Multi Channel Surface EMG
Detection and Conditioning
M. Gazzoni and U. Barone
Laboratory for Engineering of the Neuromuscular System (LISiN), Department of Electronics and Telecommunications,
Politecnico di Torino, Torino, Italy
Keywords: Multi-channel Surface EMG, Spatial Filtering, Spatial Sampling.
Abstract: The electromyogram is a compound signal comprising the electrical activity of the motor units activated
asynchronously during voluntary muscle contractions. The temporal and spatial evolution of EMG can be
sampled by surface electrodes. The basic principles and concepts about sEMG signal conditioning, spatial
filtering, and spatial sampling are introduced and discussed.
1 INTRODUCTION
The electromyogram (EMG) is a compound signal
comprising the electrical activity of the motor units
(MU) activated asynchronously during voluntary
muscle contractions. The summation of action
potentials of active MUs generates, on the skin
surface, an electrical field; the surface EMG
(sEMG). Temporal and spatial evolution of this field
might be sampled by surface electrodes
appropriately positioned above active muscle
regions. The properties of the detection system as
well as the characteristics of the circuits for the
conditioning of sEMG influence its quality and
informative content.
2 sEMG CONDITIONING
The electrode skin interface properties change
continuously due to its sensitivity to environmental,
chemical and mechanical factors which affect
electrical properties. A collection of common
disturbing events and unpredictable interfering
signals are: a) motion artifacts; b) mechanical
vibrations of cables with consequent variation of
capacitance, electric charges, and voltage drop; c)
power line interference coupling; d) ground loops; e)
fluctuations of electrode polarization; f) charge
distribution variability on skin layers;
Mechanical solutions and fabrication methods
have been studied to design surface electrodes with
low polarizable level and low noise floor within
EMG frequency bandwidth.
Several articles and technical notes about bio-
potential circuit implementations (AFE, Analog
Front End) were published. Different approaches
and solutions were presented to properly detect bio-
potentials characterized by low amplitude (order of
microvolt), extremely low frequency band (under
1kHz), high DC offset (up to ±0.5VDC) and low
SNR ([5dB 35dB]). General design criteria were
published for bioelectric data acquisition systems
(R.R. Harrison, 2007), (Bernhard Fuchs, 2002).
System-on-Chip based approaches were also
proposed (N.V. Helleputte, 2008) for advanced
biomedical applications such as miniaturized and
implantable sensing amplifiers (Wang 2006, R.R.
Harrison 2003, T. Denison 2007, R. Rieger 2006)
and wearable electronic sensors (L. Yan 2009).
Table 1 reports the main properties of typical surface
EMG signal.
Table 1: EMG Signal characteristics for a typical detection
system based on Ag/AgCl electrode with wet conductive
gel.
Property Value
Min Max
EMG bandwidth (BWEMG) 20Hz 500Hz
EMG Peak Amplitude 100μV 5mV
Total RMS Noise Voltage
(EMG)
10μVRMS
119
Gazzoni M. and Barone U..
Multi Channel Surface EMG - Detection and Conditioning.
DOI: 10.5220/0004663701190125
In Proceedings of the International Congress on Neurotechnology, Electronics and Informatics (DeNeuro-2013), pages 119-125
