Anda di halaman 1dari 13

J Appl Physiol 99: 141153, 2005.

First published October 15, 2004; doi:10.1152/japplphysiol.00494.2004.

Neck muscle fatigue and spatial orientation during stepping in place

in humans
Micaela Schmid1 and Marco Schieppati1,2
Human Movement Laboratory, Centro Studi Attivita` Motorie, Fondazione Salvatore Maugeri, Scientific Institute of
Pavia, and 2Department of Experimental Medicine, Section of Human Physiology, University of Pavia, Pavia, Italy
Submitted 10 May 2004; accepted in final form 10 August 2004

Schmid, Micaela, and Marco Schieppati. Neck muscle fatigue Less amazing, but certainly nonnegligible, effects can be
and spatial orientation during stepping in place in humans. J Appl produced by neck muscle vibration in normal subjects, in
Physiol 99: 141153, 2005. First published October 15, 2004; particular on body segment orientation and body sway during
doi:10.1152/japplphysiol.00494.2004.Neck proprioceptive input, quiet stance and on walking features. In standing subjects, neck
as elicited by muscle vibration, can produce destabilizing effects on
muscle vibration induces body tilt and increases sway, sug-
stance and locomotion. Neck muscle fatigue produces destabilizing
effects on stance, too. Our aim was to assess whether neck muscle gesting that posture is organized with respect to a body
fatigue can also perturb the orientation in space during a walking task. schema, to the construction of which the spindle input from
Direction and amplitude of the path covered during stepping in place the neck contributes together with eye and skeletal muscle (15,
were measured in 10 blindfolded subjects, who performed five 30-s 23, 32, 38, 44, 54, 55). During stepping in place, bilateral
stepping trials before and after a 5-min period of isometric dorsal neck vibration of the dorsal neck muscles produces an involuntary
muscle contraction against a load. Neck muscle electromyogram forward stepping without modifying the stepping frequency
amplitude and median frequency during the head extensor effort were and in treadmill locomotion produces an involuntary steplike
used to compute a fatigue index. Head and body kinematics were increase of walking speed (33). One-sided lateral neck muscle
recorded by an optoelectronic system, and stepping cadence was vibration, in the absence of vision, produces deviation of the
measured by sensorized insoles. Before the contraction period, sub-
trajectory during normal walking (3) and body rotation during
jects normally stepped on the spot or drifted forward. After contrac-
tion, some subjects reproduced the same behavior, whereas others stepping in place (4); under both conditions, the deviation is
reduced their forward progression or even stepped backward. The opposite to the vibrated site. On a different vein, Goodwin et
former subjects showed minimal signs of fatigue and the latter ones al. (20) originally described proprioceptive illusions induced
marked signs of fatigue, as quantified by the dorsal neck electromyo- by muscle vibration. It was later shown that neck muscle
gram index. Head position and cadence were unaffected in either vibration also elicits apparent motion of a stationary visual
group of subjects. We argue that the abnormal fatigue-induced affer- target and deviation of the perceived straight ahead, possibly
ent input originating in the receptors transducing the neck muscle by producing a change in the egocentric body-centered coor-
metabolic state can modulate the egocentric spatial reference frame. dinate system (1, 37, 62).
Notably, the effects of neck muscle fatigue on orientation are opposite Muscle vibration is certainly an artificial situation, with its
to those produced by neck proprioception. The neck represents a
major advantage being the good selectivity of its effects on the
complex source of inputs capable of modifying our orientation in
space during a locomotor task. muscle spindles primary terminations (11). Are there natu-
ral stimuli able to elicit similar effects, e.g., changes in length
locomotion; reference frame of the neck muscles connected with head posture? Head and
eyes systematically deviate toward the future direction of the
curved trajectory (21), suggesting an anticipatory role of the
AN IMPRESSIVE NUMBER OF PAPERS have recently appeared about
neck muscle for body steering while walking. Body sway
the multifarious roles of the neck muscle proprioceptive input
increases when subjects stand with their heads turned or
on the perception of body segment position and body orienta-
extended and their eyes closed (5, 58). Early observations
tion in space. A clear example of the relevance of neck input
suggested that cervical proprioceptive input could affect the
for the subjects reference frame comes from studies of neck
direction of body sway by acting on the central vestibular
vibration after cortical lesions leading to neglect of the con-
system (28). More recently, it has been shown that neck input
tralateral space and/or body. These studies clearly show that
has an influence on the direction of body displacement induced
neck vibratory treatment improves neglect symptomatology.
by galvanic vestibular stimulation (18). Moreover, natural
This clear superiority over other treatments might result from
proprioceptive stimulation (head rotation) affects the percep-
the partial (re)activation of a distributed, multisensory network
tion of the head position, object localization, and movement in
in the lesioned hemisphere (Ref. 36; for a review, see Ref. 39)
the visual space (46, 48). It has also been shown that vestibu-
that is triggered by the spindle afferent input activated by the
lar-proprioceptive interactions play a role in self-motion per-
vibration. Most likely, the parietal cortex plays a role in the
ception in addition to postural control (47).
egocentric representation of space and its calibration, since
On the other hand, disturbances of the neck musculature
many of its areas receive signals from the neck muscles and the
could bring about dizziness or unsteadiness, the so-called
labyrinth (2).
cervicogenic dizziness (6). A report by Schieppati et al. (56)

Address for reprint requests and other correspondence: M. Schieppati,

Centro Studi Attivita` Motorie (CSAM), Fondazione Salvatore Maugeri The costs of publication of this article were defrayed in part by the payment
(IRCCS), Istituto Scientifico di Pavia, Via Ferrata 8, I-27100 Pavia, Italy of page charges. The article must therefore be hereby marked advertisement
(E-mail: in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

http://www. 8750-7587/05 $8.00 Copyright 2005 the American Physiological Society 141

