cause is unresolved. We aimed to determine whether motoneurons were inhibited during a sustained fatiguing
contraction of the elbow flexor muscles and whether this inhibition was caused by the discharge of group III and IV muscle afferents.
Subjects performed brief MVCs before and after a fatiguing 2 min MVC. During maximal efforts, electromyographic responses recorded
from the elbow flexor muscles were evoked by stimulation of the corticospinal tracts at the cervicomedullary level [cervicomedullary
motor evoked potentials (CMEPs)] and by supramaximal stimulation over the brachial plexus (Mmax ). This revealed a novel decrease in
the size of the muscle response to corticospinal tract stimulation during fatigue. During the sustained MVCs, the size of CMEPs decreased
to 81
15 and 78
15% of the control value for brachioradialis and biceps brachii, respectively (mean
SEM; n 8). This recovered
within 15 sec after the fatiguing contraction. In a second set of studies, input from group III and IV muscle afferents was prolonged after
the end of the fatiguing contraction by holding the muscle ischemic with a cuff inflated above arterial pressure. Despite the maintained
discharge of group III and IV afferents, the CMEPs again recovered within 15 sec of the end of the sustained contraction. These results
show a diminished output of spinal motoneurons to stimulation of corticospinal tracts during a fatiguing MVC; however, the mechanisms
responsible for this decline are not attributable to activity in group III and IV muscle afferents.
Key words: fatigue; motoneuron (motor neuron); group III and IV muscle afferents; human; corticospinal; ischemia
Introduction
During a sustained isometric maximal voluntary contraction
(MVC) the discharge rate of motoneurons usually declines. This
decline has been observed for intrinsic muscles of the hand
(Marsden et al., 1969, 1983; Bigland-Ritchie et al., 1983a,b; Gan-
devia et al., 1990), biceps brachii (Bigland-Ritchie et al., 1986;
Woods et al., 1987), and tibialis anterior (Grimby et al., 1981;
Macefield et al., 2000). Several mechanisms have been proposed
to account for the reduction in firing rate. The most popular
explanation is that activity in group III and IV muscle afferents
reflexively inhibits spinal motoneurons. The strongest argument
in favor of a causative role for group III and IV muscle afferents is
that the firing rates of motoneurons (and voluntary force) remain
depressed as long as the muscles are held ischemic after the con-
traction ceases (Bigland-Ritchie et al., 1986, 1995; Woods et al.,
1987; Garland and Gossen, 2002). With the muscle at rest, mo-
toneuronal and supraspinal factors should recover, whereas fir-
ing of group III and IV muscle afferents, which are sensitive to
ischemia and metabolites, will be maintained while blood flow is
occluded (Kaufman et al., 1984; Hayward et al., 1991; Taylor et
al., 2000c). Further evidence for this view comes from the pres-
ence of a contraction-induced spinal inhibition between syner-
gists in the cat hindlimb (Hayward et al., 1988; Sacco et al., 1997)
as well as studies combining electrically induced contractions,
H-reflexes, and partial nerve blocks (Garland and McComas,
1990; Garland, 1991; Walton et al., 2002). However, the reflex
effects of group III and IV muscle afferents on spinal motoneu-
rons include both excitatory and inhibitory inputs (Kniffki et al.,
1979, 1981a,b), and the net effect on spinal motoneuronal excit-
ability is unclear.
Other mechanisms that could contribute to the fatigue-
induced decline in motoneuron firing rate include intrinsic mo-
toneuronal properties (Kernell and Monster, 1982a,b; Spielmann
et al., 1993; Sawczuk et al., 1997), changes in net input from other
muscle receptors [e.g., muscle spindles (Bongiovanni and Hag-
barth, 1990; Gandevia et al., 1990; Macefield et al., 1991, 1993)],
and a decline in effective descending drive to the motoneurons
(Gandevia et al., 1996; Taylor et al., 2000a).
