Exp Brain Res (1991) 84:210-218
Exp.erimental
BrainResearch
9 Springer-Verlag1991
The effect of muscle length on motor unit discharge characteristics
in human tibialis anterior muscle
D.W. Vander Linden 1, C.G. Kukulka 2, and G.L. Soderberg 2
Physical Therapy Department, University of Florida, Box J 154, Health Science Center, Gainesville, FL 32610, USA
z Physical Therapy Graduate Program, The University of Iowa
Received April 6, 1990 / Accepted October 26, 1990
Summary. Muscle length influences the contractile
properties of muscle in that when nmscle is lengthened
the relaxation phase of the muscle twitch is prolonged
and when muscle is shortened, the relaxation phase is
shorter in duration. As a result, the force exerted by
active motor units varies with muscle length during voluntary contractions. To determine if motoneuron spike
trains were adjusted to accommodate for changes in the
contractile properties imposed by shortened and lengthened muscle, motor unit action potentials were recorded
from the tibialis anterior muscle at different muscle
lengths. Twenty subjects performed isometric ramp contractions at ankle angles of 20 ~ dorsiflexion, neutral between dorsiflexion and plantar flexion, and 30 ~ plantar
flexion, which put the tibialis anterior muscle in a shortened, neutral, or lengthened condition, respectively.
During isometric contractions where torque increased at
5 % MVC/s, motor unit discharge rate at recruitment was
greater in shortened muscle than in lengthened muscle
(P < 0.05). Brief initial interspike intervals ( < 40 ms) occurred more frequently in shortened muscle than in either
neutral length or lengthened muscle. During steady contractions, motor unit discharge rate was greater per unit
torque (N.m) in shortened muscle than in neutral length
or lengthened muscle (P< 0.05). These findings indicate
that muscle length does influence the discharge pattern
of motor unit spike trains during isometric ramp contractions. Spike trains with higher discharge rates at recruitment in shortened muscle may take advantage of the
catch-like properties in muscle and be useful in taking up
the slack in the passive elements of the muscle and tendon. During steady submaximal contractions, the higher
discharge rate per unit torque (N.m) in shortened muscle
is likely due to the decreased peak tension and shorter
one-half relaxation time observed in shortened muscle,
and may indicate that the tibialis anterior muscle is operating on the steep portion of the length-tension curve
when the ankle is fully dorsiflexed.
Qffprint requests to: Dr. Darl W. Vander Linden (address see
above)
Key words: Muscle length - Motor unit discharge rate Motor control - Human tibialis anterior
Introduction
Force generation in skeletal muscle is dependent upon
the net output of the motor neuron pool and the contractile properties of the muscle fibers. During voluntary
contractions, human subjects can control the descending
input onto the motor neuron pool, which along with
afferent input, influences the recruitment and discharge
patterns of motor units for a given task. Muscle fiber
contractile properties, however, are influenced by factors
outside the subject's control, and include the contractile
history of muscle, the presence or absence of fatigue, and
muscle length (for a review see Partridge and Benton
1981). Each of these factors has received considerable
research attention yet little work has focused on the
interaction of CNS activation and changes in muscle
contractile properties due to changes in muscle length
during voluntary efforts in humans.
Bigland-Ritchie and colleagues investigated the interaction between motoneuron pool output and changes
in muscle contractile properties due to fatigue, and have
demonstrated that maximum motor unit discharge rate
decreases as muscle fatigues (Bigland-Ritchie et al. 1983;
Bellemare et al. 1983). Because fatigue results in slowing
of the contractile properties of the muscle, the slowing of
motor unit discharge rate during continuing maximum
contractions suggests that mechanisms controlling motor
neuron pool output are adjusted to accommodate for
changes in muscle (Bigland-Ritchie et al. 1986).
Others have shown that isometric torque production
is controlled differently during shortening versus lengthening contractions (Tax et al. 1990, Kato et al. 1985;
Andrew 1985). Differences in motor unit discharge patterns for isometric and anisometric contractions were
most likely due to the type of contraction rather than the
effect of muscle length, as joint angle changed only 10~
211
b
'
'
Fig. 1. Device to measure dorsiflexion
torque with subject prone. Ankle angle can
be adjusted from 20 ~ of dorsiflexion to 30~
of plantar flexion in one degree
increments; a) velcro straps, b) heel block,
c) load cell, d) patella tendon bar,
e) threaded rod, f) axis of rotation,
g) nylon strap
(digiti minimi abduction) or 20 ~ (elbow flexion, ankle
dorsiflexion) during the shortening or lengthening conditions in these studies. Such a small change in joint angle
would not change muscle length substantially, therefore
the effect o f muscle length on m o t o r unit discharge
properties could not be adequately evaluated f r o m these
experimental paradigms. Miles and colleagues (1986) did
evaluate the effect o f muscle length on m o t o r unit recruitment threshold in h u m a n masseter muscle. A l t h o u g h
they reported that passive tension is highly correlated
with recruitment threshold as muscle is lengthened, no
data regarding m o t o r unit discharge behavior were reported.
The contractile properties o f muscle (i.e. twitch contraction time and o n e - h a l f relaxation time) have long
been k n o w n to dictate the stimulation frequencies necessary to generate tetanic tension. The force-frequency
relationship for b o t h whole muscle ( R a c k and W e s t b u r y
1969), as well as for single m o t o r units (Burke et al. 1976)
has been shown to depend u p o n the type o f muscle fibers
studied. Fast twitch fibers require greater stimulation
frequencies to generate tetanic tensions than do slow
twitch fibers. It has further been shown that the stimulus
interpulse interval for most efficiently generating the
greatest tension is linearly related to the m o t o r unit's
one-half relaxation time, and is shorter for units with
shorter relaxation times (Zajac and Y o u n g 1980). Alterations in muscle contractile properties, such as occur with
changes in muscle length, in turn induce predictable
changes in the force-frequency relationship. In h o m o g e neous cat soleus muscle, b o t h twitch c o n t r a c t i o n time
and o n e - h a l f relaxation time are decreased in shortened
c o m p a r e d to lengthened muscle, thereby requiring higher
stimulation frequencies for p r o d u c i n g a given level o f
tension ( R a c k and W e s t b u r y 1969). These findings have
been verified in h u m a n tibialis anterior muscle (Marsh
et al. 1981).