showed that prolonged contraction of dorsal neck muscles, of !30 min separated the postcontraction 1 block from the recovery
bilaterally, can alter balance control through a mechanism block. The subjects preparation with the markers and electrodes took
connected to fatigue-induced afferent inflow. Strong and pro- !30 min. The entire stepping session lasted !1 h. The same subjects
longed neck muscle contractions have also been recently have also been recruited in an additional session, in a different day,
when they were asked to step in place and again perform four blocks
shown to produce postural responses (14). Fatigue-related of five subsequent stepping trials each, but without tonic isometric
sensory input may be the result of an increased outflow from contraction of the dorsal neck muscles (no-load session). This proce-
the free nerve endings as a consequence of ionic or metabolic dure was performed as a control experiment to check that the repeti-
changes (elevated interstitial potassium concentration or insuf- tion of the stepping trials per se produced no effect on body displace-
ficient oxygen availability due to reduced blood flow) (19, 52). ment and to help in interpreting highly variable data.
Muscle fatigue can indeed affect joint position sense (Refs. 12, Isometric contraction of dorsal neck muscles. A large belt was
57, 60; for a recent review, see Ref. 63), and abnormal cervical passed around the head just above the ears, and a cable was fixed to
pain-related input is able to significantly alter postural control it frontally and medially in correspondence with the forehead. The
(35). cable passed through a pulley fixed at a 1-m distance at head height.
In this study, we have tested the hypothesis that dorsal neck The pulley was part of a versatile gym training column placed in front
of the subject. Its vertical position was adjustable so that the line of the
muscle fatigue could affect the reference system for the orien- pulling action was horizontal. The other end of the cable was fixed to
tation of body position in space. In particular, the effects of a an adjustable mass. The thorax of the subject leaned against a padded
fatiguing contraction have been tested on the capacity of vertical support firmly fixed to the cable column so that the action of
stepping in place. This task reproduces several features of the mass did not displace the trunk and overall body posture. No
normal walking (8), such as the cyclic single-leg support of the contraction of the body extensor muscle chain below the support level
body and the need for accurate control of balance. It can also was required, since there was no need to push the trunk or whole body
be more easily captured by a movement-analysis system for backward, owing to the thorax support, which nullified the body
extended periods of time than by normal locomotion. As such, forward-pulling action of the mass. Therefore, the muscle action was
this task should be at least as sensitive as normal floor walking broadly limited to the neck extensors, and the subject counteracted the
to possible disorientation-producing stimuli. Perhaps, it could head-flexor torque exerted by the mass by contracting the dorsal neck
muscles. The subjects had visual feedback of the head position by way
be even more sensible than normal walking, because, contrary of a target fixed to the cable, which had to remain coplanar with a
to the latter, it lacks the advantage given by the kinetic energy reference fixed to the column. This was sufficient enough to ensure
of the body progressing in one direction. The stepping-in-place negligible head pitch during the head extensor effort against the
task has been previously employed as a test for vestibular head-flexor torque produced by the mass. For each subject, the mass
deficits and balance problems, even if its reliability in this was chosen so that it represented about one-third of the maximal
context has been repeatedly challenged (50). Because the voluntary contraction (MVC; 25.5 " 8.5 kg), i.e., 9.5 " 2.0 kg, equal
isometric contraction of the dorsal neck muscles was bilateral to 34.9 " 9.5% of MVC. MVC was the highest force value exerted
and symmetrical, we assumed that postcontraction changes in during a 10-s period of neck extension, during which the subject was
the body position during the stepping-in-place test could be encouraged to produce the strongest possible effort. This was mea-
detected along the subjects sagittal plane, much like it oc- sured by means of a mechanical dynamometer placed in series
between the head belt and the pulley. During the fatiguing periods, the
curred as a consequence of bilateral dorsal neck muscle vibra- subject had to counteract the applied mass for 5 min. At the end of this
tion, which induced an increased velocity in the stepping period, the mass and the padded support were quickly disconnected
progression along the straight-ahead direction (32). and removed.
Electromyogram recording and processing. Electrical activity of
METHODS the dorsal neck (splenius) and ventral neck (sternocleidomastoid)
muscles (from both right and left side) was recorded using bipolar
Subjects. Ten healthy subjects (9 men and 1 women, mean age 26.2 prejelled surface electrodes, with a longitudinal distance of 1 cm
yr, range 2233 yr) volunteered to participate. No subject had a between the recording spots. The electrode pairs were placed on the
history of neck pain or cervical column disease or had suffered head bellies of the dorsal neck muscles, 2 cm from the midline and 4 cm
trauma or whiplash injury. Informed consent was obtained according below the cranial insertion, and on the belly of the sternocleidomas-
to the Declaration of Helsinki. The local Ethics Committee approved toid, about halfway from its proximal and distal insertion points. The
the investigation and experimental procedures. electrodes were left in place during the entire recording session. The
Procedures. Subjects were asked to step in place at a self-selected signals were differentially amplified (#1,000), filtered (cutoff fre-
pace, blindfolded in a dim-lighted room. Before the start of the quency of 500 Hz), and acquired at a sampling frequency of 1 kHz.
acquisition session, subjects were free to step in place with eyes open Data were recorded by the Telemg Multichannel Electromyograph
to feel confident with the maneuver. Then four blocks of stepping System (Bioengineering Technology Systems). The signals were
trials were acquired, each composed of five successive trials. Each acquired for the initial 10-s period of each minute of the 5-min
trial lasted 30 s. After each trial, the subjects were repositioned at the contraction period. The analysis of the electromyogram (EMG) was
starting point by one experimenter before the next trial began. It was performed with the Acknowledge software (Biopac), and the EMG
checked that, at the beginning of each trial, the subjects started from amplitude (AMP) of the filtered and integrated signal and the signals
the same initial standing position, with the orientation of feet, trunk, median power frequency (MF) were calculated over the recorded
and head aligned. The starting signal was given by the experimenter epochs (49). The EMG activity plots of the splenius and sternoclei-
after a verbal warning signal. The four blocks were the control, domastoid muscles, during the 10 s of minute 1 and the 10 s of minute
postcontraction 1, recovery, and postcontraction 2 blocks. After the 5, are shown in Fig. 1.
control and recovery blocks, subjects performed a 5-min-long tonic Procedure of quantification of muscle fatigue based on EMG
isometric bilateral contraction of the dorsal neck muscles against a signal. Based on the changes in the AMP and MF between the
load (see Isometric contraction of dorsal neck muscles) while standing beginning and end of the contracting period, an EMG index of fatigue
on the starting spot. At the end of this period, subjects immediately was calculated for each subject using the following formula: index of
started the first trial of the postcontraction (1 or 2) block. A rest period fatigue $ ratio of AMP/ratio of MF, where the ratio of AMP $ (AMP

J Appl Physiol VOL 99 JULY 2005


Fig. 1. Examples of traces of surface electromyogram (EMG), recorded from a dorsal [splenius (SPL), left] and anterior [sternocleidomastoid (SCM), right] neck
muscle in one subject. The raw traces have been recorded during the first (top) and last minute (bottom) of the 5-min tonic isometric bilateral contraction of the
dorsal neck muscles against a load. The selectivity of the contraction is witnessed by the silence of the anterior flexor muscle (SCM, right), both at the beginning
and the end of the contraction period. The development of fatigue is accompanied by an increase in amplitude of the SPL EMG from the first to the last minute
of contraction (compare top and bottom) and by a shift toward lower frequency values of the median frequency (MDF) in the power spectrum of the EMG
(middle) observed at minute 5 of contraction.

of the 10-s period of minute 5/AMP of the 10-s period of minute 1) frequency changes, with each being computed as the ratio of the last
and the ratio of MF $ (MF of the 10-s period of minute 5/MF of the and first minute of tonic contraction. When n $ 2 clusters were
10-s perod of minute 1). imposed to the algorithm, the analysis divided the subjects into two
This index is influenced by two variables that are known to be equal-size groups. It turned out that the two groups had values smaller
affected by the fatigue phenomenon (49). To check that both variables or larger than 1.5. For simplicity, we will label the groups as fatigue
were influenced by fatigue in the present population, we plotted the and no-fatigue groups, respectively, even if the range of the indexes of
EMG median frequency vs. amplitude for each subject and for each the latter group is from 0.98 and 1.45, meaning that some subjects of
minute of the 5-min period of tonic contraction. Figure 2A shows four this group showed minimal EMG signs of fatigue.
examples of these relationships and the best-fit line for each of them. For each group, intermediate indexes of fatigue were also calcu-
In two cases, when the median frequency decreased, the amplitude lated by means of the above formula for the successive minutes, where
increased. In the other two cases, the best-fit line was less steep or flat. the minute 5 data were sequentially changed with the data of minutes
The angular coefficients of the best-fit lines were thus calculated for 2, 3, and 4. Figure 3A reports the results for the fatigue group, and Fig.
each subject and plotted against the corresponding indexes of fatigue. 3B reports results for the no-fatigue group. It can be appreciated that,
In Fig. 2B, the result of this procedure is reported. The strong linearity in the former group, the intermediate EMG indexes of fatigue mono-
of this scatter [y $ %3.56x & 3.66; r2 $ 0.91; F(1,8) $ 87.035; P $ tonically increased from the first to the last minute of the contraction
0.00001] is in favor of the appropriateness of the use of an index that period, whereas the same indexes remained stationary in the latter
combines both EMG variables. group.
Cluster analysis (K means) was performed on the EMG data of all To further validate the EMG index of fatigue against an indepen-
subjects by including in the analysis both amplitude and median dent measure of fatigue, we measured the MVC of the dorsal neck

Fig. 2. A: EMG amplitude vs. EMG median

frequency for each successive minute of the
5-min period of tonic contraction. These re-
lationships are shown for 4 subjects, and the
best-fit line for each of them is plotted. For 2
subjects (solid symbols), when the median
frequency decreased, the amplitude in-
creased. For the other subjects (open sym-
bols), the best-fit line was less steep or flat. B:
angular coefficients of the best-fit lines cal-
culated for all subjects are plotted against the
corresponding indexes of fatigue. For A and
B, the equations of the best-fit lines are re-

