EMG
Lecture 1 – Basics of EMG
EMG amplitude estimation
- EMG is a reflection of action potentials
- AP take place at the level of the muscle fibers (at
membranes of the cells)
- AP are an effect of what happens in the CNS
- Result of what happens in the motor cortex, information is
conducted by the spinal cord, MUs are activated,
information travels to the motor end plate and AP are
generated
- AP is a condition for force generation
- Change in potential difference between inside and outside
of the muscle fiber
o It will be negative outside of the fiber (AP reverses the potential difference)
o Calcium will be released from SR and allows binding of actin and myosin
EMG is an indirect effect from the action potential
- Measure it at the skin surface
- The signal has to be conducted over subcutaneous layers and skin
- The signal is pretty noisy so we have to do some analysis to get signal characteristics
EMG-amplitude estimation
- Average of the EMG signal is 0
- Express the strength as the deviation from the mean zero
- Root Mean Square (RMS) mV→ standard deviation (in statistics you use n-1,
now you use n)
- Averaged Rectified Value (ARV) mV→ mean of absolute value
- Integrated Rectified EMG (iEMG) mV*s→ integration of sum of absolute EMG
(not divided by n)
o Note that unit for iEMG in ABC is wrong
- peak-peak amplitude mV→ difference between highest positive and negative value
o not used for voluntary activity, but used for electrical stimulation or reflexes
Motor Units (MUs), MU action potentials and MUAP trains
The motor unit
- Functional building block of the muscle (it functions as a unit)
- They are either all active or non of them is active
Recording volume/pick-up area surface EMG
- Pick up information from a lot of MU, the signal is dominated from a small area of the muscle
, o Typical pick-up area surface-EMG perpendicular to fibers
o Approx. 2x2x1 cm muscle volume (also loose a 1cm because of subcutaneous fat)
- Simplified representation, you don’t only measure that area, but the signal drops off gradually
with distance and rate of decline depends on many factors
Muscle (10-300 MUs)
- Typical pick-up are surface EMG perpendicular to fibers →
o ~2 x ~1 cm 1cm = ~200 fibers 400x200 = 80.000 fibers
~1000 fibers/MU ~80 MUs in pick-up area
- You only have a sample of the muscle and not the whole muscle
- Motor units are intermingled
o Hard to record from a single motor unit, because there are always other MU nearby
- Surface EMG electrodes pick up information from many MUs simultaneously
Sequence of motor unit action potential (MUAPs)
- MUAP train
- Computer simulation: EMG of 1 motor unit (14 firings/s (average))
o Not constant, there is always variation
o Firing rate is rarely below 5 firings/s → then it will stop firing
- Summing of different MUs
o EMG summation of MUAPs within pick-up area
Number of active MUs
- Firing rate = 20 pulses per second
- Signal strength has a relation with the number of active MUs
- More motor unit active the more force produced
Increasing firing rate
- 20MUs
- Signal amplitude increases with firing rate
- Higher firing rate causes more force
Muscle force regulation
Measuring motor unit behavior
- Intra-muscular EMG
o Concentric/monopolar needle (close to MU) → clinical practice
▪ Note scale difference relative to fibers
o Fine-wire → research
- Very low inter-electrode distance
- Allows detection of a single or a few MUs
- Decomposition of EMG into MUAP trains
MU behavior: recruitment and firing rate
- 25-30 pps is about the highest firing rate of motor units
- 6-10 pps is about the lowest firing rate of motor units
- “onion skin” principle (later recruited MU lower firing rate)
- Colored lines = 1 MU (firing rate of single MU)
- Black line = muscle force
Hennemann’s “size principle”
- Increasing force gives more MU activated (recruitment)
o Recruitment is based on the size of the MU
, o Small motor neurons are easily activated, large motor neuros have a higher threshold
and are therefore later activated, but caused a higher force production
- Type I fibers → few fibers, slow, fatigue resistant (bottom/left, always activated first)
- Type II fibers → many fibers, fast, fatigable (right, activated later)
Recruitment and rate coding
- Hydraulic analogue of central drive to the alpha motoneurons
o Opening the tap = increased central drive
o More MUs produce output (recruitment)
o And each MU increases its output (firing rate)
o Earlier recruited MUs have larger output (higher firing rate)
o In the end all will be activated and reach their maximal firing rate, which are
quite similar to each other
Firing rate and force
- Increasing firing rate causes increase in force
- Beyond 30Hz an increase of stimulation frequency does not increase much
force
Why are we here?