The relationship between nmscle length and stimulus
frequency has been d e m o n s t r a t e d almost exclusively with
electrical stimulation paradigms. A l t h o u g h Miles et al.
(1986) d e m o n s t r a t e d a relationship between h u m a n
masseter m o t o r unit recruitment threshold and muscle
length, no i n f o r m a t i o n is available on recruitment or rate
coding behaviors in limb muscles. The influence o f length
and resultant changes in the muscle contractile properties
on m o t o r unit discharge characteristics have not been
adequately studied during v o l u n t a r y contractions. The
purpose o f this investigation was to evaluate the relationship between muscle length and m o t o r unit behavior in
h u m a n tibialis anterior muscle. The results reveal that
m o t o r unit discharge behaviors during isometric contractions up to 40% M V C are adjusted in association with
changes in ankle joint angle and muscle length.
Methods
Sltbjec~s
Twenty untrained male subjects (ages 21-35) with no history of
injury or orthopaedic abnormality of the lower extremities participated in this study. Participation in the study required that
subjects were not running more than 10 km a week nor involved in
resistive exercise training for the lower extremities. Recreational
activities such as basketball, bicycling, or tennis did not exclude
subjects from the study. Subjects provided informed written consent
according to guidelines established by an institutional review board,
and were compensated for participating in the study.
Instrumentation
A device was designed and fabricated to stabilize the left foot and
lower leg of subjects when they were positioned prone (Fig. l). This
device allowed for positioning of the ankle joint in one degree
increments from 20 ~ of dorsiflexion (70 ~ between tibia and bottom
surface of foot) to 30~ of plantar flexion (120~ between tibia and
bottom surface of foot). From full plantar flexion to full dorsiflexion, tibialis anterior is estimated to shorten by 3-5 cm (Brunnstrom 1966; our observations). Based on a tibialis anterior muscle
length of 29.8 cm (Wickiewicz et al. 1983) muscle length was estimated to change about 11-17% of maximum length from 30~
plantar flexion to 20 ~ dorsiflexion.
Velcro straps over the dorsmn of the foot and an adjustable
heel block held the loot securely in place. The uprights of the device
were aligned parallel to the long axis of the tibia and were secured
in place by a patella tendon bar and velcro strap. Nylon straps were
adjusted to align the lower leg perpendicular to the table, and
allowed the subject to completely relax the muscles of the thigh and
lower leg, as no muscle activity was required to maintain the lower
leg in the resting position. The knee and ankle were not subjected
to compressive forces from the device other than the weight of the
Im~er leg and the device (approximately 2 kg).
212
A load cell which measured both compressive and tensile forces
(AWU-100, Genisco Technology Corporation, Compton, CA) was
placed between the uprights and a 3/8" threaded rod to record the
force generated in dorsiflexion. The response of the load cell was
linear within +/-3.25.~
Although the calculation of torque not
was essential for the purposes of this experiment, we wished to
compare maximum torque values with those found by Marsh et al.
to help validate the use of our device to measure dorsiflexion torque
in human subjects. Briefly, torque was calculated by multiplying the
compressive force as measured by the load cell by the perpendicular
distance from the threaded rod to the estimated ankle center of
rotation. The distance of this moment arm, which varied with ankle
angle, was 19.9 cm, 18.8 cm and 15.3 cm for 70, 90, and 120 ~ of
ankle angle respectively.
Motor unit action potentials of the tibialis anterior muscle were
recorded by one of three types of indwelling, fine-wire electrodes.
Electrodes with 5 um diameter active sites 15 mn apart (Nelson and
Soderberg 1983) were used in two subjects, and although the selectivity of the electrodes was adequate, the shape and amplitude of
the unit potentials were susceptible to change as isometric dorsiflexion torque increased. Subcutaneous branched electrodes (Enoka
et al. 1988) were used in four subjects. The stability of the recording
using these electrodes was excellent, but the probability of a good
signal-to-noise ratio was poor as the electrode had to be inserted
precisely between the subcutaneous tissue and the muscle fibers. In
the remaining 14 subjects, standard bipolar 50 um stainless steel
insulated wire electrodes, constructed as described by Clamann
(1970) were used.
Electrical signals of the motor units were first amplified by a
high impedance (15 megohms at 100 Hz) on-site amplifier and then
led off for further amplification (GCS-67 Amplifier, Therapeutics
Unlimited, Iowa City, IA). The total gain for the entire amplification system ranged from 500 to 10,000, and the frequency response
was 3dB down at 40 Hz and 10 kHz. Load cell output was amplified
by a DC amplifier with a gain range of 1 to 1000 and a frequency
response of DC to 1 kHz. A 10 ms TTL pulse, coincident with the
beginning of each trial, was used to begin computer sampling for
force and interspike interval data during off-line analysis. E M G and
force signals were monitored on an oscilloscope during the experiment. and all signals were recorded on FM tape (Hewlett-Packard,
Model 3969 A, San Diego, CA) for off-line analysis.