J Appl Physiol VOL 99 JULY 2005


midpoint, could be considered a good approximation of the displace-

ments of the center of mass. In the working space, subjects were
positioned at the starting point with their mediolateral (M-L) body
axis along the working space x-axis and their anteroposterior (A-P)
body along the y-axis. Figure 5A shows an example of body midpoint
path in the working field during a stepping-in-place trial. The x values
of the body midpoint path vs. time and the y values vs. time were fitted
with a polynomial function, the order of which minimized the mean
squared error (Fig. 5B). In most cases, the algorithm chose a 10th-
order polynomial as the best fit. By plotting the x vs. y interpolating
polynomials, we obtained the interpolated path. In Fig. 5A, this is
superimposed on the acquired body midpoint path. The mean squared
errors have been taken as a measure of the body midpoint sway (in the
x- and y-axes) around the main direction of displacement. The dis-
placement of the subjects in the working space was quantified by the
A-P and the M-L displacement of the body midpoint on the x-y plane.
A-P displacements were calculated as the differences between the y
coordinates of the body midpoint at the starting point and at the final
point, whereas the M-L displacements were the differences between
the x coordinates of the same points.
To compare the postcontraction displacement changes across the
subjects despite the possible different directions (forward or back-
ward) of the postcontraction paths, an A-P index of displacement was
calculated. For each subject, it was defined as the A-P displacement of
Fig. 3. Indexes of fatigue calculated from the EMG variables (amplitude and the first trial of the first postcontraction trial (postcontraction 1) minus
median frequency) for each successive minute of the isometric neck muscle the mean value of the A-P displacement of the control trials, divided
contraction. The mean indexes are reported for the subjects showing (A) and by the sum of the absolute values of these two numbers. This
not showing (B) signs of fatigue. It can be seen that, in the former group, there
procedure weights the differences against the absolute magnitude of
was a progressive increase in the index values during the contraction period.
The same mean indexes remained stationary in the latter group. the traveled paths. A positive sign of the index meant A-P forward
displacement, and a negative sign meant A-P backward displacement.
Head movements were computed on the basis of the angle between
muscles before and after the tonic contraction period. Then, we the line joining the helmet vertex to the helmet frontal marker and the
plotted the value of MVC after the tonic contraction (as percentage) of vertical axis versor (head pitch), and on the basis of the angle between
the initial MVC) against the EMG fatigue index. To this aim, 7 of the the line joining the helmet lateral right marker to the helmet lateral left
10 subjects performed the two MVCs on a different day, the former marker and the vertical axis versor (head roll).
before and the latter immediately after a 5-min period of dorsal neck Cadence recording. The cadence was recorded by the PedarMobile
tonic isometric contraction, against the same load they sustained in the Insole System (Novel, Munchen, Germany). The analog signal pro-
first experimental session. On average, the MVC recorded during the vided by the insole sensors was sampled at a frequency of 50 Hz, and
former trial was not different from that measured during the first the data were stored in the memory card to leave a subject free to
experimental session (n $ 7; MVC $ 26.4 " 5.9 kg). Because three move without cable constraints. After each acquisition block, the data
of the original subjects could not participate in this evaluation, to were stored on a personal computer. The cadence was computed from
strengthen the validation procedure we increased the population by the number of stance phases over time.
recruiting six additional subjects (n $ 6; MVC $ 28.1 " 2.4 kg). Statistical analysis. To test the significance of the differences in the
These subjects were subjected to the same procedure for measurement computed variables and parameters across the stepping trials of the
of MVC and EMG as the original subjects, and the load was chosen control block (before isometric contraction), the postcontraction 1
to be in the same range (11.3 " 1.4 kg). Figure 4 shows that the block, the recovery block, and the postcontraction 2 block, a two-way
percent drop in MVC induced by the tonic contraction significantly repeated-measures ANOVA (blocks and trials) was used. This anal-
increased [y $ %0.12x & 1.08; r2 $ 0.43; F(1,11) $ 8, 16; P $
0.015] with the increase in the EMG fatigue index.
The index was then used for further analysis of the fatigue effects
on stepping behavior, assuming that changes in the EMG index
correlated with peripheral fatigue (29).
Kinematic recording and processing. Body movement was re-
corded by the ELITE System (Bioengineering Technology and Sys-
tems) with eight infrared cameras. The reflective spherical markers
were attached to the skin on the right and left anterior-superior iliac
spine and to an adjustable helmet in these following positions: vertex
of the head, forehead (the central point between the two temples), and
lateral left and lateral right positions (the left and right temples,
respectively). The helmet was removed during the fatiguing procedure
and then repositioned, always in the same position, during the step-
ping trials. Before the helmet was removed, a sign of its position was
made on the skin. Fig. 4. Percent drop in maximal voluntary contraction (MVC) induced by the
The markers positions were sampled at a frequency of 100 Hz. The tonic contraction (MVC immediately after contraction divided by MVC before
body movement was obtained by computing the midpoint (average of contraction period) plotted against the fatigue index for 13 subjects. Despite
the coordinates) between the left and the right anterior-superior iliac great variability, the regression was significant. The best-fit line is plotted, and
spine marker positions. The resulting point, which was labeled body its equation is reported.

J Appl Physiol VOL 99 JULY 2005


Fig. 5. A: example of body midpoint path in the working field

for one trial of one subject during the stepping in place with
eyes closed under control condition. In this case, the subject
showed a forward progression of !1 m during the 30-s stepping
trial. The step-by-step body midpoint mediolateral (M-L) os-
cillation of '10 cm is clearly evident along the x-axis. The gray
line, superimposed on the body midpoint path, is the interpo-
lated curve that minimized the squared mean error. B: projec-
tions of the instantaneous positions of the body midpoint on the
M-L (top) and anteroposterior (A-P; bottom) plane. Superim-
posed on these projections are the respective interpolated

ysis was separately performed on the two groups of subjects (fatigue distance covered in the forward direction was equal to 48.7 "
and no-fatigue groups), divided on the cluster analysis mentioned 19.3 and 34.5 " 17.2 cm (means " SE; no-fatigue and fatigue
before. To check for a significant main effect of fatigue on A-P groups, P $ 0.56).
displacement for group, both groups were analyzed together in a Qualitative effects of the isometric contraction of dorsal
three-way ANOVA (fatigue vs. no-fatigue group, blocks and trials). A
two-way repeated-measures ANOVA (blocks and trials) was used to
neck muscles on the body midpoint path. The isometric con-
test for repetition effects on A-P displacement during stepping in traction of the dorsal neck muscles did not produce the same
place in the no-load trials; all subjects collapsed. A three-way repeat- changes in the body midpoint path in all the subjects. During
ed-measure ANOVA (load vs no-load, blocks and trials) in the group the trials performed in the period after the first contraction
of five subjects who showed the highest EMG fatigue indexes was (postcontraction 1), some subjects basically reproduced the
used for assessing possible repetition effects, as opposed to fatiguing same path already seen during the control blocks. It turned out
tonic contraction effects, on A-P displacement. When appropriate, the that, in these subjects, the neck muscle EMG signals during the
Newman-Keuls test was used for post hoc comparisons. Between the isometric contraction showed no or negligible signs of fatigue
two groups, the mean indexes of fatigue and the corresponding mean and so did the index of fatigue. On the contrary, other subjects
indexes of A-P displacement were compared by means of the Stu-
dents nonpaired t-test. When necessary, a linear regression analysis
was carried out. The level of significance at which the null hypothesis
was rejected was always equal to P ' 0.05.


