 |
INTRODUCTION |
Much work on simian head
movement has focused on orienting behaviors (e.g., Crawford et
al. 1999
; Freedman and Sparks 1997) in an effort
to extend work done on saccadic eye movements in head-restrained
preparations. Studies of the saccadic system usually employ saccades
themselves as an index of the underlying neural events and kinetics
because the mechanics of this system are relatively intuitive. In
contrast, head movements are generated by more than two dozen muscles
operating on a complex multiarticular linkage endowed with substantial
inertia (Richmond and Vidal 1988; Winters 1988). The design and interpretation of experiments on such a system requires knowledge of its structural elements, which is provided
in the companion paper on muscle morphometry (Richmond et al.
2001).
Chronically indwelling electrodes in animals provide the opportunity to
assess reliably electromyographic (EMG) activity in muscles not
accessible in humans. Studies in human neck muscles have generally
relied on surface EMG from large, relatively superficial muscles
(Dee and Zangemeister 1998; Hannaford et al.
1985; Keshner et al. 1989;
Mayoux-Benhamou and Revel 1993; Mayoux-Benhamou
et al. 1997; Zangemeister et al. 1982) or
percutaneous EMG from only one or two muscles whose identities can be
hard to determine (Mayoux-Benhamou et al. 1995). A
broader range of muscles have been studied previously in cats
(Keshner 1994; Keshner et al. 1992;
Richmond et al. 1992; Thomson et al. 1994,
1996; Wilson et al. 1983), but the differing features of feline head-neck structure potentially limit the
applicability of these results to primate studies. Previous studies in
monkeys have recorded neck muscle EMG in a few neck muscles or during a
restricted subset of head movements (Bizzi et al. 1971;
Le Goff et al. 1992; Lestienne et al. 1995,
2000). However, little systematic study to date has focused on
the relationships between simian head kinematics and neck muscle
activation during a more extensive sampling of neck muscles and head movements.
In this study, the spatial and temporal patterns of EMG activity were
examined in a large number of neck muscles in monkeys free to move
their heads. Head postures and movements were generated either in a
trained protocol requiring gaze shifts to visual targets or were
generated spontaneously during orienting, tracking, feeding, expressive, and head-shaking behaviors. The inclusion of the latter head movements increased the range of kinematic patterns beyond those
which accompanied trained gaze shifts.
Some results have been reported previously in abstract form
(Corneil et al. 1996, 1999).
| |
METHODS |
Surgical and training procedures
Four male monkeys (Macaca mulatta) weighing 5.4-9.1
kg were used in these experiments according to procedures approved by the Queen's University Animal Care Committee and the guidelines of the
Canadian Council on Animal Care. Each monkey underwent two surgeries.
In both, anesthesia was induced with ketamine hydrochloride and
maintained with isoflurane. Antibiotics were administered pre- and
postoperatively, and anti-inflammatories and analgesics were
administered postoperatively.
In the first surgery, eye coils were implanted subconjunctivally
(Judge et al. 1980) to monitor gaze (eye-in-space)
position (Fuchs and Robinson 1966), and a head post was
attached to the skull by way of a dental acrylic pedestal that also
held the leads and connectors. Monkeys were trained on oculomotor tasks
(see following text) prior to the second surgery. In the second
surgery, chronically indwelling EMG electrodes were implanted in neck
muscles using a similar approach to that described in cats
(Richmond et al. 1992). Muscle layers were separated
from the dorsal midline raphe to gain access to the cleavage planes
between muscles. Up to 12 muscles in each monkey were implanted using
bipolar epimysial patch electrodes or bipolar intramuscular hook
electrodes (Table 1). Full details of the
electrode design have been described previously (Loeb and Gans
1986). In both, the recording contacts were 3 mm long,
separated by ~3 mm and were oriented perpendicularly to the long axis
of the muscle fiber fascicles. Some muscles implanted with
intramuscular hook electrodes were shielded from the potential cross-talk of adjacent muscles by suturing Silastic sheeting to the
overlying fascia. Muscle layers were approximated with a midline closure. A ground wire consisting of a single, partially bared loop of
Teflon-coated, multistranded stainless steel was stitched to
subcutaneous fascia. The leads from all implanted electrodes were
tunneled subcutaneously to the acrylic skull pedestal and soldered to
multipin connectors. By the second or third postoperative day, all
animals appeared to be making normal head movements.
Experimental procedures
Prior to EMG recording, the monkey was placed in a primate
restraining chair designed to permit unrestrained head movements. Monkeys l and f were placed in a
commercially available primate chair (Crist Instruments), modified to
allow attachment of a body harness which permitted approximately
±45° of trunk rotation in the horizontal plane. Monkeys
z and r were placed in a custom-made primate
chair that permitted the monkeys to be tethered to the chair via a
customized primate vest (Lomir Biomedical). This arrangement was more
effective at preventing trunk rotation (estimated to be ±10°)
without restraining the head or neck. We saw no evidence that patterns
of EMG activities differed in the two chairs, but a wider range of the
head positions was typically achieved when using the custom-made chair.
Spontaneous sessions were recorded in monkeys
l, f, and z in which volitional head
movements were generated during a variety of orienting, tracking,
feeding, or expressive behaviors in a well-lit room. Experimenters in
the room encouraged head movements by displaying food, verbalizing, or
hand-waving; desired movements were rewarded with food and verbal
praise. The animals appeared to behave normally and showed no signs of stress.
Monkeys z and r worked in trained
sessions in a dark, sound-attenuated room, performing an
oculomotor task in which they had been trained previously. Stimuli
consisted of 60 light-emitting diodes (LEDs) arranged at the front and
both sides of the monkey, spanning 90° to the right and left of
center, and 45° above and below center. To receive a liquid reward,
the monkey had to fixate the central LED for 
500 ms, look to one of
eight randomly selected peripheral LEDs illuminated as the central LED
was extinguished, fixate the peripheral LED for 500 ms, look back to
the central LED illuminated as the peripheral LED was extinguished, and
fixate the central LED for 500 ms. The locations of the eight
peripheral LEDs were varied between blocks, allowing a considerable
range of head postures and centrifugal (away from midline) and
centripetal (toward midline) movements to be obtained over several days.
Data collection and analysis
A flexible ribbon-cable that did not interfere with head
movements linked the EMG connector(s) to the signal processing
electronics. The signal-processing architecture used for
monkey l was different from that used for the
other monkeys. Briefly, all EMG signals for monkey
l were amplified differentially, bandwidth filtered (100-5,000 Hz) and recorded on an FM tape recorder. Segments of interest were rectified and integrated into 3.3-ms bins and digitized at 3.3 ms (see Thomson et al. 1994 for details).
Sessions for monkey l were videotaped at 60 fields/s using two shuttered cameras placed above and to one side of
the monkey, allowing estimation of the position of the head on the body.
