Journal of Electromyography and Kinesiology 14 (2004) 463474

www.elsevier.com/locate/jelekin
Spatio-temporal evaluation of neck muscle activation
during postural perturbations in healthy subjects
D. Falla
a,
, A. Rainoldi
b,c
, R. Merletti
b
, G. Jull
a
a
Division of Physiotherapy, The University of Queensland, Brisbane, Qld 4072, Australia
b
Centro di Bioingegneria, Dip. di Elettronica, Politecnico di Torino, Turin, Italy
c
Department of Physical Medicine and Rehabilitation, University of Tor Vergata and Fondazione Don Gnocchi, Rome, Italy
Abstract
The purpose of this study was to examine the spatio-temporal activation of the sternocleidomastoid (SCM) and cervical exten-
sor (CE) muscles with respect to the deltoid muscle onset during rapid voluntary upper limb movement in healthy volunteers. The
repeatability and reliability of the spatio-temporal aspects of the myoelectric signals were also examined. Ten subjects performed
bilateral and unilateral rapid upper limb exion, abduction and extension in response to a visual stimulus. EMG onsets and nor-
malised root mean square (nRMS) values were calculated for the SCM and CE muscles. Subjects attended three testing sessions
over non-consecutive days allowing the repeatability and reliability of these measures to be assessed. The SCM and CE muscles
demonstrated feed-forward activation (activation within 50 ms of deltoid onset) during rapid arm movements in all directions.
The sequence and magnitude of neck muscle activation displayed directional specicity, however, the neck exor and extensor
muscles displayed co-activation during all perturbations. EMG onsets demonstrated high repeatability in terms of repeated mea-
sure precision (nSEM in the range 1.95.7%). This was less evident for the repeatability of nRMS values. The results of this study
provide a greater understanding of cervical neuromotor control strategies. During bilateral and unilateral upper limb perturba-
tions, the SCM and CE muscles demonstrate feed-forward co-activation. It seems apparent that feed-forward activation of neck
muscles is a mechanism necessary to achieve stability for the visual and vestibular systems, whilst ensuring stabilisation and pro-
tection of the cervical spine.
# 2004 Elsevier Ltd. All rights reserved.
Keywords: Neck muscle; Electromyography; Postural control; Feed-forward
1. Introduction
Perturbation studies have been used extensively to
explore the neural mechanisms underlying balance con-
trol. Internal perturbations such as a rapid movement
of the arm produce reactive compensatory forces
opposing those produced by the arm movement [5].
These forces are transferred to body segments eliciting
a series of postural adjustments. The origin of the pos-
tural adjustments is twofold. Firstly, the movement
causes a change in body geometry. This may result in a
shift of the centre of gravity, which causes dis-
equilibrium. Secondly, the forces generated to initiate a
movement induce reaction forces [25].
Detection of the spatio-temporal aspects of myo-
electric signals during postural perturbations can be
utilised to analyse muscle recruitment and enhance our
understanding of neuromotor control strategies. Feed-
forward muscle activation is one central nervous
system (CNS) strategy, which can be detected with
analysis of the temporal aspects of the electromyo-
graphic (EMG) signal. Feed-forward postural respon-
ses are a mechanism employed by the CNS to regulate
motor control of muscles and contribute towards the
maintenance of stability [1]. Extensive evidence of feed-
forward activation of limb and trunk musculature in
response to perturbations is available in the literature
[7,8,12,1618]. However, there is a paucity of studies
that have investigated spatio-temporal aspects of neck
muscle activation during postural perturbations.
Gurnkel et al. [13] demonstrated that backward accel-
eration of the head precedes an internal perturbation

Corresponding author. Tel.: +61-7-3365-4529; fax: +61-7-3365-

2775.
E-mail address: d.falla@shrs.uq.edu.au (D. Falla).
1050-6411/$ - see front matter #2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jelekin.2004.03.003
of upper limb exion, which coincides with an increase in
trapezius and splenius capitus muscle activity. The
authors suggested that this anticipatory muscle acti-
vation provided generalised postural xation of the
cervical spine. This study was supported by the work of
van der Fits et al. [25] which demonstrated primary
activation of the dorsal postural muscles including
the neck extensors, with rapid upper limb exion.
This result conrmed that the basic nature of the
direction-specic organisation of postural adjustments,
readily identied in the lower limb and trunk muscles,
is also present in the neck.
The present study aimed to provide a more extensive
analysis of neck muscle activation in response to inter-
nal postural perturbations. This study sought to exam-
ine the temporal and spatial response of the
sternocleidomastoid (SCM) and the cervical extensor
(CE) muscles during rapid voluntary upper limb move-
ments in three directions (exion, extension and abduc-
tion) in healthy volunteers. Internal perturbations were
selected to assess neck muscle responses for the follow-
ing two reasons. Firstly, internal perturbations allow
identication of the response to the perturbation that is
temporally and spatially predictable by the CNS. Sec-
ondly, internal perturbations allow evaluation of the
nal output of the motor command from the CNS
without the inuence of aerent information resulting
from the perturbation [14].
The second aim of this investigation was to examine
the repeatability and reliability of the spatio-temporal
aspects of the EMG signals obtained from the SCM
and CE muscles during rapid upper limb perturbations.
2. Methods
2.1. Subjects
Ten right-handed healthy volunteers aged between
25 and 35 years (mean 29.2, SD 3.4) participated in
this study. Subjects were included if they were free of
neck pain, had no past history of orthopaedic disorders
aecting the neck and no history of neurological dis-
orders. Subjects attended three testing sessions over
non-consecutive days.
