1 23

European Journal of AppliedPhysiology ISSN 1439-6319 Eur J Appl PhysiolDOI 10.1007/s00421-012-2498-2

Uneven spatial distribution of surfaceEMG: what does it mean?

Alessio Gallina, Roberto Merletti &Marco Gazzoni

1 23

Your article is protected by copyright and

all rights are held exclusively by Springer-

Verlag. This e-offprint is for personal use only

and shall not be self-archived in electronic

repositories. If you wish to self-archive your

work, please use the accepted author’s

version for posting to your own website or

your institution’s repository. You may further

deposit the accepted author’s version on a

funder’s repository at a funder’s request,

provided it is not made publicly available until

12 months after publication.

ORIGINAL ARTICLE

Uneven spatial distribution of surface EMG: what does it mean?

Alessio Gallina • Roberto Merletti •

Marco Gazzoni

Received: 19 June 2012 / Accepted: 10 September 2012

� Springer-Verlag 2012

Abstract The aim of this work is to show how changes in

surface electromyographic activity (sEMG) during a repet-

itive, non-constant force contraction can be detected and

interpreted on the basis of the amplitude distribution pro-

vided by high-density sEMG techniques. Twelve healthy

male subjects performed isometric shoulder elevations,

repeating five times a force ramp profile up to 25 % of the

maximal voluntary contraction (MVC). A 64-electrode

matrix was used to detect sEMG from the trapezius muscle.

The sEMG amplitude distribution was obtained for the force

levels in the range 5–25 % MVC with steps of 5 % MVC.

The effect of force level, subject, electrode position and

ramp repetition on the sEMG amplitude distribution was

tested. The sEMG amplitude was significantly smaller in the

columns of the electrode grid over the tendons (repeated

measures ANOVA, p \ 0.01). The barycentre of the dis-

tribution of sEMG amplitude was subject-specific (Kruskal–

Wallis test, p \ 0.01), and shifted caudally with the increase

of force levels and cranially with the repetition of the motor

task (both p \ 0.01, repeated measures ANOVA). The

results are discussed in terms of motor unit recruitment in

different muscle sub-portions. It is concluded that the sEMG

amplitude distribution obtained by multichannel techniques

provides useful information in the study of muscle activity,

and that changes in the spatial distribution of the recruited

motor units during a force varying isometric contraction

might partially explain the variability observed in the acti-

vation pattern of the upper trapezius muscle.

Keywords Electromyography � Muscle �Motor unit recruitment � Methods

Abbreviations

EMG Electromyography

sEMG Surface electromyography

MU Motor unit

MVC Maximal voluntary contraction

IZ Innervation zone

RMS Root mean square

ANOVA Analysis of variance

Introduction

For the assessment of upper limb and shoulder girdle

movements, the study of the activity of the trapezius

muscle is of paramount importance because of its role in

the stabilization of the scapula. In this muscle, subject-

specific patterns of activation, in terms of timing and

contraction intensity, were reported during standardized

motor tasks and functional activities such as industrial

work (Balogh et al. 1999; Mork and Westgaard 2005) and

musical performance (Fjellman-Wiklund et al. 2004).

These inter-individual differences in muscle activation may

be due to the activation of specific portions of the muscle.

The selective activation of muscle sub-portions was

observed within some muscles (English and Segal 1993)

and it was confirmed (1) in multi-functional muscles

depending on the direction of the exerted force (Wickham

and Brown 2012), (2) in painful conditions (Tucker et al.

2009), and (3) under voluntary control when a feedback on

muscle activation was provided to the subject (Holtermann

et al. 2009).

Communicated by Toshio Moritani.

A. Gallina (&) � R. Merletti � M. Gazzoni

Laboratory for Engineering of the Neuromuscular System

(LISiN), Politecnico di Torino, Torino, Italy

e-mail: alessio.gallina@delen.polito.it

123

Eur J Appl Physiol

DOI 10.1007/s00421-012-2498-2

Author's personal copy

The trapezius muscle has different lines of action, being

able to apply forces to the scapula in various directions

(Johnson et al. 1994). Selective activation of independent

muscle sub-portions was reported for this muscle (Holter-

mann et al. 2009, Falla and Farina 2008a). These works

used intramuscular EMG or sets of bipolar detection sys-

tems that allow to detect the activity of a limited muscle

volume. To map the whole muscle activity with higher

spatial resolution, it is possible to use high-density sEMG

technique (HD-EMG) with tens of electrodes spaced no

more than one centimeter (reviewed in Merletti et al.