For monkeys f, z, and r,
digitized signals of the EMG activity and the gaze (eye-in-space) and
head (head-in-space) positions derived from the magnetic coil system
were recorded simultaneously. The coil system (CNC Engineering) yokes
the two horizontal fields together, hence the relationships between
induced current and horizontal coil position was linear over a range of
±90° from center. The search coil and a tube for the fluid reward
were secured to the head pedestal and did not interfere with normal
head movements or vision. The EMG ribbon-cable led to preamplifiers and
low-pass filters (MAX274 integrated IC filter,
fc = 8 kHz, roll-off = 24 dB/octave,
Maxim Electronics) that filtered out the coil frequencies. Data were
then fed into an Analog Preprocessor and Timer (Aztec Associates) that
enabled computer-programmable amplification, filtering (100- to
5,000-Hz bandwidth), rectifying and digitizing of the signals into 2-ms
bins. The amplification was adjusted for different channels to yield a
maximal peak-to-peak output voltage of ~5 V. The amplitudes reported
here correspond to raw EMG signals with ~10 times larger peak-to-peak
amplitudes (Bak and Loeb 1979). Bins with amplitudes 
5
µV were assumed to represent noise; this value was generally used as
a limit below which the muscle was classed as silent and data points
were excluded from correlation analyses. The EMG signals and the
horizontal and vertical gaze and head position signals were digitized
at 500 Hz with a Pentium computer running a real-time data acquisition
system (REX version 5.4) (Hays et al. 1982). The
behavioral sessions for monkeys f, z,
and r were videotaped using a single infra-red camera in conjunction with an infra-red light source, synchronized by a software-controlled trial counter placed in the camera's field of view.
Behaviors of interest from spontaneous sessions (monkeys
l, f, and z) were selected by
inspecting the videotaped sessions. In the trained sessions
(monkeys z and r), analyses were
performed only on correct trials. Segments or trials in which the
monkey adopted a posture unsuitable for these experiments (i.e., head resting on neck plate or torso twisted away from the frontal plane) were not analyzed. Head velocity and acceleration traces were obtained
by differentiation and double-differentiation, respectively, of
position signals. The onsets and offsets of orienting head movements
were determined when the head velocity crossed a 10°/s threshold. The
onset and offset of translating or torsional movements were determined
by examining the videotaped sequences along with the head position
signals. The "zero" position to which both horizontal and vertical
planes were referenced occurred when the monkey's head was pointed
straight ahead without inclination or declination, such that the search
coil mounted on the head, and hence the frontal plane of the head, was
oriented parallel to the front panel of the coil frame.
When appropriate, the temporal aspects of the EMG signals in some
muscles were determined by quantifying the time of activation and
silencing. Although no strict quantitative criteria were used, sudden
changes in EMG activity levels were easily delineated. To quantify the
magnitude of the EMG response during stable postures, the EMG signal
was averaged over the duration of the postural segment. To quantify EMG
response magnitude during head movements, the EMG signal was first
smoothed with a 50-ms running average (i.e., ±25 ms from the point
under consideration), selected because it approximates the dynamics of
muscle-force development and reduces the variability arising from the
stochastic nature of the EMG signal (Loeb and Gans
1986). EMG response magnitude was quantified by taking the peak
of this smoothed signal over a period ranging from 75 ms before
movement onset to the time of peak head velocity to capture an estimate
of the EMG contributing to the accelerational torques required to
produce the movement. For very transient EMG signals during rapid head
movements in which this method was inappropriate (e.g., bursts in Fig.
6), the magnitude of the EMG signal was obtained by integrating the
area under the EMG curve, without smoothing.
Overall, we analyzed a total of 1,053 postural segments (110 from
monkey l, 67 from monkey f,
670 from monkey z, 206 from monkey
r), 2,311 movement segments (359 from monkey
f, 1,811 from monkey z, 131 from
monkey r), and 70 complex segments composed of
multiple movement phases (i.e., head shakes, translations, feeding
behaviors, etc.; 57 from monkey f, 13 from
monkey z). Movement data from monkey
l were not analyzed quantitatively, but movements were
inspected to ensure that the EMG patterns were qualitatively similar to
results from the other monkeys. For the sake of simplicity given the
number of recorded muscles, analyzed data from a given muscle are
presented only if that muscle served as an agonist or antagonist to the
posture or movement in question.
| |
RESULTS |
All monkeys held a variety of postures and generated head
movements during either spontaneous (monkeys l,
f, and z) or trained (monkeys
z and r) sessions. It was impossible to obtain
perfectly comparable records from monkey to monkey because different
muscles were implanted (Table 1), but consistent patterns of neck
muscle activation were observed across multiple segments and monkeys. To illustrate this point, most figures display multiple postures or
movements from monkey z and one of either
monkey f or r. Small differences were
often seen in the activities of muscles from one sequence to the next;
where appropriate, we identify idiosyncratic observations to
distinguish them from the more general patterns. We present EMG
patterns associated with postures and movements in the horizontal
plane, postures and movements in the vertical plane, and during forward
translations or vigorous head shakes.
EMG activity during postures or movements in the horizontal plane
TURNED POSTURES.
All monkeys frequently held their heads in turned postures. The
eccentricity of head postures analyzed here ranged up to ±70° from
center for monkey z, ±60° for
monkey f, ±50° for monkey
r, and ±40° for monkey l. EMG
activity was negligible in all recorded muscles when the head was held
at the central "zero" position and when the head was turned away
from the preferred direction of the muscle. Modestly turned postures
<20° from center were associated with consistent activity only in
the ipsilateral suboccipital muscles obliquus capitis inferior (OCI)
and rectus capitis posterior major (RCP maj) (Fig.
1, A and B,
3rd and 4th columns). Larger turned postures
~20-50° from center were associated with stronger activation in
the ipsilateral suboccipital muscles, and a lower level of activity in
ipsilateral splenius capitis (SP cap; Fig. 1, A and
B, 2nd and 5th columns). These same
ipsilateral muscles were activated strongly in extreme postures >50°
from center, and sternocleidomastoid (SCM) contralateral to the side of
turning also became active (Fig. 1, A and B, 1st
column). The EMG activity during turned postures was usually <50
µV/bin for the suboccipital muscles and <20 µV/bin for the larger
muscles, although these values were exceeded at extreme postures.
View larger version (36K):
[in this window]
[in a new window]
|
Fig. 1.