Procedure
Subjects first performed two 3-second maximum dorsiflexion contractions at ankle angles of 70~ 90, and 120 ~ One minute of rest
between trials was given to minimize fatigue. The maximum force
during the plateau portion of the brief efforts was used as the
maximum dorsiflexion force at each angle. After the fine wire
electrodes were inserted into the central portion of the belly of the
tibialis anterior muscle, the following procedure was used to evaluate motor unit behavior at each of 3 ankle positions. A series of
ramp up-hold-ramp down isometric contractions, similar to those
used by Tanji and Kato (1973) were performed in which subjects
traced a torque template on an oscilloscope screen. Dorsiflexion
torque was increased at a rate of 5% MVC/s to a target level of 10%,
20%, 30%, or 40% MVC, the torque was held steady for 8 s, and
then allowed to decrease by 5% MVC/s back to baseline (Fig. 2).
The force generated by the weight of the foot while the subject
was at rest at 90 degrees was subtracted from the force reading by
zeroing the load cell with the subject at this position. At 30 ~ of
plantar flexion and 20 ~ of dorsiflexion, the torque generated by the
weight of the foot decreased by an estimated 0.11 and 0.04 N.m
respectively for a subject with a body weight of 75 kg. This change
in the contribution of the weight of the foot was considered to be
negligible for the purposes of this study, where 10% MVC contractions generated an average of 4 and 2 N.m of dorsiflexion torque
at 120~ and 70~ respectively.
Maximum active dorsiflexion force at 120 ~ was calculated by
subtracting the resting force generated by the passive elements of
llelUllIIILl
200 r
5 sec
5 sec
I
lNrn
Fig. 2. Tibialis anterior EIVIG recorded from intramuscular electrode (upper trace). Dorsiflexion torque from 1026 MVC ramp
up-hold-ramp down task (lower trace,)
the dorsiflexor muscles from the total force generated. Maximum
active dorsiflexion force at 70 ~ was calculated by summing the force
needed to overcome the passive force of the plantar flexor muscles
at rest and the force generated in dorsiflexion. For ramp trials, the
subjects baseline force at rest was aligned with the baseline of the
ramp template on the oscilloscope. Hence, only the generation of
active force at 120 ~ resulted in the force trace being deflected
upward to follow the template on the oscilloscope screen. At 70 ~,
force produced to overcome the passive tbrce from the plantar
flexors resulted in the force trace being deflected upward.
Two trials at each ramp-hold-ramp target level were recorded
with one minute of rest between trials to minimize fatigue. When
the ankle was moved to a new angle, subjects had an additional
3-5 rain of rest. The order of contraction levels and ankle angles was
randomized for each subject. The lower leg of the subject was in the
device for up to one hour during the recording of motor units. No
complaints regarding ischemia of the leg were reported, but in a few
cases, pressure of the velcro strap over the dorsum of the foot
caused ischemic symptoms in the subject's foot. In those cases, the
velcro straps were loosened for a few minutes until the symptoms
subsided, and the experiment was continued. Because complaints of
ischemia were limited to the foot and did not include the lower leg,
it is unlikely that the physiologic state of the TA was altered such
that discharge patterns of motor units were changed. The device in
no way constrained the circulation of the muscles of the posterior,
lateral or anterior compartments of the lower leg.
Data analysis
A window discriminator with a delay line (DDIS--1 Dual Window
Discriminator (BAK Electronics, Rockville, MD) was used for
off-line identification of motor units. A microcomputer (IBM PC)
and a 12-bit A/D converter (Dash-16, Metrabyte Corporation,
Taunton, MA) were used to calculate the interspike intervals (ISis)
from the acceptance pulses received fi'om the window discriminator.
Throughput capability of the A/D converter was 50,000 samples/s.
Improvement in the signal to noise ratio of a motor unit train
was accomplished by filtering with either a second-order Butterworth high-pass filter (low frequency cutoff at 100, 250, 500, 1000,
1500, or 2000 Hz) or a differentiator circuit (Clamann and Lamb
1976). Unit identification was obtained by adjustment of dual time
and amplitude windows of the discriminator, and by observing the
entire waveform on an oscilloscope after passing through a delay
line (Fig. 3A). The acceptance pulses of the window discriminator
(Fig. 3B) were used to trigger the computer to read the computer's
213
C
Window A
~ ~
ill
f
,
f
e
f
i
i
|
~
I
o
I III rlll~ll
r'I I'rT
i
i
I
f
f
I
i
I
100~V
II
B
r!I!
]' If! II
tl
I lllll
D
0.5rnsec
500m s e c
t
J
200
,uV
...~,,~,,,l~li,~J~,,J~i,_,1,1-~,-,.-,,,,L~.,,..m.,.,L, ,-"~**" ~t
,
[i!i
,
IIIIIII11
I rl ]1'
r'
J200,uV
2 0 0 ,uV
500msec
100m s e c
Fig. 3A-D. Discrimination of a single motor unit from the EMG
record. A Motor unit potentials multiplexed with time and amplitude windows. Discriminator windows A and B adjusted to allow
only potentials from a large unit to pass through windows. Windows A, B are ANDED so potential must pass through both
windows for discriminator to output acceptance pulse. B Train of
motor unit potentials from the same unit as in A (upper trace).
Lower trace contains acceptance pulses from discriminator for each
unit that met time and amplitude criteria for both windows A and
B. C Example of one motor unit potential in a train which was
superimposed upon a smaller unit, and subsequently did not pass
through time and amplitude windows of discriminator. D Example
of motor unit potential not from train of interest that passed
through windows and caused the discriminator to output an acceptance pulse. Interspike interval data were corrected for spurious
misses or contaminations (as in C or D) of motor unit trains prior
to analysis
clock (resolution of 1 ms) as well as sample the dorsiflexion torque.
ISis were determined from the clock readings and all data were
transferred to floppy diskette for storage and further analysis.
