Surg Radiol Anat (2008) 30:137–143
DOI 10.1007/s00276-007-0289-y
O R I G I N A L A R T I CL E
Anatomical optimization of skin electrode placement to record
electromyographic activity of erector spinae muscles
Mathieu Panchoa de Sèze · Jean-René Cazalets
Received: 7 July 2007 / Accepted: 29 November 2007 / Published online: 9 January 2008
© Springer-Verlag 2007
Abstract Fine analysis of body movements is now tech-
nologically feasible, together with simultaneous recording
of multiple muscle activity. This is especially true for the
trunk and back muscles during human walking. However,
there have been few anatomic studies of the area where
deep back muscle activity is recordable by skin electrodes.
We therefore attempted to optimize skin electrode location
for recording erector spinae muscle activity at diVerent
levels of the back. For this purpose, 20 dissections of the
posterior wall of the trunk were performed. The cutaneous
plane was reclined on both sides to expose the superWcial
muscles of the posterior wall of the trunk. We dissected
then plane-by-plane until we exposed the erector spinae
muscles. The widths of the fascial spinal muscle insertions
were measured at spinal levels easily identiWed clinically,
i.e., C7, T3, T7, T12 and L4. Electromyographic assess-
ment of the electrode location at these levels was performed
in three subjects. Erector spinae muscle activity proved
possible to record on several areas of the posterior wall
through a superWcial muscle aponeurosis. We propose a
protocol for placing skin electrodes to record erector spinae
muscle activity based on clinical anatomical references.
Keywords Anatomy · Electromyography ·
Erector spinae · Muscle · Rehabilitation · Fascia
M. P. de Sèze (&)
Laboratoire d’anatomie UFR2, Université Bordeaux 2,
146 rue Léo-Saignat, 33076 Bordeaux cedex, France
e-mail: Mathieu.de-seze@chu-bordeaux.fr
J.-R. Cazalets
Université Bordeaux 2/CNRS, 146 rue Léo-Saignat,
bat 2, zone nord, 33076 Bordeaux cedex, France
Introduction
Fine analysis of body movements is now technologically
feasible, together with simultaneous recording of multiple
muscle activity. Up to 32 muscles activities can be simulta-
neously recorded by skin electrodes during walking [5]. An
outcome of these new wireless electromyographic tech-
niques is the emergence of studies on the role of the trunk
and of the back muscles during human walking [26].
Several functional studies investigated the lumbar and abdomi-
nal muscles, to understand how the muscles stabilize the
thorax on the pelvis during human walking [1, 4, 6, 7, 9,
12]. During walking, the vertebral column is globally stabi-
lized from the pelvis to the head [17] through the involve-
ment of the deep group of the back muscles [25]. These
deep muscles extend from the pelvis to the skull [8, 11].
The most superWcial are the erector spinae (ES) muscles
that are covered by the cervical and thoraco-lumbar fascia
separating them from the intermediate and superWcial
planes of the back muscles [2].
Several studies have proposed locations for skin elec-
trodes to record the muscular activity of the low back and
abdominal muscles [1, 4, 6, 7, 9, 12]. The surface electrode
positions used to record multiWdus muscle activity are
homogeneous and located 1 cm laterally to the spinous pro-
cess of L5 [1, 9, 22]. In contrast, the recording sites pro-
posed to monitor ES muscle activity are less homogeneous,
both at the intersegmental level and within a considered
segment. The variety of segmental level proposed spans
from T1 [15] to L4 [19], including T9 [15, 23], T12 [4, 19],
L1 [1, 9], L2 [15, 19, 22] and L3 [4, 23]. One reason for the
variety of proposed locations is the length of the ES, which
lines the spine on almost all its length [11]. A second rea-