The present studies were designed to test motoneuronal excit-
ability during sustained maximal contractions by stimulation of
descending motor paths at the cervicomedullary level. This
method can activate the same corticospinal axons as transcranial
magnetic stimulation of the motor cortex as judged by collision
experiments (Ugawa et al., 1994; Gandevia et al., 1999; Taylor et
al., 2002). It causes a monosynaptic EPSP in motoneurons of
elbow flexors (Petersen et al., 2002), and the evoked responses are
not influenced by conventional presynaptic inhibition (Nielsen
Received July 8, 2003; revised Sept. 14, 2003; accepted Sept. 16, 2003.
This work was funded by the National Health and Medical Research Council of Australia (#3206).
Correspondence should be addressed to Dr. S. C. Gandevia, Prince of Wales Medical Research Institute, Barker
Street, Randwick, New South Wales, Australia 2031. E-mail: s.gandevia@unsw.edu.au.
Copyright © 2003 Society for Neuroscience 0270-6474/03/2310224-07$15.00/0
10224 • The Journal of Neuroscience, November 12, 2003 • 23(32):10224 –10230

and Petersen, 1994). Maintained muscle ischemia was used to
prolong the firing of group III and IV afferents after the end of the
sustained MVC (Kaufman et al., 1984; Taylor et al., 2000c) and
thus to reinvestigate the proposition that these afferents produce
significant inhibition of motoneurons during fatiguing voluntary
contractions.
Preliminary results have been published previously (Butler et
al., 1999).
Materials and Methods
The main experiments were performed on eight subjects (three females)
who were studied on two occasions. Additional control experiments (de-
scribed below) were also performed on eight subjects (four females) on
two occasions. Three of the subjects performed both the main and con-
trol experiments. Subjects were healthy and ranged in age from 30 to 58
years. The protocols were approved by the local ethics committee, and
the studies were conducted according to the Declaration of Helsinki. All
subjects gave informed written consent.
Experimental task and recordings
Many of our experimental procedures and recording techniques are sim-
ilar to those described previously (Gandevia et al., 1999; Taylor et al.,
2000c). The test fatiguing contraction was a sustained maximal voluntary
effort (MVC) of the right elbow flexors lasting 120 sec. The elbow was
flexed to 90 o and strapped to an isometric myograph (Fig. 1 A) (Allen et
al., 1995). Before the test contraction a series of brief MVCs (lasting 2–3
sec) were performed at1 min intervals. After the test contraction, brief
MVCs were performed (at 15, 30, 45, 60, 75, 90, 120, 135, 180, 195, 240,
255, 300, and 315 sec after the sustained contraction) to assess the recov-
ery of force and the electromyographic (EMG) responses to stimulation
(Fig. 1 B). Subjects received visual feedback of force and constant verbal
encouragement to contract maximally throughout every contraction.
This protocol was repeated several days later with one change: a sphyg-
momanometer cuff wrapped around the upper arm was inflated to 300
mmHg from a source of compressed air with 15 sec remaining in the test
contraction. Inflation of the cuff was maintained for 120 sec into the
recovery period during which eight additional brief recovery MVCs were
performed every 15 sec. After the cuff was deflated the subjects per-
formed more brief recovery MVCs (at 15, 30, 45, 60, 75, 90, 120, 135, 180,
195, 240, 255, 300, 315, 360, and 375 sec after cuff deflation).
Surface electrodes over brachioradialis and biceps brachii recorded
EMG responses (Ag–AgCl electrodes, 10 mm diameter). The signals were
amplified (100 –300), filtered (1.6 Hz–1 kHz; CED 1902 amplifiers),
and recorded via a CED 1401 interface with Signal software (sampling
rate 5 kHz) (Cambridge Electronic Design, Cambridge, UK).
Stimulation
Brachial plexus stimulation. Stimuli were deliv-
ered to the brachial plexus at Erb’s point so that
the maximal M wave could be monitored
(Mmax). The stimulus intensity was at least 50%
above that required to produce a maximal re-
sponse in both brachioradialis and biceps
brachii. The cathode was in the supraclavicular
fossa, and the anode was over the acromion.