Occasionally, the active potential of the unit of interest would
superimpose on a smaller potential (Fig. 3C) and/or spuriously
change shape, or a second unit would enter the window (Fig. 3D).
A visual inspection of each spike train was made in 0.5 2 s segments
on a storage oscilloscope to determine if such contaminations of the
acceptance pulse record were present. If the discharge of a unit was
missed, the ISI was measured from the oscilloscope screen, and the
computer file modified to reflect the correction. If a spurious discharge of a second unit was detected, a similar correction was made
to the computer file containing the ISI data. No contaminations or
misses were observed in 30% of all spike trains. In 50% of the spike
trains, three or fewer contaminations were observed, while in the
remaining 20% of the trials, four or more contaminations were
found in spike trains of up to 216 events. More contaminations were
found in those trials at higher force levels, as a greater number of
units were active at those levels. When a unit's amplitude and
waveform were not stable or could not be reliably discriminated
from other units in the recording, the unit was not used in the
analysis.
When dorsiflexion torque was increasing at 5% MVC/s, all
recorded units continued to discharge after once becoming active.
During no trials did units discharge more than once, become quiet,
and then begin to discharge again. Initial discharge rate was defined
as the reciprocal of the first interspike interval. Recruitment threshold was defined as force at which the first spike occurred for each
motor unit train.
Results
Descriptive statistics
O n e - h u n d r e d thirty-stx units from 16 of the 20 subjects
in the study were analyzed. I n four subjects, m o t o r units
could n o t be reliably identified in a n y o f the trials. O n l y
rarely was it possible to follow a single u n i t across different a n k l e positions a n d c o n t r a c t i o n levels. All units
recorded were therefore considered u n i q u e a n d entered
into the p o o l e d data for statistical analyses.
Because the w a v e f o r m o f m o t o r u n i t s sometimes
c h a n g e d as isometric t o r q u e was increasing or decreasing, not all o f the units could be reliably d i s c r i m i n a t e d
t h r o u g h o u t the entire task. M o s t units (95 o f the 136)
could be d i s c r i m i n a t e d t h r o u g h o u t the entire r a m p uph o l d - r a m p d o w n trial. A n a d d i t i o n a l 21 units could be
reliably d i s c r i m i n a t e d t h r o u g h o u t the r a m p u p a n d
steady p o r t i o n s of the trial b u t n o t d u r i n g the r a m p d o w n
p o r t i o n o f the trial. Some units (20 out o f 136) c o u l d be
reliably d i s c r i m i n a t e d only w h e n the t o r q u e was held
steady. As a result, 116 units ( n = 4 1 shortened, n = 3 3
neutral, n = 42 lengthened) were used to evaluate a n k l e
position o n m o t o r u n i t discharge rate at a n d shortly after
recruitment. All 136 units (n = 47 shortened, n = 44 neu-
214
]00
&
12
9O
MM741A
10
o_ 80
7O
8
Z
6
o
60
.~ 50
g
4O
0)
o
4
30
o9
n
~- 2o
2
10
0
0
o
2
4
6
8
10
I2
14
16
18
20
22
24
26
28
Fig. 4. Representative plot of a motor unit's instantaneous
discharge rate and corresponding dorsiflexion torque
during the ramp up-hold-ramp d o w n task. The unit was
recruited at 3.06 N.m of torque with a recruitment
discharge rate of 7.7 pps. The mean discharge rate during
steady-state torque was 17.9 pps at a mean torque of 7.48
N . m (40% MVC). The unit was derecruited at 5.23 N . m
of torque with a derecruitment discharge rate of 4.1 pps
3O
Time (s)
tral, n = 4 5 lengthened) were used to evaluate the effect
of ankle position on motor unit discharge at steady
torque. Ninety-five units ( n = 3 6 shortened, n = 2 7 neutral, n = 32 lengthened) were used to evaluate the effects
of ankle angle on motor unit discharge at derecruitment.
Mean maximum dorsiflexion torque was 20.0_+5.7
N.m for shortened, 4 5 . 1 + 5 . 9 N.m for neutral and
45.2+_5.7 N.m for lengthened muscle. In lengthened
muscle the contribution of the passive dorsiflexor elements accounted for 2.9 N.m of the 45.2 N.m measured.
In shortened muscle, 5.6 N.m of the 20.0 N.m measured
was needed to counteract the torque generated by the
passive elements of the plantar flexor muscles. The maximum voluntary dorsiflexor torque values were slightly
lower than those reported by Marsh and colleagues
(1981) which as estimated from their Fig. 3 were about
27 N.m, 46.5 N.m and 47 N.m for 70, 90 and 120 ~ angles,
respectively. These differences may be due to differences
in subjects, however, the device used by Marsh and colleagues may not have prevented the hip flexors from
generating some force that would have contributed to the
force measured by the load cell. Since our device was
referenced to the lower leg, activity of the hip or knee
muscles could not contribute to the forces measured by
the load cell.
Mean recruitment threshold for the 116 units
analyzed at recruitment were 1.6 N.m ( n = 4 1 ) , 4.6 N.m
( n = 3 3 ) , and 4.5 N.m ( n = 4 2 ) for muscle in shortened,
neutral and lengthened conditions, respectively. Recruitment threshold was significantly less in shortened muscle
than in either neutral length or lengthened muscle
unit was derecruited at 5.23 N.m of torque with a derecruitment discharge rate of 4.1 pps.
Mean durations for the recruitment ISI were 84.7 ms,
95.8 ms, and 123.4 ms for muscle in shortened, neutral
and lengthened conditions (Fig. 5A). An ANOVA revealed that the mean recruitment interval (ISI1) for
shortened muscle was significantly smaller than for
lengthened muscle ( P < 0.05). In lengthened muscle the
mean of ISI2 (130.1 ms) was significantly larger than ISI2
of neutral length muscle (102.2 ms) (P < 0.05). N o significant differences were found among muscle lengths for the
mean of ISI3 in the spike train (Fig. 5A).