Body midpoint path under control condition. The covered

path during the stepping-in-place trials was quite variable from
subject to subject, even during the block of control trials.
Nevertheless, it was possible to identify two dominant behav-
iors. Some subjects moved forward some distance along their
sagittal plane (Fig. 6, subject A). Other subjects remained
almost in place, and the movement of their body midpoint
generated a winding line covering a circumscribed area around
the starting point (Fig. 6, subject B). The behavior of each
subject was quite consistent within the trials of the control
block. The sway of the body midpoint, which was connected to
the pelvis movement during the shift of the body weight from
one foot to the other, was evident in all subjects and trials. This
was filtered out by the interpolating procedure that gave the
body midpoint path along the heading direction. Across the
subjects, large A-P displacements corresponded to long inter-
polated paths. This denotes that during the stepping-in-place
trials the predominant movement was along the A-P plane,
since, if a trial were characterized by conspicuous shifts of the
body midpoint from the straight-ahead direction, the length of
its interpolated path would be greater than the mean distance of
its A-P displacement. When all the control trials from all Fig. 6. Examples of different body midpoint displacements taken from 2
subjects, different from that of Fig. 5, during the stepping in place under
subjects were gathered into two groups, one showing and one control condition. Although subject A (A) exhibited a forward progression of
not showing EMG signs of fatigue (on the basis of the fatigue !40 cm, subject B (B) almost stepped on the spot throughout the 30-s trial. E,
index) and separately averaged, the mean value of the A-P Starting point (at y $ %140 cm and x $ 30 cm).

J Appl Physiol VOL 99 JULY 2005


Fig. 7. Examples of body midpoint paths in the working field

for 2 subjects (A and B) during the stepping in place with closed
eyes under each of the 4 experimental conditions (from top to
bottom: control, postcontraction 1, recovery, and postcontrac-
tion 2). For each condition, the 2 columns correspond to the first
and last (5th) stepping trials of each subject. Under all condi-
tions, the stepping trials began at the same spot (open circle in
each graph at x $ 30 and y $ %140 cm). Subject A reversed his
forward displacement in the first trial after the first period of
isometric dorsal neck muscle contraction; in the first trial after
the second contraction period, his displacement was directed
backward. At the 5th stepping trial of postcontraction 2, the
backward displacement disappeared. Subject B moved back-
ward from the beginning, but his backward displacement was
much larger in the first trials of the postcontraction blocks 1 and
2. Also in this case, the last trials of the postcontraction blocks
were similar to the control trials.

showed a different body midpoint path from that of the control other hand, the subject who during the control trials remained
block trials. Figure 7 shows the paths of the first and the fifth in place (subject B) moved backward after contraction, taking
trial of each block for two representative subjects (subjects A away from the starting point. The effects of the isometric
and B), in whom the neck muscle EMG during the isometric contraction were also reproduced in the first trial of the period
contraction showed clear-cut signs of fatigue. The graphs of after the second contraction (postcontraction 2). As shown in
the first trial after the contraction period (postcontraction 1) the bottom graphs of Fig. 7, both subjects moved backward,
showed different paths between the first and fifth trial of the farther from the starting point.
control block. In particular, the subject who, during the control Mean A-P displacement of body midpoint in the two groups
trials, moved forward (subject A) in this first trial remained of subjects: the effect of fatigue. The mean differences in the
roughly in place, just behind the starting point. During the trials A-P displacement and in the traveling direction (forward or
after the rest period (recovery block), the paths tended to backward) were separately analyzed in the subjects showing
recuperate the feature of the control trials without really repro- and not showing signs of fatigue based on the magnitude of the
ducing it. This means that, even after a rest period of 30 min, EMG index of fatigue (see METHODS). The subjects could in fact
this subject was unable to reach a complete recovery. On the be clustered into two groups, which had significantly different
J Appl Physiol VOL 99 JULY 2005
mean indexes of fatigue (fatigue group, index of fatigue of
(1.5, 2.302 " 0.31; no-fatigue group, index of fatigue of
'1.5, 1.146 " 0.19; P ' 0.001).
The corresponding mean indexes of displacement (A-P in-
dex) were for the fatigue group %0.49 " 0.33 and for the
no-fatigue group 0.0 " 0.17 (P ' 0.05). Figure 8 shows the
A-P indexes calculated for the first stepping trial after the first
isometric contraction period vs. the fatigue index of each
subject. It is possible to appreciate that, when the fatigue index
assumed values lower than 1.5, the A-P index was very close
to zero, whereas when the fatigue index was greater than 1.5,
the A-P index assumed much lower, negative values. It can
also be seen that, despite a considerable variability, there was
a significant inverse relationship between the two indexes [the
equation of the best-fit line was y $ %0.41x & 0.46; R2 $ 0.56;
Fig. 8. For each subject, the A-P index has been plotted vs. the fatigue index
(see METHODS). An A-P index equal to zero indicates no changes in the
F(1,8) $ 10.473; P $ 0.011].
displacement along the A-P axis, whereas a negative value indicates a back- Figure 9 shows the mean values of the body midpoint A-P
ward displacement after the first isometric contraction period. Across the displacements of the fatigue (Fig. 9A) and no-fatigue groups
subjects, the A-P indexes showed greater negative values as a function of (Fig. 9B) for each of the successive stepping trials. When both
increasing fatigue indexes. F, Subjects of the no-fatigue group; , subjects of
the fatigue group.
groups were analyzed together in a three-way ANOVA (fatigue

Fig. 9. A and B: mean values (&SE) of the body

midpoint A-P displacements for all subjects of
the fatigue (A) and no-fatigue (B) group for each
of all successive stepping trials under all exper-
imental conditions. Reduction of the A-P dis-
placement is obvious in the first trials of the
postcontraction periods 1 and 2 (Post C1 and
Post C2, respectively; shaded bars) in the fatigue
but not in the no-fatigue group. C and D: mean
values (&SE) of the body midpoint M-L dis-
placements for all the subjects, grouped as
above. M-L displacements were not affected by
the contraction periods 1 and 2 in either group.
E and F: neither the first nor second contraction
period produced significant M-L deviations of
the path, as assessed by the SD (along the
x-axis) of the interpolated path. Contr, control