Electromyographic (EMG) patterns in 4 muscles whose activity varied
with different turned postures. Records are obtained from trained
sessions in monkey z (A)
and spontaneous sessions in monkey f
(B) over a range of postures turned to the left or
right. Each record is 500 ms long. SP cap, spleniud capitis; OCI,
obliquus capitis inferior; RCP maj, rectus capitis posterior major;
SCM, sternocleidomastoid. The R or L in front of each muscle
abbreviation denotes the right or left muscle, respectively.
|
|
We computed the mean level of EMG activity during stable postures from
either trained sessions (monkeys z and
r) or spontaneous sessions (monkeys l
and f). Figure 2A
illustrates the results of this analysis for monkey
z. Some variability in the EMG levels in Fig. 2A
may be related to small trunk or shoulder rotations or torsional head
rolls, which could alter the head-re-body position without altering the
position of the head coil in space. The suboccipital muscles (Fig.
2A, middle 3 columns) were recruited consistently when the
head was held only a few degrees away from center. Activation in
ipsilateral SP cap (Fig. 2A, rightmost column) and
contralateral SCM (Fig. 2A, leftmost column) was observed in
progressively more eccentric head positions. Similar patterns of neck
muscle activation were observed in the other monkeys (Fig.
2B).
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 2.
Correlation of mean EMG activity with eccentricity of turned posture.
A: data for 5 muscles from monkey
z from trained sessions: R-SCM, L-RCP maj, L-OCI, R-OCI,
and R-SP cap. Each point in A plots the
level of EMG activity averaged over a single stable postural segment
(duration range, 250-950 ms; mean, 410 ms) as a function of the
eccentricity of the turned posture. Points with mean EMG values <5
µV are not shown since such values were associated with negligible
muscle recruitment.  , in A and B,
statistically significant (P < 0.05) linear
regression lines for the points shown. B: composite plot
of the regression lines for muscle activities related to the
eccentricity of turning for all monkeys using data obtained from
trained sessions (monkeys z and
r) or spontaneous sessions (monkeys
l and f). The regression lines for
muscles related to leftward turned postures have been flipped to allow
for comparison with muscles related to rightward turned postures so
that positive values on the abscissa represent the preferred direction
of the muscle. Only turned postures in which the head posture in the
vertical plane was within ±10° of vertical center were included.
|
|
TURNING MOVEMENTS.
Muscle recruitment during trained sessions.
The same muscles active during turned postures (OCI, RCP maj, SP cap,
and SCM) were recruited synchronously ~10-50 ms before head
movements during trained gaze shifts (Fig.
3). Suboccipital muscles OCI and RCP maj
were active during small turns (Fig. 3, A and
C), and the larger muscles SP cap and SCM were
additionally active during larger turns (Fig. 3, B and
D). Occasionally, this initial activation was a phasic
burst followed by lower tonic levels typical of the posture held at the
end of the movement (e.g., Fig. 3, A and
C); in other segments, no distinct phasic burst was
apparent (Fig. 3, B and D). Antagonist
muscles whose mechanical turning actions were away from the turn were
essentially not recruited (Fig. 3).
View larger version (40K):
[in this window]
[in a new window]
|
Fig. 3.
Horizontal head position (Hh) and velocity (dHh) traces, and EMG
patterns from monkey z (A
and B) and monkey r
(C and D) during 4 leftward turning
movements. All data are obtained from head movements generated during
trained sessions. Thick solid lines on the EMG traces denote the
smoothed EMG activity averaged over ±25 ms (see METHODS);
peak levels of EMG activity for later analyses are derived from this
smoothed activity. Vertical dashed lines in each figure delineate the
onset and offset of each turn.
|
|
Centrifugal head movements generated during trained sessions were
characterized by close linear correlations between movement amplitude,
peak velocity and peak acceleration (Fig.
4, A and B
for monkey z, C and
D for monkey r). The
relationship between amplitude and velocity, the head "main
sequence" (Fig. 4, A and C), is
commonly observed during head movements generated during orienting gaze
shifts (Freedman and Sparks 1997;
Zangemeister et al. 1981). These kinematic relationships
were reflected in the close linear relationship of the EMG activities
of agonist muscles to the peak velocity (and hence also to the
amplitude and acceleration) of the head turn (Fig.
5). The sequence of muscle recruitment
was qualitatively similar to that observed for progressively larger
turned postures: suboccipital muscles were preferentially active during
small slow turns, whereas larger, multiarticular muscles became active
additionally during larger faster turns (monkey
z: Fig. 5, A-F, monkey
r: G).
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 4.
Relationship of the kinematics of head turns associated with trained
centrifugal, horizontal gaze shifts in monkeys
z (A and B) and
r (C and D).
A and C correlate peak velocity to
movement amplitude; B and D correlate
peak acceleration to peak velocity. The different symbols denote
various head movement amplitudes, as indicated in the key. Solid lines
denote statistically significant linear regression lines
(P < 0.05). Only trials in which the initial head
position was within 10° of center and in which the movement deviated
<20° of angular deviation from horizontal contributed data to this
figure.
|
|
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 5.
Relationship of EMG activity to peak head velocity during trained
centrifugal, horizontal gaze shifts in monkeys
z (A-F) and
r (G). Data obtained from the same set of
movements as in Fig. 4, using the same symbol key to denote head
movement amplitude. Solid lines denote statistically significant
(P < 0.05) linear regression lines.
A-E plot the peak smoothed EMG values for each trial
vs. peak horizontal head velocity. The EMG activity was smoothed as
shown in Fig. 3, and the peak EMG activity was taken from the period
extending 75 ms before movement onset to the time of peak head velocity
(see METHODS). The regression lines for each muscle are
summarized in F for monkey
z, and in G for monkey
r.
|
|
Muscle recruitment during spontaneous sessions.
Untrained head movements generated during spontaneous sessions
(monkeys l, f, and z)
exhibited a broader range of kinematics, with more dissociation between
amplitude, velocity and acceleration, permitting us to identify
spatiotemporal features of EMG recruitment associated with movements of
varying kinematics. The 16 movements displayed in Fig.
6 were all obtained from spontaneous
sessions and contrast the activation of agonist and antagonist OCI
muscles. This figure is organized to facilitate the comparison of
movements with similar amplitudes but different dynamics for two
different monkeys. The EMG activation for the regularly paced movements in Fig. 6A was qualitatively similar to that which
accompanied head movements during trained gaze shifts: agonist activity
was present in the absence of antagonist activity. In contrast, during the very abrupt movements in Fig. 6B that depart markedly
from the kinematic relationships shown in Fig. 4, EMG activities had reciprocally phasic profiles in which the initial phasic burst of the
ipsilateral agonist muscle was followed some 30-50 ms later by a
phasic burst in the contralateral antagonist muscle. Commonly, the
reciprocal agonist and antagonist bursts adopted a triphasic or
multiphasic profile of multiple bursts as the head settled into the
turned posture (e.g., 2nd to 4th columns in Fig.
6B, 1st and 3rd columns of
Fig. 6D). These subsequent bursts were not easily related to
movement kinematics.