- Neural drive determines MU recruitment and firing rate
- Muscle force and EMG amplitude are determined by recruitment and firing
rate
So:
- EMG is expected to reflect neural drive from the CNS
- EMG is expected to reflect muscle force
And
- EMG is expected to be muscle specific
Variability of MU firing
α-motoneuron action potential generation
- Incoming AP generate Excitatory Post-Synaptic Potentials at α-motoneuron
- Subsequent EPSPs summate
- When summed EPSPs exceed threshold AP occurs
- AHP = after-hyper-polarization
- At end of AHP firing is possible but less likely than at rest, because it is more
negative than the resting potential
time of firing depends on CNS drive, AHP and synaptic noise
, α-motoneuron firing
E = descending drive in noise units (1 = 1SD of noise) relative to firing
threshold
Interval = 1/firing rate
- High drive → more variation
- Lower drive → distribution is wider
- Effect of noise is limited in higher drive than at lower drive
- The lower the drive the more irregular the firing rate
Summary MU firing
- The AHP of the α-motoneuron AP limits the maximal motor unit firing rate
- Synaptic noise causes non-regular firing of motor units
- Each motor unit has a slightly different AHP and different synaptic noise, consequently motor
units fire largely independently
Variability of the EMG signal
Independent and random firing of MUs
- Peaks and valleys of AP may add up (green)
- At other instants peaks or valleys cancel (red)
Random negative and positive superimposition (phase cancellation)
- Neurophysiology has to deal with bi- or multiphasic signals (e.g. motor unit action potentials)
So
- Their mean value tends to zero, also after summation
- In quantification of the data this poses a problem
Summation of MUAPs to EMG: model 1
- 2 seconds force profile
- EMG units (MUAPs) all have positive size = 1
- 1 MUAP adds 1 force unit
- No random effects
Perfect match, but irrealistic because MUs fire irregularly
Summation of MUAPs to EMG: model 2
- Same force profile (2sec)
- MUAPs now fire irregularly, still with mean size = 1
- 1 MUAP adds 1 force unit
The force can still be estimated despite the random firing, but statistical precision is lower
Still irrealistic because EMG contributions have zero mean
Lecture 1 – Basics of EMG
EMG amplitude estimation
- EMG is a reflection of action potentials
- AP take place at the level of the muscle fibers (at
membranes of the cells)
- AP are an effect of what happens in the CNS
- Result of what happens in the motor cortex, information is
conducted by the spinal cord, MUs are activated,
information travels to the motor end plate and AP are
generated
- AP is a condition for force generation
- Change in potential difference between inside and outside
of the muscle fiber
o It will be negative outside of the fiber (AP reverses the potential difference)
o Calcium will be released from SR and allows binding of actin and myosin
EMG is an indirect effect from the action potential
- Measure it at the skin surface
- The signal has to be conducted over subcutaneous layers and skin
- The signal is pretty noisy so we have to do some analysis to get signal characteristics
EMG-amplitude estimation
- Average of the EMG signal is 0
- Express the strength as the deviation from the mean zero
- Root Mean Square (RMS) mV→ standard deviation (in statistics you use n-1,
now you use n)
- Averaged Rectified Value (ARV) mV→ mean of absolute value
- Integrated Rectified EMG (iEMG) mV*s→ integration of sum of absolute EMG
(not divided by n)
o Note that unit for iEMG in ABC is wrong
- peak-peak amplitude mV→ difference between highest positive and negative value
o not used for voluntary activity, but used for electrical stimulation or reflexes
Motor Units (MUs), MU action potentials and MUAP trains
The motor unit
- Functional building block of the muscle (it functions as a unit)
- They are either all active or non of them is active
Recording volume/pick-up area surface EMG
- Pick up information from a lot of MU, the signal is dominated from a small area of the muscle
, o Typical pick-up area surface-EMG perpendicular to fibers
o Approx. 2x2x1 cm muscle volume (also loose a 1cm because of subcutaneous fat)
- Simplified representation, you don’t only measure that area, but the signal drops off gradually
with distance and rate of decline depends on many factors
Muscle (10-300 MUs)
- Typical pick-up are surface EMG perpendicular to fibers →
o ~2 x ~1 cm 1cm = ~200 fibers 400x200 = 80.000 fibers
~1000 fibers/MU ~80 MUs in pick-up area
- You only have a sample of the muscle and not the whole muscle
- Motor units are intermingled