Ethical approval for the study was granted by the
Institutional Medical Research Ethics Committee. All
studies were conducted in accordance with the declar-
ation of Helsinki.
2.2. Instrumentation and measurements
Visual commands to move were provided by light
emitting diodes (LED) xed to an adjustable visual dis-
play board positioned at the subjects eye level. A red
LED provided the stimulus to move for bilateral
shoulder movements. For the unilateral arm movement
task, a red and green LED provided the prompt for the
right and left arm, respectively. The electrical signal
providing the stimulus to move was recorded simul-
taneously with the EMG signals to generate the trigger
information for further data reduction. The time
between stimuli was randomly varied and controlled by
the investigator.
Myoelectric signals were detected bilaterally from the
sternal head of the SCM, CE, anterior, middle and
posterior deltoid muscles using bipolar disposable elec-
trodes (1 cm disc electrodes, inter-electrode distance
2 cm) (Myotronics, Noromed) following careful skin
preparation [4] and using published guidelines for elec-
trode placement. For the SCM muscles, the electrodes
were positioned at 1/3 the distance from the sternal
notch to the mastoid process at the distal end overlying
the muscle belly [9]. For the CE muscles, the electrodes
were positioned over the distal half of the distance
between the base of the occiput and the spinous pro-
cess of the seventh cervical vertebra. For the anterior,
middle and posterior deltoid electrodes were positioned
in the lower half of the distance from the acromion to
the deltoid tuberosity [9,20]. The ground electrode was
placed on the upper thoracic spine.
Signals were amplied (gain 2000), passed through
a 10450 Hz bandwidth lter (40 dB/decade slope on
each side) and sampled at 2048 Hz (ASE16-16 channel
amplier, LISiN Centro di Bioingegneria, Politecnico
di Torino, Italy). The samples were digitised by a 12 bit
A/D converter and stored on a personal computer.
2.3. Procedure
Prior to commencement of experimental trials, EMG
data were collected for 10 s during standardised man-
euvers for normalisation of the EMG amplitude. With
the subject positioned in supine lying, the reference vol-
untary contraction for the SCM involved a combined
movement of cranio-cervical and cervical exion to lift
the head so that it just cleared the bed and was held
isometrically [11]. For the CE muscles, the subject was
positioned in prone lying and performed an isometric
cervical extension contraction so that the forehead just
cleared the bed. Each reference voluntary contraction
was repeated three times with an interval of 1 min
between repetitions.
During the experimental trials, subjects performed a
series of unilateral and bilateral rapid shoulder exion,
abduction and extension movements keeping the elbow
straight, in the relaxed standing position with their feet
placed shoulder width apart. Subjects were instructed
to move, upon the light command, as fast as possible
and to focus on the speed of movement rather than on
the distance. Rapid shoulder exion and abduction
were performed to approximately 60
v
and shoulder
464 D. Falla et al. / Journal of Electromyography and Kinesiology 14 (2004) 463474
extension to 40
v
[16]. Each subject initially performed
two to three practice repetitions to allow EMG signal
quality to be assessed and to ensure consistency of the
speed and distance of arm movement between repeti-
tions. Each subject then performed three repetitions in
each direction in a randomised order. The experimental
protocol was repeated over three non-consecutive days.
2.4. Data management and statistics
Previously developed software for muscle activity
detection from EMG signals was utilised to determine
the ono timing of muscles (for a complete descrip-
tion of the algorithm used, refer to the work of Merlo
et al. [19]). Several methods for ono timing estima-
tions are proposed in the literature. These methods are
based on the single-threshold method [24] or on the
signal envelope [6,15,23]. In this study, the recent
method described by Merlo et al. [19] was adopted
since it oers the best performance in terms of onset
detection. This program is based on identifying single
motor unit action potentials from the interference
EMG signal using the continuous wavelet transform. A
manifestation variable is computed as the maximum of
the outputs of a bank of matched lters at dierent
scales. A threshold is applied to the manifestation vari-
able to detect EMG activity. With the method pro-
posed by Merlo et al. [19], the SD of onset estimate
reduces from 26 to 17 ms when the signal to noise ratio
varies from 2 to 8 dB, while the single threshold
approach leads to a SD of onset estimate always great-
er than 130 ms and results in the worst performance.
Subsequent to identication of EMG onsets, two
measures were calculated; (a) the reaction time of del-
toid muscle, dened as the time from stimulus to the
onset of deltoid muscle activity and (b) relative latency
dened as the latency between the onset of the agonist
limb muscle and each of the neck muscles. Fig. 1 pre-
sents an example of raw EMG traces recorded from
the SCM, CE and deltoid muscles obtained from a
subject performing left and right rapid unilateral upper
limb exion movements. Fig. 2 illustrates identication
of the ono timing of the CE and deltoid muscles dur-
ing rapid unilateral upper limb exion movements.
Neck muscle latencies and deltoid reaction times
were calculated for the three contractions for each con-
dition. Any EMG onset occurring between 150 ms
before and 50 ms after the agonist arm muscle EMG
onset was considered feed-forward activation [2,16].
The data from subject 6 for all contractions on the
third day and data from subject 9 for the unilateral
extension contractions on the third day were excluded
due to errors in data collection. As a consequence of
such data purging, all contractions from subject 6 and
the unilateral extension contractions from subject 9
were omitted in the assessment of repeatability.
To evaluate the EMG amplitude, root mean square
(RMS) values were calculated for the EMG bursts
(time interval between the on and o timing of a
muscle) identied in the initial signal processing pro-
cedure. For normalisation, RMS values obtained
during the experimental trials were expressed as a per-
centage of a maximum 1 s RMS value obtained across
the three repetitions of the reference voluntary contrac-
tions for each muscle.