son is the existence of two large anatomic areas where it
seems possible to record ES muscle activity by minimizing
123
138
the overlapping by the activity of the superWcial muscles.
Indeed, these areas correspond to the region where the thor-
aco-lumbar fascia covering the ES muscles is not covered
itself by muscular Wbers of superWcial muscles [2, 11].
Thus, recording of the ES activity had been proposed later-
ally at the spinous process of L1 at distances of 2 and 3 cm
[1, 15, 19] but also halfway between the spinous process of
L1 and the lateral side of the back [9], and at a point
marked by the intersection of a horizontal line crossing the
spine at the L2 level with the 12th rib [22]. The aim of this
study was therefore to describe the areas in which the thor-
aco-lumbar fascia is not covered by muscular Wbers, to pro-
pose locations where skin electrodes could be placed for
recording ES muscle activity.
Materials and methods
This study included two parts, one anatomical work per-
formed on cadavers and one electromyographic work per-
formed on healthy volunteers without history of low back
pain or spine deformity.
Anatomic work
We performed 20 dissections of the posterior wall of the
trunk from 6 fresh and 14 embalmed cadavers (9 males and
11 females). Owing to technical problems, the age, weight
and height of the subjects were not available. Before dis-
section began, the distance between the two acromions and
the width of the pelvis were measured in 11 subjects (5
males and 6 females), to establish a relation between the
external morphology and the width of the fascial muscle
insertions. Dissection began by a median incision on the
Fig. 1 Posterior view of the
superWcial muscles of the back.
a Dissection view. b Schematic
posterior view. Lat lateral direc-
tion, Cr cranial direction, in gray
fascial area
123
Surg Radiol Anat (2008) 30:137–143
spinous line. The cutaneous plane was reclined on both
sides to expose the superWcial muscles and their fascial
insertions (Fig. 1). The trapezius muscles were dissected
Wrst on their inferior and lateral sides. Progression was then
caudo-cranial by ascending resection of the spinal inser-
tions of these muscles until the rhomboideus muscle was
visible (Fig. 2). Dissection of the rhomboideus muscles dis-
played the serratus posterior superior (SPS) muscles, which
covered the thoraco-lumbar fascia and the deep muscles
(Fig. 3). The latissimus dorsi muscles were Wrst dissected
on their superior sides. This exposed the serratus posterior
inferior (SPI) muscle covering the thoraco-lumbar fascia
and the lumbar part of the ES muscles. The Xeshy and fas-
cial parts of these muscles could thus be diVerentiated. The
widths of the fascial spinal insertions were measured at
easily identiWed spine levels, i.e., C7 (identiWed by its
prominent spinous process at the base of the neck), T3
(medial angle of the scapula), T7 (inferior angle of the
scapula), T12 (identiWed by its prominent spinous process
at the thoraco-lumbar junction) and L4 (level of iliac crest).
Electromyographic work
We then hypothesized locations for skin electrodes at
diVerent spine levels. Electromyographic assessment of the
electrode location at these levels was performed in three
subjects. Tripolar surface electrodes (Thought Technology,
Canada; 10 mm diameter, 10 mm interelectrode distance)
applied to lightly abraded skin were used to record muscle
activity. The signal was ampliWed (£10,000) using 6 unit
KineMyo ampliWers (Kine, Iceland). The system consists of
surface EMG units that are snapped to the triodes. The sig-
nal is ampliWed and digitized (1,562 Hz) only a few milli-
meters above the skin, thereby picking up a clean signal
Surg Radiol Anat (2008) 30:137–143 139