2010). Some studies in literature investigated with this

technique the topography of the EMG distribution in the

trapezius muscle (Kleine et al. 2000; Holtermann and

Roeleveld 2006) showing heterogeneous distribution of the

muscle activity.

The aim of this study was to further investigate the EMG

activity of the trapezius muscle by means of HD-EMG

covering a larger portion of the muscle and with a higher

density of the electrodes with respect to previous works to

verify (1) how the anatomical characteristics of the muscle

influence the EMG amplitude distribution, (2) if the EMG

amplitude distribution during a task is subject-specific, and

(3) how the EMG amplitude distribution changes in the

trapezius along the repetition of a simple isometric force-

varying motor task.

Methods

Subjects

Twelve healthy male subjects participated in the study

[age (mean ± SD): 29.8 ± 6.1, height: 181.7 ± 6.0 cm,

weight: 75.2 ± 8.0 kg]. All the subjects were pain-free at

the time of the experiment, and reported no complaints of

pain in the neck-shoulder region in the previous month.

Subjects provided a written, informed consent before

beginning the experimental session, and the study was

approved by the local ethics committee.

Protocol

Subjects were seated upright in a custom-made chair

designed for shoulder elevation measures. A strap with a

plastic protection for the shoulder was secured on the subject

right shoulder and connected to a load cell fixed to the floor.

The subject performed three maximal voluntary contrac-

tions (MVC) of 5 s each, separated by a 2-min rest in

between. A visual feedback of the exerted force was pro-

vided to the subject. The maximum of the three force mea-

sures was considered the reference MVC. The subject was

asked to perform five force-varying isometric contractions

using a visual feedback on force, following a triangular

profile from 0 % MVC up to 25 % MVC in 15 s with 4-s rest

in between. The force level below 25 % MVC was selected

to simulate the force demand during daily activities and the

number of ramps was limited to five to avoid fatigue. To

familiarize the subject with the requested task and the

feedback, a training session of at least 10 min was provided.

During the training, the subject was asked whether he was

feeling comfortable with the strap; not much variations of

the length of the strap were allowed in order to make the

subject to feel as comfortable as possible during the force

exertion. The position and orientation of the scapula were

not constrained by the experimental set-up.

Surface EMG acquisition

Surface EMG signals were detected using a two-dimen-

sional grid of 64 electrodes (SPES-MEDICA, 1-mm

diameter, 8-mm interelectrode distance). The electrodes

were arranged in a grid of five columns (cranio-caudal

direction) and 13 rows (medio-lateral direction), with the

first corner electrode [1, 5] missing. The electrode grid was

Fig. 1 Position of the electrode grid on the trapezius muscle. The

electrode grid was positioned on the basis of some anatomical

reference points: the acromion, the C7 vertebra and the position of the

innervation zones. The position of the innervation zone (light greyarea) was identified using a linear electrode array in two different

locations (dark grey circles) of the muscle. The electrode grid was

positioned between the innervation zone and the spine. The fourth

row of the electrode grid was aligned with the line connecting C7 to

acromion; the missing electrode was positioned cranially and

laterally. The electrode [1 1] (i.e. first row, first column) is the most

cranial and medial

Eur J Appl Physiol

123

Author's personal copy

placed on the trapezius muscle with the rows in the

direction of the muscle fibers as described in the following

(Fig. 1); to avoid the innervation zones (IZ) under the

detection area, the position of the main IZ was identified by

visual inspection of the signals detected using a linear

electrode array covering the entire muscle, as described by

Barbero et al. (2011). The IZ position was identified on the

line linking the acromion and C7 vertebra and on a parallel

line 5 cm caudal with respect to the previous one. The

electrode grid was placed medially with respect to the

identified IZ with the fourth row positioned on the line

linking acromion and C7, to align the muscle fibers with

the rows of the matrix (Farina et al. 2008). The region of

the skin where the matrix was located was slightly abraded

with abrasive paste. The matrix was fixed to the skin by

adhesive tape and a reference electrode was placed at the

wrist. Surface EMG signals were acquired in monopolar

configuration, amplified 500 times, band-pass filtered

(-3 dB bandwidth, 10–500 Hz), sampled at 2,048 samples/s

and converted to digital data by a 12 bit A/D converter

(EMG-USB amplifier, LISiN and OT-Bioelettronica, Italy).