The means and SD for the nRMS values, deltoid
reaction times and SCM and CE relative latencies were
calculated from the three repetitions performed by each
subject for each muscle and each condition. The
Wilcoxon non-parametric matched paired test was used
to verify whether signicant dierences existed between
muscles, contractions, and sides within the group of
subjects. The signicance level was set at p < 0:05.
Fig. 1. Raw EMG traces from the sternocleidomastoid (SCM), cervical extensor muscles (CE) and deltoid (DELT) obtained from a subject per-
forming left and right rapid unilateral upper limb exion movements.
D. Falla et al. / Journal of Electromyography and Kinesiology 14 (2004) 463474 465
2.5. Repeatability of amplitude and time variable
estimates
The repeatability of deltoid reaction times, relative
latencies and nRMS values of the SCM and CE mus-
cles were examined. For each condition, the subject
performed three repetitions and the experimental pro-
tocol was repeated three times over non-consecutive
days. Repeatability estimates were obtained by evaluat-
ing intraclass correlation coecients (ICC) and the
normalised standard error of the mean (nSEM). Pre-
vious work examining sEMG repeatability has ident-
ied the advantages of considering both the ICC and
nSEM [21,22]. A low nSEM indicates a reliable esti-
mate of the mean, that is, minimal experimental noise
while analysis of variance (ANOVA) provides infor-
mation about the factors contributing to the total vari-
ance (days, trials, subjects, measurement conditions).
In particular, ICC is the portion (in %) of variance due
to subjects: its value indicates whether the variations of
the measurements reect subject-to-subject variations
or random experimental noise introduced by dierent
days or trials [9].
3. Results
3.1. Spatial and temporal characteristics
When rapid movements of the arm were performed
in response to a visual stimulus, the SCM and CE mus-
cles were activated within 50 ms of the onset of the
deltoid muscle, thereby fullling the criterion for
feed-forward activation with movement. The neck
exor and extensor muscles demonstrated co-activation
across all conditions.
Table 1 presents the means and SD for the deltoid
reaction times. Deltoid reaction times were signicantly
greater for unilateral upper limb movements across all
conditions compared to bilateral movements.
Fig. 3 presents the means and SD of the relative
latencies recorded for the SCM and CE muscles in per-
formance of rapid bilateral arm exion, abduction and
extension. For bilateral upper limb exion, a trend was
present to suggest onset of CE muscle activity (right
9:3 18:7 ms, left 15:4 17:8 ms) prior to the SCM
(right 22:0 14:3 ms, left 21:5 9:5 ms). During bilat-
eral upper limb abduction, a similar trend was evident
to support activation of the CE muscles (right
23:1 8:4 ms, left 30:8 5:3 ms) prior to the SCM
Fig. 2. Raw EMG traces from the cervical extensor (CE) and deltoid (DELT) muscles obtained from a subject performing (A) right and (B) left
rapid unilateral upper limb exion movements. Previously developed software [19] was utilised to determine the ono timing of muscles. EMG
amplitude, was calculated for the time interval between the on and o timing of each muscle.
Table 1
Reaction times (ms) for the deltoid muscles with each movement direction (mean SD). Dierences between bilateral and unilateral movements
are also presented. Reaction times for unilateral contractions were signicantly greater than bilateral. Wilcoxon paired test, N 54 for exion and
abduction and N 48 for extension
Condition Flexion Abduction Extension
Left Right Left Right Left Right
Bilateral 215 29 215 29 196 36 198 36 202 32 205 33
Unilateral 237 21 241 16 227 35 213 30 241 24 232 32
Unilateralbilateral 22 12

25 15

31 32

15 22

38 22

22 22

p < 0.01.

p < 0.05.
466 D. Falla et al. / Journal of Electromyography and Kinesiology 14 (2004) 463474
(right 34:9 13:5 ms, left 29:5 18:3 ms). For bilateral
upper limb extension, onset of the SCM muscles (right
17:1 15:8 ms, left 18:5 15:2 ms) occurred prior to the
CE muscles (right 24:1 14:4 ms, left 32:2 9:2 ms).
Activation of the contralateral CE muscles was not
detected during the unilateral movement tasks in eight
out of the 10 subjects. Fig. 4 presents the means and
SD of the relative latencies recorded for the bilateral
SCM and ipsilateral CE muscles in performance of uni-
lateral exion, abduction and extension movements.
No dierences in the relative latencies of the SCM and
CE muscles were identied during unilateral exion
and abduction movements. For unilateral upper limb
extension performed on the right side, the right SCM
(18:2 7:7 ms) was activated prior to the right CE
muscle (28:6 14:7 ms). No signicant dierence was
identied for the onset of left and right SCM muscles
for unilateral upper limb movements in all three direc-
tions. EMG onsets for SCM and CE were compared
between bilateral and unilateral arm movements,
revealing no dierence.
3.2. Signal amplitude behaviour
Fig. 5 presents the nRMS values obtained for the
SCM and CE muscles during each condition and
Table 2 presents the signicant dierences between
conditions. The main signicant ndings are sum-
marised as follows.
Overall, the SCM muscles demonstrated the greatest
activation during bilateral and unilateral upper limb
exion compared to bilateral and unilateral extension
and abduction. For the left SCM muscle, there was no
dierence in the magnitude of activation during unilat-
eral and bilateral arm exion manoeuvres. This was
also the case for the right SCM muscle for compar-
isons between bilateral exion and right unilateral ex-
ion. However, for the right SCM, greater activation
was evident during bilateral arm exion compared with
unilateral arm exion performed on the left.