Fig. 2 Posterior view of the

rhomboideus muscles. a Dissec-
tion view. b Schematic posterior
view. Lat lateral direction, Cr
cranial direction, in gray fascial
area, E. sc elevator scapulae, R
rhomboideus, ES erector spinae

Fig. 3 Posterior view of the

deep muscles. a Dissection
view. b Schematic posterior
view. Lat lateral direction, Cr
cranial direction, in gray fascial
area, IC iliocostalis, L longissi-
mus, R rhomboideus, S splenius,
SPS serratus posterior superior,
TR trapezius

minimizing crosstalk, and is then wireless-transmitted to a
base station. To optimize signal detection of muscle activi-
ties, the location of the surface electrodes was carefully
chosen from other studies for trapezius [15], latissimus
dorsi [22] and ES muscles at C7 [15] and L4 [19], and from
our own anatomical study for the ES at T3 and T12 (Fig. 4).
Electrode placement for the latissimus dorsi muscle was
over the muscular curve at T12 and along a line connecting
the most posterior point of the posterior axillary fold and
the S2 spinous process. For the superior part of the trape-
zius muscle, it was over the muscular curve as recom-
mended [15, 22]. For the ES, we plotted two horizontal
points 2 and 4 cm from the spinous process of C7 and L4,
respectively. We then drew a line joining these two points
and placed the electrodes on the line at each spinal level
that are to be studied. Subjects gave informed consent and
procedures were approved by the local ethics committee, in
accordance with the ethical standards in the Declaration of
Helsinki.
Statistical analysis
Since no diVerence was found between embalmed and fresh
cadavers, on the width of muscular and facial tissues in the
lumbar area [27], data recorded from the two procedures
were pooled here. Quantitative data were analyzed by aver-
aging and standard deviation, and were compared with a
bilateral paired student test using the statistical function in
Excel (Microsoft, WA, USA).
Results
The interacromial distance and the width of the pelvis were
39.5 § 5.9 cm and 36.1 § 3.1 cm, respectively. Muscle
measurements are shown in Table 1. The muscle insertions
were organized symmetrically without any signiWcant
diVerence in width between each side (P < 0.05). Thus, the
skin resection revealed a large fascial insertion for both LD.
This insertion circumscribed a diamond-shaped area with a
large vertical axis and a small horizontal axis at the level of
the iliac crest, i.e., at L4 (Fig. 5). In 15 subjects, spinal fas-
cial insertions of the trapezius muscles circumscribed a tri-
angular area with an inferior base located at T4 (Fig. 6). In
the other 5 subjects, there was a diamond-shaped area with
a large vertical axis and a small horizontal axis centered on
the C7 spinous process, as is commonly found (Fig. 1). The
spinal insertions of the rhomboideus and SPS muscles
recovered the ES and splenius muscles. They showed nar-
row spinal insertions on the spinous process of C7. The
insertions of the rhomboideus muscles on the spinous
process of C7 were fascial in 18 subjects (90%) and were

123
140 Surg Radiol Anat (2008) 30:137–143

Fig. 4 Anatomic location of

skin electrodes to record erector
spinae muscle activity. a Ana-
tomic schema. b Sample of
EMG recording in a subject
walking at 5 km/h.
LD latissimus dors,
Trap superior part of the
trapezius

Table 1 Spinal insertions of

Spinal Muscles Spinal insertion nature Width of the insertions t test
muscles
levels Mean § standard deviation (cm)

Left Right

C7 Trapezius Aponeurotic 3.8 § 1.1 3.9 § 0.9 NS

Rhomboideus Muscular Wber (n = 2) 0.3 § 0.6 0.3 § 0.6 NS
Aponeurotic (n = 18)
Serratus posterior superior Muscular Wber (n = 1) 0.0 § 0.2 0.0 § 0.2
Aponeurotic (n = 19)
T3 Trapezius Aponeurotic 3.9 § 1.2 3.9 § 1.1 NS
Rhomboideus Aponeurotic 3.3 § 1.3 3.4 § 1.2 NS
Serratus posterior superior Aponeurotic (n = 8) 5.4 § 0.9 5.4 § 0.8 NS
T7 Trapezius Muscular Wber 3.9 § 1.1 4.1 § 1.2 NS
T12 Latissimus dorsi Aponeurotic 4.1 § 0.7 4.4 § 0.8 NS
Serratus posterior superior Aponeurotic 5.9 § 0.9 6.1 § 0.9 NS
L4 Latissimus dorsi Aponeurotic 8.6 § 1.5 8.6 § 1.7 NS

Fig. 5 Posterior view of the spi-

nal latissimus dorsi insertions.
a Dissection view. b Schematic
posterior view. Lat lateral direc-
tion, Cr cranial direction, in gray
fascial area, LD latissimus dorsi
muscular Wbers, AA fascial area,
GM gluteus maximus muscle,
EAO external abdominal oblique
muscle

