Transcript ELECTROMYOGRAPHY Mario Lamontagne PhD Scope of this presentation      Introduction Background Recording Technique Analysis of the EMG signal Applications APA 6903 INTRODUCTION  The electromyographic (EMG) signal offers a great source of.

ELECTROMYOGRAPHY
Mario Lamontagne PhD
1
Scope of this presentation





Introduction
Background
Recording Technique
Analysis of the EMG signal
Applications
APA 6903
2
INTRODUCTION
 The electromyographic (EMG) signal offers a
great source of information to both clinicians
and researchers
 EMG can be used to detect gait or joints
pathologies, to assess a rehabilitation program,
to measure the functionality of sport equipment
and to implement an effective biofeedback
therapy.
APA 6903
3
INTRODUCTION
Surface EMG is also widely used in an effort to
understand a number of research issues:
• Muscles coordination around a joint
• Relationship between muscular force and muscle
electrical activity
• Neuromuscular adaptations after joint surgery
following a rehabilitation program.
APA 6903
4
Background
Nervous System
The muscle unit action potential detected by
electrodes in the muscle tissue or on the surface
of the skin.
Central nervous system (CNS) activity initiates a
depolarisation in the motoneuron.
APA 6903
5
BACKGROUND
CNS
APA 6903
6
BACKGROUND
CNS
APA 6903
7
BACKGROUND
CNS
APA 6903
8
BACKGROUND
Motoneuron
APA 6903
9
BACKGROUND
SYNAPSE
APA 6903
10
MOTOR UNIT
A single axon leading to
a muscle is responsible
for the innervation of as
few as 3 or as many as
2000 individual muscle
fibres.
A neuron and the
muscle fibres are
referred to as motor unit
APA 6903
11
MOTOR UNIT
 One a-motoneuron plus all the muscle fibers it
enervates
 Innervation ratio varies with number of fibers
per motor unit (large leg muscles have many
fibers per motoneuron for stronger responses,
facial and eye muscles have few fibers and
therefore permit finer movements but weaker
contractions)
 All-or-none rule – once a motoneuron fires all
its muscle fibers must fire
 Graded muscle responses occur because of
orderly recruitment of motor units, i.e., lowest
threshold motor units fire first followed by next
lowest threshold. Highest threshold and
strongest motor units fire last.
APA 6903
12
MU ACTION POTENTIAL
 When an action potential reaches the muscle at
localized motor points (AKA innervation points)
sarcoplasmic reticulum and t-tubule system carries the
message to all parts of the muscle fiber
 A rapid electrochemical wave of depolarization travels
from the motor point causing the muscle to contract
 Followed by a slower wave of repolarization and a brief
refractory period when it cannot contract
 The wave of depolarization can be sensed by an
electrode and is called the electromyogram (EMG). The
repolarization wave is too small to detect.
APA 6903
13
MU ACTION POTENTIAL
A neuron and the muscle
fibers are referred to as motor
unit (MU)
APA 6903
14
MU ACTION POTENTIAL
The nerve impulse is transmitted in a
nerve axon as schematically shown down
below
Triphasic Signal
A
B
Voltage
+
APA 6903
15
MU ACTION POTENTIAL
A dipole +
- is moving along a
volume conductor. A differential amplifier
records the difference between the potentials
at point A and B on the conductor.
A
B
Voltage
+
Triphasic Signal
APA 6903
16
MU ACTION POTENTIAL
The dipole is moving along the conductor. The
potential A is getting more negative.
+
Triphasic Signal
A
B
Voltage
-
APA 6903
17
MU ACTION POTENTIAL
More the dipole is moving between the
potentials more the signal is positive
A
B
Voltage
+
Triphasic Signal
APA 6903
18
MU ACTION POTENTIAL
Finally, the connector B registers the positive
end of the dipole and the connector A is returning
to zero. The result of the amplification becomes
negative
A
B
Voltage
+
Triphasic Signal
APA 6903
19
MU ACTION POTENTIAL
The triphasic curve has some similarity with an
action potential which passes through a nerve
axon.
A
B
Voltage
+
Triphasic Signal
APA 6903
20
MU ACTION POTENTIAL
Number of MU varies with the type and
function of muscles.