Across all movements, both unilateral and bilateral,
greatest nRMS values for the CE muscles were
recorded during bilateral arm abduction. For both the
left and right CE muscles, greater nRMS values were
recorded during unilateral exion and abduction com-
pared to extension (all p < 0:05). No signicant dier-
ence was identied for right and left CE muscle
activation between unilateral abduction and exion
movements. Across all conditions, lowest nRMS values
were recorded for the left and right CE muscles during
unilateral arm extension movements.
3.3. Repeatability of amplitude and time variable
estimates
Table 3 documents the results of ICC and nSEM for
the deltoid muscle reaction times and relative latencies
of the SCM and CE muscles obtained during unilateral
and bilateral rapid upper limb exion, extension and
abduction perturbations. According to the classi-
cation of Bartko [3], low levels of repeatability were
identied for each muscle as dened by the ICC. Excel-
lent values of nSEM (from 1.9% to 5.7%) were found
for both the relative latencies of SCM and CE and the
reaction times of the deltoid muscle, providing evidence
of high repeatability for these variables in terms of
repeated measure precision (reliability).
Table 4 presents the repeatability results of the
nRMS values obtained for the SCM and CE and the
absolute RMS values for the deltoid muscles during
unilateral and bilateral arm movements. Compared to
the repeatability of reaction times and relative laten-
cies, generally greater values of ICC were identied
indicating for the amplitude higher levels of repeat-
ability, particularly for the deltoid muscle during bilat-
eral arm movements. The level of measurement
Fig. 3. Relative latencies (ms) for the left and right sternocleidomastoid (SCM) and cervical extensors (CE) muscles during bilateral upper limb
movements (mean SD, N 60 contractions). indicates signicant dierence p < 0:05.
D. Falla et al. / Journal of Electromyography and Kinesiology 14 (2004) 463474 467
precision identied for the detection of EMG onsets
was smaller for the estimation of nRMS values (nSEM
between 3.8% and 25.1%).
4. Discussion
This study aimed to evaluate the temporal and
amplitude parameters of neck muscle activation in
association with rapid voluntary movement of the
upper limb. With bilateral and unilateral upper limb
movements in all directions, the SCM and CE muscles
demonstrated feed-forward activation (all relative
latencies < 50 ms). Furthermore, the neck exor and
extensor muscles displayed co-activation during all per-
turbations. The sequence and magnitude of neck mus-
cle activation displayed direction specicity.
The results of this study demonstrate a trend
towards onset of the CE muscles prior to activation of
the SCM with bilateral upper limb exion. The reverse
occurred with rapid bilateral upper limb extension,
with onset of the SCM muscles prior to the CE mus-
cles. The sequence of muscle onset was therefore con-
sistent with the demands for control of the reactive
forces and displacements of the centre of gravity.
Earlier onset of the CE muscles with shoulder exion
is consistent with the requirement for control of the
exion moment with this movement and conversely the
earlier onset of the SCM muscles with shoulder exten-
sion is consistent with the requirement to control the
extension moment conrming the presence of direc-
tional specicity of neck muscle activation during rapid
upper limb movements. Bilateral shoulder abduction
may be expected to be rather ineective as a postural
perturbation, however, feed-forward neck muscle acti-
vation was still evident with this perturbation.
The asymmetry of unilateral movements produces
additional acceleration towards the ipsilateral side and
a moment around the vertical axis towards the moving
limb [26]. Similar to the muscle response with bilateral
movements, the neck exor and extensor muscles dis-
played feed-forward activation with unilateral move-
ments in all three directions. As mentioned, EMG
activity was not consistently present in the con-
tralateral CE muscles during unilateral upper limb
movement. The consistent activation of the SCM mus-
cle bilaterally may reect a better capability to counter
the torsional eect of unilateral arm movement. The
SCM muscle was also activated prior to the CE mus-
cles during unilateral upper limb extension, conrming
the directional specicity of activation identied during
bilateral upper limb extension.
van der Fits et al. [25] determined that increasing
task load, by adding weight to the moving arm or by
performing bilateral arm movements rather than uni-
lateral movements, reduced the latency of limb and
trunk postural muscle activity. Performing bilateral
upper limb movements did not alter the temporal
sequence of neck muscle activation identied in our
study. However, greater EMG amplitude was recorded
during bilateral arm movements for the SCM and CE
muscles. This most likely reects the greater forces
imposed on the cervical spine compared to unilateral
perturbations.
The present results corroborate the existence of feed-
forward activation of the neck muscles with rapid vol-
untary upper limb exion identied in pervious
Fig. 4. Relative latencies (ms) for the left sternocleidomastoid
(SCML), right sternocleidomastoid (SCMR) and cervical extensors
(CE) muscles during unilateral upper limb movements (mean SD,
N 54 contractions for exion and abduction and N 48 contrac-
tions for extension). indicates signicant dierence p < 0:05.
468 D. Falla et al. / Journal of Electromyography and Kinesiology 14 (2004) 463474
research [13,25,26]. Gurnkel et al. [13] showed back-
ward head acceleration 40:5 12:2 ms prior to unilat-
eral upper limb exion, which was associated with
activation of the splenius capitus and trapezius prior to
activation of the deltoid muscle. Although our results
did not demonstrate activation of neck muscles prior to
the onset of deltoid muscle, neck muscle activation
occurred within 50 ms of the deltoid onset indicating
that their contraction cannot be reexly mediated and
must be centrally initiated in a feed-forward manner
[2,16]. Diculty measuring the precise timing of EMG
onsets and dierences in the method of detecting EMG
onsets is the most likely explanation for minor dissim-
ilarities between results of dierent studies. For
example, in the study conducted by Gurnkel et al.