Stimuli were rectangular pulses (100
sec du-
ration) delivered via a Digitimer DS7 constant-
current stimulator (modified for outputs up to
1 A; Digitimer Ltd., Welwyn Garden City, UK).
Corticospinal tract stimulation. Stimuli were
also delivered to corticospinal tracts close to the
cervicomedullary junction. Responses evoked
by cervicomedullary stimulation are termed
cervicomedullary motor evoked potentials
(CMEPs). For two of the eight subjects, cortico-
spinal stimulation was achieved using an elec-
trical stimulus passed between two 9 mm Ag–
AgCl electrodes filled with conductive paste and
glued to the skin over the left (cathode) and
right (anode) mastoid processes (Ugawa et al.,
1991; Gandevia et al., 1999). The pulse was 100
sec duration, and the
intensity range was 360 –540 V (Digitimer, D180). For six subjects,
CMEPS were evoked using a Magstim 200 with a double-cone coil from
which the plastic casing had been removed (Magstim Company Ltd.,
Dyffed, UK). This allowed the center of the coil to be closely opposed to
the back of the skull. Usually the coil center was positioned2 cm below
the inion, but sometimes it was moved to the right by 1–2 cm to
produce adequate responses. Current direction was downward in the coil
(Ugawa et al., 1994; Taylor et al., 2000b). Stimulus intensities were be-
tween 80 and 100% stimulator output to evoke responses during brief
control MVCs of 60 –70% Mmax. At this intensity the latency of the
responses was 8.4
0.6 msec (mean
SD) for biceps and 11.2
1.1
msec for brachioradialis. Responses were monitored at high gain to en-
sure that the onset of the CMEPs was not contaminated by activation of
axons distal to the cell body of the motoneurons.
For the main experiment, during the brief control and recovery MVCs,
stimuli to either the brachial plexus or the corticospinal tracts were de-
livered alternately at the peak force of each MVC. During the sustained
MVC, corticospinal tract and brachial plexus stimuli were delivered in
pairs (5 sec apart) every 15 sec throughout the 2 min contraction.
Additional control experiments: submaximal brachial
plexus stimulation
Because the size of the compound muscle action potential can change
with exercise (Cupido et al., 1996; Taylor et al., 1999), the size of the
CMEPs was always normalized to the compound muscle action potential
(M wave) produced at a nearby time by supramaximal stimulation of the
brachial plexus (i.e., CMEPs were expressed as a percentage of Mmax).
This procedure has been used previously (Gandevia et al., 1999; Taylor et
al., 2000c). A concern with this normalization was that the sustained
muscle contraction and the maintained ischemia for 2 min may have
different effects on the sarcolemmal function of different-sized muscle
fibers such that the change in size of Mmax may not reflect the changes in
the muscle fibers recruited in a submaximal CMEP (Enoka et al., 1992).
If this occurred, it would not be appropriate to normalize a submaximal
muscle action potential to Mmax. Therefore, additional control experi-
ments were performed in eight subjects using both supramaximal and
submaximal stimulation of the brachial plexus to test whether a sustained
MVC with maintained muscle ischemia produced differential effects on
the maximal muscle action potential (100% Mmax) compared with the
submaximal muscle action potential (65% Mmax, a level that would be
comprised of larger rather than smaller motor units). The experiment
focused on responses in brachioradialis because this muscle is distal to
the sphygmomanometer cuff. The two protocols of the main experiment
described above were repeated except that the corticospinal stimulus was
Figure 1. Experimental setup and protocol. A, Experimental setup. B, Two protocols. Brief control MVCs were performed during
which stimuli were delivered to the brachial plexus (white arrows) or to the corticospinal tracts (black arrows). The subject then
performed a 2 min sustained MVC during which stimuli were delivered alternately to the corticospinal tracts and the brachial
plexus. Brief recovery contractions were performed afterward. In the second protocol (bottom panel), the forearm muscles were
held ischemic for 2 min after the sustained MVC (region labeled CUFF).
Butler et al. • Motoneuron Inhibition during Fatigue J. Neurosci., November 12, 2003 • 23(32):10224 –10230 • 10225