J Appl Physiol VOL 99 JULY 2005


vs. no-fatigue group, blocks and trials) to check for a signifi- Mean M-L displacement of body midpoint in the two groups
cant main effect on A-P displacement for group, no significant of subjects: the effect of fatigue. The M-L displacements,
differences were found between groups [F(1) $ 0.66; P $ computed as the difference between the x coordinates of the
0.44]. There was no effect for trials [F(4) $ 0.79; P $ 0.54] starting and final points, were always small in the control trials
and a marginal effect for blocks [F(3) $ 2.65; P $ 0.07]. in the two groups. This behavior remained unchanged in all
However, there was an interaction between blocks and trials subsequent trials, independently of the experimental condition
[F(12) $ 2.44; P $ 0.008] and between groups, blocks, and and group (Fig. 9, C and D). The two-way ANOVA applied to
trials [F(12) $ 1.88; P $ 0.045]. In particular, the post hoc test the fatigue group gave neither significant main effects [blocks,
showed no difference between the trials of the control blocks of F(3) $ 0.631, P $ 0.609; trials, F(4) $ 0.573, P $ 0.686] nor
the two groups (P ( 0.02 for all comparisons). But there were significant interaction [F(12) $ 1.903; P $ 0.058]. The two-
significant differences between the two groups in the first trials way ANOVA applied to the no-fatigue group gave neither
of the second blocks (postcontraction 1, P $ 0.031) and in the significant main effects [blocks, F(3) $ 1.603, P $ 0.240;
first trials of the fourth blocks (postcontraction 2, P $ 0.0017). trials, F(4) $ 1.835, P $ 0.172] nor significant interaction
[F(12) $ 0.342; P $ 0.977]. To make sure that the subjects did
The fatigue group was characterized by a reduced A-P
not proceed along a winding trajectory during the stepping-in-
displacement in some postcontraction 1 trials as well as in
place trials to eventually stop at a point lying along the initial
some postcontraction 2 trials with respect to the control trials.
straight-ahead line, the mean standard deviation of the inter-
Two-way ANOVA showed that, although there were no dif- polated path was computed. As shown in Fig. 9, E and F,
ferences between blocks [F(3) $ 1.999; P $ 0.168] or between neither contraction condition produced significant changes in
trials [F(4) $ 0.313; P $ 0.865], there was an interaction the mean standard deviation. The two-way ANOVA applied to
between blocks and trials [F(12) $ 2.319; P $ 0.019]. The the mean standard deviation values of the interpolated path in
post hoc test showed that the significant reduction in A-P the fatigue group gave neither significant main effects [blocks,
displacement was present for the mean values of the first trials F(3) $ 0.205, P $ 0.891; trials, F(4) $ 0.176, P $ 0.947] nor
of the postcontraction 1 (compared with all control trials) and significant interaction [F(12) $ 0.417; P $ 0.949]. The two-
2 (compared with all control and recovery trials) blocks (P ' way ANOVA applied to the no-fatigue group gave neither
0.03 for all comparisons). The successive trials of postcontrac- significant main effects [blocks, F(3) $ 1.23, P $ 0.341; trials,
tions 1 and 2 showed a progressive but incomplete recovery F(4) $ 2.146, P $ 0.122] nor significant interaction [F(12) $
toward control mean values. 1.949; P $ 0.513]. Therefore, we conclude that subjects did
In the subjects showing only mild or no sign of fatigue, the not significantly deviate from a straight pathway under any
A-P displacements remained unchanged across the different experimental condition.
blocks and trials [2-way ANOVA; blocks, F(3) $ 0.849, P $ Body sway during stepping in place. There were no a priori
0.493; trials, F(4) $ 0.775, P $ 0.557]. There was a significant reasons for excluding the possibility that fatigue increased the
interaction between blocks and trials [F(12) $ 2.029; P $ body sway during stepping in place and that such instability, in
0.042]: the post hoc test showed a significant effect limited to turn, could be at least in part be responsible for the observed
the difference between the first trial of the last block (postcon- changes in the path of the postcontraction stepping trials. The
traction 2 blocks) and the last trial of the control block (P $ body midpoint sway around the main body midpoint direction
0.030). of displacement was evaluated on the basis of the mean
A-P displacement of body midpoint under no-load condition. squared error made by the path interpolation. This parameter
When the same subjects stepped in place (4 blocks of 5 estimated how much the actual path differed from the interpo-
subsequent stepping trials each), but this time without any lated path and, therefore, to what extent the subjects swayed on
interposed period of head extensor effort (no-load session), no the spot by moving the body weight from one foot to the other
effect of the repetition of the stepping trials on the body A-P during stepping in place. As shown in Fig. 10, A and B, fatigue
displacement was seen. The two-way ANOVA applied to the did not increase body sway, neither in the A-P axis nor on the
M-L axis. The two-way ANOVA applied to the A-P data of the
A-P displacement of all block and trials in all subjects col-
fatigue group gave neither significant main effects [blocks,
lapsed (n $ 10) gave neither a significant main effect [blocks,
F(3) $ 0.404, P $ 0.753; trials, F(4) $ 0.786, P $ 0.551] nor
F(3) $ 2.295, P $ 0.100; trials, F(4) $ 1.116, P $ 0.364] nor
significant interaction [F(12) $ 0.779; P $ 0.668]. The two-
a significant interaction [F(12) $ 1.05; P $ 0.409]. The way ANOVA applied to the M-L data of the fatigue group gave
ANOVA was then applied to the subjects belonging to the neither significant main effects [blocks, F(3) $ 2.013, P $
fatigue group alone, with the aim of comparing their A-P 0.166; trials, F(4) $ 1.013, P $ 0.430] nor significant inter-
displacement between the two experimental conditions (iso- action [F(12) $ 1.907; P $ 0.057].
metric contraction vs. no load). The three-way ANOVA (con- Head position during stepping in place. The position of the
dition, block, trial) gave no significant effect for conditions head was measured to check whether the isometric contraction
[F(1) $ 0.008; P $ 0.932] and trials [F(4) $ 0.213; P $ of the dorsal neck muscles provoked flexion or extension of the
0.930] but a main effects for blocks [F(3) $ 3.603; P $ 0.028]. head, since these modifications could be one of the possible
There was a significant interaction (condition, block, trial) causes of the fatigue-induced variations of the path. The mean
[F(12) $ 1.983; P $ 0.034]. In particular, the post hoc test values of pitch angle as well as roll angle did not show
indicated that the first trials of the second and fourth block of differences among the trials (Fig. 11). The two-way ANOVA
the load condition exhibited a smaller A-P displacement (sec- applied to the pitch angle data of the fatigue group gave neither
ond block, P $ 0.049; fourth block, P $ 0.001) than the significant main effects [blocks, F(3) $ 1.715, P $ 0.217;
corresponding trials of the no-load condition. trials, F(4) $ 1.68, P $ 0.951] nor significant interaction
J Appl Physiol VOL 99 JULY 2005
[F(12) $ 1.041; P $ 0.428]. The two-way ANOVA applied to
the pitch angle data of the no-fatigue group gave neither
significant main effects [blocks, F(3) $ 1.749, P $ 0.210;
trials, F(4) $ 0.647, P $ 0.637] nor significant interaction
[F(12) $ 1.037; P $ 0.432]. The two-way ANOVA applied to
the roll angle data of the fatigue group gave neither significant
main effects [blocks, F(3) $ 1.780, P $ 0.204; trials, F(4) $
0.309, P $ 0.867] nor significant interaction [F(12) $ 0.547;
P $ 0.872]. The two-way ANOVA applied to the roll angle
data of the no-fatigue group gave neither significant main
effects [blocks, F(3) $ 0.508, P $ 0.684; trials, F(4) $ 0.201,
P $ 0.934] nor significant interaction [F(12) $ 1.952; P $
0.051]. Therefore, the isometric contraction of the dorsal neck
muscles, even when it reached a fatiguing level, did not affect
the position of the head during stepping in place.
Stepping-in-place cadence. The mean stepping cadence un-
der control conditions was not significantly different between
the fatigue and the no-fatigue groups (54.0 " 3.5 and 59.3 "
5.2 steps/min, respectively). There were minor effects of iso-
metric neck muscle contraction on stepping cadence. The
two-way ANOVA applied to the fatigue group gave neither
significant main effects [blocks, F(3) $ 2.172, P $ 0.144;
trials, F(4) $ 0.185, P $ 0.943] nor significant interaction
[F(12) $ 1.645; P $ 0.111]. The two-way ANOVA applied to
the no-fatigue group gave a significant main effects for trials
[blocks, F(3) $ 0.217, P $ 0.883; trials, F(4) $ 3.347, P $
0.036] but no significant interaction [F(12) $ 0.849; P $
0.602]. The significant changes between trials in cadence in the
Fig. 10. Body midpoint sway around the main body midpoint direction of no-fatigue group were due to a very minor augmentation from
displacement was obtained by the interpolation of the actual path and the
computation of its mean squared error (MSE). The average value (&SE) of the first to the last trials (from 59.7 to 60.3 steps/min, all blocks
MSE is shown for all trials of the subjects belonging to the fatigue group. collapsed), i.e., to less than one additional step over the entire
Fatigue (either Post C1 or Post C2 trials; shaded bars) did not increase body period (30 s) of acquisition. To further check that even any
sway either in the A-P (A) or M-L axis (B). nonsignificant variation in cadence in the fatigue group could
be one of the possible causes of the path alteration and of the
A-P displacement reduction during the first trial of postcon-

Fig. 11. Mean values (&SE) of the head

pitch angle (A and B) and head roll angle (C
and D) for all the subjects of the fatigue (A
and C) and no-fatigue (B and D) groups for
each of all successive stepping trials under all
experimental conditions. The mean angles
were not different among the trials of either