[13], EMG onsets were determined by visual inspection
of the raw and rectied EMG signals, which precludes
direct comparison with our results.
In addition to demonstrating feed-forward activation
of neck muscles during unilateral upper limb exion,
the results of this study revealed the existence of feed-
forward neck muscle activation with unilateral exten-
sion, unilateral abduction and bilateral upper limb
movement in all three directions. Furthermore, the
results highlight directional specicity of neck muscle
activation in response to moments imposed by the per-
turbation. There is still uncertainty as to whether feed-
forward neck muscle activity is contributing to stability
of the neck or helping to control the position of the
centre of gravity. As suggested by Gurnkel et al. [13],
given that anticipatory postural movements measured
at the hip, ankle and knee serve to regulate position of
the trunk, it seems unlikely neck muscle activity can
contribute substantially to minimise body movement.
This hypothesis was substantiated by the results of van
der Fits et al. [25] who demonstrated that position
dependency of agonist activity was restricted to more
caudally located postural muscles. This suggests that
muscles of the lower limb and trunk are especially
involved in the maintenance of equilibrium, i.e. keeping
the centre of mass within the limits of the support sur-
face. In contrast, neck muscle activation was inde-
pendent of the starting position suggesting neck muscle
activity probably serves to oppose the reaction forces
induced by arm movements [25].
Interestingly, the results of our study demonstrated
co-activation of the neck exor and extensor muscles
during perturbations in all directions. Co-activation of
Fig. 5. Normalised root mean square (nRMS) values (%) obtained for the sternocleidomastoid (SCM) and cervical extensors (CE) muscles
during left (L) and right (R) unilateral (U) and bilateral (B) upper limb exion (FLE), extension (EXE) and abduction (ABD).
D. Falla et al. / Journal of Electromyography and Kinesiology 14 (2004) 463474 469
the cervical exor and extensor muscles would seem
essential to ensure adequate stabilisation of the head, a
necessary function for vision. Since no analysis of head
motion was involved in the current study, the extrapol-
ation of our ndings to head movement and the impli-
cations for cervical stabilisation are potentially limited.
However, in a previous study [25], 80% of subjects
demonstrated less than 3
v
of head displacement with
rapid upper limb exion, which demonstrates minimal
head movement with this type of postural perturbation.
Feed-forward activation of neck muscles is most likely
a mechanism employed to maintain stability of the ves-
tibular system and visual eld and oer protection to
the cervical spine, rather than acting as a compensation
to the perturbation.
4.1. Repeatability of EMG onsets and nRMS values
Because of the low values of between-subject nSEM,
results for the deltoid reaction times and relative laten-
cies of the SCM and CE muscles, demonstrate
reliability that supports their use in clinical and
Table 2
Results of the Wilcoxon non-parametric matched paired test to detect dierences in the normalised root mean square values obtained for the ster-
nocleidomastoid (SCM) and cervical extensor (CE) muscles between bilateral (BI) and unilateral (UNI) rapid upper limb exion (FLEX), abduc-
tion (ABD) and extension (EXT) of the right (R) and left (L) arm.

indicates signicant dierence p < 0:05,

indicates signicant dierence
p < 0:01. Refer to Fig. 4 for the normalised root mean square values obtained for the SCM and CE muscles during each condition
BIFLEX BIABD BIEXT UNIFLEXR UNIFLEXL UNIABDR UNIABDL UNIEXTR UNIEXTL
LSCM
BIFLEX
BIABD
BIEXT
UNIFLEXR
UNIFLEXL
UNIABDR
UNIABDL
UNIEXTR
UNIEXTL
RSCM
BIFLEX
BIABD
BIEXT
UNIFLEXR
UNIFLEXL
UNIABDR
UNIABDL
UNIEXTR
UNIEXTL
LCE
BIFLEX
BIABD
BIEXT
UNIFLEXR
UNIFLEXL
UNIABDR
UNIABDL
UNIEXTR
UNIEXTL
RCE
BIFLEX
BIABD
BIEXT
UNIFLEXR
UNIFLEXL
UNIABDR
UNIABDL
UNIEXTR
UNIEXTL
470 D. Falla et al. / Journal of Electromyography and Kinesiology 14 (2004) 463474
research evaluations. As previously documented
[21,22], when the between-subject variability is compa-
rable to (or less than) the within-subject variability, the
degree of repeatability dened by the ICC becomes
meaningless (very low or negative). The low values
obtained in this study for both between- and within-
subject nSEM for the deltoid reaction times and rela-
tive latencies of the SCM and CE demonstrate that
these variables could be estimated with very high
repeated measures precision. However, it also shows
that there was very little variation in the reaction times
and relative latencies obtained across subjects and
trials, so the measure might not be able to detect dier-
ences among uniform groups (such as asymptomatic
subjects). The EMG onsets of the SCM and CE mus-
cles might therefore be best utilised in providing a ref-
erence range for a uniform group of subjects (e.g.
normative data).