180
160
Shortened In = 41 ur, ts)
Neutral (n = 33 unLts)
Lengthened (n = 4.2 units)
~, 140
m
Q_,
u3
Q
g
120
100
80
A
60
i
I
ISl 1
i
ISI 2
ISI 3
Interspike interval
160
ISI1 < 4 0 m s ( n = 16 units)
[
140
~
Ii~
120
lOO
(P<0.05).
.c
Motor unit discharge behavior
Instantaneous discharge rate and dorsiflexion torque
data are plotted against time from a representative trial
shown in Fig. 4. In this example, the unit was recruited
4.26 s into the trial at 3.06 N.m of torque with a recruitment discharge rate of 7.7 pps. Discharge rate then increased to 15.4 pps by 5.23 s at 3.82 N.m of torque, and
assumed a mean discharge rate of 17.9 pps at a steady
torque of 7.48 N.m, which represented 40% MVC. The
- ----- -- -
116 units
~
:~
B
60
40
20
0
f
~
Z
'
ISI1
-----
'
IS12
Shortened (n = 10 ualts)
Neutral In = 3 units)
Lengthened (n = 3 units)
'
ISI3
Intersplke ~nterval
Fig. 5A, B. Mean interspike intervals at recruitment (ISII), ISI2 and
ISI3 for three muscle lengths. Error bars are standard error of
mean. A Data from 116 units that were discriminated at recruitment. B D a t a from 16 units where ISI1 was less than 40 ms
215
15
co
~
i::J)
14
o Muscle shortened
A Muscle neutral
[] Muscle lengthened
u~
25~
13
co
~o
Fig. 6. M e a n i n s t a n t e n o u s firing rates o f m o t o r units a n d
m e a n c h a n g e in dorsiflexion t o r q u e f r o m recruitment for
ISI4-ISI9. E a c h d a t a point represents a different ISI;
m e a n a n d s t a n d a r d error o f m e a n for six ISis are s h o w n
for each condition. Slope o f linear regression lines:
2.94 p p s / N . m for s h o r t e n e d muscle, 2.62 p p s / N . m for
neutral muscle, a n d 1.00 p p s / N . m for lengthened muscle
12
t~._~
c ~
10
I
0.5
I
1.0
1.5
I
2.0
I
2.5
3.0
Mean change in dorsiflexion torque (Nm) from
recruitment for discharges 4 - 9
Further analysis of the data revealed that in some
spike trains, the initial interspike interval was brief, with
a duration of 40 ms or less. Brief recruitment ISis were
observed in 10 of 41 units (24%) in shortened muscle, in
3 of 34 units (9%) in neutral muscle, and in 3 of 43 units
(7 %) in lengthened muscle (Fig. 5B). A Chi-square analysis revealed that differences in the frequency of occurrence of brief initial interspike intervals were significant
across muscle lengths (P<0.05). In those spike trains
where the recruitment ISI was brief, IS[2 was quite long,
while ISI3 was shorter than ISI2 (Fig. 5B). Brief motor
unit discharges at recruitment were verified visually using
a fast sweep speed on an oscilloscope to confirm the
discharges had the same amplitude and shape, and were
therefore from the same unit.
Motor unit discharge behavior during the first 10
discharges was evaluated as torque was increasing during
the ramp up portion of the task. Motor unit discharge
rate typically increased rather rapidly during the first 10
discharges (interspike intervals decreased in duration).
and then the rate at which discharge increased slowed as
dorsiflexion torque continued to increase. Fig. 4 demonstrates this typical pattern of motor unit discharge behavior during increasing torque. Pooled results were based
on 99 units in which dorsiflexion torque increased at least
3 % MVC from the time of recruitment to steady torque.
Data on the mean instantaneous discharge rate for interspike intervals 4 to 9 at each muscle length are presented
in Fig. 6. The mean instantaneous discharge rate generally increased for intervals 4 through 9 as dorsiflexion
torque increased for all muscle lengths. A linear regression was used to describe the relationship between mean
instantaneous discharge rate and dorsiflexion torque for
ISI4 through ISI9. Mean discharge rate increased
2.94 pps/'N.m in shortened nmscle, 2.62 pps/N.m in neutral muscle, and 1.00 pps/N.m in lengthened muscle. The
slope of this relationship was significantly greater for
neutral versus lengthened muscle (P<0.05), and approached significance for shortened versus lengthened
muscle (P= 0.056). There were no significant differences
in the slope of the regression lines between shortened and
neutral muscle.
Data from 136 units were used to determine the effect
of muscle length on motor unit discharge rate during the
steady torque portion of the trial. The relationships between mean motor unit discharge rate and dorsiflexion
torque during the hold portion of the task are presented
for three muscle lengths in Fig. 7. When the tibialis
anterior muscle was shortened (ankle angle = 70~ motor
3O
o_ 25
v
g
Ankle angle = 70
20
15
~5 10
A
0
I
I
I
I
I
P
p
I
3O
25
AnKle angle = 90
O
2O
15
co 10
5
0
I
I
30
e~
25
Ankre angle = 120
20
c~
~w
~- 10
5
O
0
I
r
I
I
i
5
10
15
20
25
30
Dors~fiexlon torque (i'qm)
Fig. 7 A - C . Relationship between m e a n tibialis anterior m o t o r unit
firing rate at steady t o r q u e a n d dorsiflexion torque. A A n k l e at 70 ~
muscle shortened; Y = 11.09 + 0.67X. B A n k l e at 90 degrees, muscle
neutral; Y = I 0 . 6 5 + 0 . 3 4 X . C A n k l e at 120 ~ muscle lengthened:
Y = 11.09 + 0 . 2 7 X
216
240
,~ntISI
T~entISI
220
E
Muscle shortened
Z
200
(2r
180
g
160
O
Recrurtrrent
Derecrultment
~
140
I
120
A
100
0
80
I
I
I0
60
2
Q3
-40
Shortened
Neutral
Lengthened
Neutral length muscle
8
6
Muscle [ength
Derecrultment
F i g . 8. M e a n r e c r u i t m e n t a n d d e r e c r u i t m e n t i n t e r s p i k e i n t e r v a l s f o r
Recruitment
e~
units in shortened, neutral and lengthened tibialis anterior muscle.