Muscles Number of muscle’s fibers/Neuron
Platysmus
Long Digital Flexor
Tibialis Anterior
Gastrocnemius
25
95
609
1775
APA 6903
21
MU ACTION POTENTIAL
Motor Unit Recruitment
Once an action potential reaches a muscle fiber, it
propagates proximally and distally. This is called
motor action potential (MAP).
A motor unit action potential (MUAP) is
spatiotemporal summation of MAPs for an entire
MU.
APA 6903
22
MU ACTION POTENTIAL
An EMG signal is the algebraic summation of
many repetitive sequences of MUAPs for all
active motor units in the vicinity of the recording
electrodes
MUAP1
MUAP2
MUAP3
MUAP4
APA 6903
23
MU ACTION POTENTIAL
Muscle tension
MU Recruitment
The order of MU
recruitment is according
to their sizes. The
smaller ones are active
first and the bigger ones
are active last.
MU 4
MU 3
MU 2
MU 1
MU 1
MU
2
MU 3
MU 4
APA 6903
24
MU ACTION POTENTIAL
MUAP vs. Force
– For a voluntary contraction, muscle’s force
depends on the number of MU and the
frequency of activation
– Muscle’s force is proportional of the crosssectional area of the active muscle fibers.
– Muscle force during isometric action is
around 30 N/cm2
APA 6903
25
Recording Techniques
The
preamplifier
increases
the amplitude
A differential
wide variety
of electrodes
are available
to
of the difference signal between each of detecting
measure
the
electrode
and
theelectrical
common muscle
ground. output
The advantage of
microelectrode
and needle
electrode
the •differential
preamplifier
is to
improve(not
thepractical
signal- for
to-noise
ratio ofstudies)
the measurement.
movement
• Surface electrodes (SE) and Intramuscular wire
electrodes (IWE) are commonly used in movement
studies
APA 6903
26
Recording Techniques
Differential amplifier
Ground electrode
Cable
Leads
Electrodes
APA 6903
27
27
Recording Techniques
FREQUENCY
EMG Signal
RESPONSE
Detection Summary
dynamic
range
is the
amplification
range ofrather
an electrical
Bipolar
electrodes
(active
electrode
than
• •electrode
pairs
inlinear
parallel
with
fibres
•frequency
responses
of
amplifier
and
recording
systems
device
passive
electrodes
•
midway
between
motor
point
myotendonous
must
match
frequency
spectrum
ofand
the
EMG
signal
ability
of computers
a differentialuse
amplifier
to perform
 typical
A/D
either +/–10
V oraccurate
+/–5 V
•since
“raw”
surface
have
a frequency
spectrum
from
(belly
of EMGs
muscle)
subtractions
(attenuate
common
mode
noise)
Distance
between
electrodes
10
to 20 mm
 •junction
amplifiers
usually
have
+/–10
V or more,
oscilloscopes
and
20• toapproximately
500
Hz,measured
amplifiers
recording
must
have
usually
(y=20
log10ifx)system
multimeters
(+/–200
V decibels
orand
more)
apart
2 in
cm
apart,
better
electrodes
are
same
frequency
response
or
EMG
amplifiers
should be
>80wider
dB (i.e., VS/N of 10000:1,
 tape
or minidisk
recorders
have
+/–1.25
fixed
together
to
reduce
relative
movements
• Bandwidth
of
20-500
Hz
•since
relative
movements
electrodes
cause
low
thesignals
difference
between
twoof
identical
1
V sine
waves
 EMG
must
be amplified
usually
1000x
or
more but
notfrequency
• leads
be
immobilized
to (signal
skin overload)
“artifacts,”
high-pass
filtering
is necessary
(10 to 20 Hz
because
0.1 mV)
toowould
highshould
to
amplifier
“saturation”
• CMRR
greater
than
100aredB
most
modern
EMGresolution
amplifiers
>100
dB (too few
cutoff)
if too
low,
numerical
comprised
• ground
electrode
placed will
over
electrically
neutral area
significant
digits,
from
12
bit toonly
8 bit or
less)frequencies as high
•Since
surface
EMG
signals
have
•usual
Noise
less
than
2mV
bone
as 500 Hz, low-pass filtering is desirable (500 to 1000 Hz
• •N.B.