ISBN: 978-989-8565-80-8
Copyright
c
2013 SCITEPRESS (Science and Technology Publications, Lda.)
ADC
NBI T
12
NBIT
ii
FSR
LSB
kLSBS
FSR: Full Scale Range
S/H
AFE
Impedance
Matching
Active
circuit
(optional)
EMG channel
Reference
Electrode
OR
EMG channel
skin
Conductive gel
electrode
Z
EMG
EMG
COMMON MODE FEEDBA CK
HOLD
DC offset
Re moval
circuit
INA
V
chB
V
cm
AFE
Anti-aliasing Fi lter
~
~
~
V
chA
V
chA
V
chB
+
-
>500Hz
V
EMG
Z
AFE
V
EMG
= G
AFE
·(V
chA
–V
chB
)
Z
AFE
Z
AFE
>> Z
EMG
within [20Hz-500 Hz]
Figure 1: Summary of one channel EMG amplifier chain until the digitizing process. Three main blocks describe the full
signal conditioning flow. The dotted blocks are optional and point out to specific improvement of the amplifier. The
impedance matching circuit minimizes the mismatch between the equivalent impedances of the electrode-gel-skin contacts
in order to reduce the false differential voltage detection (voltage divider effect). An example of Analog front End amplifier
(AFE) implementation is reported by functional blocks. Specifically, a monolithic Instrumentation amplifier (INA)
performs the differential signal extraction and amplification (VEMG=Gain·[VchA-VchB]). The common mode trend Vcm
of the input signals is rejected the common mode component Vcm principally caused by power line interference coupling
and motion artefacts. Specifically, low noise amplifier chain (AFE) and high resolution digitizing process (ADC) are strictly
required to design a high resolution EMG detection system according to EMG signal characteristics and Electrode-Gel-Skin
properties.
The most important building block for EMG
recording is the signal conditioning chain (see
Figure
1). Integrated circuits selection, configuration and
dimensioning should be performed in order to
maximize the signal quality until the digitizing
process.
The primary aspects to handle during EMG
amplifier design are:
Efficient techniques for removal of DC due to
electrode polarization effect.
Very flat Differential Voltage gain within EMG
bandwidth.
High accuracy voltage gain setting (<1%) and
good linearity within full voltage dynamic.
Low gain mismatch among channels (<0.5%).
AFE transfer function with very low group delay
within EMG frequency band.
Very high Common Mode Rejection within EMG
bandwidth (CMRR > 90dB).
High Power Supply rejection (PSRR > 80dB).
Very high Input impedance (|Z
AFE
|> 100M)
within EMG frequency band.
Negligible referred-to-input Total Noise floor
level with respect to Electrode-skin interface
noise
EMG
(e.g. 1μV
RMS
within EMG bandwidth
[20Hz-500Hz]).
Very low harmonic distortion of the EMG Power
spectrum.
High accuracy (<0.05%), very low noise
(<3μVPP) voltage reference for A/D conversion.
Programmable sampling frequency (>1ksps/ch)
and simultaneous digitizing process of EMG
signals.
High performances, optical isolating interfaces to
guarantee patient safety (IEC-60601).
Secondary aspects should also be evaluated, to
optimize the circuit performances. The main
characteristics to be focused on are: a) Low power
and low voltage operating conditions (e.g. Battery
based power supply system); b) Innovative small-
sized integrated circuit suitable for wearable medical
device development; c) Use of advanced materials
and new technologies to improve the behaviour of
the electrode-gel-skin interface.
NEUROTECHNIX2013-InternationalCongressonNeurotechnology,ElectronicsandInformatics
120
3 ELECTRODE
CONFIGURATION
AND LOCATION
The importance of a standardization of electrode
configuration and location for the reproducibility
and correct interpretation of sEMG measurements
has been widely recognized. The European
concerted action SENIAM showed that a large
variety of sensors, sensor placement procedures, and
equipment are used in the European laboratories to
detect sEMG signals (Hermens et al. 1999) and this
is still one of the major issues in clinical surface
EMG.
3.1 Monopolar Detection
The ideal configuration for the detection of the
potential distribution on the skin is to move on the
skin a point electrode measuring the voltage with
respect to a remote reference where the potential is
zero (monopolar detection).