Single differential spatial filtering was performed by soft-

ware along the matrix rows, resulting in a 13 9 4 single

differential channels, and signals were digitally band-pass

filtered at 20–400 Hz. Channels with contact problems or

short circuits were identified through visual inspection of

the raw EMG signals; both bad channels and the missing

channel in the top-lateral position of the matrix were

reconstructed by interpolation of the neighbouring chan-

nels. The force signal was recorded using a load cell

connected to the shoulder strap, conditioned using the

EMG-USB amplifier and low-pass filtered at 8 Hz. Raw

EMG signals collected from one subject are shown in

Fig. 2.

Data processing

Signals recorded during the ascending phase of the five

ramps have been considered (15 s each). The root mean

square (RMS) was calculated for each single differential

channel on epochs of 1.25 s (corresponding to a 2 % MVC

force step in the ideal conditions of a perfect linear force

ramp) centred on the time instants corresponding to 5, 10,

15, 20, and 25 % MVC. The barycentre of the RMS dis-

tribution has been calculated over the two lateral columns

for the analysis of cranio-caudal distribution of EMG

amplitude. RMS values of the two lateral columns

(2 9 13) were averaged (1 9 13), and the barycentre was

processed along the resulting vector.

Statistical analysis

Statistical analysis was performed with the software

Sigmaplot 12. Before each statistical test, the assumption

of normality of the data was checked using the Shapiro–

Wilk test.

The analysis of medio-lateral differences in the ampli-

tude distribution has been performed on the mean ampli-

tude values estimated on the columns of the matrix. The

repeated measures analysis of variance (ANOVA) was

performed, considering force level and matrix column as

factors.

Fig. 2 Example of raw SD

EMG signals (subject number

11, 4th ramp). The epoch shown

is in correspondence of the peak

of force (25 % MVC). The

differential spatial filter was

applied along the matrix rows

that are aligned with the fiber

direction. The innervation zone

is lateral to the electrode grid

(i.e. below the fourth column in

this plot). Two action potentials

belonging to two different

motor units are clearly visible in

the cranial (left of the matrix,

rows 1–5) and in the caudal

portion (right of the matrix).

Action potentials have lower

amplitude in the medial position

(1st column) than the lateral one

(4th column)

Eur J Appl Physiol

123

Author's personal copy

To test if the cranio-caudal distribution of EMG activity

is dependent on the subject, the effect of the subject on the

average position of the barycentre was tested with the

Kruskal–Wallis test.

The influence of the force level and of the ramp repe-

tition on the cranio-caudal position of the barycentre was

tested using the repeated measure ANOVA test (repeated

measures on subjects, factors: force level and ramp ordinal

number). Holm–Sidak post hoc test was applied when

appropriate.

The level of significance was set to p = 0.05 for all the

statistical analysis.

Results

The assumptions of normal distribution and equality of

variance were verified for the medio-lateral amplitude dis-

tribution and for the cranio-caudal distribution (p \ 0.05)

while the average position of the barycentre showed a non-

gaussian distribution.

Influence of medio-lateral position on sEMG amplitude

distribution

Figure 3 shows the influence of the medio-lateral position

and the interaction with the force level on RMS distribu-

tion; data are represented as mean ± standard deviation.

Statistical test (repeated measures ANOVA) stated that

RMS values were significantly dependent on both column

(p \ 0.01, F = 60.21) and force level (p \ 0.01, F =

79.33); moreover, the interaction between force and col-

umn was significant (p \ 0.01, F = 21.29). According to

Holm–Sidak post hoc test, all columns (p \ 0.01) and all

force levels (p \ 0.05) were significantly different from

each other.