View larger version (41K):
[in this window]
[in a new window]
|
Fig. 6.
Horizontal head position (Hh), velocity (dHh) and acceleration (ddHh)
traces, and EMG patterns for 8 movements from each of
monkey z (A and
B) and f (C and
D) during regularly paced (A and
C) or brisk horizontal (B and
D) movements. Same format as Fig. 3. All 16 head
movements are derived from spontaneous sessions. The terms
"regularly-paced" or "brisk" refer to head movements that are
either kinematically similar or very different, respectively, to the
trained head movements shown in Fig. 4. The figure is organized to
enable comparison of approximately amplitude-matched movements. Thus
the movements in A are comparable to the movements in
the same column in B for monkey
z, as are the movements in the respective columns of
C and D for monkey
f. Note the differences in the phasic nature of the
profiles in B and D compared with the
respective movements in A and C. Note as
well that the time scale can differ for each movement, as do the scale
bars for the kinematic and EMG signals.
|
|
The magnitude of the initial agonist and antagonist bursts related well
to the dynamics of the head movement. Figure
7 quantifies the area under the EMG burst
in the agonist and antagonist OCI muscles as a function of the peak
acceleration of the head during spontaneous sessions compared with the
accelerations observed during regularly-paced movements (Fig. 7, gray
regions). For each vertical line in Fig. 7, the lengths above and below
the dashed zero line denote the magnitudes of the agonist and
antagonist bursts, respectively, for a given movement. Large bursts in
the antagonist muscle only accompanied head movements that accelerated faster than regularly paced head movements and increased progressively in magnitude for movements accelerating at more than
~3,000°/s2. Not surprisingly, agonist muscle
activity accompanied all head movements and increased in magnitude for
faster movements. Similar relationships between agonist and antagonist
burst magnitudes and head acceleration were observed in RCP maj, SP
cap, and SCM.
View larger version (32K):
[in this window]
[in a new window]
|
Fig. 7.
Relationship between the integrated magnitude of the agonist and
antagonist muscle activity to the peak acceleration of the head. All
data are obtained from spontaneous sessions and are divided into
movements between 10 and 40° left from monkey
z (A), 40-70° left from
monkey z (B), and
10-30° right from monkey f
(C). Each solid vertical line represents data from a
single head movement, plotted as a function of the peak head
acceleration. The length above the dashed horizontal line expresses the
integrated magnitude of the agonist burst, and the length below the
dashed horizontal line expresses the integrated magnitude of the
antagonist burst. The magnitude of the agonist muscle activity was
calculated by integrating the area under the EMG signal from 50 ms
before to 50 ms after the time of the peak agonist burst. The magnitude
of the antagonist muscle activity was calculated by integrating the
area under the EMG signal over 100 ms after the time of the peak
agonist burst (see METHODS). The gray shaded boxes denote
the range of accelerations typical for head movements during trained
gaze shifts (the box in C was derived from data from
monkey r). Note that significant
antagonist activity is seen only for movements with accelerations much
larger than the gray region, implying that antagonist activity
accompanied only particularly brisk movements.
|
|
Centrifugal versus centripetal turning movements.
The spatiotemporal patterning of neck muscle activation varied
not only with the size and dynamics of the turn, but also with the
initial position of the head. Figure
8 compares
representative centrifugal (monkey z: Fig. 8,
A and B; monkey f:
E and F) and centripetal (monkey
z: Fig. 8, C and D; monkey
f: G and H) head turns, matched
approximately for size and speed but starting from different initial
head positions. Synchronous activation in ipsilateral agonist muscles
preceded the onset of centrifugal turns (Fig. 8, A,
B, E, and F). In contrast, a distinct
spatial and temporal EMG activation sequence was observed during
centripetal turns. First, centripetal movements were preceded by
silencing of the tonically active contralateral muscles that
contributed to the initial eccentric position (Fig. 8, C,
D, G, and H). Second, activation in
ipsilateral atlantoscapularis anterior (AS ant) preceded or was
synchronous with the start of centripetal turns (Fig. 8C; this muscle was only implanted in monkey z).
Notably, AS ant was rarely active during centrifugal turns (Fig.
8A) and was not active during turned postures. Third,
activation in ipsilateral OCI occurred only just before or slightly
after the onset of centripetal turns but was delayed relative to the
time of onset for centrifugal turns (Fig. 8, C,
D, G, and H). Fourth, a burst in
ipsilateral SP cap lagged the burst in ipsilateral OCI by up to 50 ms
or more (Fig. 8, D, G, and H).
View larger version (42K):
[in this window]
[in a new window]
|
Fig. 8.
EMG patterns from monkeys z
(A-D) and f
(E-H) during large leftward
(A, C, E, and
G) or rightward (B, D,
F, and H) turns in which the initial head
position was either central (i.e., A, B,
E, and F are centrifugal turns) or
contralateral to the direction of the head turn (i.e., C,
D, F, and H are centripetal
turns). All data are taken from spontaneous sessions. The paired turns
in A and C, B and
D, E and G, and
F and H were matched as closely as
possible for kinematics. Same format as Fig. 3, except that the
smoothed EMG average has not been applied.
|
|
We studied systematically the temporal sequences of EMG activity
depending on initial head position in monkey z,
which generated enough similarly sized turns from different initial
head positions to permit analysis (monkey f did
not contribute enough data to permit a similar analysis). Figure
9 plots the relationships between the
time of activation or silencing of EMG activity and the onset of the
turn as a function of initial horizontal head position. Activation in
ipsilateral OCI was delayed progressively relative to the onset of the
turn when turns began at more eccentric initial positions (Fig. 9,
A and B). This resulted in an increasing lag between the activation of ipsilateral OCI relative to the silencing of
contralateral OCI (Fig. 9, C and D). The more
eccentric the initial head position, the greater the lead of the burst
in ipsilateral AS ant (Fig. 9E) and the greater the lag of
ipsilateral SP cap activation (Fig. 9F) relative to the
burst in ipsilateral OCI. Although the relationships shown in Fig. 9
were somewhat variable (perhaps due to small body or shoulder rotations
or torsional head rolls), all were significant at the P < 0.05 level.
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 9.
Plots of the relationships between the relative timing of the onset of
either the turn or the muscle as a function of initial horizontal head
position, shown for either leftward head turns (A,
C, and E) or rightward head turns
(B, D, and F). All data
are taken from 41 rightward and 36 leftward turns between 45 and 65°
generated by monkey z during spontaneous
sessions. Each point plots the value of the measure described in the
graph title as a function of the initial horizontal head position, with
positive values denoting initial head position to the right. Solid
lines denote statistically significant (P < 0.05)
linear regression lines for the points shown. Short, dashed lines
denote the 0 values of the various axes. Slow head turns generated
during tracking movements, extremely fast head turns, and head turns
with vertical components >10° were excluded from this analysis. Each
plot has a different numbers of points because each interval could not
be measured for every turn.
|
|
EMG activity during postures or movements in the vertical plane
VERTICAL HEAD POSTURES.