o Hard to record from a single motor unit, because there are always other MU nearby
- Surface EMG electrodes pick up information from many MUs simultaneously
Sequence of motor unit action potential (MUAPs)
- MUAP train
- Computer simulation: EMG of 1 motor unit (14 firings/s (average))
o Not constant, there is always variation
o Firing rate is rarely below 5 firings/s → then it will stop firing
- Summing of different MUs
o EMG summation of MUAPs within pick-up area
Number of active MUs
- Firing rate = 20 pulses per second
- Signal strength has a relation with the number of active MUs
- More motor unit active the more force produced
Increasing firing rate
- 20MUs
- Signal amplitude increases with firing rate
- Higher firing rate causes more force
Muscle force regulation
Measuring motor unit behavior
- Intra-muscular EMG
o Concentric/monopolar needle (close to MU) → clinical practice
▪ Note scale difference relative to fibers
o Fine-wire → research
- Very low inter-electrode distance
- Allows detection of a single or a few MUs
- Decomposition of EMG into MUAP trains
MU behavior: recruitment and firing rate
- 25-30 pps is about the highest firing rate of motor units
- 6-10 pps is about the lowest firing rate of motor units
- “onion skin” principle (later recruited MU lower firing rate)
- Colored lines = 1 MU (firing rate of single MU)
- Black line = muscle force
Hennemann’s “size principle”
- Increasing force gives more MU activated (recruitment)
o Recruitment is based on the size of the MU
, o Small motor neurons are easily activated, large motor neuros have a higher threshold
and are therefore later activated, but caused a higher force production
- Type I fibers → few fibers, slow, fatigue resistant (bottom/left, always activated first)
- Type II fibers → many fibers, fast, fatigable (right, activated later)
Recruitment and rate coding
- Hydraulic analogue of central drive to the alpha motoneurons
o Opening the tap = increased central drive
o More MUs produce output (recruitment)
o And each MU increases its output (firing rate)
o Earlier recruited MUs have larger output (higher firing rate)
o In the end all will be activated and reach their maximal firing rate, which are
quite similar to each other
Firing rate and force
- Increasing firing rate causes increase in force
- Beyond 30Hz an increase of stimulation frequency does not increase much
force
Why are we here?
- Neural drive determines MU recruitment and firing rate
- Muscle force and EMG amplitude are determined by recruitment and firing
rate
So:
- EMG is expected to reflect neural drive from the CNS
- EMG is expected to reflect muscle force
And
- EMG is expected to be muscle specific
Variability of MU firing
α-motoneuron action potential generation
- Incoming AP generate Excitatory Post-Synaptic Potentials at α-motoneuron
- Subsequent EPSPs summate
- When summed EPSPs exceed threshold AP occurs
- AHP = after-hyper-polarization
- At end of AHP firing is possible but less likely than at rest, because it is more
negative than the resting potential
time of firing depends on CNS drive, AHP and synaptic noise
, α-motoneuron firing
E = descending drive in noise units (1 = 1SD of noise) relative to firing
threshold
Interval = 1/firing rate
- High drive → more variation
- Lower drive → distribution is wider
- Effect of noise is limited in higher drive than at lower drive
- The lower the drive the more irregular the firing rate
Summary MU firing
- The AHP of the α-motoneuron AP limits the maximal motor unit firing rate
- Synaptic noise causes non-regular firing of motor units
- Each motor unit has a slightly different AHP and different synaptic noise, consequently motor
units fire largely independently
Variability of the EMG signal
Independent and random firing of MUs
- Peaks and valleys of AP may add up (green)
- At other instants peaks or valleys cancel (red)
Random negative and positive superimposition (phase cancellation)
- Neurophysiology has to deal with bi- or multiphasic signals (e.g. motor unit action potentials)
So
- Their mean value tends to zero, also after summation
- In quantification of the data this poses a problem
Summation of MUAPs to EMG: model 1
- 2 seconds force profile
- EMG units (MUAPs) all have positive size = 1
- 1 MUAP adds 1 force unit
- No random effects
Perfect match, but irrealistic because MUs fire irregularly
Summation of MUAPs to EMG: model 2
- Same force profile (2sec)
- MUAPs now fire irregularly, still with mean size = 1
- 1 MUAP adds 1 force unit
The force can still be estimated despite the random firing, but statistical precision is lower
Still irrealistic because EMG contributions have zero mean