Effect of the subject on the cranio-caudal sEMG

amplitude distribution

Figure 4a shows the RMS distribution of four representa-

tive subjects; each map represents the average of RMS

calculated on all ramps and force levels for one subject.

Statistical analysis (Kruskal–Wallis test) proved that there

is a main effect of the subject on the position of the

barycentre (p \ 0.01, H = 46.41).

Effect of task repetition and force level

on the cranio-caudal semg amplitude distribution

Figure 4 shows examples of the shift of the barycentre in

representative subjects during force increase (B) and rep-

etition of consecutive ramps (C). The position of the

barycentre of all subjects at different force levels and in

different ramps is shown in Fig. 5. A repeated measures

ANOVA showed that the position of the barycentre was

significantly affected by both force level (p \ 0.01,

F = 11.25) and the number of the ramp (p \ 0.01,

F = 6.45). No interactions were detected among these two

factors (p = 0.2). The barycentre at 5 % MVC was sig-

nificantly more cranial than 15, 20 and 25 %; similarly, at

10 % was significantly more cranial than both 20 and 25 %

MVC (Holm–Sidak post hoc test, p \ 0.05). For what

concerns, the effect of the repetition of ramps, the bary-

centre in the first ramp was significantly more caudal than

that of the fourth and the fifth ramp (Holm–Sidak post hoc

test, p \ 0.05).

Discussion

Large inter-subject variability of sEMG patterns and

localized sEMG activity within the muscle were described

in literature during occupational activities in the trapezius

muscle.

The aim of this study was to investigate the EMG activity

of the trapezius muscle during low level ramp contractions

by means of HD-EMG to verify (1) how the anatomical

characteristics of the muscle influence the EMG amplitude

distribution, (2) if the EMG amplitude distribution during a

task is subject-specific, and (3) how the EMG amplitude

distribution changes in the trapezius along the repetition of

a simple isometric force-varying motor task.

Fig. 3 RMS values of the EMG activity as a function of force level

and medio-lateral position of the detection system. The data are

reported as mean and standard deviation on all subjects (N = 12).

Columns of the electrode grid are represented on the X axis (1 is the

most medial), whereas the Y axis contains the RMS amplitude values.

A main effect of both column and force level was proven with

statistical test (repeated measures ANOVA, p \ 0.01)

Eur J Appl Physiol

123

Author's personal copy

The original findings of this study are that, regardless of

the subject examined, the channels in the lateral portion of

the electrode grid (i.e. far from the tendon) showed the

highest EMG amplitude, and are the most affected by the

force level. The cranio-caudal distribution of EMG was

subject-specific, and shifts of the barycentre of this distri-

bution occurred during force increase (caudal direction)

and consecutive ramps (cranial direction).

In the following a more detailed analysis of the results is

reported.

Influence of medio-lateral position on sEMG amplitude

distribution

The amplitude of the EMG signal detected by an electrode

grid positioned between the innervation zone and the spine,

increases moving from medial to lateral position for all

force levels (Fig. 3). A significant interaction was found

between force level and medio-lateral distribution of the

amplitude of the sEMG signal. This means that, consider-

ing the absolute RMS values, the increase of the amplitude

related to force production is more represented in the lat-

eral channels than in medial ones. Considering that (1) the

trapezius muscle fibers run in medio-lateral direction,

almost parallel to the rows of the electrode grid (then

changes in the amplitude among the columns would be

related to anatomical characteristics of the active MUs

rather than changes in the recruited MU pool), and (2) the

electrode grid was placed medially to the innervation zone,

but the position of the tendon has not been considered, it is

possible that the medial portion of the electrode grid was

placed on the tendon. Above the tendons, the muscle fiber

action potentials extinguish with a consequent dominance

of non-travelling components (end of fiber effects) in

the monopolar sEMG signal (Merletti et al. 2001). Spatial

filters, such as single differential, attenuate the contribu-

tion from non-travelling components and EMG signal

amplitude decreases in correspondence of the tendons

(Merletti et al. 2001). The signals plotted in Fig. 2 show

action potentials with the lowest peak amplitude in the

most medial columns, supporting this hypothesis. If the

approximate direction of the muscle fibers is known,

bi-dimensional surface EMG techniques allow in deter-

mining the presence of tendons under the detection area,

possibly reducing the errors of amplitude estimation due to

anatomical factors.