there should
beon
only
ground
electrode
Electrode
located
theone
midline
of the
muscleper
cutoff)
bellyuse a “band-pass filter” (20 to 500 Hz)
person
•therefore
APA 6903
28
Recording Techniques
 impedance is the combination of electrical
resistance and capacitance
 all devices must have a high input
impedance to prevent “loading” of the input
signal
 if loading occurs the signal strength is
reduced
 typically amplifiers have a 1 MW input
resistance, EMG amplifiers need 10 MW or
greater
 10 GW amplifiers need no skin preparation
APA 6903
29
Recording Techniques
 Dry skin provides insulation from
static electricity, 9V battery
discharge etc.
 unprepared skin resistance can be
2 MW or greater except when wet
or “sweaty”
 if using electrodes with < 1 GW
input resistances, skin resistance
should be reduced to < 100 kW
 Vinput = [ Rinput/(Rinput + Rskin) ] VEMG
APA 6903
30
Recording Techniques
 telemetry has less encumbrance and permits
greater movement space
 radio telemetry can be affected by interference
and external radio sources
 radio telemetry may have limited range due to
legislation (e.g., IC, FCC)
 cable telemetry (e.g., Delsys) can reduce
interference from electrical sources
 telemetry more expensive than directly wired
systems
 telemetry has limited bandwidth (more channels
reduces frequency bandwidth)
APA 6903
31
Analysis of the EMG signal
In the time domain:
RAW
• the root-mean squared (RMS)
value or also called Linear
Envelop)
• the average rectified value
Onset
Peak
• Both are appropriate and
provide useful measurements of
the signal amplitude
• Muscle onset (time)
• Peak amplitude of RMS
APA 6903
32
EMG: In the time domain
 same as taking the absolute value of
the raw signal
 mainly used as an intermediate step
before another process (e.g.,
averaging, linear envelope and
integration)
 can be done electronically and in
real-time
APA 6903
33
EMG: In the time domain
Averaged EMG
simple to compute
can be done in real-time
averaged EMG is a “moving average” of
a full-wave rectified EMG
must select an appropriate “window
width” that changes with sampling rate
easy for determining levels of
contraction
APA 6903
34
EMG: In the time domain
Linear Envelope EMG
 requires two-step process: full-wave rectification
followed by low-pass filter (4-10 Hz cutoff)
 can be done electronically (but adds a delay)
 reduces frequency content of EMG and thus lowers
sampling rates (e.g., 100 Hz) and memory storage
 easy to interpret and to detect onset of activity
 can be ensemble-averaged to obtain patterns
 difficult to detect artifacts
 useful as a control (myoelectric) signal
APA 6903
35
EMG: In the time domain
Ensemble-Averaged EMG
usually applied to cyclic activities and linear envelope EMGs
requires means for identifying start of cycle or start and
end of activity
• foot switches or force platforms can be used for gait
studies
• microswitches, optoelectric or electromagnetic sensors
for other activities
• can also use a threshold detector of a kinematic or
EMG channel
each “cycle” of activity must be time normalized
APA 6903
36
EMG: In the time domain
Ensemble-Averaged EMG
amplitude normalization is often done
• to maximal voluntary contraction (MVC)
• to submaximal isometric contraction
• to EMG of a functional activity
mean and standard deviations for each proportion of cycle
are computed
mean and s.d. or 95% confidence interval may be
presented to show representative contraction during activity
cycle
easier to make comparisons among subjects
“grand” ensemble-averages (average of averages) for
comparisons among several experimental conditions
APA 6903
37
EMG: In the time domain
Integrated EMG (iEMG)
 important for quantitative EMG relationships (EMG vs.