Error bars are standard error of the mean
~2
P~
Bs
o
10
unit discharge increased 0.672 pps for each N . m increase
in steady state dorsiflexion torque. In the neutral condition, m o t o r unit discharge increased 0.338 pps for each
N . m torque increase, and in the lengthened condition
m o t o r unit discharge increased 0.271 pps for each N . m
increase in dorsiflexion torque. The slopes of the linear
regressions were significantly different (P < 0.05) a m o n g
muscle lengths and a contrast analysis revealed that the
slopes of the regressions lines were different between the
shortened and the lengthened conditions and between the
shortened and neutral conditions (P<0.05). No difference was found between neutral and lengthened conditions.
Ninety-five units were used to evaluate discharge
behavior when units ceased discharging during the decreasing ramp portion of the task. Mean derecruitment
interspike interval durations were 187.8 ms, 155.6 ms,
and 209.3 ms for units in shortened, neutral length, and
lengthened muscle. U n d e r all lcngth conditions, m o t o r
units ceased discharging with larger interspike intervals
than at recruitment (P<0.05), as shown in Fig. 8. At
each muscle length, the mean discharge rate during decreasing torque is less than during the increasing torque
portion of the trial in relation to dorsiftexion torque
(Fig. 9).
Discussion
The most important finding of our study was that muscle
length influences the discharge pattern of m o t o r units
during voluntary isometric r a m p contractions in h u m a n
tibialis anterior muscle. Although previous investigators
have reported that m o t o r unit recruitment threshold
changes with muscle length (Miles et al. 1986), the
present work represents the first attempt to study the
effects of muscle length on the discharge characteristics
of the m o t o r unit spike train during voluntary contractions in man. In addition to the influence of muscle length
at recruitment, m o t o r unit discharge per unit torque
(N.m) were influenced by muscle length as well. Wc can
E
Lengthened muscle
Z
8
o
6
o~
~-
)erecrultmen[
Recruitment
4
co
o
2
cE
0
0
i
I
i
i
i
i
5
10
15
20
25
30
35
Mean ~nstantaneous discharge rate (pps)
Fig. 9A-C. Plots of mean dorsiflexion torque and mean instantaneous firing rate for units during increasing ramp and decreasing
ramp portions of task demonstrating hysteresis of discharge rate.
A Muscle shortened; B muscle neutral: C muscle lengthened
summarize the results of the effects of muscle length on
m o t o r unit discharge characteristics as follows:
1. Discharge rate at recruitment is greater in shortened
muscle than in neutral or lengthened muscle.
2. The rate at which m o t o r units increase their discharge
rate is greater per unit torque (N.m) in neutral length
muscle than in lengthened muscle during increasing r a m p
contractions.
3. During steady-state torque contractions, m o t o r units
in shortened muscle exhibit a greater increase in discharge rate per unit torque (N.m) than do m o t o r units
in neutral length or lengthened muscle.
hlitial discharge pattern
The mean recruitment ISI was found to be significantly
less in shortened nmscle than lengthened muscle. In addition, the initial discharge pattern (first 3 interspike intervals) in shortened muscle more frequently exhibited a
brief initial ISI followed by a much longer ISI (Fig. 5b),
a pattern which may assist in taking up slack in the
passive elements of muscle (Stein and Parmiggiani 1979).
Our finding indicates that the central nervous system
may adjust the output of alpha motoneurons to dis-
217
charge more rapidly at recruitment when muscle is shortened, to accommodate for the physical properties of the
shortened muscle at rest. The influence of the contractile
state of the muscle upon the motor neuron pool has been
reported previously for fatigue conditions (BiglandRitchie et al. 1986).
Our finding that muscle length influences the initial
discharge pattern of motor units is further evidence that
the control mechanisms for generating muscle force is
much more complex than previously thought. Previous
studies of anisometric tasks demonstrated a difference in
motor unit activity in shortening, lengthening, and isometric contractions (Andrew 1985: Tax et al. 1990, Kato
et al. 1985). Although these previous studies clearly
showed that discharge patterns differ depending upon the
task, the change in joint angle and subsequently muscle
length was quite small during shortening and lengthening
contractions. Therefore, differences in discharge patterns
that these investigators reported were likely due to the
type of contraction rather than a change in muscle
length. Our findings complement the work of these
previous studies and indicate that muscle length is yet
another variable which may influence motoneuron output during voluntary contractions.
Discharge pattern durin9 increash~g and steady-state
torque
During the ramp-up portion of the isometric contraction,
motor unit discharge rate increased more rapidly in neutral muscle than in lengthened muscle (P < 0.05) (Fig. 6).
The rate of increase in shortened muscle (2.94 pps/N.m)
was similar to neutral muscle and although appeared to
be greater than the rate of increase in lengthened muscle,
was not statistically different (P> 0.05). The smaller increase in discharge rate for lengthened muscle may be
due to the passive elements of the lengthened muscle
being on stretch. In neutral and shortened muscle, the
passive elements are slack and a portion of the active
tension is needed to pull the series elastic elements of the
muscle and tendon taut before the remainder of the
active tension is transmitted to the limb segment. Hence,
motor unit discharge rate increased more rapidly in neutral and shortened muscle for each N.m of torque measured at the distal limb segment during the increasing
ramp.