Several muscles displayed low estimates of the nSEM
for the nRMS values (see Table 4), which are compara-
ble with repeatability results of EMG amplitude
obtained from neck muscles during isometric con-
ditions obtained in a previous protocol [10]. The low
estimates of the nSEM for the nRMS values oers
Table 3
Repeatability of deltoid (DELT) reaction times and relative latencies of the sternocleidomastoid (SCM) and cervical extensors (CE) during bilat-
eral (BI) and unilateral (UNI) rapid upper limb exion (FLEX), abduction (ABD) and extension (EXT). Intraclass correlation coecient (ICC)
reports variance due to subjects; variance contributions due to days and trials are reported from analysis of variance results. Normalised standard
error of the mean (nSEM) reports deviation within- and between-subjects
Movement Muscle ICC Days Trials Within-subjects
nSEM
Between-subjects
nSEM
BIFLEX LSCM 29.6 43.4 27.0 3.7 3.8
RSCM 31.6 40.6 27.8 4.0 4.1
LCE 29.4 40.0 30.6 4.2 4.0
RCE 39.7 31.9 28.4 5.0 5.5
LDELT 23.8 44.5 31.7 4.6 4.0
RDELT 24.9 38.7 36.4 4.4 3.8
BIABD LSCM 36.9 14.8 48.3 5.1 4.7
RSCM 24.7 13.5 61.8 4.8 3.7
LCE 32.7 11.4 55.9 4.3 3.9
RCE 30.1 17.3 52.6 4.9 4.1
LDELT 28.6 17.0 54.4 5.6 4.6
RDELT 30.1 6.8 63.1 5.3 4.4
BIEXT LSCM 43.6 26.0 30.4 4.2 5.0
RSCM 31.2 40.7 28.1 4.6 4.5
LCE 40.1 24.2 35.7 3.8 4.2
RCE 48.2 23.8 28.0 4.1 5.3
LDELT 53.6 12.7 33.7 4.3 5.4
RDELT 48.2 25.3 26.5 4.4 5.7
UNIFLEXR LSCM 5.7 31.7 74.0 5.0 1.9
RSCM 21.5 12.2 66.3 4.9 3.3
RCE 38.2 6.8 55.0 4.3 4.0
RDELT 25.0 12.5 62.5 4.9 3.6
UNIFLEXL LSCM 9.2 12.6 78.2 4.3 2.3
RSCM 18.7 27.0 54.3 4.3 3.2
LCE 11.9 19.9 68.2 4.3 2.6
LDELT 9.7 14.9 75.4 4.6 2.6
UNIABDR LSCM 24.4 24.4 51.2 4.3 3.4
RSCM 34.5 22.4 43.1 4.4 4.2
RCE 27.9 24.4 47.7 4.6 3.8
RDELT 38.4 11.5 50.1 4.3 4.1
UNIABDL LSCM 45.5 6.0 60.5 4.7 5.2
RSCM 46.3 10.4 43.3 4.7 4.9
LCE 11.9 19.9 68.2 4.3 2.6
LDELT 34.9 2.2 67.3 5.9 5.3
UNIEXTR LSCM 16.4 36.1 47.5 4.4 3.4
RSCM 33.2 20.4 46.4 4.5 4.1
RCE 20.2 30.1 49.7 4.2 3.3
RDELT 40.3 25.1 34.6 4.3 4.7
UNIEXTL LSCM 38.3 1.9 59.8 4.2 4.0
RSCM 30.0 6.0 64.0 4.9 4.0
LCE 21.9 27.7 50.4 4.4 3.6
LDELT 26.8 17.3 56.1 4.6 3.8
D. Falla et al. / Journal of Electromyography and Kinesiology 14 (2004) 463474 471
some support for their use in future clinical and
research applications as providing a reference range for
asymptomatic subjects. Further research is indicated to
determine the utility of both EMG onsets and nRMS
measures in demonstrating dierences between asymp-
tomatic and symptomatic subjects.
5. Conclusion
The results of this study provide a greater under-
standing of cervical neuromotor control strategies and
a reference framework to be used for comparison with
symptomatic subjects. During bilateral and unilateral
upper limb perturbations, the SCM and CE muscles
demonstrate feed-forward activation. The temporal
sequence and magnitude of this activation displayed
directional specicity, however, the neck exors and
extensor muscles were co-activated during perturba-
tions in all directions. It seems apparent that feed-for-
ward activation of neck muscles is a mechanism
necessary to achieve head stability for the visual and
vestibular systems, whilst ensuring stabilisation and
protection of the cervical spine.