All monkeys held their heads in different vertical postures, but
the range of postures varied. Monkeys f,
l, and r held vertical head postures ranging from
30° in inclination to 20° in declination. Monkey
z held a wider range of postures from 55° in inclination to 50° in declination. The analysis of head postures in the vertical plane was more difficult than in the horizontal plane because some
muscles active in vertical postures were also active in turned postures
(e.g., RCP maj, SP cap, and SCM). Further, our implantation regime was
not identical in all monkeys, hence we occasionally report results
obtained from only one monkey.
In all four monkeys, very little activity was recorded when the
head was in the central position (Fig.
10, 3rd column; the R-OCI
activity in Fig. 10A is due to a small rightward turn).
Complexus (COM) became increasingly active as head inclination
increased, whereas biventer cervicis (BC), which lies immediately
medial to COM, did not display strong tonic postural activity related to head inclination (Fig. 10A, 1st and 2nd
column). Obliquus capitis superior (OCS, monkey
l only) and RCP maj (Fig. 10A, monkeys
z and r) were also recruited when the head was
inclined modestly. Rhomboideus capitis (RH cap) was recruited with
increasing head inclination in monkey f (Fig. 10B,
1st and 2nd column) but not monkey
z (data not shown). SP cap (3 monkeys) and AS ant
(monkey z only) also became increasingly active
at progressively larger angles of inclination (Fig. 10, A
and B). OCI was recruited only at extreme angles of
inclination (Fig. 10A, 1st column). Maximal activation
of all muscles when the head was held in an inclined posture seldom
exceeded 50 µV/bin.
View larger version (39K):
[in this window]
[in a new window]
|
Fig. 10.
EMG patterns from monkeys z
(A) and f (B) in muscles
whose activity varied for differing postures in the vertical plane. All
data are from spontaneous sessions. Records are shown over a range of
inclined and declined postures. Activity in some muscles (e.g., R-OCI
in 3rd column of A, and L-OCI and L-RCP
maj in 4th column of A) is due to small
horizontal turns.
|
|
The pattern of muscle activation for declined head postures
varied between animals but generally included low activity levels in at
least one head extensor muscle. In monkey z but
not f, SCM was recruited for declined head postures (Fig.
10A, 4th column; the L-OCI and L-RCP maj activity was due to
a small leftward turn). COM, but not BC, was active more consistently
for declined postures in monkey z but not
f (Fig. 10, 4th column). The only muscle in monkey f that showed any activity related to the
degree of head declination was the rostral compartment of left
trapezius (TRAP, Fig. 10B, 4th column), but this pattern was
not observed for TRAP in monkey z (not shown).
Thus recruitment patterns for declined postures were consistent within
but not across monkeys, perhaps due to the differences in the magnitude
of the declined postures.
An attempt was made to construct a plot similar to Fig. 2 for
postures in the vertical plane using data from trained sessions in
monkeys z and r. However, the range of
head postures spanned only ±20° in the vertical plane because
monkeys tended to use eye movements more than head movements to fixate
targets spanning ±45° in the vertical plane (see Freedman and
Sparks 1997). This range was insufficient to provide a
definitive plot of muscle activation versus head position in the
vertical plane, particularly given the confounding dependency of EMG
activity with turned postures in some muscles.
HEAD MOVEMENTS IN THE VERTICAL PLANE.
Muscle recruitment during trained sessions.
Figure 11 shows four inclining head
movements and three declining head movements generated by
monkeys z and r. For
larger inclining movements (Fig. 11, A and
E), moderate synchronous activation in COM and BC
preceded the onset of the inclining movement by ~20-40 ms. Peak
activation of these muscles during inclining movements never exceeded
50 µV/bin. The activity profile consisted of an initial phasic
component in both muscles, followed by a tonic component predominantly
in COM typical of the posture held at the end of the movement. COM
activity was bilateral in monkey r, but a
similar determination could not be made for BC because the R-BC
electrodes in this monkey failed. The magnitudes of the phasic
components of COM and BC scaled to the size of the movement (Fig. 11,
B and F). Further, EMG activity during
inclining movements was limited to agonist muscles; antagonist muscles
(R-SCM in Fig. 11, A and B) were not
recruited.
View larger version (34K):
[in this window]
[in a new window]
|
Fig. 11.
Vertical head position (Hv) and velocity (dHv) traces, and EMG patterns
from monkey z
(A-D) and monkey
r (E-G) during movements
of various amplitudes and velocities in the vertical plane. All data
are from trained sessions. Same format as Fig. 3.
|
|
The only muscle recorded in monkey z that
was recruited during declining movements was SCM (Fig.
11D; not recorded in monkey r). Activity in this muscle rarely exceeded 10 µV/bin
and appeared only during declining movements larger than 10-15°
(Fig. 11C). No activity was observed in the antagonist
muscles (COM and BC) during declining movements (Fig. 11,
C, D, and G).
Like horizontal movements, the kinematics of vertical head movements
during trained sessions displayed a "main sequence" relationship between peak head velocity and amplitude and a linear relationship between peak acceleration and peak velocity (Fig.
12, A, B,
and G). Muscles tended to be recruited in proportion to
these kinematics and only for directions in which they could contribute
positive work: COM and BC were active during inclining head movements, and SCM was active during declining head movements (Fig. 12,
C-F and H).
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 12.
Relationship of the kinematics and EMG activity during trained
movements in the vertical plane from monkeys
z (A-F) and
r (G, H). A
correlates peak velocity to movement amplitude for
monkey z; B and
G correlate peak acceleration to peak velocity for
monkeys z and r,
respectively; C-E correlate smoothed
peak EMG activity to peak head velocity for different muscles in
monkey z; F and
H summarize the statistically significant
(P < 0.05) regression lines for the relationships
between various muscles and peak head velocity for
monkeys z and r
respectively. Same format as Figs. 4 and 5. Only trials in which the
initial head position was within 10° of center and in which the
movement deviated <20° from vertical contributed data to this
figure.
|
|
Muscle recruitment during spontaneous sessions.