Effect of the subject on the cranio-caudal sEMG

amplitude distribution

The position of the barycentre was significantly dependent

on the subject. As the rows of the matrix were parallel to

the trapezius muscle fibers, each row detected the activity

of a pool of MUs only partly intermingled between rows

(i.e. cranial-caudal direction). Shifts of the EMG activity

related to recruitment of MUs in discrete portion of the

muscle are predominantly represented in the cranio-caudal

direction (Farina et al. 2008; Madeleine et al. 2006; Kleine

et al. 2000). In this study, the barycentre calculated over

the two (averaged) lateral columns, was used to describe

the distribution of active MUs under the detection system;

Fig. 4 Maps of EMG amplitude distribution. a The maps averaged

over all conditions of four subjects. b The caudal shift of the

barycentre at different force levels in the first ramp in a representative

subject. c The cranial shift of the barycentre at a given force level

(15 % MVC) in consecutive ramps in a representative subject. In all

maps, the 64 RMS values were interpolated with a factor of 10

(processing was done on the real data). The colorbar of each map

shows the range of amplitude values of the map. The horizontal blackline is the barycentre, always calculated over the two lateral columns;

the dashed rectangle was drawn to facilitate the observation of the

shift of the barycentre

Eur J Appl Physiol

123

Author's personal copy

this was preferred over computing the barycentre of the

whole map, as we have previously shown that force-related

EMG changes are less represented in the medial columns

of the grid. Literature is consistent about individual, sub-

ject-dependent patterns of activation of the trapezius

muscle in time (Balogh et al. 1999; Mork and Westgaard

2005; Fjellman-wiklund et al. 2004; Mathiassen and

Aminoff 1997), possibly due to the redundancy of muscles

with similar motor functions in this area, such as the

levator scapulae, the rhomboideus major and the rhom-

boideus minor. On the basis of the results of our work, the

findings of the articles quoted above could also be

explained by a subject-specific distribution of motor unit

activation within the trapezius muscle, as it can be noticed

in the maps in Fig. 4a. The activation of discrete portions

of a muscle during motor tasks is widely discussed in the

literature (English and Segal 1993) and selective activation

of muscle sub-portions has often been reported (Holter-

mann et al. 2009; Wickham and Brown 2012; Watanabe

et al. 2011); persons may take advantage of this uneven

recruitment for developing individual motor strategies. In

fact, localized amplitude peaks in the map correspond to

the activity of a MU pool whose territory is limited to a

portion of the muscle (Vieira et al. 2011; Zhou et al. 2011;

Staudenmann et al. 2009), and this can reflect that MUs

within a muscle are recruited to exert force in a specific

direction (Desmedt and Godaux 1981). It can be speculated

that subjects with different distribution of muscle activity

produced forces in different directions, even if this could

not be verified in this experiment. A number of other

factors might have influenced our results. The placement of

the electrode grid was performed according to anatomical

landmarks easy to detect; it is unlikely that cranio-caudal

misplacements are responsible for this variability. Possible

confounding factors can be related to the difficulty to

control and standardize the position of the scapula, and of

the shoulder considering that the only constrain was rep-

resented by the length of the strap that connected the

shoulder to the load cell regulated according to the subject

comfort. Slight differences in scapula rotation, as well as

elevation or protraction due to individual postures might

have been responsible for the activation of different sub-

portions of the muscle (Johnson et al. 1994). Moreover,

compensations (i.e. left trunk bending) were visually

checked by one operator, but it may not have been possible

to detect minimal movements. However, the assessment

of motor behaviour in common life situations, such as

workplace and rehabilitation settings, have to cope with

individual postures and compensations. Yet, it is important

to note that the inter-subject variability of the EMG

Fig. 5 Position of the barycentre as a function of the force level (leftpanel) and consecutive ramps (right panel). In both plots, the

barycentres is expressed as rows. Each line is a different subject.