force, EMG vs. work)
 best measure of the total muscular effort
 useful for quantifying activity for ergonomic research
 various methods:
• mathematical integration (area under absolute
values of EMG time series)
• root-mean-square (RMS) times duration is similar
but does not require taking absolute values
• electronically (see next page)
APA 6903
38
Other Techniques
 auto-correlation (correlate signal with
itself shifted in time, gives signal
characteristics)
 cross-correlation (correlate signal with
another EMG signal, tests for crosstalk)
 zero-crossings (the more crossings the
greater the level of recruitment)
 peak counting (number of peaks above
a threshold)
 single motor unit detection
 double differential amplifier (velocity of
propagation)
APA 6903
39
ECG Crosstalk
 ECG crosstalk occurs when recording
near the heart (ECG has higher
voltages then EMG)
 EEG crosstalk when near scalp (rare)
 difficult to resolve
• use right side of body (away from heart)
• move electrodes as far away from heart as
possible
• “signal averaging” (average many trials)
• indwelling electrodes
APA 6903
40
Muscle Crosstalk
 one muscle’s EMG is picked up by
another muscle’s electrodes
 can be reduced by careful electrode
positioning
 can be determined by cross-correlation
 difficult to distinguish crosstalk from
synergistic contractions
 biarticular muscles have “extra” bursts
of activity compared to monoarticular
muscles (if so crosstalk is not a
problem)
APA 6903
41
Analysis of the EMG signal
In the Frequency domain:
• Spectral Density
–Median Frequency
–Mean Frequency
• Wavelet
This represents the
frequency contents of
EMG signal.
APA 6903
42
Analysis of the EMG signal
Frequency Spectrum
 useful for determining onset of muscle fatigue
 mean or median frequency of spectrum in
unfatigued muscle is usually between 50-80 Hz
 as fatigue progresses fast-twitch fibres drop
out, shifting frequency spectrum to left
(lowering mean and median frequencies)
 mean frequency is less variable and therefore
is better than median
 useful for detecting neural abnormalities
APA 6903
43
Interpretation of the EMG signal
 EMG is a tool not without its hidden
weaknesses
 These problems have the potential to
mask any benefit obtained from the
recorded information.
APA 6903
44
Anecdotal Demonstration
Adrian R. M. Upton conducted
an anecdotal demonstration of
the difficulty of documenting
brain death by placing EEG
electrodes in an upside-down
bowl of lime Jell-O (reported in
The New York Times, March 6,
1976, p. 50).
APA 6903
45
Interpretation of EMG
 As with EEG traces, the interpretation of the
recorded EMG should be conducted with care.
 However, with proper use, the surface
electromyogram is a powerful and effective tool
for both clinical evaluation and research.
APA 6903
46
Applications in Orthopaedics
Recent
technological
Most
of the
applicationsdevelopment
of sEMG and in sEMG
moved
research
from
imEMG
are based
on:the laboratory to the
field
applications.
• Muscle
activation and timing
• Muscle contraction profile
• Muscle strength of contraction
• Muscle fatigue.
APA 6903
47
APPLICATIONS IN SPORT MEDICINE
Muscle Activation and Timing1
Objective:
 Examine the neuromuscular response to
functional knee bracing relative to anterior tibial
translations.
Design:
 During randomized brace conditions,
electromyographic data with simultaneous
skeletal tibiofemoral kinematics and GRF were
recorded from four ACL deficient subjects to
investigate the effect of the functional brace
during activity.
Ramsey, D. K., Lamontagne, M., Wretenberg, P., Valentin, A., Engström, B., & Németh, G.
(2003). Electromyographic and biomechanics analysis of anterior cruciate ligament deficiency and
functional knee bracing. Clin Biomech (Bristol, Avon). 2003 Jan;18(1):28-34
APA 6903
48
Rectus femoris (mV)
APPLICATIONS IN SPORT MEDICINE
1
Muscle Activation and Timing1
A1
A2
A3
A1
A2
A3
Semitendinosus (mV)
0
1
A1
A2
A3
0
Gastrocnemius (mV)
 Kinematic and kinetic measure-ments
were synchronously recorded with the
EMG signal. The EMG data from the
RF, S, BF, and LG were integrated for
each subject in three separate time
periods: 250 ms preceding foot-strike
and two consecutive 125 ms time
intervals following foot-strike.
1
1
A1
A2
A3
Vertical & posterior
ground reaction force (BW)
Methods:
Biceps femoris (mV)
0
0
4
0
Time (s)
APA 6903
49
APPLICATIONS IN SPORT MEDICINE
Muscle Activation and Timing1
Rectus femoris
Biceps femoris
0.05
0.05
Group means
0.01
0.00
Group means
Rectus femoris
femoris
Biceps
0.02
 With
0.05 brace, ST activity
No brace
significantly
decreased 17%
0.04
Brace
21%
prior to footstrike*
0.03
 whereas BF significantly
0.02
decreased
44% during A2,
0.01
(P<0.05).