Table 4
Repeatability of absolute root mean square values for the deltoid (DELT) muscles and normalised root mean square values for the sternocleido-
mastoid (SCM) and cervical extensors (CE) during bilateral (BI) and unilateral (UNI) rapid upper limb exion (FLEX), abduction (ABD) and
extension (EXT)
Movement Muscle ICC Days Trials Within-subjects
nSEM
Between-subjects
nSEM
BIFLEX LSCM 17.8 41.9 40.3 10.0 7.8
RSCM 38.2 35.2 26.6 9.4 11.3
LCE 47.5 19.7 32.8 8.5 10.8
RCE 7.6 41.8 50.6 9.2 6.7
LDELT 14.3 56.1 29.6 9.4 7.9
RDELT 68.6 11.4 20.0 8.3 14.5
BIABD LSCM 66.3 14.8 18.9 8.4 13.6
RSCM 42.5 19.2 38.3 8.7 9.4
LCE 54.5 26.4 19.1 9.1 12.2
RCE 49.9 27.3 22.8 9.1 11.5
LDELT 67.5 16.2 16.3 7.7 12.6
RDELT 57.9 17.6 24.5 8.1 12.2
BIEXT LSCM 57.3 23.9 18.8 8.5 12.6
RSCM 51.9 11.8 36.3 8.2 10.6
LCE 27.5 48.7 23.8 8.5 8.6
RCE 24.9 39.3 35.8 9.0 7.6
LDELT 13.0 85.7 27.3 13.9 6.6
RDELT 16.3 68.3 15.4 10.8 10.2
UNIFLEXR LSCM 57.0 22.2 20.8 9.4 15.2
RSCM 31.5 20.1 38.4 9.9 11.2
RCE 7.8 48.4 43.8 10.3 6.9
RDELT 45.6 40.9 13.5 11.5 17.3
UNIFLEXL LSCM 47.1 16.2 36.7 9.2 11.2
RSCM 50.6 28.1 21.3 11.8 15.7
LCE 23.1 35.6 41.3 9.1 7.1
LDELT 38.4 12.2 49.4 9.4 8.9
UNIABDR LSCM 81.5 2.0 20.5 7.7 16.5
RSCM 39.6 18.5 41.9 9.4 10.1
RCE 47.2 112.5 40.3 8.3 9.7
RDELT 56.1 33.3 10.6 8.7 13.3
UNIABDL LSCM 35.7 21.4 42.9 10.7 11.0
RSCM 37.5 25.4 37.1 9.1 9.5
LCE 21.4 46.4 32.2 9.9 9.2
LDELT 65.1 21.8 13.1 8.7 13.6
UNIEXTR LSCM 62.9 8.6 28.5 8.4 13.4
RSCM 35.9 27.6 36.5 9.3 9.9
RCE 45.6 16.0 38.4 1.0 8.5
RDELT 19.5 51.4 29.1 13.3 15.9
UNIEXTL LSCM 65.1 8.8 26.1 8.8 15.7
RSCM 51.0 10.8 38.2 8.6 10.9
LCE 25.6 28.6 45.8 8.1 7.7
LDELT 36.9 39.8 23.3 10.8 13.5
472 D. Falla et al. / Journal of Electromyography and Kinesiology 14 (2004) 463474
Acknowledgements
The authors wish to thank A. Merlo and D. Farina
for providing the algorithms used for muscle activity
detection.
References
[1] H. Alaranta, M. Moroid, L. Elmqvist, J. Held, M. Pope,
P. Renstrom, Postural control of adults with musculoskeletal
impairment, Crit. Rev. Phys. Rehabil. Med. 6 (4) (1994)
337370.
[2] A.S. Aruin, M.L. Latash, Directional specicity of postural mus-
cles in feed-forward postural reactions during fast voluntary arm
movements, Exp. Brain Res. 103 (1995) 323332.
[3] J.J. Bartko, The intraclass correlation coecient as a measure of
reliability, Psychol. Rep. 19 (1) (1966) 311.
[4] A. Bottin, P. Rebecchi, Impedance of the skin-electrode interface
in surface EMG recordings, Proceedings of the XIV Congress of
the International Society of Electrophysiology and Kinesiology,
Vienna, Austria, 2002.
[5] S. Bouisset, M. Zattara, Biomechanical study of the program-
ming of anticipatory postural adjustments associated with volun-
tary movement, J. Biomech. 20 (1987) 735742.
[6] B.J. Connemann, P.P. Urban, V. Luttkopf, H.C. Hopf, A fully
automated system for the evaluation of masseter silent periods,
Electroencephalogr. Clin. Neurophysiol. 105 (1) (1997) 5357.
[7] P.J. Cordo, L.M. Nashner, Properties of postural adjustments
associated with rapid arm movements, J. Neurophysiol. 47 (2)
(1982) 287302.
[8] A.G. Cresswell, L. Oddsson, A. Thorstensson, The inuence of
sudden perturbations on trunk muscle activity and intra-abdomi-
nal pressure, Exp. Brain Res. 98 (1994) 336341.
[9] D. Falla, P. DallAlba, A. Rainoldi, R. Merletti, G. Jull, Identi-
cation of innervation zones of sternocleidomastoid and scalene
muscles: a basis for clinical and research electromyography
applications, Clin. Neurophysiol. 113 (1) (2002) 5763.
[10] D. Falla, P. DallAlba, A. Rainoldi, R. Merletti, G. Jull, Repeat-
ability of surface EMG variables in the sternocleidomastoid and
anterior scalene muscles, Eur. J. Appl. Physiol. 87 (2002)
542549.
[11] D. Falla, G. Jull, P. DallAlba, A. Rainoldi, R. Merletti, An
electromyographic analysis of the deep cervical exor muscles
during cranio-cervical exion, Phys. Ther. 83 (2003) 899906.
[12] W.G. Friedli, M. Hallett, S.R. Simon, Postural adjustments
associated with rapid voluntary arm movements 1. Electromyo-
graphic data, J. Neurol. Neurosurg. Psychiatry 47 (6) (1984)
611622.
[13] V.S. Gurnkel, M.I. Lipshits, F.G. Lestienne, Anticipatory neck
muscle activity associated with rapid arm movements, Neurosci.
Lett. 94 (12) (1988) 104108.
[14] P. Hodges, Motor Control of Transversus Abdominis for Stabil-
isation of the Lumbar Spine, Department of Physiotherapy, The
University of Queensland, Brisbane, Australia, 1996.
[15] P. Hodges, B.H. Bui, A comparison of computer-based methods
for the determination of onset of muscle contraction using elec-
tromyography, Electroencephalogr. Clin. Neurophysiol. 101 (6)
(1996) 511519.
[16] P.W. Hodges, C.A. Richardson, Feedforward contraction of
transversus abdominis is not inuenced by the direction of arm
movement, Exp. Brain Res. 114 (2) (1997) 362370.