123
Surg Radiol Anat (2008) 30:137–143
Fig. 6 Posterior view of the spi-
nal trapezius insertions. a Dis-
section view. b Schematic
posterior view. Lat lateral direc-
tion, Cr cranial direction, n gray
fascial area, Tr inferior part of
the trapezius, Trm middle part of
the trapezius
muscular in the other 2 (10%). At this level, the insertions of
the SPS muscles were fascial in 19 subjects and muscular in
the other one. The insertions of the rhomboideus and SPS mus-
cles were always fascial on the spinous process of T3. The
SPS muscles had insertions on the spinous process of T3
only in 8 subjects (Fig. 3). The fascial spinal insertions of
the SPI muscles were systematically wider than that of the
LD muscle. The fascial spinal insertions of the SPI muscles
were systematically wider than those of the LD muscles.
Thus, in 90% of the subjects, the deep muscles were cov-
ered only by fascial structures at C7 across 10% of the
interacromial distance. In the remaining 10% of subjects,
the medial zone of the deep muscles was covered by Xeshy
muscular Wbers across 3.5% of the interacromial distance.
At T3, the deep muscles were covered only by fascial struc-
tures across 9.5% of the interacromial distance. At T12 and
L4, the deep muscles were covered only by fascial struc-
tures across 24 and 53% of the half-width of the pelvis,
respectively.
Three healthy male university students were recruited for
the electromyographic study. The subjects’ age, height and
weight ranged from 25 to 44 years, 169–182 cm and 65–
75 kg, respectively. All subjects were right-leg dominant
and had normal body fat (each body mass index, calculated
by weight expressed in kilograms divided by the square of
the height expressed in meters, was lower than 28 kg/m2), no
low back pain history and no clinically detectable scoliosis
or kyphosis. The electromyographic traces shown in Fig. 4
present the muscle activity in one subject during forward
walking. In the three subjects recorded, the rhythmic activity
at C7 and T3 clearly exhibited a motor proWle diVerent from
that recorded in the superior parts of the trapezius. The
activity recorded at T12 and L4 exhibited a motor proWle
diVerent from that recorded in the LD.
Discussion
These Wndings show that ES muscle activity can be
recorded in several locations of the posterior wall through a
superWcial muscle aponeurosis. Some of these locations
141
have already been proposed [13, 15, 18, 20, 23, 28]. How-
ever, our results detail the width of the fascial spinal inser-
tions of the superWcial muscles more than anatomical
textbooks do [2, 3, 8, 11, 16]. Such details may be helpful
to locate sites where ES muscle activity can best be
recorded. In other words, they may be helpful to optimize
the placement of the skin electrodes to record activity of the
ES muscles. From C7 to T3, the area of electromyographic
recording of the ES muscles is limited laterally by the line
junction (i.e., the line where the muscular Wbers are inserted
on the tendinous Wbers) of the trapezius and rhomboideus
muscles. At C7, it extends laterally over 20% of the interac-
romial distance. In 10% of subjects, it is limited medially
by the Xeshy Wbers of the rhomboideus and/or SPS over a
width equal to 3.5% of the interacromial distance. These
results are consistent with photographic evidence in an ana-
tomic atlas [24]. They are also consistent with a previously
proposed site to monitor ES activity located 2 cm laterally
to the spinous process of T1 [15]. The electromyographic
activity of the trapezius diVers from the activity recorded
laterally to T3 level. This is consistent with the absence of
Xeshy muscular Wber recovering ES muscles at the same
level and we thought that it would be possible to monitor
the ES muscle laterally to the T3 spinous process over a
width equal to 9.5% of the interacromial distance. Between
T5 and T11, electromyographic recording of the ES is
impossible, because there is no fascial window above them.