The monopolar detection provides the whole
information which can be recorded from the
detection volume but it is mainly used in research
applications because of its lack of spatial selectivity
(recording of the contribution of sources that are
near or far from the electrodes) and its sensitivity to
common mode signals. Spatial filtering techniques
have been proposed to detect surface EMG signals
tendon
5 mm
end plate
45 mm
1234 5
muscle
detection
points
Monopolar detection
NDD detection
Single differential detection
Double differential detection
123 45 123 45
123 45
123 45
20 ms
5 mm
Figure 2: Example of the spatial selectivity of different
spatial filters with respect to propagating and non
propagating components. It is possible to observe: a) the
higher spatial selectivity of NDD in transversal direction
with respect to mono-dimensional spatial filters. 2) the
enhancement of end-of-fiber effects in the case of NDD
filter with respect to the reduction obtained by SD and DD
filters.
enhancing the spatial selectivity of surface
recordings by limiting the detection volume
(Broman et al.1985; Disselhorst-Klug et al. 1997;
Farina et al. 2002a).
3.2 Spatial Filtering
Spatial filters in surface EMG detection are based on
the linear combination of signals detected by a
number of electrodes placed over the skin with a
defined geometry with the purpose of attenuating
specific spatial frequencies with respect to others.
One of the most common goals is the attenuation of
non-propagating components of either physiological
(such as the end-of-fiber effects and remote sources)
or external origin (power line interference) which
are present in the monopolar signals.
The effect of the spatial filter on the detected signal
depends on the weights assigned to each electrode,
on the geometry of the electrode configuration, on
the electrode shape and size.
3.2.1 One-dimensional Spatial Filters
The simplest and most widely used spatial filter is
the bipolar or single differential (SD), which records
the difference between the potentials detected by
two electrodes placed at a fixed distance (inter-
electrode distance, IED).
Despite the simplicity of the bipolar
configuration, the effect of its transfer function as a
spatial filter requires an accurate analysis for a
correct EMG detection and interpretation.
The differential detection system output depends
on the spatial frequency of the input and sEMG
spectral parameters obtained with different inter-
electrode distances and electrode sizes cannot be
compared (Fuglevand et al. 1992).
SD detection provides the rejection of common
mode signals; a further enhancement of spatial
selectivity can be achieved by using a more selective
spatial filter.
One of the most important classes of spatial
filters in sEMG recording is represented by the
Laplace filters. The simplest Laplacian filter is the
double differential (DD) filter that is constituted by
three equally spaced electrodes, the central electrode
weighted with a factor –2 and the others with +1.
More complex detection configurations to provide
selectivity high enough to separate single motor unit
action potentials (MUAPs) from the interference
EMG signal are represented by two-dimensional
spatial filters.
MultiChannelSurfaceEMG-DetectionandConditioning
121
3.2.2 Two-dimensional Spatial Filter
The propagation of a MUAP along the muscle fibres
results, on the skin surface, in a propagating two-
dimensional distribution spatially low-pass filtered
by the volume conductor. Since the optimal spatial
filter is the one closest to the inverse of the volume
conductor filter, two-dimensional spatial high-pass
filters have been proposed (Disselhorst-Klug et al.
1997; Reucher et al. 1987a, 1987b).
Reucher et al. proposed the normal double
differentiating filter (NDD-filter) realized by five
cross-wise arranged electrodes whereby the central
electrode is weighted with the factor –4 and the
surrounding electrodes with the factor +1. NDD-
filter improves the spatial selectivity in all directions
(Figure 4), and allows the separation of the activities
of single MUs even at maximum voluntary
contraction (Reucher 1987a, 1987b). Moreover the
2-D systems are less sensitive to fiber orientation
and the electrode placement is less critical with these
types of filters.
3.2.3 Spatial Filtering and Inter-electrode
Distance
The inter-electrode distance (IED) is regarded as one
of the most relevant properties of the sEMG
detection systems. Although it affects sEMG signal
characteristics, a high variability and a wide range of
values for IEDs (4-48 mm) can be found in literature
(Hermens 1999).
In literature it has often been suggested that a
decrease of inter-electrode distance (IED) would
increase the spatial selectivity of the detection
system. In literature no evidence can be found for
this. Roeleveld (Roeleveld et al. 1997a) performed an
experimental study investigating the effect of the IED
variation (from 6 to 84 mm) in the bipolar detection
on the contribution of motor unit potentials to the
surface EMG. The contribution of superficial and
deep motor units to the recorded SEMG signal was
found to be unrelated to IED as long as IED < 40
mm while for IED exceeding 40 mm the
contribution of deeper motor-units to SEMG is
greater than the contribution of superficial ones.