[17] F.B. Horak, P. Esselman, M.E. Anderson, M.K. Lynch, The
eects of movement velocity, mass displaced, and task certainty
on associated postural adjustments made by normal and hemi-
plegic individuals, J. Neurol. Neurosurg. Psychiatry 47 (9) (1984)
10201028.
[18] W.A. Lee, T.S. Buchanan, M.W. Rogers, Eects of arm acceler-
ation and behavioral conditions on the organization of postural
adjustments during arm exion, Exp. Brain Res. 66 (2) (1987)
257270.
[19] A. Merlo, D. Farina, R. Merletti, A fast and reliable technique
for muscle activity detection from surface EMG signals, IEEE
Trans. Biomed. Eng. 50 (3) (2003) 316323.
[20] L. Nannucci, A. Merlo, R. Merletti, A. Rainoldi, R. Bergamo,
G. Melchiorri, D. Lucchetti, I. Caruso, D. Falla, G. Jull, Atlas
of the innervation zones of upper and lower extremity muscles,
Proceedings of the XIV Congress of the International Society of
Electrophysiology and Kinesiology, Vienna, Austria, 2002.
[21] A. Rainoldi, J.E. Bullock-Saxton, F. Cavarretta, N. Hogan,
Repeatability of maximal voluntary force and of surface EMG
variables during voluntary isometric contraction of quadriceps
muscles in healthy subjects, J. Electromyogr. Kinesiol. 11 (6)
(2001) 3342.
[22] A. Rainoldi, G. Galardi, L. Maderna, G. Comi, L. Lo Conte,
R. Merletti, Repeatability of surface EMG variables during vol-
untary isometric contractions of the biceps brachii muscle, J.
Electromyogr. Kinesiol. 9 (2) (1999) 105119.
[23] R. Shiavi, C. Frigo, A. Pedotti, Electromyographic signals
during gait: criteria for envelope ltering and number of strides,
Med. Biol. Eng. Comput. 36 (2) (1998) 171178.
[24] A.J. Thexton, A randomisation method for discriminating
between signal and noise recordings of rhythmic electromyo-
graphic activity, J. Neurosci. Methods 66 (2) (1996) 9398.
[25] I.B.M. van der Fits, A.W.J. Kilp, L.A. van Eykern, M. Hadders-
Algra, Postural adjustments accompanying fast pointing move-
ments in standing, sitting and lying adults, Exp. Brain Res. 120
(1998) 202216.
[26] M. Zattara, S. Bouisset, Posturo-kinetic organisation during the
early phase of voluntary upper limb movement. 1. Normal sub-
jects, Journal of Neurology, Neurosurgery and Psychiatry 51
(1988) 956965.
Deborah Falla received her Ph.D. in Physio-
therapy from The University of Queensland,
Australia in 2003. She is currently a research
ocer in the Cervical Spine and Whiplash
Research Unit at The University of Queens-
land. Her research focus involves the inte-
gration of neurophysiological and clinical
research to evaluate the pathophysiology
underlying muscle impairment in people with
a history of neck pain. Her research also
focuses on investigating the physiological
mechanisms that underpin the ecacy of therapeutic exercise for the
treatment of patients with neck pain.
Alberto Rainoldi was born in Torino, Italy
and graduated in Physics (1991) from Torino
University. He obtained the Ph.D. in Physi-
cal Medicine and Rehabilitation (2002) at
the Faculty of Medicine, Universita` di Roma
Tor Vergata, Italy. Since 1996, he has
been with the Laboratory for Neuromuscu-
lar System Engineering and Motor Rehabili-
tation at Politecnico di Torino, where his
research activity is focused on: the assess-
ment of proper methodologies useful to
increase the standardisation and the repeatability of EMG measures,
allowing their clinical use; the feasibility study of a technique of mus-
cle ber type estimation based on surface EMG; the assessment of
D. Falla et al. / Journal of Electromyography and Kinesiology 14 (2004) 463474 473
EMG signal modication in dierent clinical pathologies. He is also a
scientic consultant of the S. Maugeri Foundation (Pavia, Italy) for
the research projects in the eld of muscle fatigue.
Roberto Merletti graduated in Electronics
Engineering from Politecnico di Torino,
Italy, and obtained the M.S. and the Ph.D.
in Biomedical Engineering from The Ohio
State University. Since 1984, he is Associate
Professor of Biomedical Instrumentation at
the Department of Electronics of Politecnico
di Torino. From 1989 to 1994, he was
Associate Professor at the Department of
Biomedical Engineering of Boston University
and Research Associate at the Neuromuscu-
lar Research Center of the same University. In 1996, he founded the
Laboratory for Neuromuscular System Engineering at Politecnico di
Torino, where he is currently Director. He is Coordinator of a pro-
ject sponsored by the European Community and of a project spon-
sored by the European Space Agency. His research focuses on
detection, processing and interpretation of surface EMG, on electri-
cal stimulation and neuromuscular control.
Gwendolen Jull is an Associate Professor and Head of the Division of
Physiotherapy at The University of Queensland where she leads the
Cervical Spine and Whiplash Research Unit. Her particular research
and clinical interests are in establishing the pathophysiological pro-
cesses involved in the pain and muscle systems of patients with neck
disorders, both of insidious and traumatic onset (whiplash), with the
subsequent development of research based physical therapy interven-
tions. The research is also concerned with testing the ecacy of spe-
cic exercise programs for the management of neck pain in
randomised controlled trials as well as investigating the physiological
mechanisms of eect of therapeutic exercise for the cervico-brachial
region.
474 D. Falla et al. / Journal of Electromyography and Kinesiology 14 (2004) 463474