Indeed, at these levels, the Xeshy muscle Wbers of the trape-
zius and LD muscles completely cover the ES muscles
(Fig. 1). Therefore, previous recordings of ES activity at T9
[4, 15, 23] may have been inXuenced by the activities of the
trapezius and LD muscles. At the lumbar level, the LD and
SPI muscles cover the ES muscles. The area where it is
possible to record the electromyographic activity of the ES
is limited by the fascial insertion of these muscles. That
insertion extends laterally from the spinous line to 24 and
53% of the half-width of the pelvis at T12 and L4, respec-
tively. These results are consistent with previous Wndings
[2] and photographic evidence in an anatomic atlas [24].
They are also consistent with the previous sites used to
record the activity of the ES muscles at T12 [19], L1 [1], L2
123
142
[15, 19], L3 [4, 23] and L4 [19]. In contrast, recordings of
ES activity very laterally at L1 and L2 [9, 22], may have
been inXuenced by the muscle activity of the interspersed
Xeshy Wbers of the LD muscle.
A limitation of this study is that we did not diVerentiate
the ES and splenius muscles at the thoraco-cervical level,
nor did we diVerentiate the multiWdus and ES muscles in
the thoraco-lumbar region. However, the multiWdus mus-
cles are superWcial only at the low lumbar level [2], and
their activity has been recorded near the spinous process
only at L5 [1]. Moreover, while the multiWdus muscles act
more locally to prevent disk disruption [10, 21, 25], the
longissimus and iliocostalis muscles are usually considered
to be physically active simultaneously during walking and
dynamic postural adjustments [21]. Since the common glo-
bal function of the ES muscles is dynamic trunk stabiliza-
tion [22, 23], their activity is of speciWc interest during
locomotor tasks such as walking.
Another limitation concerns the small number of sub-
jects whose electromyographic activity was studied in vivo.
However, that part of the study was designed only to con-
Wrm the low level of overlap in electromyographic activity
between the superWcial and ES muscles. Physiological vali-
dation of the sites would require comparison between sur-
face recordings obtained at each site and intramuscular
reference recordings, as is the case for other muscles [14].
While that study will be performed subsequently, our aim
here was to deWne the most appropriate areas for recording
ES muscle activity with skin electrodes.
Conclusion
The present Wndings suggest a simple skin electrode conWgu-
ration for recording deep muscle activity by minimizing over-
lapping by the activity of the superWcial muscles of the
posterior wall of the trunk. The conWguration is based on clin-
ical anatomical references, i.e., C7 (identiWed by its promi-
nent spinous process located at the base of the nape); T3
(located at the level of the medial angle of the scapula); T12
(identiWed by its prominent spinous process located at the
thoraco-lumbar junction); and L4 (located at the level of the
iliac crest). It is simple to build (Fig. 4). After marking two
points 2 and 4 cm laterally from the spinous process of C7
and L4, respectively, the points are joined by a line and the
electrodes are placed on it at each spine level to be studied.
References
1. Anders C, Wagner H, Puta C, Grassme R, Petrovitch A, Scholle
HC (2007) Trunk muscle activation patterns during walking at
diVerent speeds. J Electromyogr Kinesiol 17:245–252
123
Surg Radiol Anat (2008) 30:137–143
2. Bogduk N (2005) Les muscles lombaux et leurs fascias. In: Anat-
omie du rachis lombal et sacré. Elsevier, Amsterdam, pp 131–
167
3. Bouchet A, Cuilleret J (1983) Anatomie topographique, descrip-
tive et fonctionnelle. Tome 2: Le cou et le thorax. Simep, Paris,
pp 931–943
4. Callaghan JP, Patla AE, McGill SM (1999) Low back three-
dimensional joint forces, kinematics, and kinetics during walking.
Clin Biomech (Bristol, Avon) 14:203–216