The spatial and temporal recruitment patterns of EMG activity during
both inclining and declining movements in spontaneous sessions were
affected by the more variable movement kinematics. There were not
enough similarly sized movements of different dynamics to permit a
quantitative analysis similar to that shown in Fig. 7, hence we present
the EMG patterns from representative examples of head movements in the
vertical plane and discuss the features that were common to movements
of varying kinematics (Fig. 13). As
with trained movements, slower inclining movements were preceded by
moderate bursts of activity in COM and BC (Fig. 13, A,
B, and D). Activity in muscles other than COM was
observed in some but not all regularly paced inclining movements,
probably due to small coincident turns. Particularly abrupt inclining
movements were characterized by large phasic bursts that preceded
movement onset by ~20-50 ms in BC and COM, and were also observed
consistently in other muscles not recruited during the slower inclining
movements, such as OCI, SP cap and RH cap (Fig. 13, C and
E). RCP maj was typically activated more strongly or more
extensively than OCI during inclining movements (Fig. 13C);
this dissociation between RCP maj and OCI was unique to inclining as
opposed to turning movements. Further, bilateral recruitment was
commonly observed in SP cap and OCI even though this was never observed
during turning movements. In monkey f in
particular (Fig. 13E), an interval of silence followed these
initial bursts, during which a burst of activity was commonly but not
always observed in SCM and TRAP (Fig. 13E; the burst in
R-OCI was most likely secondary to a small leftward turn).
View larger version (36K):
[in this window]
[in a new window]
|
Fig. 13.
EMG patterns from monkeys z
(A-C) and f
(D and E) during inclining movements of
20° with different dynamics. All data are from spontaneous sessions,
and all began near the center vertical position. Note the time scale
differs for different movements.
|
|
For declining head movements, large (>20 µV/bin) bursts of muscle
activity were observed only during abrupt reversals of inclining movements (Fig. 14). Synchronous bursts
of activity in extensors alternating with synchronous bursts of
activity in flexors (particularly SCM and TRAP; shown for
monkey f) preceded and presumably accounted for
the triphasic peaks of acceleration (vertical lines in Fig. 14A). In contrast, little activity was present during the
initial slow declining phase of the sequence illustrated in Fig.
14B even though its peak velocity was similar to that
occurring during the subsequent reversal of a fast inclination.
Further, these large bursts of activity were generated while the head
was still inclining, presumably resulting in high force output in the
actively lengthening fibers (see Zajac and Gordon 1989
for review).
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 14.
EMG patterns from monkey f during 2 sequences of head movements in the vertical plane, composed of both
inclining and declining movements. A shows a sequence in
which a rapid 12° inclining movement was followed immediately by a
13° declining movement. B shows a sequence in which a
quick 12° inclining movement was generated in the middle of a 30°
declining movement. Hv, dHv, and ddHv label traces of vertical head
position, velocity and acceleration respectively. The vertical
short-dashed lines are aligned on peak inclining accelerations of the
head, and the vertical solid lines are aligned on peak declining
accelerations of the head. Note that the time scale differs in
A and B.
|
|
Oblique inclining head movements.
Inclining head movements with a turning component were typically
associated with activity in those ipsilateral muscles previously described to be strongly active during turns. In addition, COM, BC, and
TRAP were more active during oblique inclining movements with
contralateral as opposed to ipsilateral turning components (Fig.
15, A and B;
oblique movements generated by monkey r were not
sufficiently eccentric to permit analysis). Activities in R-COM, R-BC,
and L-TRAP were consistently larger for contralaterally directed
oblique head movements (i.e., leftward for R-COM and R-BC, rightward
for L-TRAP). Oblique declining movements were not analyzed similarly
because the sample size was small, and the movements were more variable
in their kinematics.
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 15.
A and B: horizontal (Hh) and
vertical (Hv) position traces and EMG patterns from
monkey z during either an oblique 25°
left 15° up head movement (A) or an oblique 25°
right 15° up movement (B). These examples were
selected for comparison since they both had inclining components of
15° and approximately equal peak velocities (135-145°/s) but
oppositely directed turning components. Note that R-COM and R-BC were
more active during the left-up movement than the right-up movement,
while the opposite pattern was seen in L-trapezuis (TRAP).
C-E: relationship between the peak smoothed EMG
activity during 45 oblique upward head movements for R-COM
(C), R-BC (D), and L-TRAP
(E) plotted as a function of the vector of the oblique
head movement. The location of each diamond in C-E
denotes the endpoint of the oblique upward head movement relative to 0, and the size of each diamond denotes the magnitude of the smoothed peak
EMG activity. All data are taken from trained sessions.
|
|
Feeding behaviors and head shakes
All monkeys moved their heads during eating or head shaking. The
neck muscle activities during these movements attested to the larger
range of highly orchestrated recruitment patterns that the monkey is
capable of generating. During these movements, EMG patterns became
quite specialized, and previously inactive muscles were recruited. We
did not focus quantitatively on such movements because the magnetic
search coil system provided calibrated measurements only for horizontal
and vertical rotations, not for torsional rotations nor translational movements.
The EMG patterns shown in Fig. 16
were generated during a representative forward translation along the
occipito-nasal axis, which monkey f used to crane
for offered food. The unique feature of this type of movement was the
strong and increasing bilateral activation in SCM (Fig. 16), which
occurred without significant activity in other muscles that were
typically synergistic (OCI and SP cap) or antagonistic (COM) with SCM.
TRAP was also active during this movement, although compared with SCM,
the activity in TRAP was more discrete and did not increase greatly
during the movement. Similarly strong activity was observed in R-SCM in
monkey z during craning movements, although this
muscle was not implanted bilaterally.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 16.
EMG activity during a head movement made by monkey
f in which the head translates forward for food during a
spontaneous session. The position signals were derived from coil
signals and hence do not measure head translation.
|
|
All monkeys spontaneously generated head-shaking sequences
consisting of multiple oscillations in which the head was rapidly turned from one side to another over a period of 0.5-2 s. Head shakes
were associated with the highest levels of EMG activity recorded in all
muscles, regardless of whether the muscle linked the skull to the
spinal column (OCI, SP cap, RCP maj, COM), the skull to the shoulder
girdle (TRAP, RH cap), or the shoulder girdle to the spinal column (AS
ant, AS post). Figure 17A
shows a typical example consisting of seven oscillations over a period
of 1 s. All muscles displayed discrete bursting profiles, and had
activity levels >60-100 µV/bin. A closer examination of the
temporal aspects of the EMG activity revealed that some muscles burst
twice per cycle, whereas other muscles burst only once per cycle (Fig.
17, B and C). In those muscles bursting twice per
cycle, the magnitude of the two bursts differed. The synergies
identified for rapid turns could be discerned during head shakes, but
coactivation was stronger during shakes than turns (e.g., note the
bilateral activation in the OCI muscles in Fig. 17B). Strong
activation in many muscles often occurred 20 ms prior to the start of
the turn toward the preferred direction (i.e., R-OCI and R-SP burst
before right turns; L-OCI, L-RCP maj burst before left turns: Fig.
17C), while these muscles lengthened.
View larger version (55K):
[in this window]
[in a new window]
|
Fig. 17.