Circles represent the mean position of the barycentre at each force

level processed on the ramps. In the right panel, instead, the ramp

number is represented in the abscissa, the data are reported as mean of

the position of the barycentre calculated on the force levels (N = 5).

In both panels, squared markers were used to identify the position of

the barycentre averaged across subjects. Standard deviation is not

shown for clarity. A caudal shift of the barycentre can be observed

when force increases; on the contrary, the barycentre moves cranially

as a function of the number of ramp (repeated measures ANOVA,

p \ 0.01 for both)

Eur J Appl Physiol

123

Author's personal copy

distribution observed in this group of subjects may not be

present in other tasks, such as those involving open kinetic

chain.

Effect of force level on the cranio-caudal sEMG

amplitude distribution

Variations in the cranio-caudal distribution of EMG

amplitude were analyzed during force increase and in

consecutive ramps (Figs. 4b, c, 5). Both factors influenced

the shift of the barycentre, but their interaction was not

significant. Force production requires recruitment of MUs,

which may be localized in a discrete muscle region

(Desmedt and Godaux 1981; English and Segal 1993; Zhou

et al. 2011). In this work, the barycentre of EMG amplitude

was found to shift caudally with the increase of force.

Other authors analyzed variations of sEMG in the cranial-

caudal sub-portions of the trapezius muscle in force-

varying contractions (Holtermann and Roeleveld 2006;

Kleine et al. 2000; Troiano et al. 2008). Holtermann and

Roeleveld (2006) reported inhomogenieites in the activa-

tion of the trapezius muscle at different force levels, but the

position of the newly recruited MUs was not analyzed in

detail. The other studies reported no changes in cranio-

caudal distribution of EMG amplitude in ramps up to 50 %

MVC. The different results may be explained by the

smaller portion of muscle investigated in these studies

(interelectrode distance between the most cranial and the

most caudal electrode: 32 mm in Troiano et al. (2008),

60 mm in Kleine et al. (2000), and 96 mm in the present

work). Instead, the results similar to ours have been

obtained by Holtermann et al. (2008) who analyzed the

conduction velocity in the cranial and the transverse por-

tions of the trapezius muscle. No significant changes of

muscle fiber conduction velocity in the cranial portion were

found during 0–90 % MVC ramps, whereas it was evident

in the transverse portion of the muscle. As conduction

velocity is expected to increase with fresh MUs recruit-

ment, the results of Holtermann et al. (2008), together with

ours, suggest that MUs are preferentially recruited in the

transverse portion of the muscle when force exerted is

increased in the interval 0–25 % MVC. However, these

results cannot be generalized to higher force contractions

as Kleine observed shift of the sEMG distribution in the

cranial direction comparing 50 and 100 % MVC. The

observed caudal shift of the barycenter during force

increase may appear contradictory: given the fact that MUs

of different sub-portions of the trapezius muscle may be

recruited independently (Holtermann et al. 2009; Falla and

Farina 2008a, b), the recruitment of upper portions of the

trapezius muscle would have produced a force vector more

in line with scapula elevation. However, Palmerud et al.

(1998) showed that transverse fibers of the trapezius

muscle, together with rhomboid, may vicariate the role of

the upper trapezius in isometric tasks that require shoulder

elevation. Another possible interpretation is that tracking a

profile on the screen is an unusual task, and a precise

modulation of the force exerted is required. Subjects might

have taken advantage of co-contraction of synergic and

antagonist muscles, stiffening the joint for obtaining a

better performance (Osu et al. 2002).