3.2.4 Detection System Orientation and
Location on the Muscles
The orientation defines the direction of a mono-
dimensional detection system with respect to the
direction of the muscle fibers; the location defines
the position of the detection system on the muscle.
The SD and DD detection systems are usually
placed in the direction of the muscle fibers but they
could also be placed in the transversal direction.
When electrodes are arranged parallel to the muscle
fibres, the filters are referred as longitudinal (LSD
and LDD) while when arranged transversally to the
muscle fibers, the filter is known as transversal
(TSD and TDD). LSD and LDD result in better
longitudinal and poorer transversal selectivity with
respect to TSD and TDD.
In literature, the most common locations of the
detection system on a muscle are the following: a) in
the center of the muscle, b) on the muscle belly; c)
somewhere between the innervation zone and one
tendon; d) on the motor point.
Although the transfer function of the spatial filter
is independent of the electrode location on the
muscle, the motor end-plates (where MUAPs
generate) and the muscle-tendon junctions (where
MUAPs extinguish) are two positions that must be
carefully considered.
The signal generated by a single fiber and
detected with a bipolar system placed along the fiber
direction, symmetrically with respect to the end-
plate of the fiber, provides a zero voltage. Since the
innervations of MUs in a muscle are concentrated in
one or two locations, this electrode location
corresponds to signals that are small, noisy, and
sensitive to small displacements between electrodes
and muscle (Masuda et al. 1985) and must be
avoided in practical applications. Similar
observations apply to muscle-tendon junction. For
this reason, the location of a pair of electrodes is
critical and should be optimized by placing the
electrodes between the innervation zone(s) and a
muscle-tendon junction. Figure 3 depicts this
situation and shows the importance of a correct
electrode placement.
EMG global variables, such as amplitude,
spectral characteristic frequencies, and estimates of
conduction velocity, are heavily altered when
electrode pairs are placed on or near the innervaton
zone (Jensen et al., 1993, Lateva et al., 1993, Roy et
al.,1986).
These considerations imply the identification of
the innervation zone(s) before the electrode pairs are
applied: this task can be achieved by means of an
electrode array.
3.2.5 Spatial Filtering and Cross-talk
The signal detected on a muscle and generated by
another active muscle is referred as cross-talk. Many
applications of sEMG require the simultaneous
NEUROTECHNIX2013-InternationalCongressonNeurotechnology,ElectronicsandInformatics
122
detection of sEMG from many muscles to evaluate,
for example, the muscular co-ordination pattern
(Koh and Grabiner 1993; Winter et al. 1994). In
these cases, it is mandatory to reduce the cross-talk
from near muscles.
Spatial high-pass filters enhance the signals
propagating along the fibers of MUs located close to
the recording electrodes and suppress the
contributions of more distantly located sources. At
some distance from the sources the contributions due
to end-of-fiber effects become predominant with
respect to the propagating components because the
latter decay in space more slowly than the first. It
could be concluded that high-pass spatial filtering
would reduce cross-talk and different kinds of
spatial filters have been applied for cross-talk
reduction. However, the experimental results have
shown that cross-talk is not reduced by spatial high-
pass filtering (van Vugt and van Dijk 2001).
This discrepancy between theory and practice is
justified by the fact that the model used for the
description of spatial filter theory considered a
potential distribution generated on the skin by a
propagating MUAP neglecting the MUAP
generation and extinction phenomena. It is known
that the generation and extinction phenomena
produce a non-propagating potential distribution on
the skin surface, and spatial filters have different
transfer functions with respect to propagating and
non-propagating signals. Simulations of the filter
responses to non-propagating potentials have shown
that most of the spatial filters do not reduce them
and in some cases they even enhance them
(Dimitrova et al. 2002; Farina et al. 2002b). Figure 2
shows an example of enhancement of end of fiber
potentials by NDD filter with respect to the
reduction obtained by SD and DD filters.