A: EMG activity and horizontal head position (Hh) traces
during a vigorous head shake generated by monkey
z in a spontaneous session. The vertical dashed lines
denote the approximate onset and offset of the sequence. The solid
bracket denotes the 2 cycles magnified in B and
C. The vertical solid lines in B and
C demarcate individual cycles of the head shake.
C: stylized representation of the spatial and temporal
EMG activation patterns for the magnified sequence in B.
The boxes denote the timing of the EMG activity with dark and light
boxes signifying relatively higher or lower levels of EMG activity,
respectively.
|
|
| |
DISCUSSION |
This report is the first detailed examination of
EMG activation in multiple monkey neck muscles during a large range of
head postures and movements. Similar studies in cats revealed that individual neck muscles, and even individual compartments within a
given neck muscle, are controlled by the CNS in specific and reproducible ways (Keshner 1994; Keshner et al.
1992; Richmond et al. 1992; Thomson et
al. 1994, 1996; Wilson et al. 1983) but revealed
complex and somewhat counter-intuitive patterns, as discussed in the
following text. The activity from analogous muscles in monkeys is
similarly specific and reproducible but somewhat simpler to relate to
kinematics. This discussion focuses on the similarities and differences
between cats and monkeys in terms of the likely kinetics of the various
postures and movements and the comparative architecture of the muscles
as revealed in the companion paper (Richmond et al.
2001). This approach must eventually be combined with
quantitative kinetic and biomechanical analyses of normal and perturbed
head movements, as well as single-unit recording studies in the brain
stem, to understand fully the underlying neuromuscular control.
Methodological considerations
To study head movements systematically, we examined head movements
generated during trained gaze shifts since these are measured easily
and reproduced consistently. However, had we only studied trained head
movements, we would not have observed the wealth of variation in EMG
activation that accompanied spontaneous head movements. For example,
craning movements were associated with coactivation of the two SCM
muscles (Fig. 16), which are antagonists during head turns. Head shakes
were associated with very high levels of activity in all muscles,
apparently phased to contribute to the high accelerations observed
during such rapid movements (Fig. 17). General features of these
movements were observed by comparing repeated sequences generated over
the course of weeks. If future studies are concerned with the
neuromuscular strategies underlying such movements, clever ways will
have to be devised to elicit many such movements reproducibly.
In this study, we recorded only horizontal and vertical head rotations
and not the remaining 4 degrees of freedom (df; torsional rotation and
3 directions of translation). Qualitatively, we observed motion in all
6 df during the natural repertoire of head movements. Translating
movements were frequently generated when the monkeys craned for food or
tried to visualize objects partially obstructed by barriers. Torsional
rotations of varying magnitudes were generated during orienting
movements and feeding behaviors but were generally excluded from
analysis. Other methodologies, such as three-dimensional coil systems
and analysis of reflective and fluoroscopic markers, would be better
suited to quantifying such motion and estimating the kinetic
implications to better understand the specialization of neck muscle
control for these tasks.
Postures and synergies
To hold its head in the central position, the chair-restrained
monkey requires little or no muscle activity (Fig. 10), probably because the head is aligned vertically on top of the cervical and
thoracic vertebrae so that the weight of the head is borne by
compression of underlying vertebrae. Deviations from this
"metastable" position required increased levels of muscle activity
to balance the head. Interestingly, both statically inclined and
declined postures were associated with flexor and extensor muscle
activity that was variable but sometimes included some co-contraction
of antagonist muscles (Fig. 10).
To hold its head in progressively more turned postures, the monkey was
observed to recruit progressively all of the muscles that tended to
pull the head into such turned postures. Suboccipital muscles such as
OCI and RCP maj were active for even the smallest deviations from
center to which activity in multiarticular head turners such as SP cap
and SCM was added for more turned postures. Similar trends have been
reported previously (Bizzi et al. 1971; Lestienne
et al. 1995, 2000) and have been suggested to reflect underlying kinematics: small turns are thought to be executed by
rotations at suboccipital joints whereas larger turns require additional rotations about lower cervical vertebrae (Kapandji 1974). Furthermore, some muscles (e.g., RCP maj, SP cap, and
SCM) are also involved in postures and pure movements in the vertical axes. This observation is interesting considering that horizontal and
vertical components of head movements appear to be controlled by
separate brain stem structures downstream of the superior colliculus (owls: Masino and Knudsen 1990; cats: Grantyn and
Berthoz 1987; Isa and Naito 1994, 1995;
Sasaki et al. 1999; monkeys: Cowie and Robinson
1994).
The postural activity of some monkey neck muscles was different
from that reported for homologous cat muscles (Akaike et al. 1989; Guitton and Mandl 1978; Keshner et
al. 1992; Richmond et al. 1992; Roucoux
et al. 1980, 1989; Thomson et al. 1994, 1996). The differences may be related to the posture of the cervicothoracic joints, which some of these muscles cross. Studies in cats are usually
conducted with the trunk oriented horizontally and the cervicothoracic
column in an S-shaped configuration, whereas studies in chaired monkeys
are conducted with the trunk and cervical column oriented more
vertically. In cats, BC is active tonically over most of the range of
inclined and declined midline head positions, including the central
position; in monkeys BC had little activity during tonic postures (Fig.
10). However, BC has been found to become active when monkeys hold
their heads in neutral postures during quadrupedal stance when the
cervicothoracic region is held in a more S-shaped posture like that in
which cats were studied (E. Keshner and B. Peterson, unpublished
observations). Further, large horizontal turns in cats are associated
with paradoxical activity in the contralateral COM muscle; in chaired
monkeys, COM was active only when it could contribute positive work in inclination or during oblique turns. In the monkey, the long cervical muscles that contribute to large horizontal turns may impose mostly compression forces in addition to the axial rotation of the vertical cervical column; in the cat these muscles presumably produce complex torsional forces along the S-shaped cervical column that may require some degree of cocontraction for stabilization. The preferential recruitment of extensors contralateral to oblique turns may serve to
counteract small torsional rotations (Fig. 15).
When rhesus monkeys move in their natural environment, they do so with
weight borne on all four limbs. They have been classified as
terrestrial quadrupeds, although they can resort to facultative bipedalism when it is useful (e.g., carrying food) (Napier and Napier 1967, 1985). At rest, they typically adopt a squatting body posture that frees their hands for tasks such as feeding and
grooming. The presence of the squatting posture and facultative bipedalism has led some to regard monkeys as bipeds to differentiate monkeys from obligatory quadrupeds like rodents and cats (Graf et al. 1995; Vidal et al. 1986). Both the use
and anatomy of the monkey head-neck-scapular system are distinct from
cats and humans. Compared to cats, monkeys do not use the head as the
primary prehensile organ, they have a more modest dorsal neck muscle
mass, and they have a scapula attached mechanically to the trunk via
the clavicle. Compared to humans, the orientation of the primate
scapula is more parasagittal than frontal, and monkeys retain
musculature apparently unique to quadrupeds (e.g., RH cap and somewhat
separate BC and COM) (Richmond et al. 1999a). These
observations do not preclude monkeys as animal models for human head
movements, but caution extrapolating results across species.