Effect of task repetition on the cranio-caudal sEMG

amplitude distribution

During consecutive ramps, the subjects showed different

cranio-caudal amplitude distributions, with the barycentre

moving cranially during the exercise (Figs. 4c, 5). The

magnitude of the shift was lower than that due to force

increase. Subjects with the barycentre of EMG activity

toward the cranial portion of the trapezius in the first ramp

showed almost no shifts during the repetitions. Instead, the

subjects with the barycentre of EMG activity localized

caudally at the beginning of the task showed marked cra-

nial shifts of the amplitude map barycentre during the

contraction, meaning a progressive shift of muscle activity

in the direction of the cranial portion of the trapezius. This

phenomenon could be explained as a consequence of the

optimization of the MU recruitment with practice. In fact,

lower activity of antagonist and synergic muscles was

observed in expert drummers with respect to novices (Fujii

et al. 2009; Furuya and Kinoshita 2008) and was proven to

occur during motor task learning (Osu et al. 2002). Agonist

and antagonist co-contraction is frequent in the first stages

of motor learning process, to increase joint stiffness and

decrease the influence of external perturbations. This

activity decreases in parallel with the learning process

(Fujii et al. 2009; Furuya and Kinoshita 2008). In this

experiment, the cranial portion of the trapezius may be

considered the agonist of the movement, whereas the

caudal portion (transverse fibers) may act as stabilizer of

the scapula. Despite a training session of at least 5 min

with the force feedback, the subjects have not been able to

learn in a definitive way the required motor task. The

changes in EMG activity distribution that mainly occur

between the first and second ramp are probably due to a

fast recall of the strategy learned during the 10-min training

(that occurred about 20 min before the start of the mea-

sures). However, the collected data are not sufficient to

verify this hypothesis that remains speculative. No signif-

icant interactions were found on the effects of force level

and ramp number on the barycentre position. This means

that learning effects observed in consecutive ramps did not

occur at a preferential force level.

Eur J Appl Physiol

123

Author's personal copy

Conclusions

This work contributes to clarify some of the factors

underlying the subject-specific patterns of activation of the

trapezius muscle reported in literature during standardized

motor tasks and functional activities. The present study

shows that this variability could be related, to some extent,

to changes of the spatial location of the recruited MUs due

to the level of exerted force and the adaptation to the task.

The influence of other factors possibly related to inter-

subject variability (e.g. the direction of the exerted force

and the position of the scapula) still has to be investigated.

From a methodological point of view, the results of this

work suggest that HD-EMG can help in identifying some

of the factors at the origin of the variability of trapezius

muscle sEMG patterns during daily or work activities.

Acknowledgments This work was financially supported by Fond-

azione Cassa di Risparmio di Torino, Italy, and Compagnia di San

Paolo, Torino, Italy.

Conflict of interest The authors declare that they have no conflict

of interest.

References

Balogh I, Hansson GA, Ohlsson K, Stromberg U, Skerfving S (1999)

Interindividual variation of physical load in a work task. Scand J

Work Environ Health 25(1):57–66

Barbero M, Gatti R, Lo Conte L, Macmillan F, Coutts F, Merletti R

(2011) Reliability of surface EMG matrix in locating the

innervation zone of the upper trapezius. J Electromyogr Kinesiol

21(5):827–833

Desmedt JE, Godaux E (1981) Spinal motoneuron recruitment in

man: rank deordering with direction but not with speed of

voluntary movement. Science 214(4523):933–936

English AW, Segal RL (1993) Compartmentalization of muscles and

their motor nuclei: the partitioning hypothesis. Phys Therapy

73(12):857–867

Falla D, Farina D (2008a) Motor units in cranial and caudal regions of

the upper trapezius muscle have different discharge rates during

brief static contractions. Acta Physiol (Oxf) 192(4):551–558

Falla D, Farina D (2008b) Non-uniform adaptation of motor unit

discharge rates during sustained static contraction of the upper

trapezius muscle. Exp Brain Res 191(3):363–370

Farina D, Leclerc F, Arendt-Nielsen L, Buttelli O, Pascal M (2008)

The change in spatial distribution in upper trapezius muscle

activity is correlated to contraction duration. J Electromyogr

Kinesiol 18:16–25

Fjellman-Wiklund A, Grip H, Karlsson JS, Sundelin G (2004) EMG

trapezius muscle activity pattern in string players: part I—is

there variability in the playing technique? Int J Ind Ergon

33:347–356

Fujii S, Kudo K, Ohtsuki T, Oda S (2009) Tapping performance and

underlying wrist muscle activity of non-drummers, drummers,

and the world’s fastest drummer. Neurosci Lett 459(2):69–73

Furuya S, Kinoshita H (2008) Organization of the upper limb

movement for piano key-depression differs between expert

pianists and novice players. Exp Brain Res 185:581–593

Holtermann A, Roeleveld K, Karlsson JS (2005) Inhomogeneities in

muscle activation reveal motor unit recruitment. J Electromyogr

Kinesiol 15:131–137

Holtermann A, Karlsson JS, Roeleveld K (2008) Spatial distribution

of active muscle fibre characteristics in the upper trapezius

muscle and its dependency on contraction level and duration.