J Appl Physiol VOL 99 JULY 2005


traction 1 and postcontraction 2, the A-P index was plotted vs. subjects for whom dorsal neck muscle contraction did not show
the percent cadence change in the first trial after fatigue (with or showed only minor EMG signs of fatigue.
respect to the mean cadence in the 5 control trials) for each All subjects also participated in an additional session, when
subject of the fatigue group. The regression analysis gave no they were asked to step in place and perform the same se-
significant relationship [R2 $ 0.0002; F(1,3) $ 0.00067; P $ quence of blocks and trials but without the tonic isometric
0.981]. contraction of the dorsal neck muscles (no-load session). The
repetition of the stepping trials per se did not produce signif-
DISCUSSION icant effects on the body displacement. This strengthens the
notion that the cause of the observed reduced forward progres-
During stepping in place, before the prolonged head exten- sion or backward displacement was the neck muscle contrac-
sor effort, the subjects presented a substantial variability in tion rather than some process connected with the repetition of
their behavior. Most of them had a tendency to slowly proceed the trials or with the feedback received by subjects being
along the subjects sagittal plane, and some showed a mild repositioned at the starting point after each trial.
M-L displacement. Only a few subjects really did step in place. When we searched for a graded change in the body progres-
Across the subjects, the mean progression of the body midpoint sion along the sagittal plane by plotting the A-P index of
was in the forward direction, whereas the displacement in the displacement against the EMG index of fatigue, it turned out
M-L direction was, on average, close to zero. In all cases, no that, despite a considerable variability (likely connected to
subject was aware of his or her displacement in the walking individual anthropometric characteristics, to peculiarity of the
space. The extent of the forward progression of the subjects stepping task, and to the magnifying effects of the ratios used
changed significantly after the 5-min period of tonic contrac- for computing the indexes), the relevant changes in the dis-
tion of the dorsal neck muscles. In turn, the extent of this placements along the sagittal plane after contraction were again
change depended on the degree to which the contraction evident only in the subjects in whom the contraction produced
produced muscle fatigue in the neck muscles. clear-cut EMG signs of fatigue. This finding is also in favor of
Because the EMG can furnish an indirect measure of fatigue a rather selective role of fatigue, rather than prolonged con-
in the muscle of interest under conditions of controlled isomet- traction per se, in producing the effects and also strengthens the
ric contraction (64), an EMG index was used to quantify the EMG criteria for separating the subjects into two groups.
muscle contraction level across the subjects. This index de- Body unsteadiness during stepping? Dorsal neck muscle
pended on EMG changes in both amplitude and median fre- contraction, be it fatiguing or not, did not produce abnormal
quency occurring during the period of isometric dorsal neck oscillations around the interpolated stepping path of the body
muscle contraction. Indeed, the changes in these two variables midpoint during stepping, which would happen if a subjects
were highly correlated within several subjects of our popula- body midpoint became unusually unsteady around the mean
tion, regardless of gender and weight, whereby the EMG position. This might have been assumed on the basis of a
amplitude gradually increased together with a gradual decrease previous study, in which dorsal neck muscle fatigue was able
in EMG median frequency during the contraction period. In to produce moderate but significant increases in body sway
other subjects, this correlation was not evident, since EMG during quiet stance (4). It should be noted, however, that the
amplitude and median frequency showed no sizeable changes increase in body sway during quiet stance that was described in
during the contraction period. The relationship between EMG that paper was not associated with any major change in the
index and fatigue was also shown by the significant regression, center of foot pressure (forward, backward, or sideways),
across 13 subjects, between the percent decrease in MVC which might have entrained directional effect in the body
produced by the tonic contraction and the fatigue index. On displacement during the stepping in place. However, any cor-
these bases, we assumed that the EMG index was a good respondence between body oscillations under static (quiet bi-
marker of muscle fatigue, and we used it for dividing all our pedal stance) and dynamic condition (stepping in place) must
subjects into two subgroups. be considered with caution, because in the latter case balance
In the subjects showing EMG signs of neck muscle fatigue, control heavily depends on the alternate shift of the body
the forward displacement during the stepping-in-place trials weight on a single support.
proved to be much reduced (to 32%, on average) immediately The fatigue-induced changes in the A-P displacement during
after the 5-min period of tonic contraction with respect to the the stepping-in-place task were not dependent on changes in
stepping trials performed under control conditions. Some sub- head posture. In fact, head position did not change postcon-
jects even stepped in the backward direction. This behavior traction, either in pitch or roll angle, either with or without
was obvious only in the first trial immediately after the fatigu- signs of fatigue. The fact that the head remained in place
ing period, a sign of relatively rapid recovery from the fatigue despite the fatigue of the head extensor muscles was likely
effects, in keeping with recent reports from other investigations dependent on the very low level of force necessary to coun-
(16). However, it was reproduced and even enhanced in the teract the gravity-dependent head flexion, a level of force
first trial of the second series of the stepping trials performed easily produced even by heavily fatigued muscles. Possibly,
after the second fatiguing session, where the average displace- the production of the appropriate dorsal neck muscle tone was
ment was in the backward direction. This would point to a also aided by a mechanism of minimization of the change in
residual effect of the first fatiguing contraction that had not the discharge from the gravito-inertial labyrinthine receptors.
completely vanished at the end of the recovery period of 30 Certainly, vision could not help, since the subjects were blind-
min and added to the effects of the second fatiguing contraction folded during the stepping phase. In turn, the absence of
(67). Shorter bouts of back muscle fatiguing contraction re- excessive body midpoint oscillations during stepping could be
cover earlier (41). This behavior was not obvious in those favored by a mechanism of head-referenced body stabilization,
J Appl Physiol VOL 99 JULY 2005
the importance of which in the postural control system and could also directly access the brain centers responsible for
inertial guidance of locomotion has been already emphasized shaping our reference frames for navigation. Interestingly,
by Pozzo and coworkers (53). abnormal conditions of the neck muscle (tension neck) can
Putative neural mechanisms. Neck muscles are surely cen- disrupt the normal postural response to neck muscle vibration
tral in the construction of a reference frame for movement in (40), pointing to a susceptibility of the reference system to
space. Such a role is apparently not played by the leg muscles abnormal fatigue-related inputs of neck muscle origin.
involved in the production of body movement, at least as Admittedly, fatigue can affect the discharge of many sensory
indicated by the minor effect of bilateral triceps surae vibration receptors, including muscle spindles (27, 34, 63, 64, 67). In
during walking (13). It is therefore in that light that we would turn, spindle discharge could be modulated by fatigue-induced
like to propose an explanation of the observed phenomenon. In alterations in gamma-motoneuron discharge (43, 51). Such
so doing, it seems useful to discuss the present data, together abnormal muscle spindle inflow could possibly perturb and
with those of the studies describing the effects of dorsal neck deteriorate postural control through an action on the central
muscle vibration, in the line of the process of multisensory nervous system (19). However, one may note that there was no
fusion that leads to egocentric space representation. neck muscle contraction during the stepping trials and presum-
During stepping in place, neck vibration produces involun- ably little spindle discharge, since the latter should not outlast
tary forward stepping at about 0.3 m/s without modifying the previous contraction, contrary to the discharge of the receptors
stepping frequency (32). In our case, neck muscle fatigue sensitive to fatigue-induced intramuscular changes. Any mod-
instead induced an involuntary smaller forward progression or ulation of a scarce proprioceptive input would therefore hardly
even backward progression. The difference between the effects produce the gross effects shown here in body progression,
of fatigue and vibration cannot be ascribed to the more aspe- emphasizing a direct effect of fatigue-related input as a cause
cific distribution of fatigue, since the concurrent EMG record- of the abnormal progression after the fatiguing contraction.
ing from the anterior neck muscles showed no sign of activa- Neck muscle fatigue effects on locomotion. Stepping in place
tion of these muscles. Similar to vibration, the stepping ca- with closed eyes and without acoustic cues is a nonoptimal
dence did not undergo major changes as a consequence of neck definition for a locomotor task and presents itself with peculiar
muscle fatigue. The small changes in cadence observed were characteristics from subject to subject (7). The drift from the
unrelated to the distance or direction of body displacement starting position is a sign of unreliable feedback during step-
during the stepping. Apparently, the central effects of both ping in place (61). Most likely, many sensory modalities
vibration and fatigue do not influence the rhythmic outflow
normally able to convey velocity and direction information,
from the central pattern generators implicated in this automatic
such as the foot sole receptors and their central integration, are
locomotor-like stepping activity.
below the threshold under these circumstances, or the central
Perhaps, the changes in body progression along the sagittal
mechanisms for calculating the mismatch between the succes-
plane should be attributed to the outflow from the neck muscles
sive positions of the two leg can be fooled by the minimal
to the brain centers responsible for building the body spatial
reference frame. Vibration and fatigue both produce modifica- differences of their positions between steps. The gravito-
tions in joint position sense (for vibration, see Refs. 9, 20, 22, inertial receptors would also be of no use, given that the
30, 31; for fatigue, see Refs. 12, 16, 24, 26, 57, 59). This successive head and body up and down movements do not
testifies of the capacity of the afferent input from dorsal neck seem to be influenced by the body progression and that the
muscles of accessing the central nervous system and altering linear body velocity in the midsagittal plane is negligible even
the perception of the position of the segments in space. How- when it reaches its maximum value (!3 cm/s) (25). Certainly,
ever, the fatigue state and vibration effects dramatically di- no subject was aware of his or her displacement in the working
verge as far as the modification in the presumed reference space, either before or after the fatigue-induced effects. And
frame for controlling locomotion is concerned. Fatigue reduces this comes as no surprise, since much larger body displacement
the forward distance reached by subjects who normally exhibit produced by lateral neck muscle vibration during stepping in
a forward progression and pushes clearly backward those who place went completely unnoticed by the subjects (4). There-
stepped in place. fore, the body displacement would be the expression of the
This would speak in favor of a different set of receptors and spontaneous drift in the activity of the brain centers controlling
afferent fibers under vibration and fatiguing contraction con- the body position in space. This command center, although
ditions, indirectly or directly responsible for the modulation of numb to the afferent sensory input from the evolving move-
the reference frame. In particular, the receptors responsible for ment, is, however, sensible to other afferent inputs like that
the shrinkage of our imagined space in the forward direction from fatigued neck muscles. Despite the highly variable body
or for the backward displacement of the body position displacement across subjects, the relative consistency of this
reference after fatigue would be different from those possibly displacement in the different trials within a subject could be
activated by muscle contraction per se (like tendon organs or taken as an indication that the subjects reference for the
muscle spindles). Rat experiments have shown that fatigue stepping position can be relatively stable. It seems, therefore,
activates small-fiber receptors, the same that are activated by that stepping in place after neck muscle fatigue is being
capsicine (52). Small-fiber sensory input may be the result of performed with respect to a reference system and modified in
an increased inflow from the free nerve endings as a conse- the same sense by the fatigue-induced input in every subject.
quence of ionic or metabolic changes (elevated interstitial Such a modification could be reminiscent of an error of
potassium concentration or insufficient oxygen availability due parallax, as if the brain would incorporate depth information
to reduced blood flow) (19, 52). This input, in addition to when it updates its stored representations of space after the
inhibiting muscle spindle afferent activity during fatigue (10), neck muscle fatiguing contraction (45).
J Appl Physiol VOL 99 JULY 2005