To maintain a given torque level, higher discharge
rates were required when the muscle was shortened as
compared to neutral or lengthened muscle (Fig. 7A-C).
This finding is likely related to the shift found in the
stimulus rate-tension curves reported by Rack and Westbury (1969) and Marsh et al. (1981). The shift of the
stimulus rate-tension curve to the right in shortened
muscle (Fig. 9a from Rack and Westbury 1969; Fig. 5b
from Marsh et al. 1981) predicts a greater discharge rate
would be required to generate the same amount of torque
in shortened muscle as compared to neutral or lengthened tibialis anterior during isometric contractions.
The lack of statistical differences in discharge rate at
steady torque between neutral and lengthened muscle
would suggest that as the ankle moves from 90~
~
torque-discharge rate relationship is not changed sub-
stantially, which is predicted by the torque-stimulation
rate curves from Marsh et al. 1981. Herring et al. (1984)
proposed that sarcomere number in muscle is regulated
so that the sarcomere's actin and myosin overlap is optimal when the muscle is most active during functional
activity. In untrained individuals, the tibialis anterior is
used functionally most frequently during gait to control
ankle plantar flexion eccentrically after heel strike, and
to bring the ankle into a position of slight dorsiflexion
after toe off and during the swing phase of gait (Sutherland et al. 1980). The tibialis anterior muscle is also active
to help maintain upright posture in standing by responding to small increments of postural sway to keep the
center of gravity over the base of support (Nashner
1976). Most activity then occurs when the ankle is in a
range of 85~
~ An ankle angle of 70~ the fully shortened position used in this experiment, puts the muscle
in a rather severely shortened condition which would be
to the left of the plateau in the angle-torque curve, and
may put the muscle fibers to the left of the plateau in the
length-tension curve. Finnochio and Luschei (1985) have
suggested that the plateau of tension in the lengthtension curves of muscle operating submaximally may
not be as broad as reported for tetantic contractions.
Hence, in a shortened condition during a submaximal
contraction, the muscle may be operating on a portion
of the length-tension curve where a much higher motor
unit discharge rate is necessary to produce the same
amount of force as in a neutral or lengthened condition.
The higher discharge rate per unit torque during steady
state contractions at 20~ dorsiflexion would support Finnochio and Luschei's hypothesis.
Discharge behavior during decreasing torque
The finding that the derecruitment interspike interval
was longer that the recruitment interspike interval at all
muscle lengths was not unexpected (Fig. 8). This hysteresis effect of motor unit discharge rate at recruitment
and derecruitment has been reported previously (Clamann 1970; Person and Kudina 1972; Milner-Brown
et al. 1973). Because mean torque at derecruitment was
not significantly different from mean torque at recruitment for each muscle length, the relationship between
motor unit discharge rate and torque also exhibited a
hysteresis (Fig. 9). After the ramp up-steady contraction,
the amount of torque during decreasing ramp and at
derecruitment was maintained with significantly lower
discharge rates than at and shortly alter recruitment.
This difference was observed at all muscle lengths.
A similar hysteresis between force and stimulation
rate has been reported in cat soleus and medial gastrocnemius motor units by Binder-Macleod and Clamann (1989). The authors suggested that this hysteresis
effect may be explained by a time-dependent rate of
tension development or by the catch-like property that
has been described in mammalian muscle (Burke et al.
1970). The task in the present experiments was to linearly
increase and decrease dorsiflexion torque, while motor
unit discharges were recorded. The rapid increase in
discharge rate per unit torque (N.ln) at recruitment ob-
218
served in the present experiments m a y take a d v a n t a g e of
similar catch-like properties in h u m a n muscle d u r i n g
voluntary contractions.
Funclional implications
It might be argued that the differences in discharge rates
per u n i t t o r q u e i m m e d i a t e l y after r e c r u i t l n e n t (Fig. 6)
a n d d u r i n g steady state c o n t r a c t i o n s (Fig. 7) merely represent a scaling o f these variables due to differences in
n a a x i m u m force g e n e r a t i n g capabilities at differcnt muscle lengths, a n d are therefore n o t significant findings.
W h e n the results o f Figs. 6 a n d 7 are plotted as a f u n c t i o n
of percent m a x i m u m t o r q u e for a given length, differences in discharge rate per u n i t torque are n o t evident.
We argue, however, that such n o r m a l i z a t i o n belies the
m a i n f u n c t i o n a l i m p l i c a t i o n s o f this study. These results
m u s t be viewed in terms of the d e m a n d s placed u p o n the
n e r v o u s system w h e n a muscle is activated at different
lengths. F r o m a m o t o r c o n t r o l s t a n d p o i n t , the n e r v o u s
system m u s t deliver the a p p r o p r i a t e signals to the m o t o n e u r o n s for d e v e l o p i n g the forces required in a specific
task. A h y p o t h e t i c a l example best illustrates this point.