Because rhesus monkeys are neither obligatory quadrupeds nor bipeds,
legitimate concerns may be raised regarding the squatting posture
imposed by our primate chairs. Although we could not measure the
orientation of the cervical column directly, a previous study using
X-ray photography confirmed that similar craniocervical postures were
obtained during squatting in chair-restrained and unrestrained monkeys
(Vidal et al. 1986). We reiterate that the activation
patterns observed in chair-restrained, squatting monkeys will likely
change when monkeys adopt a quadrupedal posture; biomechanical differences between the two postures are likely reflected in distinct uses of at least some of the muscles of the head-neck-scapular complex,
particularly those spanning the cervicothoracic junction.
Movements and kinetics
The head movements generated during trained gaze shifts followed
simple scaling rules for amplitude, velocity, and acceleration in both
the horizontal and vertical axes. Muscle recruitment during these
movements appeared to be associated both qualitatively and quantitatively with the presumed kinetic requirements of the movements. The kinetics reflect a mixture of elastic, viscous and inertial force
requirements related to position, velocity, and acceleration, respectively, but the covariance of these variables during trained movements makes it difficult to use such data to test hypotheses about
neuromuscular strategies. Spontaneous movements exhibited a wider range
of kinematic patterns that make it possible to consider the kinetic
strategies. This reinforced the notion that neck muscle activation in
chaired monkeys is relatively simple to relate to the mechanics and
kinetics of head movement.
VARIATIONS IN MOVEMENT DYNAMICS.
Head movements during trained gaze shifts were associated with phasic
and tonic patterns of recruitment in agonist muscles consistent with
overcoming elastic and viscous forces (Figs. 3 and 11). Very rapid
movements of the same amplitude were associated with large reciprocal
phasic bursts in agonist and antagonist muscles consistent with the
requirements of accelerating and decelerating an inertial mass (Figs.
6, 7, 13, and 14). Triphasic patterns of recruitment have been
described for rapid arm movements (Hallett et al. 1975;
Lestienne 1979; Marsden et al. 1983), and
for SCM and SP cap during head movements in humans (Zangemeister
et al. 1982). The activation of antagonistic muscles in humans
may be a consequence of instructing fast movements. Our results suggest that antagonist muscle activation may be more the exception than the
rule, particularly in trained tasks with no premium on either speed or accuracy.
During the fastest movements, the activity of many muscles was not
simply triphasic but instead could involve multiple sequences of
interleaved bursts in agonist and antagonist muscles during the latter
portions of the movement (Figs. 6, 13, and 14). We speculate that the
primary agonist and antagonist bursts were part of a preprogrammed,
open-loop neural command related to the kinetic requirements of the
movement (Fig. 7), whereas the more variable subsequent bursts in the
latter part of the movement may reflect the operation of corrective
reflexes generated by reciprocally organized neural circuitry and
perhaps reflecting vertebral as well as head motion. Such complexity
within EMG envelopes cautions against the use of methods that determine
the mean of EMG waveforms across similar trials, particularly for the
latter portions of the movements. Had we simply averaged our data
across multiple trials, these discrete multiphasic reciprocal
bursts would have averaged into a generalized pattern of amorphous
cocontraction. Although the observations reported here rely partly on
the evaluation of individual examples, an alternative technique would
be to deconstruct EMG patterns from multiple trials into the least
common waveforms to distinguish aspects of the signal that are common
to multiple movements from those appearing in an idiosyncratic fashion.
CENTRIFUGAL VERSUS CENTRIPETAL MOVEMENTS.
Comparisons of centrifugal and centripetal turning movements
matched for size and speed revealed that the initial position of the
head affected muscle recruitment. The first event preceding centripetal
turns was the silencing of previously active contralateral muscles
(Figs. 8 and 9). In addition to releasing stored elastic energy to
assist in initiating the turn, the silencing of contralateral muscles
may also avoid the high forces that would be required to stretch active
muscles: had the silencing of contralateral muscles coincided with the
activation of ipsilateral muscles, remnant forces produced by
lengthening the contralateral musculature would have resisted the
developing turning forces.
The synergy for centripetal but not centrifugal turns was found to
include AS ant, which spans from the lateral half of the scapula to the
transverse process of C1. The biomechanical
implications of this activity are not clear, but the temporal ordering
of recruitment suggests that forces developed by AS ant play a role in
the early acceleration of centripetal turns. The involvement of AS ant
in centripetal turns is surprising considering this muscle, at least in
humans, is viewed commonly as being involved in scapular elevation rather than head turns (Bull et al. 1984;
Mayoux-Benhamou et al. 1997). Experimental protocols are
often confined to centrifugal movements and would miss such activity.
The temporal activation of agonist muscles also differed between the
synchronous activation during centrifugal turns and the staggered
recruitment during centripetal turns. Although this report is the first
to study temporal relationships in detail, previous studies in monkeys
have reported some variability in either the magnitude (Bizzi et
al. 1971) or timing (Lestienne and Liverneaux
1988; Lestienne et al. 1995, 2000) of muscle
recruitment that appeared related to initial head position. A more
complete kinetic analysis will be required to understand such details, which may relate to stability in other axes such as torsion or translation. The later recruitment in ipsilateral OCI and SP cap may
also serve to delay their activation until they reach a more advantageous position in the force-length relationship. It is obvious
from our results that the central command to individual neck muscles is
highly specialized and depends systematically on initial position. Thus
it seems inappropriate to produce ensemble EMG activity profiles by
averaging normalized EMG activity across multiple neck muscles (e.g.,
Lestienne et al. 2000).
Muscle architecture
Some differences in muscle function between cats and monkeys may
relate to differences in the architecture of individual muscles in
addition to the postural differences noted in the preceding text. For
example, RCP maj in the cat is almost a pure head extensor. In monkey,
its skull insertion is distributed more laterally (Richmond et
al. 2001), contributing a turning moment that is reflected in
its recruitment during horizontal postures and turns. The gradual loss
of distinctive muscle functions that seems to be associated with the
tendency to hold the neck and torso in a vertical orientation may be a
driving force for the simplification of the primate neck musculature
with evolution. For example, COM and BC in the cat are anatomically and
functionally distinct (Richmond et al. 1992). In the
monkey, COM functions somewhat similarly to BC and often merges
anatomically with the anterolateral edge of BC (Richmond et al.
2001). In humans, both muscles appear to have been subsumed into the single muscle semis
TOP
ABSTRACT
INTRODUCTION