J Electromyogr Kinesiol 18(1):16–25

Holtermann A, Roeleveld K, Mork PJ et al (2009) Selective activation

of neuromuscular compartments within the human trapezius

muscle. J Electromyogr Kinesiol 19:896–902

Holtermann A, Roeleveld K (2006) EMG amplitude distribution

changes over the upper trapezius muscle are similar in sustained

and ramp contractions. Acta Physiol 186:159–168

Johnson G, Bogduk N, Nowitze A, House D (1994) Anatomy and

actions of the trapezius muscle. Clin Biomech 9(1):44–50

Kleine B, Schumann N, Stegeman DF (2000) Surface EMG mapping

of the human trapezius muscle: the topography of monopolar and

bipolar surface EMG amplitude and spectrum parameters at

varied forces and in fatigue. Clin Neurophysiol 111:686–693

Madeleine P, Leclerc L, Arendt-Nielsen L, Ravier P, Farina D (2006)

Experimental muscle pain changes the spatial distribution of

upper trapezius muscle activity during sustained contraction.

Clin Neurophysiol 117:2436–2445

Mathiassen SE, Aminoff T (1997) Motor control and cardiovascular

responses during isoelectric contractions of the upper trapezius

muscle: evidence for individual adaptation strategies. Eur J Appl

Physiol Occup Physiol 76(5):434–444

Merletti R, Rainoldi A, Farina D (2001) Surface electromyography

for noninvasive characterization of muscle. Exerc Sport Sci Rev

29:20–25

Merletti R, Botter A, Cescon C, Minetto MA, Vieira TM (2010)

Advances in surface EMG: recent progress in clinical research

applications. Crit Rev Biomed Eng 38(4):347–379

Mork PJ, Westgaard RH (2005) Long-term electromyographic

activity in upper trapezius and low back muscles of women

with moderate physical. J Appl Physiol 99:570–578

Osu R, Franklin DW, Kato H et al (2002) Short- and long-term

changes in joint co-contraction associated with motor learning as

revealed from surface EMG. J Neurophysiol 88:991–1004

Palmerud G, Sporron H, Herberts P, Kadefors R (1998) Consequences of

trapezius relaxation on the distribution of shoulder muscle forces: an

electromyographic study. J Electromyogr Kinesiol 8(3):185–193

Troiano A, Naddeo F, Sosso E, Camarota G, Merletti R, Mesin L

(2008) Assessment of force and fatigue in isometric contractions

of the upper trapezius muscle by surface EMG signal and

perceived exertion scale. Gait Posture 28:179–186

Tucker K, Butler J, Graven-Nielsen T, Riek S, Hodges P (2009)

Motor unit recruitment strategies are altered during deep-tissue

pain. J Neurosci 29:10820–10826

Staudenmann D, Kingma I, Daffertshofer A, Stegeman DF, Van

Dieen JH (2009) Heterogeneity of muscle activation in relation

to force direction: a multi-channel surface electromyography

study on the triceps surae muscle. J Electromyogr Kinesiol

19:882–895

Vieira TMM, Loram I, Muceli S, Merletti R, Farina D (2011) Postural

activation of the human medialis gastrocnemius muscle: are the

muscle units spatially localised? J Physiol 589(2):431–443

Watanabe K, Kouzaki M, Moritani T (2011) Task-dependent spatial

distribution of neural activation pattern in human rectus femoris

muscle. J Electromyogr Kinesiol 22(2):251–258

Wickham JB, Brown JM (2012) The function of neuromuscular

compartments in human shoulder muscles. J Neurophysiol 107(1):

336–345

Zhou P, Suresh NL, Rymer WZ (2011) Surface electromyogram

analysis of the direction of isometric torque generation by the

first dorsal interosseous muscle. J Neural Eng 8(3):036028

Eur J Appl Physiol

123

Author's personal copy