A m o t o r task m i g h t require the p r o d u c t i o n o f 5 N . m
of steady dorsiflexion t o r q u e a n d the task m i g h t be d o n e
in 20 ~ ankle dorsiflexion (muscle shortened) versus 30 ~ o f
p l a n t a r flexion (muscle lengthened). P r o d u c t i o n of the
desired t o r q u e will require different m e a n discharge rates
from m o t o r units for the two j o i n t positions. F r o m the
regression e q u a t i o n s of Fig. 7, these rates would corresp o n d to 14.44 a n d 12.44 pps for s h o r t e n e d versus lengthened c o n d i t i o n s respectively. I n a d d i t i o n , to overcome
the slackened passive elements of a s h o r t e n e d muscle,
m o t o r units would be expected to be recruited at faster
discharge rates a n d to increase their discharge at faster
rates as torque increased. F r o m the s t a n d p o i n t o f the
nervous system, a greater drive to the m o t o n e u r o n pool
is required to p r o d u c e the same 5 N . m o f t o r q u e for the
shortened versus l e n g t h e n e d c o n d i t i o n . If the task required 25 N . m o f torque, the lower force g e n e r a t i n g
capabilities o f the s h o r t e n e d muscle w o u l d result in failure to p e r f o r m the task n o m a t t e r how h a r d the n e r v o u s
system a t t e m p t e d to drive the m o t o n e u r o n s . F r o m a
f u n c t i o n a l s t a n d p o i n t therefore, the n e r v o u s system m u s t
adjust to a n d provide for a different level o f m o t o n e u r o n
c o n t r o l w h e n faced with a s h o r t e n e d versus l e n g t h e n e d
muscle.
Acknowle@ement~. This work was supported in part by a grant
from the Foundation for Physical Therapy and by NIH grant
NS2499l.
References
Andrew PD (.1985) Motor unit activity under low tensions as muscle
changes length. Am J Phys Med 64:235 254
Bellemare F, Woods JJ, Johansson R, Bigland-Ritchie B (1983)
Motor-unit discharge rates in maximal voluntary contractions
of three human muscles. J Neurophysiol 50:1380 1392
Bigland-Ritchic B, Johansson R, Lippold OCJ, Smith S, Woods JJ
(1983) Changes in motoneurone firing rates during sustained
maximal voluntary contractions. J Physiol 340:335 346
Bigland-Ritchie BR, Dawson NJ, Johansson RS, Lippold OCJ
(1986) Reflex origin for the slowing of motoneurone firing rates
in fatique of human voluntary contractions. J Physiol
379:45l 459
Binder-Macleod SA, Clamann HP (1989) Force output of cat motor
units stimulated with trains of linearly varying frequency.
J Neurophysiol 61:208 217
Brunnstrom S (1966) Clinical kinesiology, 2nd edn. FA Davis Company, Philadelphia
Burke RE, Rudomin P, Zajac FE (1970) Catch property in single
mammalian motor units. Science 168 : 122-124
Burke RE, Rudomin P, Zajac FE (1976) The effect of activation
history on tension production by individual muscle units. Brain
Res 109:515 529
Clamann HP (1970) Activity of single motor units during isometric
tension. Neurology 20 : 254-260
Clamann HP, Lamb RL (1976) A simple circuit for filtering single
unit action potentials from elcctrograms. Physiol Behav
17:149-151
Enoka RM, Robinson GA, Kossev AR (1988) A stable, selective
electrode for recording single motor unit potentials in humans.
Exp Neurol 99:761-764
Finocchio DV, Luschei ES (1985) Active length-tensioncharacteristics of naturally activated muscle: measurement and implications. In: Titze IR, Scherer RC (eds) Vocal fold physiology.
Demer Center for tile Performing Arts, Denver, pp 283-291
Herring SW, Grimm AF, Grimm BR (1984) Regulation of sarcomere number in skeletal muscle: a comparison of hypotheses.
Muscle Nerve 7:161-173
Kato M, Murakami S. Yasuda K (1985) Behavior of single motor
units of human tibialis anterior muscle during voluntary shortening contraction under constant load torque. Exp Neurol
90 : 238-253
Marsh E, Sale D, McComas AJ, Quinlan J (1981) Influence of joint
position on ankle dorsiflexion in humans. J Appl Physio1:160 167
IVliles TS, Nordstrom MA, Turker KS (1986) Length-related
changes in activation threshold and wave form of motor units
in human masseter muscle. J Physiol 370:457-465
Milner-Brown HS, Stein RB, Yemm R (1973) Changes in firing rate
of human motor units during linearly changing x.oluntary contraction. J Physiol 230:371-390
Nashner LM (1976) Adapting reflexes controlling the human posture. Exp Brain Res 26:59 72
Nelson RM, Soderberg GL (1983) Laser etched bifilar fine wire
electrode for skeletal muscle motor unit recording. Electroencephal Clin Neurophysiol 55:238 239
Partridge LD, Benton LS (1981) Muscle, the motor. In: Brooks VB
(ed) Handbook of physiology, Sect I. The nervous system, Vol
II. Motor control, Part 1. Am Physiol Soc, Bethesda MD, pp
43-106
Person RS, Kudina LP (1972) Discharge frequency and discharge
pattern of human motor units during voluntary contraction of
muscle. Electroencephal Clin Neurophysiol 32 : 471 ~483
Rack PMH, Westbury DR (1969) The effects of length and stimulus
rate on tension in the isometric cat soleus muscle. J Physiol
204:443 460
Stein RB, Parmiggiani F (1979) Optimal motor patterns for activating mammalian muscle. Brain Res 175:372 376
Sutherland DH, Olshen R. Cooper BA. Woo SLY (1980) The
development of mature gait. J Bone Joint Surg 62A: 336-353
Taniji J. Kato M (1973) Firing rate of individual motor units in
voluntary contraction of abductor digiti minimi muscle in man.
Exp Neurol 40:771 783
Tax AAM, Denier van der Gon JJ. Gielen CCAM, Kleyne M (1990)
Differences in central control of m. biceps brachii in naovement
tasks and force tasks. Exp Brain Res 79:138-142
Wickiewicz TL, Roy RR, Powell PL, Edgerton VR (1983) Muscle
architecture of the human lower limb. Clin Orlho Rel Res
179 : 275-283
Zajac FE, Young JL (1980) Properties of stimulus trains producing
maximum tension-ttme area per pulse from single motor units
in medial gastrocnemius muscle of the cat. J Neurophysiol
43:1206 1120