Although different techniques have been tested
to reduce crosstalk, this issue is not yet satisfactorily
solved.
Matrix column: 1 2 3 4 5
distal
proximal
SD channels
norm.val: 0.8 mV
distal
proximal
SD channels
norm.val: 0.8 mV
100 ms
30 degrees
110 degrees
RMS map
IZs
IZs
Figure 3: Example of topographical information obtained from multi-channel detection systems. sEMG signals have been
recorded during an elbow flexion using a grid of electrodes (13 rows and 5 columns with one missed electrode, 8 mm ied)
during a progressive elbow flexion. The SD SEMG signals (on the left) and the interpolated RMS distribution estimated on
one epoch 250 ms long (on the right) are reported for two elbow angles (30 degrees on the top and 120 degrees on the
bottom (0 degrees correspond to maximum extension)). The positions of the innervation zones and of the tendon are
highlighted. From the images on the right it is possible to identify the two areas (left and right) of activity corresponding to
the two heads of the biceps brachii and the different positions and shift of the innervation zones and tendon for the two
biceps brachii heads.
MultiChannelSurfaceEMG-DetectionandConditioning
123
3.3 Spatial Sampling
The surface EMG signal evolves in time and space,
and it can be described as a three dimensional signal
with one temporal and two spatial (the skin plane)
dimensions. Sampling the EMG potential
distribution by placing a number of detection
systems in different locations over the skin allows
studying how the surface EMG signal evolves in
time and space.
If a spatial filter (one-dimensional (SD or DD),
two-dimensional (NDD or other type)) is applied to
each detection point, the potential distribution is
spatially filtered and also spatially sampled.
3.3.1 Spatial Sampling in One
Dimension: the Linear Arrays
The first systems performing a spatial sampling of
sEMG were proposed by DeLuca, Merletti, and
Masuda. They proposed linear arrays of electrode
placed along the fiber direction to estimate the
velocity of propagation of action potentials, to
identify some anatomical characteristics such as the
innervation zone location and the muscle fiber
length and to investigate in detail the processes of
generation, propagation, and extinction of the
MUAPs along the muscle fibers.
Roeleveld et al. (1997b) proposed the use of two
electrode arrays located both longitudinally and in
the transversal direction with respect to the muscle
fibers to estimate MU depth.
Linear electrode arrays have also been applied to
obtain guidelines for the standardization of the
sEMG recording when a global analysis of the signal
is performed.
3.3.2 Spatial Sampling in Two Dimensions
The spatial distribution of voltage on the skin above
the muscle can be detected with a grid of electrodes
that provides two dimensional (2D) sampling in
space. If the grid covers a large part of the muscle it
provides spatial information that is largely
independent of the temporal information. The time
evolution of the voltage distribution on the skin can
be tracked by sampling in time.
Multi-channel sEMG is an interesting non-
invasive methodology to: 1) obtain muscle
anatomical information (such as the location of
innervation zones, tendon endings, and the direction
of the muscle fibers), 2) to obtain a topographical
representation of muscle activity, and 3) to
decompose the surface EMG signal into the
constituent single MU action potential trains if
electrode grids with small electrode sizes and inter-
electrode spacing (High Density EMG, HD-EMG)
are used (Zwarts et al. 2003).
3.3.3 Anatomical Information and
Topographical Representation
of Muscle Activity
The knowledge of fiber direction and innervation
zone location is relevant, for instance, for defining
the optimal locations for estimating EMG variables
in isometric and dynamic contractions. The
topographical representation of muscle activity
allows studying the regional variations in the degree
of muscle activation with time. This is particularly
important in dynamic contractions. Figure 3 shows
the distribution of single differential (SD) sEMG
RMS at two different elbow angles of isometric
contraction of the biceps brachii and demonstrates
that a single sampling point is not representative of
the spatially heterogeneous muscle activity and the
activity it detects depends in a strong way on the
geometrical factors.