5. Cappellini G, Ivanenko YP, Poppele RE, Lacquaniti F (2006)
Motor patterns in human walking and running. J Neurophysiol
95:3426–3437
6. Carlson H, Nilsson J, Thorstensson A, Zomlefer MR (1981) Motor
responses in the human trunk due to load perturbations. Acta Phys-
iol Scand 111:221–223
7. Chen WJ, Chiou WK, Lee YH, Lee MY, Chen ML (1998) Myo-
electric behavior of the trunk muscles during static load holding in
healthy subjects and low back pain patients. Clin Biomech (Bris-
tol, Avon) 13:S9–S15
8. Chevallier J (1998) Anatomie 2: Appareil locomoteur. Medecine
sciences Xammarion, Paris
9. Danneels LA, Coorevits PL, Cools AM, Vanderstraeten GG, Cam-
bier DC, Witvrouw EE, De CH (2002) DiVerences in electromyo-
graphic activity in the multiWdus muscle and the iliocostalis
lumborum between healthy subjects and patients with sub-acute
and chronic low back pain. Eur Spine J 11:13–19
10. Donisch EW, Basmajian JV (1972) Electromyography of deep
back muscles in man. Am J Anat 133:25–36
11. Drake R, Vogl W, Mitchell A (2006) Musculature du dos. In:
Gray’s anatomie pour les étudiants. Ed française, Elsevier Masson
SAS, pp 45–65
12. Granata KP, Wilson SE (2001) Trunk posture and spinal stability.
Clin Biomech (Bristol, Avon) 16:650–659
13. Hodges PW (2001) Changes in motor planning of feedforward
postural responses of the trunk muscles in low back pain. Exp
Brain Res 141:261–266
14. Ivanenko YP, Poppele RE, Lacquaniti F (2004) Five basic muscle
activation patterns account for muscle activity during human loco-
motion. J Physiol 556:267–282
15. Ivanenko YP, Poppele RE, Lacquaniti F (2006) Spinal cord maps
of spatiotemporal alpha-motoneuron activation in humans walking
at diVerent speeds. J Neurophysiol 95:602–618
16. Kamina P, Rideau Y (2000) Myologie des membres. Maloine,
Paris, pp 14–15
17. Konz RJ, Fatone S, Stine RL, Ganju A, Gard SA, Ondra SL (2006)
A kinematic model to assess spinal motion during walking. Spine
31:E898–E906
18. Kumar S, Zedka M, Narayan Y (1999) EMG power spectra of
trunk muscles during graded maximal voluntary isometric contrac-
tion in Xexion-rotation and extension-rotation. Eur J Appl Physiol
Occup Physiol 80:527–541
19. Lamoth CJ, Meijer OG, DaVertshofer A, Wuisman PI, Beek PJ
(2006) EVects of chronic low back pain on trunk coordination and
back muscle activity during walking: changes in motor control.
Eur Spine J 15:23–40
20. Lariviere C, Gagnon D, Loisel P (2000) The comparison of trunk
muscles EMG activation between subjects with and without
chronic low back pain during Xexion-extension and lateral bend-
ing tasks. J Electromyogr Kinesiol 10:79–91
21. MacDonald DA, Moseley GL, Hodges PW (2006) The lumbar
multiWdus: does the evidence support clinical beliefs? Man Ther
11:254–263
22. Ng JK, Parnianpour M, Richardson CA, Kippers V (2001) Func-
tional roles of abdominal and back muscles during isometric axial
rotation of the trunk. J Orthop Res 19:463–471
Surg Radiol Anat (2008) 30:137–143
23. Radebold A, Cholewicki J, Panjabi MM, Patel TC (2000) Muscle
response pattern to sudden trunk loading in healthy individuals and
in patients with chronic low back pain. Spine 25:947–954
24. Rohen J, Yokochi C, Lütjen-Drecoll E (1999) Anatomie Humaine:
Atlas photographique d’anatomie systématique et topographique.
Maloine, Paris, pp 209–215
25. Solomonow M, Zhou BH, Baratta RV, Burger E (2003) Biomechan-
ics and electromyography of a cumulative lumbar disorder: response
to static Xexion. Clin Biomech (Bristol, Avon) 18:890–898
143
26. Sutherland DH (2001) The evolution of clinical gait analysis part
l: kinesiological EMG. Gait Posture 14:61–70
27. Vialle R, Wicart P, Drain O, Dubousset J, Court C (2006) The wilt-
se paraspinal approach to the lumbar spine revisited: an anatomic
study. Clin Orthop Relat Res 445:175–180
28. Wolf SL, Bobinac-Georgievski A, Braus V, Montani M (1997)
Electrokinesiologic measurement of trunk sagittal mobility and
lumbar erector spinae muscle activity. J Rehabil Res Dev 34:470–
478
123