 RF0.00activity
significantly
1
2
3
increased 21%
A2 (P<0.05).
Timein
interval
 No consistent
reductions in
Semitendinosus
0.05
anterior
translations were
No brace
0.04
Brace
17%
evident.
Group means
0.05
0.04
0.03
0.02
0.02
0.01
0.05
0.05
No brace
3
No brace
Time interval Brace
0.04 SemitendinosusBrace
0.04
1
No brace
Brace
0.02
0.02
2
3
0.04
0.03
Lateral gastrocnemius
21%
No brace
44%
Brace
*
*
0.02
0.01
1
0.01
0.01
0.00
0.00
2
0.00
3
1
Time interval
2
3
Time interval
11
22
33
Time interval
interval
Time
Semitendinosus
Lateral
gastrocnemius
0.05
0.05
0.04
0.04
17%
*
No brace
No brace
Brace
Brace
0.03
0.03
0.03
APA 6903
0.02
1
Time interval
0.05
0.03
0.03
*17%
Group
Groupmeans
means
Group means
*
0.00
2
0.01
0.00
*44%
0.03
Group means
Rectus femoris
0.04
* 21%
Group means
Results:
0.03
Group means
No brace
Brace
0.04
No brace
Brace
0.02
0.02
50
APPLICATIONS IN SPORT MEDICINE
Muscle Activation and Timing1
Conclusion:
 Joint stability may result from proprioceptive feedback
rather than the mechanical stabilising effect of the
brace. As a result of bracing, we observed decreased
S and BF activity but increased RF activity. We
suggest increased afferent input from knee
proprioceptors and brace-skin-bone interface modifies
EMG activity.
APA 6903
51
Applications in Orthopaedics
Gender Difference for a cut motion
 Male and Female elite football players
 Control speed
 Cue given at 1.2m from the FP
See EMG Data
APA
6903
2
1
C
52
Applications in Orthopaedics
Muscle Fatigue1
Surface EMG can be used as muscle
fatigue indicator
We investigated possible differences in muscle
fatigue and recovery of knee flexor and extensor
muscles in patients with a deficient anterior cruciate
ligament compared with patients with a normal
anterior cruciate ligament.
Tho, K., Németh, G., Lamontagne, M., & Eriksson, E. (1997). Electromyographic Analysis of Muscle Fatigue in
Anterior Cruciate Ligament Deficient Knees. Clinical Orthopaedics & Related Research(340), 142-151.
APA 6903
53
Applications in Orthopaedics
Muscle Fatigue1


SEMG of 15 patients with ACL deficiency was
measured while the muscles were under 80% of MVC
for 60 s and remeasured after 1, 2, 3, and 5 minutes
of rest
Knee joint was at 45 degrees of flexion.
APA 6903
54
Applications in Orthopaedics
Muscle Fatigue1
C oefficient of MF change and am plitude increase du rin g 80% M VC fo r 60s
(m od ified from Th o et al. 1997
).
Findings showed that:
• Conditions
First 60 s of contraction
Injured
Knee
Knee
> all muscles
recorded
significantlyNormal
decreased
Muscles
MPF
Coe fficient of
Amplitud e
C ha ng e
Coe fficient of
Amplitud e
C ha ng e
MF (SD)
(SD)
(% )
MF (SD)
(SD)
(% )
>
an
increase
in
LEEMG
amplitude.
Va stus Me dialis
R ectus•fem oris
Rate of decrease of MPF was significantly greater in
Vastus Late ra lis
the injured quadriceps and normal hamstrings.
Med ial Ha mstrings
• All muscles
recovered
to the
initial
MPF 228
level
after
La tera l H a mstrin gs
-0.159 (0.155)
204 (178)
80
-0.222 (0.152)
(269)
71
1 mius
min of-0.105
rest
in-0.208**
the
injured
Med . Gastrocne
(0.132)but two
62 (63) muscles
40
(0.146)
52 (53)and
33
L at. Gastrocne
mius
-0.151
(0.118)
88 (72)
63
-0.187 (0.139)
54 (61) power
28
normal
limb
recorded
an overshoot
of mean
frequency during the recovery phase.