The two dimensional spatial sampling obtained
using HD sEMG detection systems results in a three
dimensional signal, which can be used to reconstruct
the 3-D potential distribution if the Nyquist limits
are met in all the three dimensions.
3.3.4 Aliasing in Space
To meet the Nyquist theorem in space, the inter-
electrode distance (IED) must be smaller than a
threshold value. If we simplify the problem and we
consider only propagating signals, the IED threshold
value can be identified starting from the relationship
f
s
= f
t
/v between the spatial frequency (f
s
cycles/m),
the temporal frequency (f
t
, cycles/s or Hz) and the
signal propagation velocity (v, m/s). If we consider
400Hz as the highest temporal frequency of sEMG
and a propagation velocity of about 4m/s, the highest
spatial frequency is 100 cycles/m. For the Nyquist
theorem, the spatial sampling frequency should be
higher than 200 samples/m, which means IED less
than 5 mm. Some commonly used values of IED (10
mm or 20mm) imply aliasing in space but its
consequences on the sEMG signals have not been
investigated.
REFERENCES
R. R. Harrison; “A Versatile Integrated Circuit for the
Acquisition of Biopotentials,” In Custom Integrated
NEUROTECHNIX2013-InternationalCongressonNeurotechnology,ElectronicsandInformatics
124
Circuits Conference, 2007. CICC ’07. IEEE, pp 115–
122, September 2007.
Bernhard Fuchs, Sven Vogel, Dietmar Schroeder;
“Universal application specific integrated circuit for
bioelectric data acquisition,” Medical Engineering &
Physics, vol. 24, no. 10, pp 695–701, December 2002.
N. Van Helleputte, J.M. Tomasik,W. Galjan, A. Mora
Sanchez, D. Schroeder, W.H. Krautschneider, R.
Puers; “A flexible system-on-chip (SoC) for
biomedical signal acquisition and processing,” Sensors
and Actuators A: Physical, vol. 142, no.1, pp 361–
368, March 2008.
Chua-Chin Wang, Chi-Chun Huang, Jian-Sing Liou,
Kuan-Wen Fang; “A 140-dB CMRR Low-noise
Instrumentation Amplifier for Neural Signal Sensing,”
In Circuits and Systems, 2006. APCCAS 2006. IEEE
Asia Pacific Conference on, pp 696–699, December
2006.
R. R. Harrison, C. Charles; “A low-power low-noise
CMOS amplifier for neural recording applications,”
Solid-State Circuits, IEEE Journal of, vol. 38, no. 6,
pp 958– 965, June 2003.
T. Denison, K. Consoer, W. Santa, A.T. Avestruz, J.
Cooley, A. Kelly; “A 2µW 100nV /Hz Chopper-
Stabilized Instrumentation Amplifier for Chronic
Measurement of Neural Field Potentials,” Solid-State
Circuits, IEEE Journal of, vol. 42, no. 12, pp 2934–
2945, December 2007.
T. Denison, K. Consoer, A. Kelly, A. Hachenburg, and W.
Santa; “A 2.2µW 94nV /Hz, Chopper-Stabilized
Instrumentation Amplifier for EEG Detection in
Chronic Implants,” In Solid-State Circuits
Conference, 2007. ISSCC 2007. Digest of Technical
Papers. IEEE International, pp 162–594, February
2007.
R. Rieger, M. Schuettler, D. Pal, C. Clarke, P. Langlois, J.
Taylor, and N. Donaldson; “Very Low-Noise ENG
Amplifier System Using CMOS Technology,”
Neural Systems and Rehabilitation Engineering, IEEE
Transactions on, vol.14, no.4, pp 427– 437, December
2006.
Long Yan, Jerald Yoo, Binhee Kim, Hoi-Jun Yoo; “A
0.5µVrms 12µW patch type fabric sensor for wearable
body sensor network,” In Solid-State Circuits
Conference, 2009. A-SSCC 2009. IEEE Asian, pp
105–108, November 2009.