ACKNOWLEDGMENTS 22. Gregory JE, Morgan DL, and Proske U. Aftereffects in the responses of
cat muscle spindles and errors of limb position sense in man. J Neuro-
The help of M. V. Beretta and A. M. De Nunzio during data collection is
physiol 59: 1220 1230, 1988.
gratefully acknowledged. We thank two unknown referees for critical reading
23. Gurfinkel VS, Ivanenko YP, and Levik YS. The influence of head
and detailed suggestions.
rotation on human upright posture during balanced bilateral vibration.
Neuroreport 7: 137140, 1995.
GRANTS 24. Hardin JA. The effects of muscle fatigue on shoulder joint position sense.
Am J Sports Med 26: 262265, 1998.
This work was supported by grants from the Italian Ministry of Health 25. Harris LR, Jenkin M, and Zikovitz DC. Visual and non-visual cues in
(Ricerca Finalizzata 2002) and Ministry of University and Research (Progetto the perception of linear self-motion. Exp Brain Res 135: 1221, 2000.
Ricerca Interesse Nazionale 2003, Fondo Investimenti Ricerca di Base 2002). 26. Hiemstra LA, Lo IK, and Fowler PJ. Effect of fatigue on knee propri-
oception: implications for dynamic stabilization. J Orthop Sports Phys
REFERENCES Ther 31: 598 605, 2001.
27. Hill JM. Increase in the discharge of muscle spindles during diaphragm
1. Biguer B, Donaldson IM, Hein A, and Jeannerod M. Neck muscle fatigue. Brain Res 918: 166 170, 2001.
vibration modifies the representation of visual motion and direction in 28. Holtmann S, Clarke A, and Scherer H. Cervical receptors and the
man. Brain 111: 14051424, 1988. direction of body sway. Arch Otorhinolaryngol 246: 61 64, 1989.
2. Bottini G, Karnath HO, Vallar G, Sterzi R, Frith CD, Frackowiak RS, 29. Horita T and Ishiko T. Relationships between muscle lactate accumula-
and Paulesu E. Cerebral representations for egocentric space: functional- tion and surface EMG activities during isokinetic contractions in man. Eur
anatomical evidence from caloric vestibular stimulation and neck vibra- J Appl Physiol Occup Physiol 56: 18 23, 1987.
tion. Brain 124: 11821196, 2001. 30. Inglis JT and Frank JS. The effect of agonist/antagonist muscle vibration
3. Bove M, Diverio M, Pozzo T, and Schieppati M. Neck muscle vibration on human position sense. Exp Brain Res 81: 573580, 1990.
disrupts steering of locomotion. J Appl Physiol 91: 581588, 2001. 31. Inglis JT, Frank JS, and Inglis B. The effect of muscle vibration on
4. Bove M, Courtine G, and Schieppati M. Neck muscle vibration and human position sense during movements controlled by lengthening muscle
spatial orientation during stepping in place in humans. J Neurophysiol 88: contraction. Exp Brain Res 84: 631 634, 1991.
22322241, 2002. 32. Ivanenko YP, Talis VL, and Kazennikov OV. Support stability influ-
5. Brandt T, Krafzcyk S, and Malsbended I. Postural imbalance with head ences postural responses to muscle vibration in humans. Eur J Neurosci
extension: improvement by training as a model for ataxia therapy. Ann NY 11: 647 654, 1999.
Acad Sci 374: 636 649, 1981. 33. Ivanenko YP, Grasso R, and Lacquaniti F. Neck muscle vibration
6. Brandt T and Bronstein AM. Cervical vertigo. J Neurol Neurosurg makes walking humans accelerate in the direction of gaze. J Physiol 525:
Psychiatry 71: 8 12, 2001. 803 814, 2000.
7. Brenie`re Y. Why we walk the way we do. J Mot Behav 28: 291298, 34. Johansson H and Sojka P. Pathophysiological mechanisms involved in
1996. genesis and spread of muscular tension in occupational muscle pain and in
8. Brenie`re Y and Ribreau C. A double-inverted pendulum model for chronic musculoskeletal pain syndromes: a hypothesis. Med Hypotheses
studying the adaptability of postural control to frequency during human 35: 196 203, 1991.
stepping in place. Biol Cybern 79: 337345, 1998. 35. Karlberg M, Persson L, and Magnusson M. Impaired postural control in
9. Brumagne S, Lysens R, Swinnen S, and Verschueren S. Effect of patients with cervico-brachial pain. Acta Otolaryngol Suppl (Stockh) 520:
paraspinal muscle vibration on position sense of the lumbosacral spine. 440 442, 1995.
Spine 24: 1328 1331, 1999. 36. Karnath HO, Christ K, and Hartje W. Decrease of contralateral neglect
10. Brunetti O, Della Torre G, Lucchi ML, Chiocchetti R, Bortolami R, by neck muscle vibration and spatial orientation of trunk midline. Brain
and Pettorossi VE. Inhibition of muscle spindle afferent activity during 116: 383396, 1993.
masseter muscle fatigue in the rat. Exp Brain Res 152: 251262, 2003. 37. Karnath HO, Sievering D, and Fetter M. The interactive contribution of
11. Burke D, Hagbarth KE, Lofstedt L, and Wallin BG. The responses of neck muscle proprioception and vestibular stimulation to subjective
human muscle spindle endings to vibration of non-contracting muscles. straight ahead orientation in man. Exp Brain Res 101: 140 146, 1994.
J Physiol 261: 673 693, 1976. 38. Kavounoudias A, Roll R, and Roll JP. Specific whole-body shifts
12. Carpenter JE, Blasier RB, and Pellizzon GG. The effects of muscle induced by frequency-modulated vibrations of human plantar soles. Neu-
fatigue on shoulder joint position sense. Am J Sports Med 26: 262265, rosci Lett 266: 181184, 1999.
1998. 39. Kerkhoff G. Modulation and rehabilitation of spatial neglect by sensory
13. Courtine G, Pozzo T, Lucas B, and Schieppati M. Continuous, bilateral stimulation. Prog Brain Res 142: 257271, 2003.
Achilles tendon vibration is not detrimental to human walk. Brain Res 40. Koskimies K, Sutinen P, Aalto H, Starck J, Toppila E, Hirvonen T,
Bull 55: 107115, 2001. Kaksonen R, Ishizaki H, Alaranta H, and Pyykko I. Postural stability,
14. Duclos C, Roll R, Kavounoudias A, and Roll JP. Long-lasting body neck proprioception and tension neck. Acta Otolaryngol Suppl (Stockh)
leanings following neck muscle isometric contractions. Exp Brain Res 529: 9597, 1997.
158: 58 66, 2004. 41. Lariviere C, Gravel D, Arsenault AB, Gagnon D, and Loisel P. Muscle
15. Eklund G. General features of vibration-induced effects on balance. Ups recovery from a short fatigue test and consequence on the reliability of
J Med Sci 77: 112124, 1972. EMG indices of fatigue. Eur J Appl Physiol 89: 171176, 2003.
16. Elfving B, Liljequist D, Dedering A, and Nemeth G. Recovery of 42. Lekhel H, Popov K, Anastasopoulos D, Bronstein A, Bhatia K, Mars-
electromyograph median frequency after lumbar muscle fatigue analysed den CD, and Gresty M. Postural responses to vibration of neck muscles
using an exponential time dependence model. Eur J Appl Physiol 88: in patients with idiopathic torticollis. Brain 120: 583591, 1997.
8593, 2002. 43. Ljubisavljevic M and Anastasijevic R. Fusimotor system in muscle
17. Forestier N, Teasdale N, and Nougier V. Alteration of the position sense fatigue. J Peripher Nerv Syst 1: 8396, 1996.
at the ankle induced by muscular fatigue in humans. Med Sci Sports Exerc 44. Lund S. Postural effects of neck muscle vibration in man. Experientia 36:
34: 117122, 2002. 1398, 1980.
18. Fransson PA, Karlberg M, Sterner T, and Magnusson M. Direction of 45. Medendorp WP, Tweed DB, and Crawford JD. Motion parallax is
galvanically-induced vestibulo-postural responses during active and pas- computed in the updating of human spatial memory. J Neurosci 23:
sive neck torsion. Acta Otolaryngol (Stockh) 120: 500 503, 2000. 8135 8142, 2003.
19. Garland SJ. Role of small diameter afferents in reflex inhibition during 46. Mergner T, Siebold C, Schweigart G, and Becker W. Human percep-
human muscle fatigue. J Physiol 435: 547558, 1991. tion of horizontal trunk and head rotation in space during vestibular and
20. Goodwin GM, McCloskey DI, and Matthews PB. Proprioceptive illu- neck stimulation. Exp Brain Res 85: 389 404, 1991.
sions induced by muscle vibration: contribution by muscle spindles to 47. Mergner T, Huber W, and Becker W. Vestibular-neck interaction and
perception? Science 175: 13821384, 1972. transformation of sensory coordinates. J Vestib Res 7: 347367, 1997.
21. Grasso R, Prevost P, Ivanenko YP, and Berthoz A. Eye-head coordi- 48. Mergner T, Nasios G, Maurer C, and Becker W. Visual object local-
nation for the steering of locomotion in humans: an anticipatory synergy. isation in space. Interaction of retinal, eye position, vestibular and neck
Neurosci Lett 253: 115118, 1998. proprioceptive information. Exp Brain Res 141: 3351, 2001.