-0. 096 (0.073)
125 (172)
42
-0.069 (0.064)
132 (95)
76
-0.136* (0.086)
64 (119)
20
-0. 100 (0.046)
60 (112)
23
-0. 105* (0.087)
89 (141)
29
-0. 054 (0.073)
165 (184)
67
-0. 207 (0.124)
125 (132)
58
-0.266* (0.112)
119 (149)
49
* : p < 0.05 (paire d t-test)
** : p < 0.01 (paire d t-test)
APA 6903
55
Applications in Orthopaedics
Muscle Fatigue1
The findings confirmed
• the fatigue state in all the muscles, suggest
recruitment of more Type II fibers as the muscle
fatigue
• show the physiological adaptation of the quadriceps
and hamstrings to ACL deficiency.
• dissociation between low intramuscular pH and
mean power frequency during the recovery phase.
APA 6903
56
Applications in Orthopaedics
Muscle Fatigue2
We investigated the possible influence of wearing
functional knee braces on various factors of muscle
fatigue.
• Measured parameters were; MVC, Peak Velocity
(PK), power and number of repetition to muscle
fatigue during isokinetic exercise, and also muscle
fatigue during 50s isometric contraction
Lamontagne, M. & Sabagh-Yazdi, F. (1999). The Influence of Functional Knee Braces on Muscle Fatigue.
Paper presented at the XVIth of the International Society of Biomechanics, Calgary, Canada.
APA 6903
57
Applications in Orthopaedics
Muscle Fatigue2
 Two groups of healthy and ACL-deficient knee joint
subjects with an average age of 28.8 years and 26,6
years respectively volunteered to this study.
 All tests were performed on an isokinetic device (KinCom 500H) while the EMG signal was collected at
1000 Hz for six muscles (RF), (VL), (VM), (G), (MH)
and (LH).
APA 6903
58
Applications in Orthopaedics
Muscle Fatigue2
Analysis of EMG data revealed that
• no significant differences were obtained for the EMG
amplitude or the integral of the linear envelope EMG
between the groups and conditions
• During the 50s isometric exercise at 80% MVC, the
fatigue state is represented by decline of MF value of
EMG signal greater than 10 Hz
• Muscle fatigue state was obtained in all muscles
APA 6903
59
Applications in Orthopaedics
Muscle Fatigue2
Average
percentage of decline
of the median
frequency.
• Percentage
of decline
of MF
in the
Gastrocnemius was
significantly different between the groups (p<0.05).
ACL
Healthy
• Percentage of decline of median frequency in VM and G of
Muscles
VL
RF
LH
VL
RF
VM
MH
LM
G° MH
G°
ACL group
andVM
VL and
G of healthy
group
was
found
Braced statistically
9.1 27.6different
14.8
1.8 35.0
27.3
18.4 conditions.
24.9
12.3
1.7
39.3
34.5
(p<0.05)
between
• the outcomes
showed
correlation
between
the
Unbraced
12.0
22.4
9.0*
10.6*a high
43.4
24.0
8.9*
21.2
16.4
9.5*
48.0
28.5
subjective perception of fatigue and percentage of decline of
* : significantly
conditionsfor
( VL and
p <RF
0.05) muscles during the brace
thedifference
MF (rbetween
= 0.64)
condition.
° :significantly difference between groups (
p < 0.05)
• All other muscles showed very low correlation.
APA 6903
60
CONCLUSION
Factors like signal reliability, muscle synergy,
mechanisms of proprioception, muscle fatigue
mechanisms have been a great deal of interest in
movement studies but these topics certainly need
more research in order to understand muscle
function and adaptation for ordinary people and
athletes.
Lamontagne, M. (2000). Electromyography in sport medicine (Chapter 4). In
Rehabilitation of Sports Injuries (Ed. G. Puddu, A. Giombini, A. Selvanetti ),
Springer-Verlag, Berlin, Heidelberg, New York
APA 6903
61
Partly funded by:
Natural Sciences and Engineering Council of Canada
and
Let People Move
APA 6903
62