Hermens H., B. Freriks, R. Merletti, D. Stegeman, J. Blok,
G. Rau, C. (1999). Disselhorst-Klug, G. Hägg,
European recommendations for surface
electromyography, ISBN 90-75452-15-2, Roessingh
Research and Development, Enschede, NL.
Broman H., Bilotto G. and De Luca C.(1985). A note on
the non-invasive estimation of muscle fiber conduction
velocity. IEEE Trans. BME; 32:341-343
Disselhorst-Klug C., Silny J., Rau G.(1997). Improvement
of spatial resolution in surface-EMG: a theoretical and
experimental comparison of different spatial filters.
IEEE Trans. Biomed. Eng., vol. 44, pp. 567-574.
Farina D., Cescon C., and Merletti R.(2002a). “Influence
of anatomical, physical and detection system
parameters on surface EMG”, Biol. Cybern., vol. 86,
pp. 445-56.
Fuglevand AJ, Winter DA, Patla AE, Stashuk D. (1992).
Detection of motor unit action potentials with surface
electrodes: influence of electrode size and spacing.
Biol Cybern. 67(2):143-53.
Reucher H, Rau G, Silny J. (1987). Spatial filtering of
noninvasive multielectrode EMG: Part I-Introduction
to measuring technique and applications. IEEE Trans
Biomed Eng., 34(2):98-105
Reucher H, Silny J, Rau G. (1987). Spatial filtering of
noninvasive multielectrode EMG: Part II-Filter
performance in theory and modeling. IEEE Trans
Biomed Eng., 34(2):106-13
Roeleveld K, Stegeman DF, Vingerhoets HM, Van
Oosterom A. (1997a). Motor unit potential
contribution to surface electromyography. Acta
Physiol Scand. ,160(2):175-83.
Masuda T., Myano H., Sadoyama T. (1985). The position
of innervatoin zones in the biceps brachii investigated
by surface electromyography, IEEE Trans. BME 32:
36-42.
Jensen C., Vasseljen O., Westgaard R. (1993). The
influence of electrode position on bipolar surface
electromyogram recordings of the upper trapezius
muscle, Eur. J of Applied Physiol. 67:266-273.
Lateva Z., Dimitrova N., Dimitrov G. (1993). Effect of
recording position along a muscle fiber on surface
potential power spectrum, J. Electrom. and Kines.
3:195-204.
Roy S., De Luca C., Schneider J. (1986) Effects of
electrode location on myoelectric conduction velocity
and median frequency estimates, J. Appl. Physiol.
61;1510-1517.
Koh TJ, Grabiner MD. (1993). Evaluation of methods to
minimize cross talk in surface electromyography. J
Biomech. 26 Suppl 1:151-7.
Winter D. A., Fuglevand AJ, Archer SE., Crosstalk in
surface electromyography: Theoretical and practical
estimates. J Electromyogr Kinesiol. 1994;4(1):15-26.
van Vugt J.P., van Dijk J.G. (2001). A convenient method
to reduce crosstalk in surface EMG. Clin.
Neurophysiol., vol. 112, pp. 583-92.
Dimitrova N.A., Dimitrov G.V., Nikitin O.A.(2002).
“Neither high-pass filtering nor mathematical
differentiation of the EMG signals can considerably
reduce cross-talk”, Journ. Electromyogr. Kinesiol,
(4):235-46.
Farina D., Merletti R., Indino B., Nazzaro M., and Pozzo
M.(2002b). Cross-talk between knee extensor muscles.
Experimental and model results. Muscle Nerve, vol.
26, pp. 681-95.
Roeleveld K, Stegeman DF, Vingerhoets HM, Van
Oosterom A. (1997b). The motor unit potential
distribution over the skin surface and its use in
estimating the motor unit location. Acta Physiol
Scand., 161(4):465-72.
Zwarts MJ, Stegeman DF. (2003). Multichannel surface
EMG: basic aspects and clinical utility. Muscle Nerve.,
28(1):1–17.
MultiChannelSurfaceEMG-DetectionandConditioning
125