J Appl Physiol VOL 99 JULY 2005

49. Merletti R and Lo Conte LR. Surface EMG signal processing during 59. Skinner HB, Wyatt MP, Hodgdon JA, Conard DW, and Barrack RL.
isometric contractions. J Electromyogr Kinesiol 7: 241250, 1997. Effect of fatigue on joint position sense of the knee. J Orthop Res 4:
50. Norre ME, Forrez G, and Beckers A. Clinical application of vestibu- 112118, 1986.
lospinal reflex tests in peripheral vestibular disorders. Acta Otolaryngol 60. Sterner RL, Pincivero DM, and Lephart SM. The effects of muscular
Suppl 468: 337339, 1989. fatigue on shoulder proprioception. Clin J Sport Med 8: 96 101, 1998.
51. Pedersen J, Ljubisavljevic M, Bergenheim M, and Johansson H. 61. St.-Onge N and Feldman AG. Referent configuration of the body: a
Alterations in information transmission in ensembles of primary muscle global factor in the control of multiple skeletal muscles. Exp Brain Res
spindle afferents after muscle fatigue in heteronymous muscle. Neuro- 155: 291300, 2004.
science 84: 953959, 1998. 62. Taylor JL and McCloskey DI. Illusions of head and visual target displace-
52. Pettorossi VE, Della Torre G, Bortolami R, and Brunetti O. The role ment induced by vibration of neck muscles. Brain 114: 755759, 1991.
of capsaicin-sensitive muscle afferents in fatigue-induced modulation of 63. Taylor JL, Butler JE, and Gandevia SC. Changes in muscle afferents,
the monosynaptic reflex in the rat. J Physiol 515: 599 607, 1999. motoneurons and motor drive during muscle fatigue. Eur J Appl Physiol
83: 106 115, 2000.
53. Pozzo T, Berthoz A, and Lefort L. Head stabilization during various
64. Vllestad NK. Measurement of human muscle fatigue. J Neurosci Meth-
locomotor tasks in humans. I. Normal subjects. Exp Brain Res 82: 97106,
ods 74: 219 227, 1997.
65. Weerakkody NS, Whitehead NP, Canny BJ, Gregory JE, and Proske
54. Pyykko I, Enbom H, Magnusson M, and Schalen L. Effect of propri- U. Large-fiber mechanoreceptors contribute to muscle soreness after
oceptor stimulation on postural stability in patients with peripheral or eccentric exercise. J Pain 2: 209 219, 2001.
central vestibular lesion. Acta Otolaryngol (Stockh) 111: 2735, 1991. 66. Weerakkody N, Percival P, Morgan DL, Gregory JE, and Proske U.
55. Roll JP, Vedel JP, and Roll R. Eye, head and skeletal muscle spindle Matching different levels of isometric torque in elbow flexor muscles after
feedback in the elaboration of body references. Prog Brain Res 80: eccentric exercise. Exp Brain Res 149: 141150, 2003.
113123, 1989. 67. Yaggie JA and McGregor SJ. Effects of isokinetic ankle fatigue on the
56. Schieppati M, Nardone A, and Schmid M. Neck muscle fatigue affects maintenance of balance and postural limits. Arch Phys Med Rehabil 83:
postural control in man. Neuroscience 121: 277285, 2003. 224 228, 2002.
57. Sharpe MH and Miles TS. Position sense at the elbow after fatiguing 68. Yoshitake Y, Ue H, Miyazaki M, and Moritani T. Assessment of
contractions. Exp Brain Res 94: 179 182, 1993. lower-back muscle fatigue using electromyography, mechanomyography,
58. Simoneau GG, Leibowitz HW, Ulbrecht JS, Tyrrell RA, and Ca- and near-infrared spectroscopy. Eur J Appl Physiol 84: 174 179, 2001.
vanagh PR. The effects of visual factors and head orientation on postural 69. Zhang LQ and Rymer WZ. Reflex and intrinsic changes induced by fatigue
steadiness in women 55 to 70 years of age. J Gerontol 47: 151181, 1992. of human elbow extensor muscles. J Neurophysiol 86: 1086 1094, 2001.

J Appl Physiol VOL 99 JULY 2005